GENOMIC AND BREEDING RESOURCES TO PRODUCE SEEDED …

GENOMIC AND BREEDING RESOURCES TO PRODUCE SEEDED AND HIGH BIOMASS INTERSPECIFIC HYBRIDS OF NAPIERGRASS AND PEARL MILLET

By

DEV RAJ PAUDEL

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT

OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2018

© 2018 Dev Raj Paudel

To my late Mom

4

ACKNOWLEDGMENTS

I wish to express my appreciation to the members of my advisory committee: Dr.

Fredy Altpeter, Dr. Jianping Wang, Dr. Patricio Munoz, Dr. Calvin Odero, and Dr.

Salvador Gezan, who have guided, supported, and encouraged me throughout the

course of my research project. Sincere thanks to my research advisor Dr. Fredy

Altpeter for allowing me to join his group and pursue the work described here. Thank

you for all the support, guidance, and mentorship that you have provided during my

graduate studies. I am truly inspired by your professionalism and leadership role.

I am particularly indebted to my co-advisor Dr. Jianping Wang who provided me

with space in her lab to do my experiments and provided me with opportunities to

develop skills in molecular and computational biology. Your mentorship has been an

invaluable gift over the past couple of years. One day, I hope to inspire others as you've

inspired me.

I would like to gratefully acknowledge the University of Florida Graduate School

Fellowship for funding the first four years of my PhD. I am thankful to the Graduate

School Doctoral Dissertation Award for funding my final semester. I am grateful to the

Florida Plant Breeders Working Group for providing funds for this research.

I was grateful to be surrounded by many wonderful people during my journey.

Many thanks to talented undergraduates, Fan Wen, Erik Hanson, and Stephanie Maya,

who helped not only with the experiments, but also helped me learn and hone my

mentorship skills.

In addition, I would like to express my gratitude to the following people:

• Dr. Baskaran Kannan without whose help, support, and wisdom, this work

wouldn’t have been completed.

5

• Staff at Plant Science Research & Education Unit, Citra, FL for coordinating and

helping in planting, managing, and harvesting of research plots.

• Dr. Calvin Odero, Venancio Fernandez, Raphael Mereb Negrisoli, and Nikol

Havranek at the Everglades Research and Education Center for providing

support to do fieldwork in Belle Glade, FL.

• Dr. Tina Strauss, Dr. Eshan Gurung, Er. Prasan Gurung, and Dr. Laxman

Adhikari for your continued support, valuable insights in my research, and

friendship.

• Members of the Altpeter Lab: Derek Hurley, Dr. Saroj Parajuli, Dr. Ratna Karan,

Dr. Simon Gere, and Bryant Brown, for helping during field work and research

activities.

• Members of the Wang Lab: Dr. Liping Wang, Dr. Xiping Yang, Dr. Ze Peng, Dr.

Zhou Hai, Aleksey Kurashev, Dr. Yu-Chien Tseng, and Dr. Song Jian for helping

in fieldwork, lab experiments, as well as discussions related to research that

were instrumental in providing insights into some of the work described in this

document.

• Marco Sinche for kindly providing yield and flowering data of mapping

experiments.

• Dr. Rajeev Varshney for providing the pearl millet reference genome sequence.

• Dr. Karen-Harris Schultz for providing napiergrass sequences.

• Cynthia Hight for playing a phenomenal role in supporting academic endeavors

and ensuring that I was on track every semester.

6

• Members of the Plant Science Council for guidance, support, friendship, and

humor that has helped me survive graduate school.

• Nepalese community in Gainesville, FL for always being there for me.

Obtaining this advanced degree was only possible due to the sheer sacrifice,

patience, unconditional love and understanding, long-term support, and encouragement

in all aspects of my career and life from the love of my life, my dear wife, Ashmita

Guragain, my father Bala Bhadra Paudel, mother late Dhan Kumari Paudel, brother

Dipendra Paudel, sister-in-law Kalpana Sapkota, and niece Deleena Paudel. I am

thankful to my parents-in-law and family, and other relatives for persistent

encouragement and support.

Finally, this dissertation is dedicated to my true champion and idol, my mother,

late Dhan Kumari Paudel, whom I lost during the final semester. She was my confidant

and mentor. She was the epitome of love, sacrifice, simplicity, wisdom, and strength, to

whom I owe everything. Life isn't the same without you.

I hope I have made you proud.

7

TABLE OF CONTENTS page

ACKNOWLEDGMENTS .................................................................................................. 4

LIST OF TABLES .......................................................................................................... 10

LIST OF FIGURES ........................................................................................................ 12

LIST OF ABBREVIATIONS ........................................................................................... 17

ABSTRACT ................................................................................................................... 18

CHAPTER

1 BACKGROUND ...................................................................................................... 20

Introduction ............................................................................................................. 20

Botany .............................................................................................................. 21 Napiergrass Ancestry ....................................................................................... 22 Napiergrass in the United States ...................................................................... 23

Pearl Millet ........................................................................................................ 24 Interspecific Hybrids of Napiergrass and Pearl Millet (PMN Hybrids) ............... 25

Cytoplasmic Male Sterility (cms) ...................................................................... 25 Molecular Tools Applied in Plant Breeding ....................................................... 27

Genotyping by Sequencing (GBS) ................................................................... 29 Target Enrichment Sequencing ........................................................................ 30

Quantitative Trait Loci Analysis ........................................................................ 30 Biomass yield ................................................................................................... 31 Flowering Time ................................................................................................. 31

Genes Related to Flowering ............................................................................. 33 Objectives ............................................................................................................... 34

2 SURVEYING THE GENOME AND CONSTRUCTING A HIGH-DENSITY GENETIC MAP OF NAPIERGRASS (CENCHRUS PURPUREUS SCHUMACH.) ......................................................................................................... 37

Introduction ............................................................................................................. 38

Methods .................................................................................................................. 41 Napiergrass Genome Survey ........................................................................... 41 SSR Identification and Marker Development .................................................... 42

Plant Materials and DNA Extraction ................................................................. 43 Genotyping-by-sequencing ............................................................................... 43 Comparative Genomics .................................................................................... 43 Sequence Analysis and SNP Calling ................................................................ 44 Linkage Map Construction ................................................................................ 45 Comparison Between Napiergrass and Pearl Millet Genome........................... 46

8

Results .................................................................................................................... 46 Napiergrass Genome Survey ........................................................................... 46

Genotyping-by-sequencing ............................................................................... 47 SNP Calling by Various SNP Callers ................................................................ 49 Genetic Linkage Map Construction .................................................................. 49 Comparison Between Genomes of Napiergrass and Pearl Millet ..................... 51

Discussion .............................................................................................................. 52

3 MAPPING QTLS CONTROLLING FLOWER NUMBER AND FLOWERING TIME IN NAPIERGRASS ........................................................................................ 84

Introduction ............................................................................................................. 84 Materials and Methods............................................................................................ 87

Development of a Mapping Population ............................................................. 87 Phenotyping the Mapping Population ............................................................... 88

Genetic Map ..................................................................................................... 88 QTL Analysis .................................................................................................... 89 Candidate Gene Identification .......................................................................... 90

Results .................................................................................................................... 91 Phenotypes ...................................................................................................... 91

Number of flowers ...................................................................................... 91 Flowering Time .......................................................................................... 91

QTL Analysis .................................................................................................... 91

Number of flowers ...................................................................................... 91 Flowering time ............................................................................................ 92

Candidate Genes ............................................................................................. 92

Discussion .............................................................................................................. 93

Conclusion .............................................................................................................. 97

4 EVALUATE THE GENETIC BACKGROUND OF FLOWERING TIME IN A NAPIERGRASS GERMPLASM COLLECTION .................................................... 113

Introduction ........................................................................................................... 113 Materials and Methods.......................................................................................... 117

Plant Materials and Phenotyping .................................................................... 117

DNA Extraction ............................................................................................... 117 Targeted Candidate Genes ............................................................................ 118 Probe Design .................................................................................................. 118

Probe Synthesis, Selection, and Sequencing ................................................. 119 Sequence Read Trimming and Mapping ........................................................ 120 SNP Calling .................................................................................................... 120 Population Structure ....................................................................................... 120

Results .................................................................................................................. 121 Candidate Genes Related to Flowering .......................................................... 121 Probe Design .................................................................................................. 121 Sequence processing ..................................................................................... 123 Phylogenetic analysis ..................................................................................... 124

9

Genome wide association analysis ................................................................ 124 Discussion ............................................................................................................ 125

Conclusion ............................................................................................................ 129

5 GENERATION OF INTERSPECIFIC HYBRIDS BETWEEN PEARL MILLET AND NAPIERGRASS AND EVALUATION OF THEIR PERFORMANCE. ............ 142

Introduction ........................................................................................................... 142 Materials and Methods.......................................................................................... 147

Description of Male Sterile Line of Pearl Millet ............................................... 147 Production of cms Lines of Pearl Millet .......................................................... 147 Production of PMN Hybrids ............................................................................ 148 Experimental Design ...................................................................................... 149

Traits Evaluated ............................................................................................. 149 Data Analysis ................................................................................................. 150

Results .................................................................................................................. 150 Plant Height .................................................................................................... 150 Tiller Number .................................................................................................. 150

Stem Diameter ............................................................................................... 151 Leaf Length .................................................................................................... 151

Leaf Width ...................................................................................................... 152 Plant Biomass ................................................................................................ 152 Dry Biomass ................................................................................................... 153

Coefficient of Variation ................................................................................... 153 Correlation ...................................................................................................... 154

Discussion ............................................................................................................ 154

6 CONCLUDING REMARKS ................................................................................... 179

Summary .............................................................................................................. 179 Future work ........................................................................................................... 181

LIST OF REFERENCES ............................................................................................. 183

BIOGRAPHICAL SKETCH .......................................................................................... 208

10

LIST OF TABLES

Table page 2-1 Parameters used for SNP calling for each software ........................................... 58

2-2 Repetitive elements present in the napiergrass genome .................................... 61

2-3 The sequence alignment of ten napiergrass sequence contigs to the pearl millet genome ..................................................................................................... 62

2-4 Frequency of classified repeat types (considering sequence complementary) in napiergrass ..................................................................................................... 63

2-5 Primer pairs developed for napiergrass SSR markers ........................................ 67

2-6 Summary of the alignment of non-redundant tags of napiergrass (Cenchrus purpureus) to the available genomes of different species .................................. 68

2-7 Alignment of individual napiergrass reads using Bowtie2 ................................... 69

2-8 Summary of the combined linkage map of napiergrass and the percentage of gaps less than 5 cM in male and female parent linkage maps ........................... 70

2-9 Summary of napiergrass single nucleotide polymorphism (SNP) markers mapped on the combined linkage map using 9 different software pipelines ....... 71

3-1 Descriptive statistics of flowering date and number of flowers for 185 F1

hybrids of a cross (N190 N122) at Citra, FL., in 2012, 2013 (Sinche 2013) and 2016 .......................................................................................................... 109

3-2 Detailed information about the QTLs for number of flowers. LG = Napiergrass linkage groups, Peak markers represent the marker at the highest peak of QTL, LSI represents the LOD-1 support interval in cM. .................................... 110

3-3 Detailed information about the QTLs for flowering time. LG = Napiergrass linkage groups, Peak markers represent the marker at the highest peak of QTL, LSI represents the LOD-1 support interval in cM. .................................... 111

3-4 List of putative flowering related genes from the genome of pearl millet (Varshney et al. 2017). ..................................................................................... 112

4-1 Genes and publications related to flowering used in this research ................... 130

5-1 Details of the cross types used in the experiment. ........................................... 175

11

5-2 Descriptive statistics for different types of crosses. P-value represents Pr>F for Levene’s test for homogeneity of variance (center = median), CV = coefficient of variation. ...................................................................................... 176

5-3 Correlation coefficients and p-values for biomass weight and biomass-related traits for PMN hybrids evaluated in Citra, FL. ................................................... 177

12

LIST OF FIGURES

Figure page 1-1 Distribution of napiergrass. Black dots represent local napiergrass and red

dots represent napiergrass listed as invasive species ........................................ 36

2-1 Sequence variation for SNPs called in various regions of the pearl millet genome. ............................................................................................................. 72

2-2 Micro-collinearity between contigs from napiergrass to the pearl millet genome. ............................................................................................................. 73

2-3 Inversion duplication between napiergrass and pearl millet (shown in bottom figure). 74

2-4 Estimated coverage of PstI restriction sites in the pearl millet genome. ............. 75

2-5 Histogram of uniquely mapped reads to the pearl millet genome. ...................... 76

2-6 Venn diagram showing concordant napiergrass SNPs called by five reference-based SNP callers, SAMtools, GBS-SNP-CROP, GATK, FreeBayes, and TASSEL. Numbers in parenthesis after the program name shows the total number of SNPs called by each program. ................................. 77

2-7 Genetic linkage map of the napiergrass female parent N190. ............................ 78

2-8 Genetic linkage map of the napiergrass male parent N122. ............................... 79

2-9 Genotyping by sequencing single nucleotide polymorphism (GBS-SNP) marker distribution for the 14 linkage groups of napiergrass. A black bar means a GBS-SNP marker. A blue bar represents segregation distortion region. The left scale plate represents genetic distance (centiMorgan as unit). . 80

2-10 Consensus genetic linkage map of napiergrass. ................................................ 81

2-11 Circos plot of the mapped TASSEL de-novo UNEAK napiergrass markers with pearl millet reference genome. Pearl millet pseudomolecules start with “PM” and are color coded for each pseudomolecule. Napiergrass linkage groups start with “LG” and are in green color. Each line corresponds to tags that showed significant BLAST hits to the pearl millet genome (> 80% identity and > 50 bp length)............................................................................................. 82

2-12 Syntenic regions between napiergrass linkage groups and the pearl millet genome. PM01 to PM07 are pearl millet pseudomolecules, LG01 to LG14 are napiergrass linkage groups. The small dots represent significant BLAST hits of mapped UNEAK tags to the pearl millet genome (>80% identity and >50 bp length). ................................................................................................... 83

13

3-1 Histogram of number of flowers in the mapping population in 2012 in Citra, FL. X-axis represents the number of flowers and y-axis represents the count of plants. Average number of flowers of two parental lines, N122 and N190 are plotted. Data is adapted from Sinche, 2013. ................................................. 98

3-2 Histogram of number of flowers in mapping population for 2013 in Citra, FL. X-axis represents the number of flowers and y-axis represents the count of plants. Average number of flowers of two parental lines, N122 and N190 are plotted. Data is adapted from Sinche, 2013. ....................................................... 99

3-3 Scatterplot of number of flowers between 2012 and 2013 in Citra, FL. X-axis represents the number of flowers in 2012 and y-axis represents the number of flowers in 2013. Data is adapted from Sinche, 2013. ................................... 100

3-4 Histogram of number of days to first flowering in 2012 in Citra, FL. X-axis represents the number of days to flowering (FT) and y-axis represents the number of accessions. Days to first flowering of two parental lines, N122 and N190 are plotted. Data is adapted from Sinche, 2013. ..................................... 101

3-5 Histogram of number of days to first flowering in 2013 in Citra, FL. X-axis represents the number of days to flowering (FT) and y-axis represents the number of accessions. Days to first flowering of two parental lines, N122 and N190 are plotted. Data is adapted from Sinche, 2013. ..................................... 102

3-6 Histogram of first flowering date in the mapping population in 2016 in EREC, Belle Glade, FL. X-axis represents the number of days to flowering (FT) and y-axis represents the number of accessions. Days to first flowering of two parental lines, N122 and N190 are plotted. ...................................................... 103

3-7 Scatterplot of first date of flowering between different years and locations. X-axis and Y-axis represent the number of days to flowering (FT) in different years. Data for 2012 and 2013 are from Citra, FL and adapted from Sinche, 2013. Data for 2016 are from Belle Glade, FL. ................................................. 104

3-8 Scatterplot between number of flowers and days to flowering in 2012 in Citra, FL. X-axis represents the number of days to flowering (FT) in 2012 and y-axis represents the number of flowers per line in 2012. Data is adapted from Sinche, 2013..................................................................................................... 105

3-9 Scatterplot between days to flowering and number of flowers in 2013 in Citra, FL. X-axis represents the number of days to flowering (FT) in 2013 and y-axis represents the number of flowers per line in 2013. Data is adapted from Sinche, 2013..................................................................................................... 106

3-10 Linkage map of male parent N122 showing potential and stable QTLs. QTLs for number of flower (NF) are color coded in blue and that for flowering time (FT) are colored in green. A black bar means a GBS-SNP marker. Markers

14

are labeled on the left and genetic distance (centimorgan as unit) are labeled on right. LG = Linkage group. ........................................................................... 107

3-11 Linkage map of female parent N190 showing potential QTLs. QTLs for number of flower (NF) are color coded in blue and that for flowering time (FT) are colored in green. A black bar means a GBS-SNP marker. Markers are labeled on the left and genetic distance (centimorgan as unit) are labeled on right. LG = Linkage group. ................................................................................ 108

4-1 Histogram of length of flowering related genes................................................. 131

4-2 Number of probes designed per gene. ............................................................. 132

4-3 Number of probes designed as a factor of the size of the gene. ...................... 133

4-4 Distribution of the targeted flowering genes in the genome of pearl millet. Each blue bar represents a flowering gene mapped on the pearl millet genome. ........................................................................................................... 134

4-5 Number of paired-end reads per sample. ......................................................... 135

4-6 Bayesian Information Criteria (BIC) vs. number of clusters in k-means clustering suggests K=3 in the germplasm collection. ...................................... 136

4-7 Projection of the napiergrass germplasm collection using first two linear discriminants (LDs) from discriminant analysis of principal components (DAPC). The shape of the points represent grouping by DAPC (circle = group 1; triangle = group 2; square = group 3) and the colors represent the origin continent: black = Americas; blue = Asia; and cyan = Africa. ........................... 137

4-8 Evolutionary relationships of taxa. G stands for group assigned by DAPC and are in different shapes (circle = group 1; triangle = group 2; square = group 3) and the fill colors represent the origin: black = Americas, blue = Asia, cyan = Africa, and white=unknown). ......................................................................... 138

4-9 Histogram for days to flowering trait in napiergrass germplasm collection. ...... 139

4-10 QQ plots for different models using GWASpoly using 78,129 SNP markers. DTF = Days to flowering. .................................................................................. 140

4-11 Manhattan plots for different models using GWASpoly. P values adjusted with FDR at 0.05. DTF = Days to flowering. ..................................................... 141

5-1 Seeds of pearl millet dwarf cms line (A), cms high biomass pearl millet line (B), PMN hybrid (C), and napiergrass (D), respectively from left to right. ......... 158

5-2 Boxplot of plant height (cm) for four different types of crosses studied. Small colored dots represent individual plant height of the cross. Different letters in

15

blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17. ...................................................................... 159

5-3 Histogram of plant height (cm) for four different crosses. X-axis represent plant height in cm and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17. ........... 160

5-4 Boxplot of number of tillers for four different types of crosses studied. Small colored dots represent number of tillers for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17. ............................................ 161

5-5 Histogram of number of tillers for four different crosses. X-axis represent the number of tillers and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17. ........... 162

5-6 Boxplot of stem diameter (mm) for four different types of crosses studied. Small colored dots represent the stem diameter (mm) for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17................................ 163

5-7 Histogram of stem diameter (mm) for four different crosses. X-axis represent the stem diameter (mm) and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17. ... 164

5-8 Boxplot of leaf length (cm) for four different types of crosses studied. Small colored dots represent the leaf length (cm) for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17. ............................................ 165

5-9 Histogram of leaf length (cm) for four different crosses. X-axis represent the leaf length (cm) and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17. ........... 166

5-10 Boxplot of leaf width (mm) for four different types of crosses studied. Small colored dots represent the leaf width (mm) for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17. ............................................ 167

5-11 Histogram of leaf width (mm) for four different crosses. X-axis represent the leaf width (mm) and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17. ........... 168

16

5-12 Boxplot of fresh biomass per plant (kg) for four different types of crosses studied. Small colored dots represent the fresh biomass per plant (kg) for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17. ............................................................................................................. 169

5-13 Histogram of fresh biomass per plant (kg) for four different crosses. X-axis represent the fresh biomass per plant (kg) and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17. ................................................................................................ 170

5-14 Boxplot of projected dry biomass (tons per ha) for four different types of crosses evaluated at seven months of growth. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17. ............................................................................................ 171

5-15 Histogram of dry biomass (tons per ha.) for four different crosses evaluated at seven months of growth. X-axis represent the dry biomass (tons per ha.) and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17. ....................................... 172

5-16 Coefficient of variation (%) for the different traits evaluated. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17. ........... 173

5-17 Picture of four different crosses during harvest in Citra, FL. Cross A = Tift 85 × 25-17, Cross B = 787-S5 × 25-17, Cross C = MS 787 BC4 × 25-17, Cross D = P787 × 25-17. ............................................................................................ 174

17

LIST OF ABBREVIATIONS

cms Cytoplasmic male sterility

GBS Genotyping by sequencing

LG Linkage Group

NGS Next-generation sequencing

PMN Pearl millet napiergrass

SNP Single nucleotide polymorphism

SSR Simple Sequence Repeat

TES Targeted exome sequencing

18

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

GENOMIC AND BREEDING RESOURCES TO PRODUCE SEEDED AND HIGH BIOMASS INTERSPECIFIC HYBRIDS OF NAPIERGRASS AND PEARL MILLET

By

Dev Raj Paudel

December 2018

Chair: Fredy Altpeter Cochair: Jianping Wang Major: Agronomy

Napiergrass (Cenchrus purpureus Schumach) is a promising candidate for

forage and lignocellulosic biofuel production due to its high biomass potential. However,

napiergrass is listed as an invasive species in Florida due to wind dispersed seeds.

Seed production in napiergrass is compromised by low temperatures. Therefore, late

flowering genotypes, which are not able to flower before low temperatures come in

Florida, could be utilized to improve biosafety of napiergrass. However, genetic and

genomic resources for napiergrass are limited that preclude exploiting marker assisted

selection (MAS) for crop improvement. Genetic linkage map is an important tool for

MAS. In this research, we constructed the first high-density genetic map of napiergrass

by genotyping-by-sequencing a bi-parental mapping population of 185 F1 hybrids. As a

result, we mapped 1,913 single nucleotide polymorphism (SNP) markers into 14 linkage

groups of napiergrass, spanning a length of 1,410 cM with a density of one marker per

0.73 cM.

This genetic map enabled us to identify three stable and three potential

quantitative trait loci (QTLs) controlling number of flowers and flowering time,

19

respectively in the mapping population. We also identified five candidate genes related

to flowering in close proximity to the QTLs detected.

Full characterization of germplasm collections is very critical to efficiently utilize

them in breeding programs. We used targeted enrichment sequencing to characterize

the napiergrass germplasm collection and identified 78k SNPs in the collection. We

inferred the structure of the germplasm collection and constructed its phylogeny.

Genome wide association studies revealed one significant SNP for flowering time.

Napiergrass is mostly vegetatively propagated, which makes the planting

process much complicated and labor intensive. To ameliorate this, we introgressed

cytoplasmic male sterility (cms) into elite lines of pearl millet and hybridized them with

napiergrass to produce seed derived, sterile pearl millet napiergrass (PMN) hybrids. We

evaluated the biomass yield and uniformity of PMN hybrids generated by using different

parental backgrounds. There was a tremendous variation in different biomass related

traits among the crosses. These seeded-yet-sterile PMN hybrids could have a major

impact in the forage and biofuel industry if large scale production of high-quality seeds

that give rise to high yielding progenies can be developed.

20

CHAPTER 1 BACKGROUND

Introduction

Napiergrass (Cenchrus purpureus Schumach, syn. Pennisetum purpureum

Schumach) is a tropical perennial grass that originated from Africa (Singh, Singh, and

Obeng 2013). It is commonly called merker grass, elephant grass, or Uganda grass.

Napiergrass is a monocotyledonous flowering plant that belongs to the grass family

(Poaceae) and genus Cenchrus. Recently, species of Pennisetum and Odontelytrum

were merged and designated to the unified genus Cenchrus (Chemisquy et al. 2010).

Pennisetum genus consists of a heterogeneous group of over 140 species (Brunken

1977) and it includes species with basic chromosome number of 5, 7, 8 or 9 and ploidy

ranging from diploid to octoploid (Martel et al. 1997).

Napiergrass is an important fodder crop as well as an important cellulosic energy

crop due to its high dry biomass yield compared to sorghum, maize, sugarcane,

switchgrass, johnsongrass, and Erianthus (Ra et al. 2012). The first published note on

this grass was by Mynhardt, a Hungarian missionary in Barume (then north west

Rhodesia), who sent material to Zurich Botanical Gardens in 1905. However, the name

napiergrass recognizes the contribution of Colonel Napier of Bulawayo who first wrote

to the Rhodesian Agricultural Department to make them aware of the value of this plant

(Boonman 1988). Napiergrass is a short-day plant and flowering in tropical climates

occurs from autumn through winter (Singh, Singh, and Obeng 2013). It is an open-

pollinated species. The cultivars were mainly developed by vegetative propagation of

superior clones derived from natural out-crossing material (Bhandari, Sukanya, and

Ramesh 2006; Augustin and Tcacenco 1993). Germplasm collections of napiergrass

21

are maintained in several countries in Africa, Brazil, Puerto Rico, United States,

Australia, China, Pakistan, and India (Azevedo et al. 2012; Bhandari, Sukanya, and

Ramesh 2006).

Botany

Napiergrass is a perennial, rhizomatous grass, propagated vegetatively,

commonly from joints or cuttings of the canes (Thompson 1919). It tillers vigorously,

and mature tillers have 20 or more internodes with plant height ranging from 2 to 6 m.

Its leaves are long and narrow. Reproductive part of this grass is panicles,

characterized by tawny or purplish color, sessile fascicles, and sparsely plumose

bristles (Singh, Singh, and Obeng 2013). Napiergrass is protogynous where the stigma

exerts 3-4 days, starting at the top of the inflorescence, before anthesis, followed by

anthers shedding pollen for 3-4 days. The seeds are tiny in size averaging 3.8 million

seeds/kg and seed set is poor and prone to shattering with low germination rate.

Napiergrass requires a day length of 11 h or less to flower (Hanna 2004). When it is

grown in sub-temperate regions, it produces no or very few flowers because by the time

days shorten for flowering, the temperature drops too low to support flower growth.

Deep roots and rhizomes of napiergrass bind up the soil and prevent soil erosion (Juma

2014). The plant can be harvested in the first year itself unlike other grasses like

Miscanthus, that require the first year for establishment.

Napiergrass adapts to a wide range of soil types and pH range, while maximum

growth is attained on well-drained loamy soils that have high organic matter. It is

susceptible to waterlogging and traps cereal stem borer insect pest. Cereal farmers in

western Kenya control pests like stem borer using push-pull technology by establishing

a hedge of napiergrass around their cereal plots (Khan et al. 2001).

22

Napiergrass is highly heterozygous and gives rise to a very heterogeneous

population of seedlings, that are not morphologically uniform (Juma 2014). High

morphological variability of napiergrass acts as a valuable source of genetic variation.

Mass selection and hybridization has been used to develop several varieties of

napiergrass such as Uganda hairless, Cameroon, Gold Coast, and Clone 13 (Juma

2014).

Napiergrass Ancestry

Napiergrass is an allotetraploid (2n=4x=28, A’A’BB) (Jauhar 1981) with two sub-

genomes, A’ and B. The chromosomes in the A’ sub-genome of napiergrass is

homologous to the A genome of pearl millet (Cenchrus americanus (L.) Morrone, syn.

Pennisetum glaucum, 2n=2x=14) (Jauhar 1981). The two species, napiergrass and

pearl millet, can naturally hybridize to produce hybrids that are triploids with AA’ B

genomes, thus are sterile (Singh, Singh, and Obeng 2013). Based on chromosome

pairing in triploid hybrids (2n=3x=21; AA’B), it was inferred that the two species basically

share a genome (A and A’ being very similar). Genomic in situ hybridization revealed a

high level of homeology between genomes A and A’ (Reis et al. 2014). Approximately

29% of napiergrass genomes (A’B) were hybridized by the genome of pearl millet. It

was inferred that napiergrass and pearl millet had concomitantly diverged from the

common ancestor. The origin of napiergrass occurred at the interspecific hybridization

event, combining genome A from the common ancestor with genome B whose source is

still unknown (Reis et al. 2014; Jauhar and Hanna 1998). Genome A is 24% larger than

genome A’ of napiergrass. This difference in genome size could be related to genic

duplication in pearl millet and to genomic rearrangements in napiergrass. Pearl millet

has genome DNA content of 4.72 pg while the genome DNA content of napiergrass is

23

4.60 pg (Martel et al. 1997). Both species have approximately similar DNA content (pg)

but are different with regards to the basic chromosome number. Napiergrass is a

tetraploid and has about half the DNA content (1.15 pg) of the pearl millet monoploid

genome (2.36pg). Napiergrass also has smaller chromosomes than that of pearl millet

(Martel et al. 1997; Reis et al. 2014).

Napiergrass in the United States

Napiergrass was first introduced to Florida by forage breeders based on its high-

yielding forage attributes and now occurs on the banks of canals, waterways and

roadsides. Napiergrass cultivars that flower early produce a large amount of wind

dispersed seeds that enhance its invasive potential. Napiergrass can spread to natural

vegetation (D’Antonio and Vitousek 1992; Loope, Hamann, and Stone 1988; Schofield

1989) (Figure 1-1) and is listed as an invasive species by The Florida Exotic Plant Pest

Council (FLEPPC 2011). Therefore, environmental biosafety has become one of the

major concerns in napiergrass cultivation. It is also a major concern for sugarcane

cultivation as both have similar growth habits and reproduction modes (Singh, Singh,

and Obeng 2013) making it difficult to manage in the crop. In order to enhance the

environmental biosafety of napiergrass, measures to prevent invasiveness need to be

implemented. Napiergrass usually flowers as day lengths decreases and late flowering

has been observed in a number of accessions. Seed formation in napiergrass is

compromised due to low temperatures at the end of the season. To improve the

biosafety of napiergrass, it is important to select plants for late flowering and

hybridization. Therefore, the development of cultivars that are sterile or are late

flowering is important for napiergrass management in the field to mitigate invasiveness.

24

Pearl Millet

Pearl millet is a highly cross-pollinated monocot (Rajaram et al. 2013) belonging

to Paniceae tribe within the Panicoideae subfamily of the Poaceae family (Devos and

Gale 2000; Devos et al. 2000). It is a coarse grass grown primarily for grain in Africa

and Asia and as a feed and forage in America (Gupta and Mhere 1997). Pearl millet

accounts for almost half of the global millet production and is important for food security

in some of the world’s driest and hottest areas. It is grown mostly for its ability to

produce grain under hot, dry conditions on infertile soils with low water-holding capacity

(ICRISAT and FAO 1996; Rajaram et al. 2013). Pearl millet has much larger seeds

compared to napiergrass and is easy to establish in the field via seeds. Availability of a

reference draft genome of pearl millet (Varshney et al. 2017) provides an excellent

resource that can be used extensively for genomic improvement of pearl millet and

related species.

Cytogenetic studies on napiergrass and pearl millet have classified pearl millet as

primary and napiergrass as the secondary gene pool of the genus Pennisetum (Harlan

and Wet 1971; Martel et al. 1997). Primary gene pool of the genus Pennisetum includes

three species (one cultivated and two wild species) with 2n=2x=14 chromosomes.

Secondary gene pool is represented by one allotetraploid species, Pennisetum

purpureum (2n=4x=28) with A’A’BB genomes (Martel, Ricroch, and Sarr 1996). The

primary gene pool corresponds to the traditional concept of biological species and

includes spontaneous races (wild and/or weedy) as well as cultivated races. Crossing

and gene transfer is generally easy among species in the primary gene pool to form

generally fertile hybrids. Species in the secondary gene pool can hybridize with species

in the primary gene pool, but the hybrids tend to be sterile (Harlan and Wet 1971).

25

Interspecific Hybrids of Napiergrass and Pearl Millet (PMN Hybrids)

The hybrids of pearl millet and napiergrass, commonly called PMN hybrids,

combine the superior forage quality of pearl millet and the high yielding ability of

napiergrass (Gupta and Mhere 1997; Osgood, Hanna, and Tew 1997). These hybrids

are male and female sterile due to triploidy (2n=3x=21) (Gupta and Mhere 1997) and

have reduced persistence and ratooning ability as compared to napiergrass (Cuomo,

Blouin, and Beatty 1996). The resulting hybrid with AA’B genome is morphologically

more similar to napiergrass due to larger genetic contribution (66.7% chromosomes)

and dominance of napiergrass grass B genome over the pearl millet A genome for

genetic characters such as earliness, inflorescence and leaf characteristics, and seed

size (Obok, Ova, and Iwo 2012; Gonzalez and Hanna 1984). The dominance of B

genome over the A’ genome masks genetic variability (consequently phenotypic

variability) on the A’ genome. Most of the characteristics like resistance to pests,

vigorous growth, and outstanding forage yield potential are contributed by the B

genome (Hanna 1987). PMN hybrids do not set seed and thus do not pose a threat of

uncontrolled establishment through dissemination of seeds (Hanna and Monson 1980)

and are not considered invasive (Jessup 2013). Vegetative propagation of PMN hybrids

or napiergrass is a major cost factor. Development of varieties that allow large scale

production of high quality seeds giving rise to sterile plants represents a significant step

in biomass grass breeding because field establishment using seeds will allow for

automation resulting in significant cost reduction (Osgood, Hanna, and Tew 1997).

Cytoplasmic Male Sterility (cms)

Cytoplasmic male sterility (cms) results from the interaction between organellar

and nuclear genomes that renders functional pollen production obsolete (Ram, Ram,

26

and Yadava 2007). cms plants do not produce viable pollen while being completely

female fertile. cms is important in hybrid seed production, because it helps to control

pollination for commercial production of F1 hybrid seeds (Smith and Chowdhury 1983).

cms has been extensively and cost effectively exploited in a number of agronomic and

horticultural crops including pearl millet (Havey 2004). cms in pearl millet was

discovered in the 1955 (Burton 1958) and was first released as male-sterile inbred

Tifton 23A (A1 or milo cytoplasm). Other cms sources were also later identified (Burton

and Athwal 1967) and the A4 cytoplasm was found useful for forage hybrids that do not

require male-fertility restoration (Havey 2004). The availability of the cms pearl millet

‘Tift 23A’ paved the way to commercially produce seed-propagated interspecific hybrids

that can facilitate seed harvest and allows the production of hybrids from easily drilled

seeds (J. B. Powell and Burton 1966). cms can be transferred to elite varieties to make

them sterile. To facilitate this, the method of backcrossing is suitable, where established

cultivars that are deficient in one or two specific traits can be improved through

crossing. In a backcross method, F1 hybrid is repeatedly crossed with a desirable parent

to get the desired traits (Schlegel 2009; Acquaah 2012). This has been commonly used

to transfer entire sets of chromosomes from foreign cytoplasm in order to create

cytoplasmic male-sterile genotypes (Acquaah 2012). The availability of cms lines of

pearl millet will help to develop homogenous lines of pearl millet that are male sterile.

These lines can then be used to cross with near-inbred napiergrass to produce uniform

progenies of male and female sterile PMN hybrids. While producing seeds at a

commercial scale, it should be noted that napiergrass usually flowers in November –

27

December, so seed production should be established in areas that stay frost-free until

December.

Molecular Tools Applied in Plant Breeding

As a group of important forage and biofuel species, napiergrass has received

increasing attention in recent years. Molecular studies in napiergrass have concentrated

on assessing genetic diversity, cultivar identification, origin, and relatedness using

limited genetic marker technologies (López et al. 2014; Harris-Shultz, Anderson, and

Malik 2010; Kandel et al. 2016; Dowling, Burson, and Jessup 2014; Dowling et al.

2013). Despite some advances in genetic research of napiergrass, the genetic and

molecular mechanisms of agronomic traits that must be improved for commercialization

are still poorly understood. A good genetic understanding of biomass related traits such

as emergence date, flowering time control, nutrient uptake, abiotic and biotic stress

tolerance is needed to aid genetic improvement of crops which can be facilitated by a

genetic linkage map (Ma et al. 2012) that helps to understand genome structure and to

identify trait-specific molecular markers (Cai et al. 2015). In order to accelerate

molecular tool development high-throughput genotyping is required (F. Lu et al. 2013).

Molecular markers have recently gained popularity because they are not subject to

environmental influence, are available in vast numbers, and are more objective as

compared to phenotypic markers (Kim et al. 2015). Today, plant breeders have access

to several types of DNA markers and molecular breeding tools to choose from.

A DNA marker is a specific DNA sequence on a chromosome that shows

polymorphism between individuals (Andersen and Lübberstedt 2003; Agarwal,

Shrivastava, and Padh 2008; Kumar 1999). Genomic variation leading to DNA

polymorphisms can be directly linked to differences in phenotype, used to find

28

relationships between individuals in a population, and can be used as genetic markers

(Deschamps, Llaca, and May 2012; Rafalski 2002). Most crop agronomic traits are

controlled by several loci that contribute to minor phenotypic effects and are known as

quantitative trait loci (QTL)s (Falconer, Mackay, and Frankham 1996). Since, molecular

markers are stable and detectable in all tissues and are not confounded by the

environment and other effects (Agarwal, Shrivastava, and Padh 2008), it is possible to

assign chromosomal positions to individual QTLs and to determine the effect of

individual QTLs (Kumar 1999). In order to utilize these markers in a breeding program,

these markers should be tightly linked to the desired QTL. One of the major objectives

of molecular breeding in plant species is to connect phenotype to the genotype and then

use this knowledge in MAS by making phenotypic predictions based on genotypes

(Poland and Rife 2012). Various marker technologies are currently available like

restriction fragment length polymorphism (RFLP), random amplified polymorphic DNA

(RAPD), amplified fragment length polymorphism (AFLP), simple sequence repeat

(SSR), and single nucleotide polymorphism (SNP) (Agarwal, Shrivastava, and Padh

2008; Kumar 1999).

SSRs or microsatellites are short (2-6bp) repetitive DNA sequence distribute

randomly in genomes. Since SSR markers are hyper-variable, multi-allelic, often co-

dominant, highly reproducible, and readily multiplexed, they are one of the best choices

for foreground selection in marker-assisted programs (Rajaram et al. 2013). They have

been commonly used in genotyping populations, constructing linkage maps, and

identifying QTLs for desirable traits. Linkage maps create a framework for trait mapping

and QTL analysis (Rajaram et al. 2013). With the advent of next generation sequencing

29

(NGS) platforms, genotyping a large number of progenies of mapping populations

become much efficient, which further speedup the QTL identification.

In recent years, SNP markers have gained much consideration in the breeding

community as they occur in abundance and every SNP in a single copy DNA is a

potentially useful marker (Ganal, Altmann, and Röder 2009). SNPs can be categorized

according to nucleotide substitutions either as transitions (C/T or G/A) or transversions

(C/G, A/T, C/A or T/G) (Jiang 2013). SNPs are typically identified from sequences, such

as expressed sequence tag (EST) sequence data, array analysis, amplicon sequencing,

and next generation sequencing technologies (Ganal, Altmann, and Röder 2009). For

SNP genotyping, the widely used platforms include BeadXpressTM, GoldenGateTM,

Infinitum®, GeneChipTM, GenFlex TM, SNAPshotTM, TaqMan TM, SNPstreamTM,

SNPWaveTM, iPLEX GoldTM, ARRAYTM, and KASPTM among others. (Semagn et al.

2014).

Genotyping by Sequencing (GBS)

Sequence-based genotyping methods like GBS have enabled simultaneous

marker discovery and genotyping (Poland and Rife 2012). GBS uses direct genome

sequencing to produce genotyping information utilizing high throughput and multiplexing

capacities of NGS (Paux et al. 2012; Deschamps, Llaca, and May 2012). Restriction

enzyme is utilized for complexity reduction followed by multiplex sequencing, which

requires less DNA as well as avoids random shearing and size selection (Poland,

Brown, et al. 2012). The resulting sequences allow direct SNP detection, which avoids

the marker assay development stage (Deschamps, Llaca, and May 2012) and thus

make GBS an efficient genotyping method (Elshire et al. 2011). GBS has proven to be

effective for marker discovery and trait mapping in several species like wheat (Poland,

30

Endelman, et al. 2012), switchgrass (F. Lu et al. 2013), and potato (Uitdewilligen et al.

2013). Since, GBS performs SNP discovery and genotyping simultaneously, it is

particularly of high value to understudied crops lacking reference genomes (Kim et al.

2015).

Target Enrichment Sequencing

Target Enrichment Sequencing (TES) is another NGS-enabled approach, which

focuses on genes or genomic regions of interest for sequencing. In this approach,

probes are designed to enrich capturing of the target genomic regions based on

sequence homology, which are then sequenced. TES has been used in several crops

such as barley (Russell et al. 2016), switchgrass (Grabowski et al. 2016), and peanuts

(Peng et al. 2017), to identify sequence variations but hasn't been previously applied in

napiergrass.

Quantitative Trait Loci Analysis

Quantitative trait loci (QTL) refers to a specific region on chromosomes that

harbors gene(s) controlling traits. QTL analysis is performed by estimating correlation

between phenotype data with genotype (markers) data of segregating populations

(Miles and Wayne 2008). Primary types of segregating population for QTL mapping

include F2, recombinant inbred lines (RILs), BC1 (backcross 1), double haploid lines

(DHL), near-isogenic lines (NILs) and full-sib F1 (pseudo-testcross) (Schneider 2005).

Quantitative traits, such as yield, height, flowering time, pest and disease resistance, in

plants have been mapped in genomes of several species. QTLs are valuable

information to develop markers linked to traits of interests for MAS in molecular

breeding programs. Genetic variability for flowering time exists in napiergrass (Sinche

2013). Therefore, identifying flowering time related QTLs will help in identifying markers

31

linked to late flowering trait, an important trait to reduce invasiveness. Sexual

hybridization of genetically distant parents and selection of late flowering, high yielding

accessions would increase the biofuel yield and enhance the biosafety of napiergrass.

Biomass yield

In a comparison of major energy crops for ethanol production, napiergrass

showed the highest dry biomass yield than sorghum, maize, sugarcane, switchgrass,

johnsongrass, and Erianthus (Ra et al. 2012). Napiergrass yields up to 84.8 Mg ha-1 yr-1

have been obtained in Puerto Rico (Vicente-Chandler, Silva, and Figarella 1959), while

in Florida, yields ranged between 35-45 Mg ha-1 yr-1 (Woodard and Sollenberger 2012;

Erickson et al. 2012). Biomass yield is a complex trait and studies have been conducted

to identify QTLs for various components affecting yield. In sorghum, QTLs for yield

related traits such as plant height, tiller number, leaf length, leaf width, stem diameter,

and flowering time have been identified (Hart et al. 2001; Murray et al. 2008; Xiao-ping

et al. 2011). Similarly, QTLs for leaf yield, stem yield, plant height and flowering time

have also been identified in Miscanthus sinensis (Gifford et al. 2015; Atienza et al.

2003). In Miscanthus, QTLs for yield co-segregated with other traits like number of

tillers, leaf area, leaf length, and leaf width (Gifford et al. 2015). In napiergrass, biomass

yield showed high correlations with number of tillers, plant height, and stem diameter

(Sinche 2013).

Flowering Time

Floral transition is the switch from vegetative growth to reproductive growth in

plants and it primarily determines flowering time. Flowering time is a key factor in plant

adaptation and is linked to various attributes like plant height, yield, number of leaves,

etc. (Durand et al. 2012). In many species, flowering is induced in response to the

32

length of periods of lightness and darkness associated with day and night length. These

species are categorized as short-day, long-day, intermediate-day, or day-neutral based

on their day length requirement (Schlegel 2009; Bastow and Dean 2002). Plants in

which flowering is favored by day lengths shorter than the critical and corresponding

long nights are called short-day plants (eg. Glycine max, Oryza sativa, and Zea mays).

The plants in which flowering is initiated when the day length is longer than the critical

are called long-day plants (eg. Hordeum vulgare, Triticum aestivum, and Solanum

tuberosum) (Garner 1933; Schlegel 2009; Bastow and Dean 2002). Napiergrass

belongs to the short-day plants (Osgood, Hanna, and Tew 1997; Singh, Singh, and

Obeng 2013).

Many studies have detected QTLs related to flowering time or earliness in

various crops. The genetic control of flowering time is in general quantitative in nature

(H. Lu et al. 2014). For example, in rice, 15 QTLs were associated with days to

flowering (Maheswaran et al. 2000), and in tomato, three QTLs related to earliness were

identified that were associated with flowering time, fruit set time, and ripening time

(Lindhout et al. 1994). However, a previous study in maize showed that flowering time

was a complex trait and no major QTL should be expected (Buckler et al. 2009). QTLs

for flowering related traits like 50% anthesis and heading date colocalized with other

QTLs for number of tillers, tiller diameter, leaf width, and leaf area in Miscanthus

(Gifford et al. 2015). In orchardgrass (Dactylis glomerata L)., 11 QTLs for heading date

and flowering time were found to be distributed on three linkage groups where

candidate genes such as hd1 and VRN1 were annotated (Zhao et al. 2016).

33

Flowering time in napiergrass is of prime importance as it is related to biosafety

and biomass quality. Late flowering napiergrass lines serve as potential bio-safe

biofuels as their flowering is compromised by low temperatures which usually occurs in

early December in Florida. Late flowering cultivars, therefore have less potential for

invasiveness than early flowering genotypes and may produce higher yields due to a

longer period of vegetative growth. Extensive phenotypic variation in flowering time is

an indication that flowering time is quantitative in nature and several genes might be

controlling the trait. Identifying QTLs for flowering time in napiergrass will help to

shorten the breeding cycle by the successful utilization of MAS. A biparental mapping

population from a cross between early-flowering and late-flowering lines will generate

progenies segregating for flowering time and this population can be genotyped in order

to identify QTLs for flowering time.

Genes Related to Flowering

Flowering time has been studied in some grass species and genomic studies

have identified a number of genes involved in flowering. Potential candidates can be

FRI, LEAFY, CO, DNF, MADS-box, and RID1 that have some roles in flowering

regulation in other species like Arabidopsis, wheat, maize, rice (Lee, Bleecker, and

Amasino 1993; Wuxing Li et al. 2013; Suárez-López et al. 2001; Morris et al. 2010; C.

Wu et al. 2008). For example, high levels of proteins encoded by FLOWERING LOCUS

T (FT) are correlated with early flowering and the lack of these causes late flowering

(Samach 2012). Major genes involved in photoperiod of flowering are highly conserved

between rice and Arabidopsis (C. Wu et al. 2008). For example, Hd1 QTL in rice that

promotes heading under short-day conditions corresponds to a gene homolog of

CONSTANS in Arabidopsis (Yano et al. 2000); Hd17 corresponds to a homolog of

34

Arabidopsis ELF3 (EARLY FLOWERING 3) (Matsubara et al. 2012; Matsubara et al.

2008); and Hd3a encodes a protein that is closely related to Arabidopsis FT (Kojima et

al. 2002). Mining of genome sequences that are available for several grass species for

flowering related genes and their characterization can identify candidate genes in

napiergrass. These candidate genes can be used for screening germplasm collections

for identification of haplotypes that confer late flowering. In addition to this, integrating

genomics with conventional breeding will help to shorten the breeding cycle for

selection.

Objectives

The overall objectives of this research are to identify genetic components (QTL

and candidate genes) underlying flowering time in napiergrass and to develop male and

female sterile PMN hybrids that are yielding high amounts of biomass and display

uniform seed progenies.

The substantial variation in flowering time in napiergrass can be utilized for

breeding late flowering varieties. This helps to make napiergrass a bio safe genotype for

biomass production in northern Florida where a freeze event typically occurs before

flowering of these late lines. The identification of QTLs controlling flowering time will

make the pre-selection for late flowering lines in breeding materials highly efficient.

Identifying polymorphisms within flowering genes in napiergrass germplasm collection

helps to reveal allelic variation in the germplasm. Development of a male sterile and

homozygous line of pearl millet will prevent its self-pollination, thus facilitating easy

crossing with napiergrass under field conditions. These high-biomass yielding male

sterile lines can then be used for hybridization with napiergrass.

35

The specific objectives of this research were to 1) construct genetic maps of

napiergrass, 2) identify QTLs for flowering time in napiergrass, 3) evaluate sequence

variations in a napiergrass germplasm collection, and 4) develop and evaluate male and

female sterile PMN hybrids.

36

Figure 1-1. Distribution of napiergrass. Black dots represent local napiergrass and red dots represent napiergrass listed as invasive species

-50

0

50

-100 0 100 200

Longitude

La

titu

de

37

CHAPTER 2 SURVEYING THE GENOME AND CONSTRUCTING A HIGH-DENSITY GENETIC

MAP OF NAPIERGRASS (CENCHRUS PURPUREUS SCHUMACH.)

Napiergrass (Cenchrus purpureus Schumach.) is a tropical forage grass and a

promising lignocellulosic biofuel feedstock due to its high biomass yield, persistence,

and nutritive value. However, its utilization for breeding has lagged behind other crops

due to limited genetic and genomic resources. In this study, next-generation sequencing

was first used to survey the genome of napiergrass. Napiergrass sequences displayed

high synteny to the pearl millet genome and showed expansions in the pearl millet

genome along with genomic rearrangements between the two genomes. An average

repeat content of 27.5% was observed in napiergrass including 5,339 simple sequence

repeats (SSRs). Furthermore, to construct a high-density genetic map of napiergrass,

genotyping-by-sequencing (GBS) was employed in a bi-parental population of 185 F1

hybrids. A total of 512 million high quality reads were generated and 287,093 SNPs

were called by using multiple de-novo and reference-based SNP callers. Single dose

SNPs were used to construct the first high-density linkage map that resulted in 1,913

SNPs mapped to 14 linkage groups, spanning a length of 1,410 cM and a density of 1

marker per 0.73 cM. This map can be used for many further genetic and genomic

studies in napiergrass and related species.

This chapter was published in Scientific Reports and is licensed under a Creative Commons Attribution 4.0 International License.

Paudel, D., Kannan, B., Yang, X., Harris-Shultz, K., Thudi, M., Varshney, R.K., Altpeter, F. and Wang, J., 2018. Surveying the genome and constructing a high-density genetic map of napiergrass (Cenchrus purpureus Schumach). Scientific reports, 8(1), p.14419.

38

Introduction

Napiergrass (Cenchrus purpureus Schumach., syn. Pennisetum purpureum

Schumach.), also known as elephant grass, is a tropical perennial grass native to

eastern and central Africa. It is cultivated primarily for forage and widely used by

smallholder dairy farmers due to its high growth rate, leaf nutritive value, perennial

nature, persistence, ease of propagation, and broad adaptation (Bhandari, Sukanya,

and Ramesh 2006; Farrell, Simons, and Hillocks 2002; Singh, Singh, and Obeng 2013;

Chemisquy et al. 2010). As a C4 grass species, napiergrass is a promising candidate

feedstock for biofuel production due to its superior yield of biomass (Ra et al. 2012;

Anderson, Casler, and Baldwin 2008; Somerville et al. 2010). Napiergrass cultivars are

typically developed from natural out-crossings (Bhandari, Sukanya, and Ramesh 2006;

Augustin and Tcacenco 1993). It is an allotetraploid (2n = 4x = 28, A’A’BB) (Jauhar

1981) with an average amount of DNA per G1 nucleus of 5.78 pg (M. G. Taylor and

Vasil 1987). The chromosomes in the A’ genome of napiergrass are believed to be

homologous to the A genome of pearl millet (Pennisetum glaucum, 2n = 2x = 14, AA)

(Jauhar 1981). Pearl millet and napiergrass form a monophyletic group (Martel et al.

1997) and were initially classified as primary and secondary gene pool of the genus

Pennisetum, respectively (Harlan and Wet 1971; Martel et al. 1997). Recently, species

of Pennisetum and Odontelytrum were transferred to the unified genus Cenchrus

(Chemisquy et al. 2010). Pearl millet and napiergrass can hybridize to produce hybrids

called kinggrass (Dowling, Burson, and Jessup 2014) or Pearl Millet-Napiergrass (PMN)

hybrids (Hanna and Monson 1980; Burton 1944; Muldoon and Pearson 1979). These

hybrids are sterile due to triploidy (2n = 3x = 21, AA’B genome) (Gupta and Mhere

1997), thus preventing the unintended spreading into natural areas or other cropping

39

systems by wind dispersed seeds. Some PMN hybrids show high heterosis for biomass

yield and forage quality while the perennial, persistent nature is often reduced

compared to napiergrass (Singh, Singh, and Obeng 2013).

The targeted improvement of napiergrass includes identification of agronomically

superior genotypes and studies assessing genetic diversity and relatedness using

random amplification of polymorphic DNA (RAPD), amplified fragment length

polymorphism (AFLP), isozymes, and simple sequence repeats (SSRs) developed for

other species like pearl millet and buffelgrass (Pennisetum ciliare) (Lowe et al. 2003;

Bhandari, Sukanya, and Ramesh 2006; Harris-Shultz, Anderson, and Malik 2010;

Kandel et al. 2016; Dowling et al. 2013; Dowling, Burson, and Jessup 2014; López et al.

2014; Smith et al. 1993). Other than these, genetic information on napiergrass is very

meager (Smith et al. 1993). A genetic map is lacking and molecular tools are not yet

deployed in napiergrass breeding programs (Dowling et al. 2013; Negawo et al. 2017).

Development of molecular markers for detection and utilization of DNA polymorphisms

will help to understand the molecular basis of various agronomic traits (Song et al.

2015). Molecular breeding for yield components, flowering date, nutrient uptake, abiotic

and biotic stress tolerance will accelerate genetic improvement of napiergrass. This can

be greatly facilitated by having access to marker resources like SSR, single nucleotide

polymorphisms (SNPs), and genetic linkage maps. SSRs as molecular markers are very

advantageous because they are locus specific, multi-allelic, co-dominant, and easy to

detect by polymerase chain reaction (PCR) (W. Powell, Machray, and Provan 1996;

Kannan et al. 2014). SNP markers have gained increasing consideration in molecular

breeding and linkage map construction as they occur in a large number and high

40

density (Ganal, Altmann, and Röder 2009). Access to these resources will support

marker-assisted selection (MAS) by making phenotypic predictions based on the

genotype (Poland and Rife 2012).

Recently, next generation sequencing (NGS) technology has simplified linkage

map construction by using high throughput genotyping-by-sequencing (GBS), which

allows simultaneous SNP discovery and genotyping across the whole genome of the

population of interest (Elshire et al. 2011; Poland and Rife 2012; Deschamps, Llaca,

and May 2012). GBS has been effective for marker discovery, genetic mapping,

quantitative trait locus (QTL) analysis, population genetics, and comparative genomics

studies in several diploid species, and has recently gained popularity in polyploid

species such as wheat (Triticum aestivum) (Poland, Endelman, et al. 2012), switchgrass

(Panicum virgatum) (F. Lu et al. 2013), potato (Solanum tuberosum) (Uitdewilligen et al.

2013), and sugarcane (Saccharum spp.) (Yang, Sood, et al. 2017) among others.

However, the presence of highly similar homeologous copies of two genomes in

allopolyploid species complicates SNP detection which relies on delineating true allelic

SNPs from homeologous SNPs because sequences from homeologous loci mimic

allelic SNPs and can introduce false-positives. Distinguishing allelic SNPs from

homeologous SNPs relies on the use of high-stringency sequence read alignment,

specifically uniquely aligned reads (Clevenger et al. 2015). Despite challenges of using

GBS for genotyping of polyploid species, genetic mapping without a reference genome

has been carried out in switchgrass by defining linkage groups with the modulated

modularity clustering (MMC) method (Stone and Ayroles 2009) referring to the genome

of foxtail millet (Setaria italica) (F. Lu et al. 2013). A genetic map of wheat was

41

constructed by using the bin-mapping procedure with homozygous genotypes of a

double-haploid population (Poland, Brown, et al. 2012). Each program for calling

variants utilizes different models or algorithms to identify potential polymorphisms,

therefore, multiple software programs need to be evaluated in order to identify the best

SNP caller for polyploids (Clevenger et al. 2015).

Linkage maps are important tools for map-based cloning, marker-assisted

breeding, QTL identification, genome organization, and comparative genomics of

important species. A number of linkage maps have been constructed for several

grasses including pearl millet (Punnuri et al. 2016). However, so far, napiergrass SSR

markers, genetic linkage map, or reference genome assembly are lacking. The purpose

of this study was to survey the napiergrass genome and to construct a high-density

genetic linkage map. Here, for the first time, we have surveyed whole genome

sequences in napiergrass, developed SSR markers, and constructed high-density

genetic map of napiergrass to investigate its genomic and genetic architecture.

Methods

Napiergrass Genome Survey

The genomic DNA of napiergrass cultivars Merkeron and UF1 was sequenced

using Illumina Genome Analyzer and 454 GS-FLX. For Illumina reads, reads that

contained more than 50% low-quality bases (Q20) were removed and adaptor

sequences were trimmed. Quality and adapter trimming of 454 reads was done using

default Newbler v2.8 (454 Life Sciences, Roche, Branford, CT) settings. Illumina reads

were assembled using ABySS/1.3.4 (Simpson et al. 2009) with kmer size ranging from

25 to 60 at intervals of 5. The 454 reads were assembled using Newbler v2.8 (454 Life

Sciences, Roche, Branford, CT) with default parameters. The assemblies were

42

completed using CAP3 (Xiaoqiu Huang and Madan 1999). The largest 10 contigs of the

assembly were selected for further analysis. Repeats on these contigs were masked

using a comprehensive public repeat database compiled from TIGR plant repeats

(http://plantrepeats.plantbiology.msu.edu/), Plant miniature inverted-repeat transposable

elements (P-MITE) database (http://pmite.hzau.edu.cn/django/mite/), MIPS Repeat

Element Database (http://mips.helmholtz-muenchen.de/plant/recat/), and Repbase from

RepeatMasker software (http://www.repeatmasker.org/). Unique repeats were extracted

from this database by removing redundant repeats with 98% identity using CD-HIT/4.6

(Garrison and Marth 2012). SNPs and indels were called using FreeBayes/0.9.15

(Garrison and Marth 2012) excluding alleles with depth less than 20. The annotation of

SNPs was performed using SnpEff/4.0 (http://snpeff.sourceforge.net/) (Cingolani et al.

2012). In order to identify sequence similarity among the two genomes, clean reads

from Illumina and 454 were aligned to the pearl millet genome v1 (Varshney et al. 2017)

using bowtie2/2.2.5.

SSR Identification and Marker Development

The napiergrass assembly was used to identify SSR markers that

contained repeat motifs ranging in length from 1 to 6 bp. The minimum number of

repeats was 10 for Mono-, 6 for Di-and 5 for Tri-, Tetra-, Penta- and Hexa-. SSRs were

analyzed based on their types, number of repeats, and percentage frequency of

occurrences of each SSR motif. SSRs in napiergrass were detected using

MIcroSAtellite identification tool (MISA) (Thiel et al. 2003) and primers were developed

using primer3 software (Rozen and Skaletsky 2000). SSR search results were input into

scripts p3_in.pl and p3_out.pl in order to identify SSR primer pairs for napiergrass.

http://plantrepeats.plantbiology.msu.edu/

http://pmite.hzau.edu.cn/django/mite/

http://mips.helmholtz-muenchen.de/plant/recat/

http://www.repeatmasker.org/

http://snpeff.sourceforge.net/

43

Plant Materials and DNA Extraction

A mapping population of 185 F1 hybrid progenies were developed from a cross

between two napiergrass accessions (N122 and N190) described previously (Sinche

2013; Sinche et al. 2018). The 185 F1 hybrids were planted in the field at the Plant

Science Research and Education Unit (PSREU), Citra, Florida, along with the parental

accessions.

Young and healthy leaf tissues were harvested from each individual of the

mapping population. DNA extraction was done following the protocol described

previously (Dellaporta, Wood, and Hicks 1983). The extracted DNA samples were run

on a 2% agarose gel to check the quality and quantified with PicoGreen to meet the

requirements of GBS. 185 F1 plants that were confirmed to be true hybrids using SSR

markers (Sinche et al. 2018) were selected for GBS.

Genotyping-by-sequencing

GBS data was generated at the Institute of Biotechnology, Cornell University as

described previously (Elshire et al. 2011). Briefly, DNA samples were digested with the

restriction enzyme PstI followed by ligation of adapters, that consisted of Illumina

sequencing primers and barcode adapters, to the DNA fragment ends. After ligation, 95

samples were combined into a pool and PCR amplification was performed to create a

GBS library and sequenced on Illumina HiSeq 2000.

Comparative Genomics

Unique tags of napiergrass from TASSEL de-novo UNEAK/3.0 (Glaubitz et al.

2014) pipeline were used for comparative genomic analysis. CD-HIT/4.6.4 (Weizhong Li

and Godzik 2006) was used to cluster the tags. Genomes of rice (Osativa_323_v7.0),

Brachypodium (Bdistachyon_314_v3.0), maize (Zmays_284_AGPv3), sorghum

44

(Sbicolor_313_v3.0), foxtail millet (Sitalica_312_v2), switchgrass (Pvirgatum_273_v1.0),

wheat (Taestivum_296_v2), Arabidopsis (Athaliana_167_TAIR9) were downloaded from

Phytozome v11 (https://phytozome.jgi.doe.gov/pz/portal.html). The barley genome

(ASM32608v1) was downloaded from Ensembl (http://www. ensembl.org). We used

BLASTN (BLAST v2.5.0) with the default settings and an e-value cutoff of 1×10-8 to

blast the uniquely clustered tags of napiergrass against different genomes in order to

find the percentage similarity of napiergrass reads among the various grass species.

Tags of 64 bp with > 80% identity and alignment length > 50 bp to the respective

genomes were counted as a hit.

Sequence Analysis and SNP Calling

Raw data processing and SNP identification was performed using both de novo

and reference-based approaches. Common software capable of calling SNPs de novo

used in this research were TASSEL/3.0 UNEAK (Glaubitz et al. 2014), Stacks/1.24

(Catchen et al. 2013), and GBS-SNP-CROP 1.1 (Melo, Bartaula, and Hale 2016). For

the reference based approach, pearl millet reference genome v1 (Varshney et al. 2017)

was used. The reference genome consists of seven pseudomolecules. Six different

reference based pipelines were evaluated to call SNPs viz., TASSEL 4.3 (Glaubitz et al.

2014), Stacks 1.24 (Catchen et al. 2011), GBS-SNP-CROP 1.1 (Melo, Bartaula, and

Hale 2016), SAMtools 1.2 mpileup (H. Li et al. 2009), FreeBayes 0.9.21 (Garrison and

Marth 2012), and GATK 3.3 (McKenna et al. 2010). Parameters used in each software

are provided in Table 2-1. Sequence variants called were filtered with a minimum depth

of 48 per sample.

https://phytozome.jgi.doe.gov/pz/portal.html

45

Linkage Map Construction

QC-filtered SNPs were further filtered by the following standards for map

construction: 1) markers must be genotyped in at least 180 individuals; 2) Individuals

with over 10% missing data were discarded; and 3) Redundant markers were removed

by standard of similarity = 1. For each parental map construction, only single dose

markers were used. Markers segregating at a distorted Mendelian ratio (expected ratio

for ‘lmxll’ type and ‘nnxnp’ type is 1: 1, χ2 test, 0.001 < P < 0.05) were marked. The

single dose markers from the maternal and paternal parent were analyzed separately

using JoinMap 4.1 (van Ooijen 2006) and outcross pollinated family (CP) was selected

as the population type. Markers that were heterozygous in N122 and homozygous in

N190 (‘lmxll’ type) were selected to build N122 linkage groups. Markers that were

heterozygous in N190 and homozygous in N122 (‘nnxnp’ type) were selected to build

N190 linkage groups. The linkage groups were built using regression mapping

algorithm, with a minimum logarithm of odds (LOD) value at 20, and a maximum

recombination frequency at 0.40. Marker positioning calculation was performed with a

goodness-of-fit jump at 5, followed by a “ripple” procedure (value =1). Map distances

were estimated using the Kosambi mapping function. Genetic distance between SDR

markers were corrected using DistortedMap (S. Q. Xie, Feng, and Zhang 2014).

Linkage maps were drawn with MapChart (Voorrips 2002). For integrated map

construction, markers that were heterozygous in both parents (‘hkxhk’ type) were

selected to build combined linkage groups. Markers segregating at distorted Mendelian

ratio (expected ratio for ‘hkxhk’ type is 1:2:1, χ2 test, 0.001< P < 0.05) were marked. The

retained markers were then added with the markers from male and female parents to

construct a combined map. The linkage groups were built using regression mapping

46

algorithm, with a minimum logarithm of odds (LOD) value at 20, and a maximum

recombination frequency of 0.40. Other parameters were the same with linkage map

construction above. Regions showing segregation distortion (0.001< P <0.05) with more

than three adjacent loci were marked as SDR regions (Paillard et al. 2003; Z. Zhang et

al. 2016).

Comparison Between Napiergrass and Pearl Millet Genome

Consensus sequence of mapped markers from TASSEL de-novo UNEAK were

used to compare with the reference genome of pearl millet with same parameters

(BLASTN defaults with an e-value cutoff of 1x10-8). Markers that showed significant hits

to the genome sequence and /or gene models of pearl millet with > 80% identity and

alignment length > 50 bp were extracted and used for comparative genomics study. A

circos plot was drawn using circos/0.69-2 (Krzywinski et al. 2009).

Results

Napiergrass Genome Survey

Approximately 211 million raw reads from Illumina and 97 thousand raw reads

from 454 sequencing were subjected to a sequence quality check. After filtration and

trimming, 161 million clean Illumina reads and 96,000 clean 454 reads were aligned to

the pearl millet genome v1 (Varshney et al. 2017). A total of 62.5 million (38.8%) reads

were able to align with the pearl millet genome. Polymorphisms were detected between

the napiergrass and pearl millet aligned reads, of which 619,708 SNPs and 24,135

indels were identified. Most of the sequence variations (58.7% SNP) were in intergenic

regions (Figure 2-1). The clean reads were assembled into 113,789 contigs with a total

size of 44.5 Mbp and a N50 of 435 bp and a GC content 43.45%. The largest 10 contigs

of the sequence assembly, which ranged from 8,506 to 25,329 bp in length, were

47

selected as representative napiergrass genome fragments. The repeat content of the 10

longest contigs ranged from 5% to 90% with an average of 27.5% and a total of 164

repetitive elements (Table 2-2). Two contigs had no hits in the pearl millet genome due

to a high repeat content (over 80%). The rest of the contigs had one or more large hits

(>500 bp) to the pearl millet genome. The micro-synteny showed mostly collinear

relationship between the genomes of the two species (Figure 2-2). However,

chromosome inverted duplications were also observed in the pearl millet genome

(Figure 2-3), indicating that the chromosome rearrangement occurred after the

speciation of napiergrass and pearl millet. The length of stringently (>500 bp and >80%

sequence similarity) aligned regions accounted for 36.3% of the examined contig

sequences of napiergrass (Table 2-3). The total length of the alignment was 25.1%

higher in pearl millet than in napiergrass aligned regions.

From the assembled napiergrass survey sequences, 5,339 SSRs were identified.

Mono- type repeats were most common in napiergrass, followed by Tri-, Di- and Tetra-

type repeats (Table 2-4). From these identified SSRs, 1,926 were successfully used for

primer design (Table 2-5). All of the primer sequences aligned to the assembly of

napiergrass and 89% of the primers were uniquely aligned. On the other hand, the

overall alignment rate of the primers with pearl millet genome v1 (Varshney et al. 2017)

was 31% with 15% uniquely aligned. These SSR primers will undoubtedly serve as an

abundant resource for molecular markers in napiergrass.

Genotyping-by-sequencing

To construct the linkage map for napiergrass, an F1 bi-parental population was

developed, which consisted of 185 true hybrid individuals (Sinche et al. 2018). These

hybrids were subjected to GBS. A total of 549 million raw reads were generated. After

48

trimming and filtering, 512 million high quality reads were retained. The average number

of reads per sample was 2.6 million and ranged from 44 thousand to 5.4 million reads

per sample. In silico digestion of the pearl millet genome v1 (Varshney et al. 2017) with

PstI yielded DNA fragments in the range of 170-350 bp, which suggest that an

estimated average depth for the mapping population was 36.5 reads per locus per

sample (Figure 2-4), which should allow us to call the SNPs confidently at most of the

loci.

A total of 695,602 unique tags were identified from the clean reads generated

from the mapping population by using the TASSEL de-novo UNEAK pipeline. These

tags were further clustered into 182,934 non-redundant tags by CD-HIT. To examine

the sequence similarity between napiergrass and other grass species, we aligned the

non-redundant tags of napiergrass against several grass species with complete genome

sequences including rice (Oryza sativa) (Osativa_323_v7.0), Brachypodium

(Bdistachyon_314_v3.0), maize (Zmays_284_AGPv3), sorghum (Sbicolor_313_v3.0),

foxtail millet (Sitalica_312_v2), switchgrass (Pvirgatum_273_v1.0), wheat

(Taestivum_296_v2), pearl millet v1(Varshney et al. 2017) , and barley (ASM32608v1),

with Arabidopsis (Athaliana_167_TAIR9) as an outgroup control. The results showed

that the percentage of napiergrass sequence tags aligned to these grass species

ranged from 2.6% to 37.9% for barley and pearl millet genome, respectively (Table 2-6),

indicating a relatively close relationship between napiergrass and pearl millet.

49

SNP Calling by Various SNP Callers

Three de-novo SNP calling pipelines, TASSEL-UNEAK, Stacks, and GBS-SNP-

CROP identified 10,799, 6,871, and 4,521 SNPs, respectively. Reference based

pipelines were also applied by using pearl millet v1 (Varshney et al. 2017) as the

reference genome. However, the alignment rate was relatively low due to the

differences between the napiergrass and pearl millet genomes. The percentage of clean

reads aligned to the pearl millet genome using Bowtie2 ranged from 5.60% to 44.62%

with an average of 39.68%. Two samples had a small number of sequences (< 10% of

the average number of sequences per sample) and also the lowest percentage of

uniquely mapped reads (Table 2-7, Figure 2-5). Therefore, these samples were

removed from linkage map construction. Six different reference-based pipelines were

employed to call SNPs viz., TASSEL 4.3 (Glaubitz et al. 2014), Stacks 1.24 (Catchen et

al. 2011), GBS-SNP-CROP (Melo, Bartaula, and Hale 2016), SAMtools 1.2 mpileup (H.

Li et al. 2009), FreeBayes 0.9.21 (Garrison and Marth 2012), and GATK 3.3 (McKenna

et al. 2010). TASSEL 4.3, Stacks, and SAMtools identified 7,326, 4,920, 27,082 SNPs,

respectively in the mapping population, whereas FreeBayes, GBS-SNP-CROP, and

GATK that can handle ploidy identified 25,193, 2,906 and 197,475 SNPs, respectively.

The six reference-based SNP callers concordantly called only 11 SNPs (Figure 2-6,

only five programs are shown in figure due to Venn-diagram display limitations) and

207,391 non-redundant SNPs.

Genetic Linkage Map Construction

From a total of 549,944 SNPs called by both reference based and de-novo

pipelines, 287,093 SNPs were filtered for further analysis. Out of these, a total of 18,286

single-dose SNPs were genotyped in more than 180 progenies. Three individuals with

50

more than 10% missing sites were removed from further analysis. For linkage map

construction of each parental line, only the SNPs showing heterozygous in one parent

but homozygous in the other parent were selected. A total of 3,276 loci were

heterozygous in female parent but homozygous in male parent and segregated with an

expected ratio of 1:1 in the population, thus can be used for female parent linkage map

construction. Similarly, 3,417 loci were heterozygous in male parent but homozygous in

female parent and segregated with an expected ratio of 1:1 in the population, thus can

be used for male parent linkage map construction. For the female parental line, a total

of 1,606 SNPs were grouped and 899 loci were mapped on 14 linkage groups with a

total length of 1,555.17 cM averaging 1 marker every 1.72 cM (Figure 2-7). Inclusion of

segregation distorted (SD) markers increased the genetic distance of the female map by

28.13%. For the male parent, a total of 1,509 markers were grouped into 14 linkage

groups and 1,073 markers were mapped onto these 14 linkage groups with a total

length of 1,939.19 cM averaging 1 marker every 1.80 cM (Figure 2-8). Inclusion of SD

markers increased the total genetic distance of the male map by 38.41%.

A combined linkage map containing markers that segregated from both female

and male parents was constructed, which can facilitate future QTL mapping of the

population. To construct a combined linkage map, the markers showing heterozygous

on both parents in addition to male-parent heterozygous and female-parent

heterozygous markers were used. Therefore, a parent-averaged combined map was

constructed by using 378 heterozygous markers for both parents that segregated in a

1:2:1 ratio in the population, in combination with 3,417 male-parent heterozygous and

3,276 female-parent heterozygous markers. In total, 4,058 markers were grouped into

51

14 linkage groups out of which 1,913 markers were mapped. The final composite

linkage map spanned a length of 1,410.10 cM with an average of 0.73 cM between

markers. The largest linkage group was Linkage group 2 (LG 2), which spanned 142.40

cM and contained 170 markers (Table 2-8). Length of each linkage group ranged from

70.18 cM to 142.40 cM and density ranged from 0.88 to 1.77 markers per cM (Figure 2-

9, Table 2-8, Figure 2-10). Results of the χ2 test indicated that 114 (6.06%) of the 1,879

markers showed significant segregation distortion (0.001<P<0.05) on the combined

map. These distorted markers showed clustered distribution on three segregation

distortion regions (SDRs) in linkage groups LG07 and LG08 (Fig 2-9).

Among the different reference-based SNP callers, GATK called the highest

number of SNPs (197,475) followed by SAMtools and FreeBayes (Table 2-9). After

accounting for segregation ratio and missing data, SAMtools retained the largest

number of SNPs followed by TASSEL de-novo UNEAK. However, when considering the

total number of markers mapped on the combined linkage groups, TASSEL de-novo

UNEAK showed the highest percentage of SNPs mapped followed by Stacks (Table 2-

9).

Comparison Between Genomes of Napiergrass and Pearl Millet

Sequence tags of the markers that mapped on napiergrass linkage groups were

extracted and compared to the pearl millet genome. Among the 1,156 TASSEL de-novo

UNEAK tags positioned on the combined map, 663 were found to have significant

sequence similarities to the genome sequence of pearl millet. Considerable collinearity

was observed between the napiergrass and pearl millet genomes (Figure 2-11). For

each pearl millet pseudomolecule, two corresponding regions in the linkage groups

(LGs) of napiergrass genome were identified (Figure 2-11, Figure 2-12). However, some

52

pearl millet genomic regions had more than two corresponding regions on napiergrass

genome. For example, pseudomolecule 3 of pearl millet had regions corresponding to

three linkage groups LG03, LG12, and LG14 of napiergrass indicating possible

chromosomal rearrangement between the two species after speciation (Figure 2-11,

Figure 2-12).

Discussion

Despite its importance as a forage grass and its enormous potential as a biofuel

crop, molecular, genetic, and genomic studies have been severely limited in

napiergrass. Currently, there was no equivalent genome sequence in the public domain

to be used as a reference for napiergrass. In this study, an initial comparison between

the napiergrass survey sequences to 10 available grass genomes revealed that

napiergrass genomic sequences had the highest similarity with the pearl millet genome,

which could be explained by the presence of the A’A’ genome of napiergrass that is

homologous to the AA genome of pearl millet. Consequently, in this study we have

utilized pearl millet genome v1 (Varshney et al. 2017) as a reference for SNP calling

and also performed de-novo SNP calling without a reference genome. A total of 38.8%

of the napiergrass reads aligned to the pearl millet genome using Bowtie 2, which

performed better over BWA, another popular aligner (Langmead and Salzberg 2012;

Yang, Song, et al. 2017). The large portion of unaligned reads might be from the B

genome or the divergent chromosome regions of A genome between the two species.

From the genome survey comparison, the total length of all the alignments of

napiergrass reads was 25.1% longer in pearl millet indicating genic duplication or

expansion in pearl millet and genomic rearrangements between the two species during

evolution from their ancestral genome. This is consistent with a previously reported

53

genomic in situ hybridization, which verified that the pearl millet genome A was 24%

larger compared to the chromosomes of genome A’ of napiergrass (Reis et al. 2014).

For the 10 longest contigs in our assembly, average repeat content (27%) was lower

than reported from other grasses including sorghum (61%) (Paterson et al. 2009),

maize (85%) (Schnable et al. 2009), foxtail millet (46%) (G. Zhang et al. 2012), rice

(43.3%) (J. Yu et al. 2002), and pearl millet (77%) (Varshney et al. 2017). Low repeat

content in napiergrass could be attributed to the loss of genomic sequences after

hybridization. Rearrangements and loss of genomic sequences are common events

after hybridization (Kellis, Birren, and Lander 2004). Similar to other plant genomes,

long-terminal repeat (LTR) retrotransposons comprised the most abundant class (62.19

%) of repeats (Table 2-2). Significant relationships between napiergrass, pearl millet,

and P. squamulatum suggested their common origin and it was inferred that

napiergrass and pearl millet had concomitantly diverged from a common ancestor (Reis

et al. 2014; Martel et al. 2004; Martel et al. 1997) and the origin of napiergrass occurred

at the interspecific hybridization event, by combining genome A of the ancestor with

genome B of a still unknown second ancestor (Reis et al. 2014). Our study showed that

the napiergrass genome had considerable microcolinearity with the pearl millet genome

showing evidence of their relatedness and shared ancestry. Chromosome inverted

duplications on pseudomolecule 3 of pearl millet showed possible rearrangement after

speciation of napiergrass and pearl millet. Two corresponding regions on the

napiergrass linkage groups for each pearl millet chromosome corroborate the

hypothesis that these two genomes evolved from a common ancestor.

54

We developed a limited genomic assembly of napiergrass based on Illumina and

454 sequences. Nearly two thousand SSR markers were developed, which could be

immediately useful for applications in napiergrass breeding and genetics. With the

advancement of NGS, high throughput NGS-enabled genotyping technologies are

becoming readily accessible. Yet, SSR markers remain as a popular tool for genetic

studies, variety identification, monitoring of seed purity, and hybrid quality. They are

particularly important in laboratories which have limited resources and lack access to

NGS facilities or bioinformatic expertise. To our knowledge, this is the first study in

napiergrass where SSR markers were developed based on napiergrass genome

survey.

A genetic linkage map is an important tool to reveal the genome structure and to

identify marker-trait associations (Cai et al. 2015) which ultimately help in MAS (F. Lu et

al. 2013) to improve precision of selection. In this study, we used the GBS approach to

construct a combined high-density linkage map that spanned 1,473.9 cM with 1,917

markers on 14 linkage groups, which is a very critical tool for further genetic and

genomics studies of napiergrass. GBS has been extensively used for genotyping many

diploid organisms, however, SNP calling from the NGS data in allotetraploids like

napiergrass is particularly challenging due to existence of highly similar homeologous

copies, one corresponding to A genome and the other to B genome (Nagy et al. 2013).

Therefore, different strategies have been devised to construct linkage map in

allopolyploids. For example only uniquely aligned reads (single copy) were considered

for SNP calling and subsequent map construction (Trick et al. 2009; X. Zhou et al. 2014)

to avoid the collapsed alignment of homoeologous reads due to low divergence, recent

55

polyploidization event, and severe domestic bottlenecks (Pandey et al. 2012). SNP

calling in allotetraploid Brassica napus L. (rapeseed; 2n = 4x =38; AACC) was done by

utilizing only uniquely mapped reads (single copy) and a read depth minimum of three

to four reads at each potential SNP (Trick et al. 2009). Linkage map construction in

zoysiagrass (Zoysia matrella) was performed by utilizing single-dose markers after

calling SNPs using the maximum likelihood method in Stacks (Xiaoen Huang et al.

2016). Similarly, single dose markers from TASSEL de-novo UNEAK were used to

construct linkage maps in prairie cordgrass (Spartina pectinate)(Crawford et al. 2016).

In this study, we applied multiple SNP callers and strategies to maximize SNP

calling for linkage map construction for napiergrass. In the final combined genetic map,

the number of markers identified by different software varied dramatically. GATK called

the highest number of SNPs followed by SAMtools and FreeBayes initially. Both GATK

and SAMtools apply Bayesian method to compute the posterior probability for each

possible genotype and then choose the genotype with the highest probability as the

consensus genotype (X. Yu and Sun 2013). GBS-SNP-CROP and TASSEL showed a

low matching percentage, which is similar to results from previous research (Melo,

Bartaula, and Hale 2016). The number of useful markers for linkage group construction

was the highest in SAMtools (47.75%) followed by TASSEL de-novo UNEAK (35.68%).

However, the TASSEL de-novo UNEAK pipeline had the highest number of markers

mapped on the linkage groups (60.43%) followed by Stacks (13.43%). This indicated

that the network-based SNP discovery in TASSEL de-novo UNEAK and UStacks

pipeline (Kim et al. 2015) could be efficiently utilized for constructing linkage maps in

non-model species. Even though TASSEL was primarily designed for diploids, it is

56

powerful enough to give a large number of mapped markers compared to other

programs that handle polyploidy like FreeBayes, GATK, or GBS-SNP-CROP.

The SNP markers were relatively evenly distributed among the linkage groups

with more than 97.45% of marker interval being less than 5 cM. To our knowledge, this

linkage map with an average inter-marker distance of 0.7 cM is the first genetic linkage

map of napiergrass to date. A study based on an interspecific population of a cross

between pearl millet and napiergrass has been previously reported to link RAPD

markers with biomass related traits in Pennisetum (Smith et al. 1993). The large number

of markers and their even distribution in our study facilitate full-scale map coverage.

Few regions where the interval space was > 5 cM might be due to stretches of large

repeats or due to low coverage sequencing of GBS (Poland and Rife 2012; Mathew et

al. 2014). Segregation distortion is regarded as a potential evolutionary force and

including these markers for linkage map construction could increase genome coverage

as well as benefit QTL mapping (S. Xu 2008; D. R. Taylor and Ingvarsson 2003).

Including SDR markers and correcting for bias led to an increase in genetic distance

between distorted markers (S. Q. Xie, Feng, and Zhang 2014). The deviation from

expected Mendelian ratio shows disturbances in the transmission of genetic information

from one generation to the next and can be caused by chromosome loss or

rearrangements, genetic load, gametic selection, zygotic selection, or both (Faris,

Laddomada, and Gill 1998; Karkkainen, Koski, and Savolainen 1996; Bodénès et al.

2016). Napiergrass generally outcross through wind pollination that could result in high

levels of gene flow leading to genetic load. The assignment of napiergrass linkage

57

groups according to the pearl millet genome allows for future fine mapping and QTL

analysis.

In summary, this study reports for the first time a high-density genetic linkage

map using NGS-derived SNP markers, as well as the development of SSRs from

napiergrass genomic sequences. The napiergrass genome showed considerable

collinearity with the pearl millet genome and the genetic map contains 14 linkage groups

with low inter-marker interval. The results will be useful for future molecular breeding

programs such as identification of QTLs for important traits as well as MAS for the

genetic improvement of napiergrass and comparative genomics. These resources will

play a critical role in future whole genome sequencing projects and leveraging

molecular breeding of napiergrass.

58

Table 2-1. Parameters used for SNP calling for each software.

Reference based Parameters Remarks. [defaults]

TASSEL 4.3 -c 5 Min. number of times a tag must be present to be output <5> [1]

-mnMAF 0.01 Min. minor allele frequency <0.01>[0.01]

-mnMAC 100000 Min. minor allele count <100000>[10] (SNPs that pass either -mnMAF or -mnMAC will be output)

-misMat 2 Threshold genotypic mismatch rate above which the duplicate SNPs won’t be merged <2>[0.05]

-callHets When two genotypes at a replicate SNP disagree for a taxon, call it a heterozygote

Stacks -A CP CP type for genetic map

-m 3 Min. number of identical, raw reads required to create a stack <3>[3]

GBS-SNP-CROP -l 30

-sl 4:30

-tr 30

-m 32

Trimmomatic LEADING parameter

Trimmomatic SLIDINGWINDOW parameter

Trimmomatic TRAILING parameter

Trimmomatic MINLEN parameter

-rl 100

-pl 32

-p 0.01

-id 0.93

Raw GBS read length

Min. length required after merging to retain read

p-value for PEAR

Nucleotide identity value required for USEARCH read clustering

-Q 30

-q 0

-f 0

-F 2308

Phred score base call quality

Alignment quality

SAMtools flags

SAMtools flags

-mnHoDepth0 11 Min. depth required for calling a homozygote when the alternative allele depth = 0

-mnHoDepth1 48 Min. depth required for each allele when calling a heterozygote

-mnHetDepth 3 Min. depth required for each allele when calling a heterozygote

59

Table 2-1. Continued

Reference based Parameters Remarks. [defaults]

-altStrength 0.9 Across the pop. For a given putative bi-allelic SNP, this alternate allele strength is the minimum proportion of non-primary allele reads that are the secondary allele

-mnAlleleRatio 0.1 Min. required ratio of less frequent allele depth to more frequent allele depth

-mnCal 0.75 Min. acceptable proportion of genotyped individuals to retain a SNP

-mnAvgDepth 4 Min. avg. depth of an acceptable SNP

-mxAvgDepth 200 Max avg. depth of an acceptable SNP

SAMtools mpileup -uf Default

FreeBayes -C 2

--min-alternate-count Require at least <2> observations supporting an alternate allele within a single individual in order to evaluate the position [1]

-p 4 --ploidy <4> [2]

--use-best-n-alleles 4

Evaluate only the best N SNP alleles ranked by sum of supporting quality scores [all]

--min-coverage 5 Require at least <5> coverage to process a site [0]

GATK -T UnifiedGenotyper

Call SNPs and indels on a per-locus basis

-stand_call_conf 30

The min. phred-scaled confidence thresholds at which variants should be called <30> [30]

-stand_emit_conf 10

The minimum phred-scale confidence threshold at which variants should be emitted (and filtered with LowQual if less than the calling threshold) <10> [30]

-ploidy 4 Ploidy <4> [2]

-mbq 20 Minimum base quality required to consider a base for calling

-glm BOTH Genotype likelihoods calculation model <BOTH> includes SNPs and INDELs [SNP]

60


de-novo based Parameters Remarks. Defaults []

UNEAK -e PstI Restriction enzyme used

-c 5 Min. count of a tag must be present to be output [5]

-e 0.03 Error tolerance rate in the network filter [0.03]

-mnMAF 0.05 Min. minor allele frequency [0.05]

-mxMAF 0.5 Max. minor allele frequency [0.5]

-mnC 0 Min. call rate (proportion that how many taxa are covered by at least one tag)

-mxC 1 Max. call rate [1]

Stacks -m 3 Min. number of identical, raw reads to create a stack

-M 2 No. of mismatches allowed between loci when processing a single individual [2]

-n 1 No. of mismatches allowed between loci when building the catalog [1]

-t Remove, or break up, highly repetitive RAD-Tags in the ustacks program

GBS-SNP-CROP Same as reference based except, script 3 settings not required

61

Table 2-2. Repetitive elements present in the napiergrass genome.

Transposable Element Count

DNA transposon

Tc1/Mariner 8

hAT 20

PIF/Harbinger 16

EnSpm 1

CACTA 4

Polinton 1

LTR Retrotransposon

LTR 37

LTR/Copia 5

LTR/Gypsy 5

Copia 4

Retrotransposon 3

Retroelement 1

Gypsy 2

Non-LTR Retrotransposons

L1 2

Pseudogene

tRNA 1

rRNA 8

rDNA-like 3

Others

Mutator 15

Telomeric 4

MobileElement 2

Micro-like sequence 1

Low-complexity 3

Simple Repeats 12

Unspecified 6

Total 164

62

Table 2-3. The sequence alignment of ten napiergrass sequence contigs to the pearl millet genome.

No.

Napiergrass contig

Size of contig (bp)

GC content (%)

Repeat content (%)

Corresponding pearl millet pseudomolecule

Number of hits above 500bp

Number of predicted genes

Alignment length in napiergrass

Alignment length in pearl millet

Sum of expanded length in pearl millet

1 Contig1 8,506 55.3 80.3 Pg5 0 0 / / /

2 Contig3434 9,653 42.5 9.8 Pg3 5 0 5,407 5,384 -23

3 Contig5516 11,920 44.9 17.5 Pg3 11 1 6,776 13,509 6733

4 Contig5578 8,558 42.8 5.9 Pg2 3 0 5,948 6,487 539

5 Contig5588 8,595 44.6 14 C26927002 1 1 591 591 0

6 Contig5729 9,088 46 20.4 Pg5 3 0 2,250 2,246 -4

7 Contig5798 8,651 43.1 5.3 C27370090 1 0 1,385 1,385 0

8 Contig5878 25,329 43.4 12.8 Pg6 4 0 5,692 5,676 -16

9 Contig5902 13,801 46.1 16.4 Pg6 8 0 6,209 7,596 1387

10 Contig6139 14,330 42.2 93.1 Pg7 0 0 / / /

Total 10 118,431 45.09 27.5 8 36 2 34,258 42,874 8,616

63

Table 2-4. Frequency of classified repeat types (considering sequence complementary) in napiergrass.

Repeats 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 ≥20 Total

A/T - - - - - 1045 338 173 76 49 34 17 9 10 3 12 1766

C/G - - - - - 116 71 36 26 11 7 7 6 2 1 283

AC/GT - 118 43 27 18 14 6 2 1 1 1 2 1 9 243

AG/CT - 250 137 82 45 31 19 14 13 9 6 3 5 3 2 6 625

AT/AT - 180 81 55 40 17 4 5 4 8 5 1 4 12 416

CG/CG - 30 10 1 41

AAC/GTT 45 15 14 6 1 2 1 84

AAG/CTT 260 85 46 27 13 13 8 7 12 3 2 476

AAT/ATT 62 31 6 4 5 2 1 2 1 1 2 117

ACC/GGT 86 35 15 7 3 1 2 1 1 151

ACG/CGT 45 3 7 1 56

ACT/AGT 35 12 4 2 1 54

AGC/CTG 131 53 30 15 4 3 236

AGG/CCT 129 34 10 5 1 2 1 182

ATC/ATG 86 38 13 13 4 3 3 3 1 1 1 166

CCG/CGG 157 39 18 3 1 1 219

AAAC/GTTT 2 2

AAAG/CTTT 14 5 3 2 2 26

AAAT/ATTT 19 1 2 1 1 2 26

AACC/GGTT 3 1 4

AAGC/CTTG 1 1

AAGG/CCTT 1 1

AATC/ATTG 1 1

AATG/ATTC 2 1 1 4

AATT/AATT 1 1

ACAG/CTGT 5 1 1 7

ACAT/ATGT 6 2 1 1 1 5 1 1 4 22

ACCC/GGGT 1 1

ACCG/CGGT 1 1

64


Repeats 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 ≥20 Total

ACCT/AGGT 1 1

ACGC/CGTG 2 2

ACGG/CCGT 1 1

ACTC/AGTG 2 2

ACTG/AGTC 1 1

AGAT/ATCT 7 2 2 1 1 13

AGCC/CTGG 1 1 2

AGCG/CGCT 1 1

AGCT/AGCT 1 1

AGGC/CCTG 1 1 2

AGGG/CCCT 2 2

ATCC/ATGG 6 1 1 1 9

ATGC/ATGC 3 3

CCGG/CCGG 2 2

AAAAC/GTTTT 2 2

AAAAG/CTTTT 1 1 2

AAAAT/ATTTT 1 1

AAACC/GGTTT 1 1

AAAGG/CCTTT 1 1 2

AAAGT/ACTTT 1 1

AAATC/ATTTG 1 1

AACAG/CTGTT 1 1

AACAT/ATGTT 1 1

AACCT/AGGTT 1 1

AACTG/AGTTC 1 1

AAGAG/CTCTT 1 1

AAGAT/ATCTT 1 1

AAGCT/AGCTT 1 1

AAGGG/CCCTT 1 1

65


Repeats 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 ≥20 Total

AATAC/ATTGT 1 1

AATAG/ATTCT 2 1 1 4

AATCC/ATTGG 1 1

AATCT/AGATT 1 1 2

AATGG/ATTCC 1 1

AATGT/ACATT 1 1

ACACC/GGTGT 1 1 2

ACACG/CGTGT 2 2

ACCAG/CTGGT 1 1

ACCGC/CGGTG 1 1

ACGGC/CCGTG 1 1

ACGTC/ACGTG 1 1

ACTAT/AGTAT 1 1 2

ACTCC/AGTGG 2 2

AGAGG/CCTCT 2 2

AGATG/ATCTC 1 1 2

AGCCG/CGGCT 2 2

AGCCT/AGGCT 1 1

AGGCC/CCTGG 1 1

ATCCC/ATGGG 1 1

AAAAAG/CTTTTT 1 1

AAATCG/ATTTCG 1 1

AACCCG/CGGGTT 1 1

AACTCC/AGTTGG 1 1

AAGAGC/CTCTTG 1 1

AAGAGG/CCTCTT 2 2

AAGATG/ATCTTC 2 1 2 5

AAGCAC/CTTGTG 1 1

AAGGAG/CCTTCT 3 3

66


Repeats 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 ≥20 Total

AAGTAG/ACTTCT 1 1

AATAGG/ATTCCT 1 1

AATAGT/ACTATT 1 1

AATCCC/ATTGGG 1 1

AATCGG/ATTCCG 1 1

ACCAGC/CTGGTG 1 1

ACCATC/ATGGTG 1 1 2

ACCCCG/CGGGGT 1 1

ACCGTG/ACGGTC 1 1

ACGGGG/CCCCGT 1 1

ACTGCT/AGCAGT 1 1 2

AGATAT/ATATCT 1 1

AGATGG/ATCTCC 1 1

AGCAGG/CCTGCT 1 1

AGCATC/ATGCTG 1 1

AGCATG/ATGCTC 1 1

AGCCGG/CCGGCT 1 1

AGCCTG/AGGCTC 1 1

Total 1172 953 451 252 142 1263 459 243 136 84 56 31 26 16 6 49 5339

67

Table 2-5. Primer pairs developed for napiergrass SSR markers. Available at https://www.nature.com/articles/s41598-018-32674-x#Sec18

https://www.nature.com/articles/s41598-018-32674-x#Sec18

68

Table 2-6. Summary of the alignment of non-redundant tags of napiergrass (Cenchrus purpureus) to the available genomes of different species.

Genome used (species name) Number of tags

with blast hits

Percentage of tags

with blast hits (%)

Arabidopsis (Arabidopsis thaliana) 120 0.07

Purple false brome (Brachypodium

distachyon) 6,029 3.30

Barley (Hordeum vulgare) 4,751 2.60

Rice (Oryza sativa) 6,879 3.76

Pearl millet (Pennisetum glaucum) 69,385 37.93

Switchgrass (Panicum virgatum) 24,163 13.21

Sorghum (Sorghum bicolor) 14,654 8.01

Foxtail millet (Setaria italica) 40,849 22.33

Wheat (Triticum aestivum) 6,459 3.53

Maize (Zea mays) 11,972 6.54

69

Table 2-7. Alignment of individual napiergrass reads using Bowtie2. Number of raw reads Clean reads (retained

by Stacks) % of reads retained

Total reads mapped to pearl millet

Uniquely mapped % uniquely mapped

Min. 44,858 20,350 45.37% 16,658 1,140 6.00%

Max. 5,046,114 4,739,605 97.53% 3,022,185 1,902,058 45.00%

Avg. 2,893,836.07 2,696,528.11 92.71% 1,727,194.04 1,077,802.96 39.68%

Total 549,828,854 512,340,341 93.18% 328,166,867 204,782,562 39.97%

70

Table 2-8. Summary of the combined linkage map of napiergrass and the percentage of gaps less than 5 cM in male and female parent linkage maps.

Napier grass linkage group

Pearl millet syntenic pseudomolecule

Number of grouped markers

Mapped markers

Unmapped markers

Length (cM)

Density (markers per cM)

Combined map Gaps <=5 cM (%)

Female map Gaps <= 5cM (%)

Male map Gaps <=5 cM (%)

LG01 PM01 411 163 248 109.33 1.49 98.15 93.98 93.75

LG02 PM06 378 170 208 142.40 1.19 97.63 89.19 92.45

LG03 PM03 360 156 204 89.81 1.74 98.06 96.97 85.71

LG04 PM05 339 127 212 74.83 1.70 99.21 95.51 85.53

LG05 PM02 324 182 142 105.50 1.73 98.34 96.55 96.36

LG06 PM04 300 129 171 112.53 1.15 96.09 96.23 96.83

LG07 PM07 279 99 180 96.84 1.02 96.94 89.8 84.72

LG08 PM06 279 120 159 98.45 1.22 97.48 84 89.86

LG09 PM02 278 96 182 108.70 0.88 92.63 94.64 95.24

LG10 PM01 254 181 73 102.49 1.77 98.89 92.65 75.76

LG11 PM07 237 141 96 97.35 1.45 98.57 91.53 88.06

LG12 PM03 230 146 84 105.72 1.38 97.24 89.29 90.11

LG13 PM05 224 101 123 70.18 1.44 98.00 83.67 94.12

LG14 PM03 165 102 63 96.00 1.06 97.03 89.09 90.91

Total (average) 4,058 1,913 2,145 1,410.10 (1.37) (97.45) (91.65) (89.96)

71

Table 2-9. Summary of napiergrass single nucleotide polymorphism (SNP) markers mapped on the combined linkage map using 9 different software pipelines.

Software Number of

SNPs called

Total SNPs

used for map

construction

No. of SNPs on

map

Percentage of

SNPs on the

map (%)

FreeBayes 25,193 6 0 0.00

GATK 197,475 52 5 0.26

SAMtools 27,082 3,377 151 7.89

GBS-SNP-CROP 2,906 115 52 2.72

TASSEL 7,326 116 56 2.93

Stacks 4,920 447 257 13.43

GBS-SNP-CROP

de-novo

4,521 96 51 2.67

Stacks de-novo 6,871 339 185 9.67

TASSEL de-novo

UNEAK

10,799 2,523 1,156 60.43

Total 287,093 7,071 1,913

72

Figure 2-1. Sequence variation for SNPs called in various regions of the pearl millet genome.

0%

10%

20%

30%

40%

50%

60%

70%Intergenic

Up

5' UTR

Exon

Donor

Intron

Acceptor

Exon

3' UTR

Down

Intergenic

73

Figure 2-2. Micro-collinearity between contigs from napiergrass to the pearl millet genome.

74

Figure 2-3. Inversion duplication between napiergrass and pearl millet (shown in bottom figure).

75

Figure 2-4. Estimated coverage of PstI restriction sites in the pearl millet genome.

0

5

10

15

20

25

30

35

40

0 5

10

15

20

25

30

35

40

45

50

55

60

65

70

75

80

85

90

95

10

0

Mo

re

Num

ber

of

sam

ple

s

Estimated coverage (X)

76

Figure 2-5. Histogram of uniquely mapped reads to the pearl millet genome.

0

10

20

30

40

50

60

70

80

90

100

<6 // 34 36 38 40 42 44 46 48

Num

ber

of

sam

ple

s

Percentage of uniquely mapped reads (%)

77

Figure 2-6. Venn diagram showing concordant napiergrass SNPs called by five reference-based SNP callers, SAMtools, GBS-SNP-CROP, GATK, FreeBayes, and TASSEL. Numbers in parenthesis after the program name shows the total number of SNPs called by each program.

78

Figure 2-7. Genetic linkage map of the napiergrass female parent N190.

Napiergrass: Female Parent Linkage Map

dT20387dT41575dT1702 dT3027dT909C1_10467974dT38795S3_64306230dT3328dT49252dT30456dT16886dT53891dT16764dT35147dT49010dT2473dT1951dT16527C1_220011116dT50264dT29470dT32431dT9029dT50826dT47548dT45924 dT7985dT24786 dT732C1_102656250dT13338dT19473dT18485dT46948dT51201dT23838dT23364dT23237dT28498dT51779dT43458dT6993dT29597dT5196dT14180dT48800dT3431dT48497S1_17257463dT9167dT28049dT48243S1_17257559S1_17257551dT18424dT973dT8459dT33654dT48112dT53672dT12272dC_359278C1_133709193dC_339765S1_171369178dT40041dT18064dT971dT22725dT10843 dT19448dT47057C1_234266014TS1_273230575dT30844dT7314dT766dT36979S1_217117736dT30297dT3813S1_241274562

1

S6_80432248

S3_200493878

dT46838

dT13195

dT35970

dT7946dT15344dT3773dT22636dT9283dT40427

dT34184dT17674dT14387dT1832

dT11200

S6_199487878

S6_197918666

dT40470dT31571dT45101dT50977dT53887dT50932dT25232dT32392dT44224dT42117dC_193114dT18155dT16658dT51346dT35381dT10670dT15893 dC_23232dT47482 dT35323dT16487dT16409dT4135dT35318dT16616dT16875dT17908dT35582dT15707dT49861dT50319dT22787dT4190 C7_82320373dT49133dT32996dT15879dT8813S6_398797

dT29058

dT46613

dT7678

dT42236

dT1466

dT40376

dT42316

dT53493

dT3930

dT159dT590dT6548dT41013dT38488dT6810

dT50779

dT17665

2

dT28623dT35743

dT30402

dT53129C6_237238320dT7257

dT42723

dT27350

dT165S3_58228915dT17556dT24472dT11506dT38048dT7396

dT22681dT44178dT17379dT19812

S3_234961343

dT12951dT19939

S3_65392670

dT35908

dT25817

dT6688dC_42042dT2147S3_68670889dC_28051 dT5016dT22363S3_68670908dT33771dT51194dT13022

dT40851

dT11544dT42647dT42768dT50758dT6927S3_92611362dT37158dT12246dT31485dT18863dT18489dT35462dT19043dT11234dT26079dT16041dT13169dT25285dT18211dT49009dT28196dT6590dT50718dT22708dT27517S3_35602619dT3175dT35742dT43871dT2820dT492dT1427dT12717dT32542dT42678dT18556dT14307C3_245369357dT43387dT43702TS3_205733261dT24042dT52226dT48403dT15989dT23856dT42771dT207dT21277dT38486C3_202051529dT8152dT15649dT15821dT22318dT31754dT3519dT25849dT35984dT23371dT25577S3_41221675

3

S1_6210777

S5_128073888

S1_6210761

dT24515

dT17924TS5_114098156dT16773dT40995dT15817dT14651dT18480TS5_128053793 dT13827C5_156998414 C5_112498054dC_24267dT27345 dT30788dT9315S5_40839446dT4737dT31217dT32027dT46427dT32695dT43636dT31735

dT50145

dT43048C5_132196063dT20788dT32207dT24792

dT46652

dT27717dT15391dT1524dT14375dT31410dT31722

dT18219dT36949

dT24276

dT7856

dT5489

dT43892

dT48559

dT42534

dT49410dT52149

dT32206

dT6664

dT8114dT53440dT40421dT14806dT22647S5_31904873dT39799dT3053S5_26740824dT4924dT37762

dT26721dT26387dT6733dT17794S5_16898738dT37016dT47995dT34310dT51527TS5_13509703dT54119dT53202dT32125dT27794dT2435dT46426dT20384C5_3132406dT15117dT29259dT780dT15952dT26168dT36121dT32758dT30470

4

S2_16737601S2_16737644

dT5178S2_174624948dT43720dT28282dT38918S2_24418753dT12948S2_24418742dT42842dT19311dT38665dT47411dT3578dT50366dT33944dT1873dT14669dT8327dT31449dT49212 dT13764C2_21121892S2_110636956dT16449dT36036dT29465dT49950S2_76988086TS2_76988086S2_76982335dT5796dT50180dT16824dT17391dT20397dT50336dT27602dT19647dT23469dT10549dT4979dT53077dT15063dT348dT34541dT45884S2_192440020dC_212370dT51287dT10378dT46228dT11302dT31033dT45970dT2290dT3272dT10652dT44368dT9233dC_167176dT3180dT44853dT25838dC_64162dT51942dT30560dT27325dC_88689dT48120S2_246378092dT32534dT6193dT20949dT12622dT50361dT31503dT31430dT43605dT13489dT33550dT10926dT37626dT37138dT37696dT2179

5

S4_43915384S4_43915389

dT20158

dT12265

dT21313dT16746

S4_12461988

S4_88888650

S4_17912300

dT49047

dT52847

dT12456dT18515dT9337dT35551dT37277

dT1848dT23823dT49039

dT17535 dT23182S4_49830629dC_23097 S4_53822034C4_54042612 dC_15219dT20953

S4_14538257

dT33655dT18867dC_45605dT19245dT3906dT4427 dC_60915dC_107792dT51291 TS4_14538257dC_10879C4_99397060 dC_90916dC_31636dT51677dT2317TS4_171269213dT30722dT29654dT12209TS4_171269280C4_194142272 C4_194142277dT8377dT9213

6

S7_182881700

dT30583S7_112269G7_212623362dT46230dT910dT26620dT43326dT27451dT50763dC_388382dT50419dT24186dT24408TS7_180945596dT46209

S7_133861267dT8557dT14658

S7_100645462

dT37501

S7_90744125dT28918S7_201830035dT10131dT429S7_32127059dT33050S7_117332098dT36791dT7088dT526dT30134dT35232dT45831 dT53184TS7_167047574dT52369S7_102687604 C7_89265590C4_41131304 dC_221873dT40952dT3713dT22267dT12440dT32512S8_44730456

S7_89265546

7

dT50316

dT37150

dC_93403 dC_63791

dT15691

dT13837dT18102

dT3402dT46843dT23515

dT11877dT28906

dT44856C6_223591048dT1593dT26926

S6_209490431dT43186C6_200342230dT35845

dT9008

dT47714

S6_9770308

dT44581

dT4343

dT1085

TS6_16445661

dT47034dT34102dT6207TS6_395354dT26411dT42227C6_73259826C6_59409932 dT24125C6_59409930dT771dT1888 dT40500dT47247dT12621dT26512dT49424dT29318dT29815dT38935dT31316dT358dT45995

8

dT14190

dT44709

dT12313

dT27092

C2_22518425dT48390

dT741

dT38026

dT21645dT33261dT15678dT3554dC_183703dT47400dT21265dT41244dT11408S2_78894878S2_78894944dT36645dT35239dT8425dT30235dT52869S2_78894908dT22956dT22796dT36870C2_206875269dT40868dT28517dT48482dT45174dT7119 dT5487dT1149 TS1_248242945dT37165dT3390 dT44398dT22957dT8998dT3928dT24570dT1352dT2536dT7910

dT17835dT44636dT32308dT17639

dT35127dT6817dT15991dT48401

dT48669

9

dT29619

dT39834

dT42418

dT47966dT47917dT5844dT3588dT29992dT7436dT20060dC_23137dC_15179C4_182430864dT30922 dT50307dT12488dT32245dT5081dT45821dT48530dT28837dT38881dT51420dT13498dT4836dT21114dT42109dT40903dC_1926dC_9578 dC_4103dC_4097dC_9584C4_136757390dT47650dT34779dT10689dT28312dT14780dT11768dT46085dT10507dT50903dT48992S1_135580638

dT47961

dT47132dT50618dT51615S2_236302464dT45124dT24369dT33254dT10533dT48984dT40084dT23737dT52173S1_268164908dT37270C8_45958102dT27051dT31375dT36311dT33782dT46112dT40810dT37914

10

dT49930S4_131830044 G4_131830067S4_131830067S4_131830066S4_131830063dT22876dT12269C4_131830077dC_61912 dC_55022dT22567dT26116dT38883dT36340 dT10040dT26884dT10739 dT621dT17325 S7_89250999dT17578dT25711dT5186dT8749dC_10270dT649dT53247dT53022 dT12194S7_39646339S7_39646309S7_39646292dT33677dT3510dT19298dT1762dT41196dT30018dT40278S7_83881022dT7491dT38871dT29364

dT8910

dC_223673

dT43632dT34571dT3675S7_213184560dT22858dT32036dT29409

dT19025dT7711

dT39297

dT42995

dT47137

dT37923

11

dT25904

dT43814

dT8647

dT51744

dT37862dT17058

dT9730dT24050dT38555dT38857dT19291 dT3341C7_118131103dC_15863dT8512C3_216897716dT37132dT28896dT47840dT5917S3_296004281dT50258dT16793dT34879dT50794

dT5877

dT28463

dT14945

12

dT6944

dT6724

dT21137

dT10917dT15762

dT51352dT12719dT5926dT8126dT36120dT18762dT3360dT50697dT26006dT15569dT37911

dT12924dT54030 dC_79210dC_79234dT1355dT43106dT46299dT32011

dT31151dT23798dT15747dT15626dT5625

dT14825

dT22667dT12193dT8566TS5_23967052S5_34147564dT51949

dT46811

dT19739

dT6729

TS5_12313000

dT10696

dT40635

S5_7812618

dT46712

dT30275

dT49612

dT22104dT50943dT46312

13

S7_177490027

dT19814

dT24731dT51987dT21395

dT34497

S3_237947843

dT11865

dT22841

dT10571

S3_255053440S3_255053465S3_257143847dT35964dT17340dC_399018S3_253954916

dT29965dT45243dT48835

dC_1015698dT1549S3_244121865dT29691dT51116dT12496dT41167dT33642S3_276917001dT12709S7_182689743S7_182689737S7_182689721dT19423dT51331dT23909dT30367dT23052dT2763dT40596dT42183dT37207dT7307 dT31619dT21724 dT39129dT8609dT6698dT30790dT34207dT32848dT40659dT18945

dT28003

S4_86603238

14

0

5

10

15

20

25

30

35

40

45

50

55

60

65

70

75

80

85

90

95

100

105

110

115

120

125

130

135

140

145

150

155

160

165

170

175

79

Figure 2-8. Genetic linkage map of the napiergrass male parent N122.

Napiergrass: Male Parent Linkage Map

S3_66915283S3_66915258dT6838dT44555dT27134S3_66915278S1_7068544dT3437dT46011dT38476dT35347dT33533dT28618dT39822dT28250C5_27253496TS7_74798545dT31263dT6267dT30050C1_119093525 C1_119093531dT37316dT12054S1_71014695dT17414dT29057S3_284635389dT2475dT18370 dT12390dT41113dT30582dC_618193dT44577dT34797dT51180dT28611dT35867dT23729dT34784dT37028S1_193475897dT36388dT3762dT38724TS1_74139971 dT11106dT6132dT7984dT13772dT53762dT8946dT18605dT29128dT52531dT41949dT47752

dT27185

S4_21013305

dT8342dT52469dT37356dT35128dT53136dT39139dT36338dT19693dT44366dT28974S1_249594415dT5535dT8792dT42596dT48305dT32706dT49414C1_251517594dC_259973S1_251483789dT41404dT42403dT9863S1_260924412dT35453dT45297dT28492dT47753dT32523dT21677dC_2450640dT48962dT50248dT27136dT32605S1_267892930

1

dT11902dT17348dT46533dT53963dT38576C3_124558897dT10674TS6_10754586dT10631dT27467dT2555dT45550dT11037S5_31045825S6_35473070dT35938C6_39611566dT6835dT16233dT38911dT38805dT6794dT14931dT16992dT46041dT16346dT47787dT42989dT50566dT9987

dT29726

dT23975dT31396dT43235

dT2439

dT39167

dT35142

dT5958dT8821

dT25121dT39882dT30721 dT19708dT36233dT18641dT28669dT15498 dC_1941dC_5486367 dC_428936dT50064 dT41460dT5222dT43295dT31359dT44933dT33834dT46138

dT987dT17187

dT3299dT29166

dT2229

dT19151dT22571dT36501dT11509dT4155dT1086dT20499

S6_118469241dT14954dT39428dT19912dT53574dT43321dT34936dT2297dT8340dC_234325dT20274dT22921dT30440dT14807dT30703dT6973dC_196800dT18761dT36412dT24998dT25533dT29143S6_55827481

dT43184C6_200342258dT9665dT18516dT18539dT13751dT50350

dT22426

dT17747

dT45130

dT29238

dT13151

dT31763

2

S3_260011343

S3_43511354

S3_248431947

dT14717

dT29788dT42777

TS3_102980429

S3_277598650dT3273G3_243475604 S3_243475604S3_243475658dT36593dT50908 C3_273261021S3_243475784dC_75778C2_168637894dT41484dT6670dT17358dT40996dT7695dC_242221dC_150274dT17019dT43835dT17380dT52592dT24195dT5635S3_267455176C3_119371432 C2_56055562S3_12254914dT2704dT24290dT35854dT34136dT29456dT1963dT6453dT53771dT52360

dT9432TS3_132668864dT15069dT40139dT41347dT723

dT38859

dT33605dT50383dT4281dT45959

dT21342dT12737

dT48862C3_67469347dT26209dT15881

S3_215627633dT53635dT36969dC_45074S3_58384211S3_58384176dT25438dT18781dT26851dT35166TS3_52460829dT44698

S3_61665185

dT20803

dT13705

dT24779

3

S5_3204810S5_3204852 S5_3204827S5_3204849S5_3204807

dT8760dT16974dT23430C5_4168325dT21831dT36067dT50761dT941dT52160dT22371dT32024dT3345dT43dT34717

S5_15050945

S5_32287078

dT35978

dT24372

C5_64002238dT8498dT19279dT5137dT24075dT19864S5_77555777dT4800S5_52320341

S5_59674699

dT22005dT28525

dT2126

dT32675dT48427dT51146dT23969

dT7381

dT12898

dT47043

dT22927dT21823dT38799 dT9509dT27949C5_65406593dT49533dT20066dC_282880

S5_142555945dT41084dT47758dT9747dT20176dT7627dT40123C5_154569850dT32999dT717dT49728dT51649dT10471dT17979dT31973dT26005dT53506dT36750dT13412dT36706dT15816dT5428S5_116583893S5_116583902

4

S2_14223558dT30464S2_16747159dT8117dT28731dT41513dT12415dT15685C2_174576603dT1897dT53530dT29388dT7310dT48623dT19877dT23648S2_30824637dT45539dT22002dT51889dT42630dT1791dT32449dT28199dT41204dT32273dT51314dT6913dT13912dT15477dT11178 dT3792C2_69640554 dT777dT8400dT34457dT3108dT13121dT4303dT1297dT6238dT43507dT34243dT53270dT9893dT27178dT24088dT3606dC_37024dT13934dT47510dT50561dT743S2_151241666C2_151270539dT52704dT39328dT45195dT2081dT44451dT31232dT34738dT52854dT11711dT35398dT33063dT35972dT27463dT24964dT20846S2_191695057S2_191695048dT49783dT49964dT32474dT53079dT52542S2_116787266dT51978dT51060S2_201762543dT33021dT1490dT51442dT26122dT35797dT35508S2_214316973dT10485dT9291dT45475S5_46066941dC_247492dT2450S2_100034039dT14670dT38863dT38020S2_226559654S2_216176645dT9060dT25565dT44700S2_176915862dT17024dT5816dT13896dT21007dT24044dT35466

5

S4_192354592

dT43261

dT3062

dT20572

dT9434

dT50501

dT13383

dT25619dT15831dT29846dT20879 dT3030dT27514dT40535 dT18538dT48489C4_200698121dT4430dT23119dT43805dT3891dC_13461S4_116088663dT13255dT32416C4_90405411dT8320S4_29656682dT42579dT4620dT1631

dT43233dT18870dT8454dT7751

dT29230dT4496dT6466dT16151TS4_49707946dT11273dT43759S4_109511769TS4_79794887dT36845dT33420dT25340dC_124707dT31777S4_17921468dT39406dT45232dT4301 C4_13275000dT24376dT36267dT26600dT52456dT10193dT34107dT22232dT9256S4_17796331

6

S2_174467255

S1_203215691C2_174467255 dC_58121dC_58157dT1325dT24339dT33835dT10459dT21208dT41386dT16455dT38126dT20154dT24498

dT50610

dT44347dT6436dT4353

dT5027dT23432dT39900dT861dT23905dT42142S7_227510

S7_161179045 S7_161179072

dT21588

dT14775

dT34305

dT5751

dT17613

dT7877

dT7992dT22362S7_168784407S7_168784355

dT26372

S7_146917990

S7_100737594

dT46088

dT36827dT36253

C7_38012075 S7_52134705

S7_52134761

S7_52134764

dT25902dT43823

dT19286dT43692S5_1337832dT23441dT28159dT9688

dT11814dT16180dT26922S7_119895303dT8614dT29758dT10112

dT6553

dT37830S7_66630588dT39337dT44668dT25362dT28445dT9471dT3618

7

S1_256402076

dT7731dT4991dT45191dT14502dT40880dT5478dT32461

dT40798

dT4388dT8297dT15729dT33449dT28736

dT45351

dT36410

S6_161712969

dT5999

dT47337

dT20275

S5_15660406

dT21669

dT45910

S6_69712882S6_69712890dT41069dT31386dT9518dT51317dT20193dT3770dT51103dT26827dT29908dT34795dT48051dT53110dT16879C5_27253540 dT9307C6_186956079 dT47900dT37205dT12426dT18627dT32069dT46240dT43200dT47089dT38416S6_99869597S6_99869584dT1280dT3322dT8391dT52684dT30730dT33180dT35795dT41539dT33330dT1374dT29657TS2_146426838dT31286dT17087dT43816dT10476dT53934

8

dT29596

dT6813dT8719dT33885dT31744dT26674dT44870dT30975dT13445dT8898dT24568dT46387dT36424dT33344dT4180dT39395dT25394dT47535dT1174dT2510dT3073dT19937dT37466dT38885dT48964dT22920dT1264dT11516dT52321dT12612dT9304dT42843dT1869dT18940dT13681dT2286 dT30154dT16070dT30575dC_161538dT25258dT38621dT27267S2_187091485dT33297dT1716dT53659dC_66514 dT30639dT40485dT53712dT32613dT51740C2_186899428dT23775dT9800dT32303S2_15995145 S2_15995173S2_15995166dT25206dT31778dT10587dC_198183dT34859dT30984dT7580dT25812dT11183dT35258dT10669dT21093dC_299275dT51487dT334dT50663dT10724dT18564dT36456S2_225137713S1_49814862S2_225245713dT32617S2_223315521

9

dT25323

dT41707dT45868

dT46296

dT19710

dT38592

dT53592

dT36198

S8_31103785

TS4_61117193dT16762

dT39279dT4379dT2661C4_54525188dT27692

dT43430dT11454

S1_211915643

dT5896

dT8867

dT41621

dT49164

dT46261

dT44455

dT30147

dT14352dT29968

dT32403dT29535

dT27669

dT35529

dT11525

10

dT47338dT35994

TS7_92021240dT13178

dT8606

dT11256dT42952C4_131830009

dT52290

dT20991

dT36533dT30022dT52180

dT31904

dT11891TS4_175268085

dT13560

dT29868

dT45689

dT52876dT15502dT18206

dT8487

dT26938

dT31173

dT10167dT2668dT8511dT34945dT17028dT16149dT28164dT838dT34453dT47823dT5924dT49429S7_12814263

dT45454

dT32197dT46410dT30920dT31921S7_146917952dT15629dT41891dT20567dT27310

dT11287

dT20514

dT24863

dT13782

dT46507

dT40245

dT24540dT17489

dT29640dC_75137dC_162440dT35101dT35136dT36597dT50685dT10621dT9181

S7_216787106

dT32609

11

dT52537

dT16456dT26692dT43606C2_14084192 C8_42747377dT14701

dT31069

dT9771

dT2397dT41528dT11071

dT20355dT47937dT47375dT23509dT13802C2_38440300dT47931dT9367dT35753dT16483dT29644dT36437dT5577dT29458dT22493dT5545dT304dT24621dT30035dT18050dT51168dT14389 dT1961dT12204dT991dT12112dT44755dT26322dT23286dT38411dT45471dT38599dT12000G3_47707669S3_47707650G2_236219679dT27423dT35339dT40541S3_50037874dT25091dT50771dT29828dT14016dT5445C3_198785997dT574dT53719dT14813dT11551dT40850dT6981dT32941dT27447dT10535dT2057dT2770dT221dT39089dT36711dT52029dT48485dT29482dT16591dT38040C3_64621327dT11606dT332 dT23060dC_45050dT24014dT41922

dT14847

dT5281

dT20166

S3_45974916dT10443

dT20483

dT23539

12

dT36565dC_4861C5_3132373dT37861S5_2055822dT40955dT25964dT39601C5_1635981dT51406dT51571dT27547dT43154dT34088dT41728dT25168dT44314dT9903dT9904dT5307dT1844dT24760dT48033dT13687dT43027dT6561dT24969dT43643dT1444dC_121093 C5_13708967dT53121dT1436dT6043S3_123721011dT49798dT18120C5_23963287dT52425dT45026dT34144

dT31045

dT14412dT16dT48116dT46519dT10955dT46588dT5755dT46787dT49723dT33950dT48885dT40660dT7562dT22374dT24215dT29452dT40958dC_26149dT22682dT10481dT49967dT13083dT43713dT39674dT1564dT21673dT48272dT1781dT16697dT28451dT13262dT11880dT48311dT25766dT34492dT14203S5_136028687dT47563S5_132636460 S5_132636448S5_132636508dT30544dT28689

13

S1_78149353S1_78149366dT4830dT25275dT18503dT10276dT32396dT39911dT52605dC_94047dT12412dT32705

dT25225

dT31852dT48139dT14455S6_32125800dT9078dT32166dT11671dT53836

dT7886

dT22293dT6248

dT35828

dT11704

dT17003dT6792dT24471

TS3_259950719dT47792TS3_71103541dT34857dT48904dT48523dT10105dT47317dT38317dT40493dT43569

S3_258304119S3_258304047 S3_258304050S3_258304063

14

0

5

10

15

20

25

30

35

40

45

50

55

60

65

70

75

80

85

90

95

100

105

110

115

120

125

130

135

140

145

150

155

160

165

170

175

180

185

190

195

200

205

210

80

Figure 2-9. Genotyping by sequencing single nucleotide polymorphism (GBS-SNP) marker distribution for the 14 linkage groups of napiergrass. A black bar means a GBS-SNP marker. A blue bar represents segregation distortion region. The left scale plate represents genetic distance (centiMorgan as unit).

81

Figure 2-10. Consensus genetic linkage map of napiergrass.

Napiergrass: Integrated linkage map

S1_267892930S1_241274562dT36338dT39139dT53136dT35128dT3813dT37356dS_22638dT52469dS_58585St_7983dT8342St_8341TS1_57392744dT30392dT36979dT766dT36115dT7314TS1_242079107dT30844TS1_273230575C1_234266014dT47057dT19448dT10843dT22725dT971dT18064dT37631dT40041S1_171369178dC_339765C1_133709193dC_359278dT13670TS1_149192348dT45509dT12272dT53672dT48112dT33654dT8459S1_144792807dT973dT16562dT18424St_2847dT50103dT28049dS_5480S1_17257551St_1775St_1908St_3838S1_17257559dS_38214dS_33009dT48243S1_17257463dT48497dT3431dT48800dT14180dT42594C1_212950043TS1_134040623dT46801dT29597dT6993C1_117276997dT36356dT38584St_4264St_4265dT28498dT13814St_4395dT23237dT13772dT23364dT53762dT23838dT51201dT45924dT7984dT46948TS1_74139971dT11106St_11047dS_83781dT19473St_119dS_25631dT732dT13338C1_102656250dT24786dT7985St_11044dS_9398dT39765dT18485dT47548dT9029dT50826dT32431dT29470dT50264St_5341St_203dT16527St_202dT33481dS_62356dS_8179dT46699dT52420dT1951dT35867dT49010dT35147dS_20460St_8592St_26736dS_64991St_26737dT10713dT3328S3_64306230C1_10467974dT34797dT1702dT3027dC_42488dT38795dT31457dT9151dT51180dT2680dT44577dT36538St_4579dS_37868dT18370St_857St_84477St_795dT30050St_5465G8_2178054dS_38908dS_53100St_1128St_8964St_10547dS_896dT3437St_44026St_43733dT14768dT7777

LG01

S6_80432248dT46838dT35970dT13151dT29238S6_197918666dT17747dT18539dT50350dS_64625St_65272dS_36562dT22426dT11200dT29974St_63794St_63444dT2241dT47054dT14807dS_24618dT30703dC_196800dT24998dT30440St_68017dS_24309dS_50465dT41470dT8340dS_78648dT2297dT41082dS_20475dT42484S6_242118514dT40470dT31260dT42554dT45101dT50977dT25232dT15707St_52312dT13790dC_193114dT42117dT2283dT11665St_67496dT16875dS_65719St_70712dT4135dS_32733St_69789dT18264St_12035dT35582St_864dT35318dT16616St_58109dS_35557dT16409dS_66127TS8_83771289dT35381dC_23232dT47482dT10670dT16487dT16658dT22201dT21000S7_154422455dT7220dT15893dT661dT41813dT7888dT38698dT42580dT5181dT50319St_45778St_30560dT5239dT39882dT30721dC_100348C2_129356854dT4190C7_82320373dT22787St_38861dT28192dT36233dT31759St_68287dC_428936dT28669dT5222dT37647dT43295dC_193851dT19708dT42197dT49133St_66871dS_16944St_68737dS_66207dS_19764dT46586dS_50880dS_50638St_63927St_63925dT46613dT38873dS_65237dS_10139dT43417dS_27918dT37638dT9987dT1466dT50566dT47787dT16346dS_68159dT3930dT6794dT14931dT53493dT46041dS_33008dS_58337dT16992dT38805dT38911dT16233dS_64711dT6835dT17665C6_39611566dT35938dT6810S6_35473070dT38488S5_31045825dT11037dT6548dT41013dT590dT45550dT2555dT159dT27467dT50779dT10674dT10631TS6_10754586C3_124558897dT38576dT53963dT46533dT17348dT11902

LG02

St_36922St_27049St_27050dT24472S3_85079300dT11506dT38048St_36360St_36361dT44178dT7396S3_234961343S3_65392670St_35372dT35908St_31668TS3_40995143dT26209dT2147dC_42042S3_68670889St_36963S3_68670908dT5016dC_28051dT22363dT33771dT51194St_27334St_27337dT13022dT40851dT42647dT42768St_26265TS3_109270808dT50758dT6927St_27574St_26191St_27576dT24195TS3_34404935S3_92611362dT52592dT12246dT24290dT2704St_27619St_27618S3_12254914dT31485St_27764dS_63784St_28126dT20803dT22019dT18489dT18863dT37926S3_99632078dT13705dS_55508dT26079dT31367dT16041dT42473dS_37174dT13169dC_150274St_33851S3_37058907St_30077St_35238St_33736TS3_274563801TS3_274563794TS3_274563751St_31749dT24779TS3_123337609St_12442dT40996TS3_219893756dT50718dT7695dT28196dS_67157St_38576St_27151dT50908dS_40283dT22708dT5249St_32104dS_2204dT6670dT41484S3_35602619S3_243475658dT1934S3_243475784C2_168637894dC_75778C3_273261021St_34784dT43871dT36593dT2820G3_243475604S3_243475604dT492dS_51063dT3273dT35742dT4631S3_277598650dT12717dS_24557St_32307dS_60100dT40657dT19568dS_10652TS3_102980429dS_68998dT32542dT20435St_32659St_30275dT43387St_32493dT24042St_50788dT25239dS_20958dT15989dT42777dT29788dS_53679dT21277dT15721C3_202051529dT24470dT8152dT15649dT38486dT14717dS_69156dT18085S3_258397449S1_55738760S3_16736775S3_41221675dT33250S3_260011343

LG03

S5_3204810S5_3204852S5_3204827S5_3204849S5_3204807St_55016dS_69164St_83036St_55233S5_34645373St_56474dS_12036St_56231dT8760St_57688St_50301dS_3673dS_33240dS_55772St_56111dT22629C5_4168325St_52503dT36067dT21831dT50761dT941dT50341dT23430St_52805St_54858dS_40368St_54860dT41977dT22371St_53749dS_24959St_53600dT36706St_53555dT31058dT13412dT36750dT53506dS_64191dT26005dT17979dT717dT32999St_54089C5_154569850dT20176dT15952dT7627dT780dT9747dT47758S5_92980877C5_3132406St_57085dT15117dT46426dT20384St_55733St_55732dS_40895dS_35332St_54510St_57106dC_282880dT10717St_53813dT20066dT27794C5_65406593dT9509dT21823dT32419S5_65361310St_56639dT2435dT50835dT15164dT48427dT5772dT23323dT30226dT8428TS5_114098156S5_40839446dT41901dT40995dT34274dT28525dT48365dT18480dT9315St_57290S5_63556385dS_51396TS5_128053793dT13827dC_24267C5_112498054C5_156998414St_51734S5_63556383dT27345dT30788dT14651dT15817St_50448dS_62152St_49453dT4737St_50451dT22005dT16773dT47857dT27775dT24515S5_59674699dT51425dT51S5_52320341dT24372S1_6210777

LG04

S2_30030493dT13183dT41204dT51314dT6913dT11178dT8400dT3792C2_69640554dT777dT34457dT13912dT15477dT32273dT3108dT28199S2_16737601dT1791S2_16737644dT42630dT51889St_22811dS_12677dT22002St_23394S2_174624948dT5178dT43720dT28282dT40697dT38918S2_24418753dT12948S2_24418742dT42842dT19311St_16460dT38665dT33063dT47411St_7438St_33827St_33826dT3578dT50366dT33944dT1873dT14669dT8327St_17032dT31449dT13764dT49212S2_110636956C2_21121892dS_26130St_24033dS_49336dT16449St_41249dT36036St_12768dT29465dT49950St_18051St_13306St_75341S2_76988086TS2_76988086S2_76982335dT5796S2_14223558dS_3168dT50180dT16824St_14790dT1183dS_68166dS_22487dT1114dT17391dT20397dT50336dT27602dT19647dS_66177dT23469dT10549dT4979S2_214301707dT53077dT15063dT348dS_7312St_14914dT34541St_15285St_15284dT45884S2_216176645S2_192440020St_16949St_18971dC_212370dS_17901dT51287dS_50616dT10378dT46228dT11302dT31033dT9291dT45970St_19647dT2290St_15938dT3272dT10652dT44368dT9233dS_69530dC_167176S2_226559654S2_230011940dT3180St_20919dT38020TS2_230011891TS2_230011940dS_292dT44853dT10355St_20865St_20885dC_64162dT25838dS_43657dC_88689dT51942dT30560dT27325dT48120dT25304St_22003St_13438dT19734TS2_210211525dT32534dT6193dT20949TS2_102571886dT12622dT50361dT5902dT44728dS_68599St_58349C2_79892636dC_299059dT31503dT2426dT31430St_17805dT12632dT43605dT44700dT25565dT13489dT33550St_22721dT47765dT10926S7_61864734dT32441dT640dT13896dT21007dT37138dT37626dT24945dT37696dT2179

LG05

dT42156dT36401dT53206dT12209dT49934TS4_171269213dT22615dT2317dT23827dS_49354dT27674St_48501dT51677dC_31636C4_99397060dC_90916S4_43915384dT27183S4_43915389TS4_199335191dT21313dT30293dT47539St_40435TS4_17374522dT52456dT23773S4_17921468St_40403dT45232dS_43008dT39406dS_62092dS_59686St_44404St_40987S4_56719430dT12456dS_16097dT49039dT9886dT1848St_39084dT38946TS4_56235778St_47561St_47563dT35551TS4_79794887dT28999St_47706TS4_79798648TS4_79798642St_47414dT33420dC_124707C3_186584463dT37277St_47413dT30894dT9337dT36845dT18515dT52450TS4_114218088St_32476S4_55762918dT10273S4_71863256dT52847dT5808St_46189dT30598S4_109511769St_46150dT11273dT16151dT6466S4_143679490St_54552dT8454dT18870dT43233dT4620dT42579dS_55849S4_29656682dT3285dT21909dT8320C4_90405411St_41693dT32416dT13255dS_39004St_42645St_13970dT29340dS_38178St_48484St_31330St_31333St_42445St_56407dS_59418St_42444dC_13461dT32924St_47777dS_57611St_47778dS_25947dT43805dS_29894dT23119C4_200698121dT27514dT3030dT40535dT20879dT18538dT48489dT29846dT15831dT25619dT3891dT22674dT12425S4_199859831

LG06

S7_219034180

dT1325dC_58157dT24339dT33835dT10459dT41386dT21208dT38126dT20154S7_182881700dT16455dT24498dT16860dT6436dT4353dS_40842dT39900dT23432dT861dT23905S7_227510dT30583dT42142S7_161179045S7_161179072S7_112269G7_212623362dT46230dT34305dT910dT14775dT26620dT43326dT27451dT50763dC_388382dT11554dT24186dT50419dT24408S7_162635983dT46209S7_100737594S7_133861267dT8557S7_52134764dT36827dT46088dT14658C7_146983287dT36253S7_100645462C7_38012075dT25902dT43823dT37501dT43692dT19286S5_1337832dT14025S7_90744125dT28159dT9688dT47365dT23441dT11814dT26922dT16180dT10131S7_32127059dT36791S7_119895303dT33050dT8614dT7088dT526dT30134dT35232dT45831dT53184dT29758TS7_167047574dT52369S7_102687604C7_89265590C4_41131304dC_221873dT40952dT10112dT3713dT22267dT12440S7_66705818dT32512dT47732S8_44730456S7_89265546S7_136957144

LG07

St_62484St_63298S6_15199694S6_37610339dT45995dT2684dT31673dS_29023dS_7808dT7834dT52932dT19776dT29318dT1374dT41539dC_9265dT33330TS2_146426838dT32668dS_68537St_67003dT25459St_67005St_69926St_8239dT771St_71627dT35795dT40500St_60317St_65388dT33180St_65387dS_43802dT52684dT47247C6_59409930dT24125C6_59409932C6_73259826dT47034dS_59431dT1281TS6_101968558dS_15188dT454dT17688dT22268TS6_16445661C6_105814832dT29814dT52515dT31219dT13258dT52774dT9518dT32504dT38233St_45444St_60942dT4343S6_59035298S6_199704613S6_241915017St_62949S6_241915045St_59677dT45221dT23999dT7731dT4991St_19657St_67385dT45191dT14502dT40880dS_49183dT4808St_64847dT47341dT5478dS_30976dT39816dT32461St_67353TS6_200342427TS6_200143111dT46758dT47729St_63770St_65266St_7565dT51264dS_53085dT330S6_209490431dT15729dS_31075dS_7074dT33449St_63375St_70658dT26926dT8297dT1593C6_223591048dT44856dT28906dT11877dT23515dT46843dT18102dT13837dT15691dT47337dT5999dC_93403dC_63791dT37150dT50316

LG08

dT27092

dT14190

dT44709TS8_54685526

dT12313

dT18940S5_74991349dT30575dT47221dT38621dT9800dT45548dT36456dT2536dT9428dT1352dT18564dT24570dT1670dT7119dT5487dT50663dT44398dT3390dT48482dT37165dT10724dT1149TS1_248242945dT8998dT22957dT3928dT45174dT26839dT21093dC_299275dT28517dT40868St_13976C2_206875269dT36870dT334S2_78894908St_58417dS_34607dS_44017dT52869dS_13200C2_143807176dS_24485dT36645dT191dS_13845dT50464dS_35262dS_13192dT51911S2_225137713S1_49814862dT40583St_16188dS_64130S2_225245713dT17835St_33671dS_57503St_21261dT44636dT32617dT32308dS_23912St_24952S2_144268311dS_56758S2_144268276dS_14880dS_57817dS_41244St_25765dT40973dT29502dS_30742dT14696dT20326dT35127dT52863dS_47193St_17971St_17970dS_68589St_20466St_20650dT48611dS_60240St_18499St_80242

LG09

dT37914dT46112dT41558dT40810dT33782dT36311dT31375dT18082dT32644dS_15838dT27051C8_45958102dT35420dS_28564St_43790dT40690dT37270dT4159S1_268164908S1_268948076dT9656dT27080St_8337dS_20051dS_5062TS1_270056181dT23820dT52173dT23737dT40084dT48984dT10533dT33254dT26907St_7976St_7494dT24369dS_35489S2_236302464dT45124dT20774dT47132dT51615dT50618dT43193S1_256031621dT47532TS1_55816576dT144dT40320dT6229dC_220532dT50746dT45573dS_30445dT7085dT20685dT17967dT6028dT36552dT49336dT23044dS_59001dT47746dT29658dC_105219dT2822dT39692dT40261C1_149814711dT31390dT43343S1_235015034S1_235015054dT42355dT12855dT34010dT5814dT41014dT30197dT5876St_3875dT12762St_9029St_9025dT49435C1_231986176dT40418dS_54187dT20183dT48992dS_8637dT50903dT10507dT46085dT11768dT14780dT28312St_9993St_8983dT47682dT10689St_45245dT34779dT47650C4_136757390dC_9584dC_4097dC_4103dC_9578dC_1926dT40903dS_6690dT1078TS4_5133343dT42109dT21114dT4836dT13498dS_2186dT51420dT38881dT28837dT48530dT45821dT5081dT32245dT50307dT30922dT12488dS_16469C4_182430864dT41621dC_15179dC_23137dT20060St_11615St_46107dT16799dT7436dT5896St_36519dT8867St_39105dT29992dS_55275dS_63047S1_211915643dT3588St_46582dT33996dT5844dT47966dT43430dT47917St_70710C4_54525188St_46842St_46843dS_68339dT42418dT44125dT47502St_41094dS_42852dT53273dT5708dT53333dT39834dT30409dT51639dT11173dT29622dT13526dT25986dT49885dT19710dT17365dT53592dT36198dT38592

LG10

dT34863dT43981dT32288dT34945dT16149dT8511dT2668dT49930dT26938St_26088dS_48139dT31173dT30424dT18438S4_131830044dT44002dT15502dT52876St_72633dT45689G4_131830067dT29868S4_131830067S4_131830066TS4_175268085S4_131830063dT13560dT11891dT22876dT31904dT12269dC_55022dT22567dT38883dT28123dT26116dT10040dS_61343dT36340dT26152dT10739dT26884dT621dT17578St_80208S7_89250999dT17325dT25711dS_21720St_16173dS_353dS_12928dT5186dT8749dC_10270dT15950TS8_38834569dT649dT53247dT53022dT12194S7_39646339S7_39646309S7_39646292dT33677dS_17159dT3510dS_36568dS_18627dT19298St_35836dT1762dT41196St_76782dT30018dT1053dT40278S7_83881022dS_4342dT7491St_81113S4_175268085dS_51373S4_175268116dT32609dT9181dT10621St_83873dT29364dT20560dT35101dT37141dT35136dT49802dT29640dT42825dC_162440dT43565dC_75137dS_53783St_116030dS_89120St_116029dT50685dT43419dT53109dS_217dT10058dT31974dT53392dT42256dT24540dT34174dT47569St_75947dT952dT54084TS7_195845045TS7_195845122dT21991dT44682dT46307dT8910dT24512dT12987dC_223673St_77447St_77432dT43632dT34571dT3675S7_213184560dT22858dT32036dT29409dT19025dT7711dT39297dT42995dT47137dT37923

LG11

dT52537dT16456dT26692dT43606C2_14084192C8_42747377dT14701dT31069St_83499dT9771dT2397dT41528dT11071dS_70286dT20355dT47937dT47375dT29957St_23648dT23509dT13802C2_38440300dT47931dT9367dT35753dT16483dT29644dT36437dT5577dT54067dT36404dT29458dT37862dT5545dT304dT24621St_18279dT17058dT30035dT18050dT51168St_19470dT14389dT1961dT12204dT27243dT991dT12112dT29731dT19317dT44755dT37721dT26322dT39135dT23286dT38411dT45471dT38599dT12000dT4552dS_3981G3_47707669dS_27715dS_68831St_28245St_28242dT26266S3_47707650dT9730dS_58321dT13363dT27423dS_2736dT38555dT24050St_26083dT35339dC_15863dT3341C7_118131103dT40541S3_146735713S3_50037874dT25091dT5917dT28896dT37132dT8512dT50771dT29828dT47840dT14016C3_216897716dT38857dT5445dT19291C3_198785997dT574dT53719dT14813dT11551dT24476S3_58732195dT4006dT40850dT34879dT6981dT32941dT27447dT50794dT10535dT2057dT9107dT2770dT6635dT39970St_37822dT221dT39089dT36711dT52029dT48485dT29482dT16591dT38040C3_64621327dT11606dT332dT23060dS_63173dS_2101dC_45050dT2954dT24014dT5877dT41922dT14847dT5281dT20166S3_45974916dT20483dT23539dT13850dT28463dT4710dT14945

LG12

dT19752St_57083dT6944St_51502St_52583dS_53273dS_30098dT47563dT31151dT15747dT14203dT34137S5_136028687dT8126dS_23166St_51337dT36120dT51352dT18762dT3360St_50213dT12719dT34794dT25766dT34959dT15762dT19372dT34492dT43443dT41039dT45156dS_63438dT44382St_57235dT20917dS_5311dT7338dT48311dT33358dT11880dS_66625St_68482dS_4586St_52257dS_46016dT28451dT15626dT16697dS_12408dT16760dS_26459dS_49110dT14825dT13865dT22667dT12193dS_63968dT8566St_54452TS5_23967052dT36565S5_34147564dT51949S5_875176dC_4861C5_3132373TS5_12313000S5_2055822dT10696dT51406dT51571dT34088dS_26201dT41728dT25168S5_7812618dS_29656dT48033dT40635dT9904dT9903dT44314dT3855dS_21070St_77252dT46712TS5_6700729St_57050St_56490dS_15609St_59001St_77253dT30275dT53121dC_121093C5_13708967dT1436dT49612dT50943dT22104dT46312

LG13

S7_177490027

dT19814dT24731dT51987dT21395S3_258304063S3_258304050S3_258304047dT34497S3_258304119S3_237947843dT11865S3_259562492TS3_202192213dT22841dT43569dT40493dT27298dT27432dT47317TS3_52779446TS3_52779443dT48523dT10571dT10105dT48904TS3_71103541dT43079dT47792S3_255053440dT4657S3_255053465S3_257143847dS_45674dT35964dT17340dT6792dT17003dC_399018St_32898S3_253954916dS_3349dS_1370dT29965dS_33295dT11704dT45243dT48835dT35828dT6248dT22293dS_335dC_1015698dT1549S3_244121865dT29691dT51116dT12496dT41167dT33642S3_276917001dT12709S7_182689721dT317dT9078dT19423dS_45055TS6_161742923dT51331dT23909dT36064dT30367dS_40112St_31374TS3_228287887dT7231dS_37949dS_36977dS_27204dT2763dT40596dT37207dT8609dT7307dT31619dT21724dT39129dT6698dT30790dT42183dT12412dT34207dT32705dT40659dT52605dT18945St_30603dT27593dT39911S4_86603238dS_31623St_32380

LG14

0

5

10

15

20

25

30

35

40

45

50

55

60

65

70

75

80

85

90

95

100

105

110

115

120

125

130

135

140

82

Figure 2-11. Circos plot of the mapped TASSEL de-novo UNEAK napiergrass markers with pearl millet reference genome. Pearl millet pseudomolecules start with “PM” and are color coded for each pseudomolecule. Napiergrass linkage groups start with “LG” and are in green color. Each line corresponds to tags that showed significant BLAST hits to the pearl millet genome (> 80% identity and > 50 bp length).

83

Figure 2-12. Syntenic regions between napiergrass linkage groups and the pearl millet genome. PM01 to PM07 are pearl millet pseudomolecules, LG01 to LG14 are napiergrass linkage groups. The small dots represent significant BLAST hits of mapped UNEAK tags to the pearl millet genome (>80% identity and >50 bp length).

84

CHAPTER 3 MAPPING QTLS CONTROLLING FLOWER NUMBER AND FLOWERING TIME IN

NAPIERGRASS

Introduction

Napiergrass (Cenchrus purpureus Schumach) is a tropical perennial grass that

originated from Africa (Singh, Singh, and Obeng 2013) and is an important fodder crop

widely used as feed for dairy cows (Farrell, Simons, and Hillocks 2002). In addition, due

to its high biomass potential, napiergrass is considered as a promising crop for

cellulosic biofuel with higher dry biomass yield compared to sorghum (Sorghum bicolor),

maize (Zea mays), sugarcane (Saccharum sp.), switchgrass (Panicum virgatum),

johnsongrass (Sorghum halepense), and Erianthus (Ra et al. 2012). Napiergrass was

introduced to the United States in 1913 (Burton 1990). Being non-edible, napiergrass

escapes the 'food versus fuel' debate as the biofuel feedstock and has competitive

advantage over tree species because it can be harvested for biomass in the first year

after planting. Furthermore, the lignin content, which is considered a hindrance to the

fermentation of biomass to ethanol, is much lower in napiergrass than woody biomass,

10% in napiergrass compared to 20-30% in woods (Tong, Smith, and Mccarty 1990;

Mckendry 2002). In addition, napiergrass tolerates multiple harvests after which it

shows better ratooning ability than energycane (Cuomo, Blouin, and Beatty 1996). This

supports a constant feedstock supply and minimizes transportation costs of biomass.

Napiergrass is a short-day plant and flowering in tropical climates occurs from

autumn through winter (Singh, Singh, and Obeng 2013). Early flowering cultivars

produce abundant wind dispersed seeds, which contribute to the high potential of

invasiveness or weediness (D’Antonio and Vitousek 1992; Loope, Hamann, and Stone

85

1988; Schofield 1989). Therefore, The Florida Exotic Plant Pest Council has listed

napiergrass as an invasive species (FLEPPC 2011). Controlling flowering or modifying

flowering time of napiergrass can minimize invasiveness and boost its potential as

biofuel feedstock. Late flowering will reduce the total seed production and low

temperatures during late season may even compromise development of viable seeds

(Grabowski et al. 2016), thus reducing invasiveness. Significant genotypic variation for

agronomic traits have been documented in napiergrass (Sinche et al. 2018). Flowering

time showed a considerable variation within an F1 population of napiergrass, thus there

is a good reason to believe that improvements of flowering time in napiergrass can be

achieved using genetic approaches. Genome editing tools could be utilized to regulate

flowering time (Jung et al. 2018; Jung and Müller 2009) and to avoid unintended

spreading of napiergrass. However, routine transformation and genome editing

protocols still need to be developed for napiergrass (S. Zhou et al. 2018; J. Wang et al.

2017). So far there are no quantitative trait loci (QTL) mapping reports for agronomic

traits in napiergrass. Even in this genomic era, there is no noticeable genomic data

publicly available for napiergrass. This presents a challenge to improve agronomic traits

in napiergrass breeding utilizing marker assisted selection (MAS). Therefore, more

genomic resources and a better understanding of the genetic basis of flowering time is

necessary in napiergrass.

Most studies on napiergrass were limited to assessing genetic diversity and

relatedness using random amplification of polymorphic DNA (RAPD), amplified

fragment length polymorphism (AFLP), isozymes, and simple sequence repeats (SSRs)

developed for other species like pearl millet (Cenchrus glaucum) and buffelgrass

86

(Pennisetum ciliare) (Lowe et al. 2003; Bhandari, Sukanya, and Ramesh 2006; Harris-

Shultz, Anderson, and Malik 2010; Kandel et al. 2016; Dowling et al. 2013; Dowling,

Burson, and Jessup 2014; López et al. 2014; Smith et al. 1993). Environmental

biosafety of napiergrass can be increased by utilizing molecular markers that are linked

to specific traits of interest such as late flowering. This MAS allows effective selection of

breeding materials. Identification of QTLs related to flowering will support MAS

programs for breeding napiergrass with delayed or less flowering characteristics,

limiting invasive potential of napiergrass. This may enhance the potential for utilizing

napiergrass as a forage and biofuel crop.

QTL mapping tries to identify stretches of DNA that are closely linked to genes

underlying the trait of interest by performing statistical analysis of genome-wide

molecular markers and traits measured in progeny of controlled crosses (Stinchcombe

and Hoekstra 2008). The advancement of next generation sequencing (NGS) has

hugely facilitated QTL identification and mapping. High density genetic maps developed

by genotyping-by-sequencing (GBS) have been used successfully to identify genes

related to flag leaf traits in wheat (Triticum aestivum) (Hussain et al. 2017), winter

hardiness and fall dormancy in Medicago sativa (Adhikari et al. 2018), bunch fruit weight

and height in palm (Elaeis guineensis) (Pootakham et al. 2015), and bloom date in

peach (Prunus persica) (Bielenberg et al. 2015). QTLs related to flowering time have

also been identified in other species as well. In pearl millet, QTLs for flowering time co-

mapped with QTLs for stover yield, grain yield, and biomass yield in LG4 and LG6

(Yadav et al. 2003). Similarly, three QTLs on LG2, LG3, and LG4 were identified for

grain yield across variable post-flowering moisture environments in pearl millet (Bidinger

87

et al. 2007). In wheat and barley, vernalization (Vrn) and photoperiod (Ppd) genes were

involved in flowering time variations (Cockram et al. 2007). Transcription factors such as

AP2 and agamous-like MADS-box were detected in Adzuki bean for QTLs related to

flowering time, maturity, and seed coat color (Y. Li et al. 2017). In rice, several different

QTLs for flowering time were identified and cloned in early and late flowering cultivars

that may have been involved in adaptation to cold regions (Izawa 2007; Xue et al.

2008). However, no QTL analyses in napiergrass have been reported so far. Whether

the orthologs in other species underlying flowering time also control the flowering time in

napiergrass is unknown.

The variation of flowering time in napiergrass accessions can be exploited to

understand the genetic basis of flowering in napiergrass. QTL analyses in napiergrass

is enabled by the generation of the first genetic map of napiergrass, which contains

1,913 SNP markers called from GBS of a pseudo-F2 mapping population, derived from

a cross between an early flowering line and a late flowering line of napiergrass (Paudel

et al. 2018). This map facilitates the identification of QTLs for various traits including

flowering related traits that can be utilized in the future for MAS. The main objective of

this research was to use this first genetic map of napiergrass to identify markers linked

to genes controlling flowering time and flower number through QTL analyses.

Materials and Methods

Development of a Mapping Population

A mapping population of 185 F1 hybrids was developed by crossing an early

flowering accession N122 and a late flowering accession N190 of napiergrass (Sinche

88

2013). Accession N190 produced a higher biomass and had a reduced number of thick

tillers compared to N122 (Sinche 2013).

The 185 F1 hybrids along with the two parents, N122 and N190, and an

established cultivar Merkeron were phenotyped as replicated single row plots at PSREU

in 2011 and 2012. Clones of the whole population were also planted at the Everglades

Research and Education Center (EREC) in Belle Glade, FL in 2015 and flowering date

was recorded in the following year after planting.

Phenotyping the Mapping Population

The field experimental design followed a randomized complete block design

(RCBD) with three replicates of 187 lines (185 F1 hybrids and 2 parents). One block

contained a single plant as a replicate of each line. Within each block the lines were

randomly assigned to estimate block effects. The first flowering date and flower

numbers of each line were recorded in October ~ December of 2012, 2013 on the plants

established at PSREU. The first flowering date was also recorded on the plants

established in EREC in October ~ December of 2016. The flowering date, defined as

the date when the first flower was visible, was documented weekly from the first week of

October to the first week of December. Flowering traits in 2012 and 2013 at PSREU

were obtained from a previous study (Sinche 2013). Flowering time (FT) was calculated

as the number of days between the first appearance of the flower and vernal equinox

(March 20) for the specific year (Lambert, Miller-Rushing, and Inouye 2010).

Genetic Map

The genetic map was generated based on the SNPs from GBS of the 185 lines

of the population (Paudel et al. 2018). Briefly, GBS was used to genotype the 185 F1

89

individuals and the two parents. SNPs were identified using various software tools to

construct a linkage map for maternal and paternal parent by employing a pseudo test-

cross strategy (Paudel et al. 2018). For the female parental line, a total of 899 SNP loci

mapped on 14 linkage groups with a total length of 1,555.17 cM were used. Similarly, for

the male parental line, a total of 1,073 markers that were grouped into 14 linkage

groups spanning a length of 1,939.19 cM were used (Paudel et al. 2018).

QTL Analysis

For QTL detection we chose the composite interval mapping method on

WinQTLcart 2.5 (S. Wang et al. 2005). Mean values of each trait across three replicates

from different years were used for the QTL analysis. The walking speed chosen for all

traits was 2 cM. A forward and backward stepwise regression method with a probability

of 0.1 and a window size of 10 cM were utilized to determine cofactors. LOD thresholds

used to determine the significance of identified QTLs was identified by using the

thousand-permutation test to each data set (p ≤ 0.05) (Churchill and Doerge 1994).

Adjacent QTLs on the same chromosome for the same trait were considered different

when the support intervals did not overlap (Haggard, Johnson, and St. Clair 2015). The

95% confidence interval was calculated for each QTL considering a 2-LOD support

interval (van Ooijen 1992). The QTL span was delimited using LOD-1 support interval

(LSI). The contribution rate (R2) was calculated as the percentage of phenotypic

variance explained by each QTL in proportion to the total phenotypic variance. QTLs

were named according to McCouch et al. (McCouch et al. 1997). Specifically, the QTLs

detected for number of flowers on the linkage map constructed for early flowering parent

N122 were designated “qNFE” (qtl Number of Flowers Early) followed by a linkage

90

group number and QTL number on the same linkage group for the same trait, separated

by a dash “-“. Similarly, the QTLs detected for flowering number based on the linkage

map constructed for late flowering parent N190 were named “qNFL”. The QTL for

flowering time on linkage map for early flowering parent N122 were named as “qFTE”

(qtl Flowering Time Early) and the QTLs for flowering time on map for late flowering

parent N190 were given “qFTL” (qtl Flowering Time Late). QTLs with a positive or

negative additive effect for a trait imply that the increased or decreased phenotypic

value is contributed by the QTL, respectively.

We categorized the identified QTLs as either stable and potential QTLs. QTLs

detected based on the phenotypic data from more than one year were considered as

stable, while those detected based on only one year’s phenotypic data were considered

as potential QTLs. The positions of the QTLs identified on each linkage map were

indicated by using MapChart 2.3 (Voorrips 2002).

Candidate Gene Identification

Sequences were extracted from the sequence tags, which generated the SNP

markers flanking QTL region. The extracted sequences were then BLASTed against the

pearl millet genome v1.1 (Varshney et al. 2017) to identify nucleotide matches of these

sequences to identify the QTL sequence intervals. Gene models and KEGG annotation

(Varshney et al. 2017) of pearl millet genes within the identified QTL sequence interval

were extracted. The function of each gene model was checked for its role in flowering

and a list of potential candidates was curated.

91

Results

Phenotypes

Number of flowers

The number of flowers of the 185 F1 individuals, and their parental lines, early

flowering line (N122) and a late flowering line (N190) ranged from 0 to 46 with an

average of 6 in 2012 (Figure 3-1, Table 3-1) and from 5 to 256 with an average of 61 in

2013 (Figure 3-2, Table 3-2). Many F1 individuals actually didn’t flower before the

harvest date (Dec. 6th, immediately before the first predicted frost) in 2012.The

correlation between number of flowers for 2012 and 2013 was 0.62 (Fig 3-3). The

heritability estimates for number of flowers ranged from 79% to 85% (Table 3-1).

Flowering Time

The days to flowering of the whole F1 population including the two parental lines,

ranged from 219 to 270 days with an average of 240 days in 2012 (Fig 3-4) and from

220 to 258 days with an average of 238 days in 2013 (Table 3-1, Fig 3-5). The average

number of days to flowering in 2016 was 233 days (Fig 3-6, Table 3-1).

The correlation between days to flowering for 2012 and 2013 was 0.34, while that

between 2012 and 2016 was 0.22. Correlation of days to flowering between 2013 and

2016 was 0.26 (Fig 3-7). The heritability for the days to flowering were estimated

ranging from 60% to 87% (Table 3-1). Days to flowering was negatively correlated with

number of flowers (r = - 0.42 for 2012 and r = -0.64 for 2013, Fig 3-8 and Fig 3-9).

QTL Analysis

Number of flowers

Three stable QTLs (qNFE-1-1, qNFE-1-2, and qNFE-1-3) were identified on LG1

of the genetic map for early parent N122 for number of flowers (Table 3-2, Figure 3-10).

92

Three potential QTLs (qNFE-2-1, qNFE-6-1, and qNFE-6-2) were identified on LGs, 2

and 6 (Table 3-2). The most important number of flowers QTL (qNFE-1-2) for N122

parent (R2=0.20, LOD=10.76) was found in LG1and was located at 123.7 ~ 126.8 cM.

The same linkage group harbored the two additional stable QTLs (qNFE-1-1 and qNFE-

1-3) that explained 15-17% of variance with LOD score ranging from 8.13 to 9.10 at

peak intervals from 133.5 cM - 134.6 cM and 116.9 cM - 118.2 cM. This suggests that

LG1 plays an important role for number of flowers in napiergrass.

Five potential QTLs for number of flowers were identified on LG 1, 4, and 5 of

the N190 map (Fig 3-11, Table 3-2). These potential QTLs had a PVE of 6-10% with a

LOD value ranging from 3.18 - 4.78. Four out of those five QTLs had additive effects in

favor of trait value. One potential QTL (qNFL-11-2) had a negative effect (Table 3-2).

Flowering time

A total of six potential QTLs were detected for FT. Two potential QTLs out of the

six were detected on the data of year 2013 on LG 1 of N190 and two potential QTLs

were detected on the same LG of N122 with R2 ranging from 0.11- 0.14 and LOD score

of 5.40-6.77. Two additional potential QTLs were identified on LG 7 of map N122. The

R2 for these potential QTLs ranged from 0.09 - 0.12 with LOD scores ranging from 3.91-

5.57. Both QTLs showed negative effect of small value (R2 = 0.07) on the trait.

Candidate Genes

Candidate genes were only identified in the genome regions corresponding to the

three stable flower number QTLs: qNFE-1-1, qNFE-1-2, and qNFE-1-3. Based on

BLAST analysis of sequences flanking the QTL marker against the pearl millet genome

(Varshney et al. 2017), three genome intervals spanning a genome size of 4 Mb, 12.37

93

Mb and 1.4 Mb, respectively, were identified on pseudomolecule 1 of the pearl millet

genome, corresponding to the three QTLs. A total of 624 gene models were identified

within the three QTL intervals of the genome (160 gene models on qNFE-1-1, 397 on

qNFE-1-2, and 67 on qNFE-1-3). 295 of these models had KEGG annotation results (71

on qNFE-1-1, 191 on qNFE-1-2, and 33 on qNFE-1-3) (Varshney et al. 2017). This

region harbored potential candidate flowering genes such as AGAMOUS, DELLA, Floral

homeotic protein DEFICIENS, PPM1 (MIKCC MADS-domain protein), WRKY, and

SERK1 (Table 3-4).

Discussion

Napiergrass is an important forage crop with high potential as a cellulosic biofuel

feedstock. Breeding of non- or late flowering varieties that have an extended vegetative

growth is not only important for obtaining a high yield but also for reducing invasiveness.

So far, genes involved in flowering regulation have not been identified in napiergrass,

which has severely hindered the improvement of flowering time traits in napiergrass

breeding. Over the last few years, mapping of QTLs for economically important traits

and genome assembly has largely facilitated the breeding programs by the

development of MAS (Y. Wang et al. 2011). Recently, a high-density genetic map of

napiergrass has been developed (Paudel et al. 2018) that will ease identifying QTLs.

Exploring QTLs is important because many studies have identified candidate loci near

QTLs. For example in cereals, a QTL controlling heading date variation was close to loci

controlling photoperiod and vernalization (Laurie et al. 1994; Bezant et al. 1996). In rice,

four genes controlling flowering time (Hd1, Hd6, Hd3a, SE5) were identified from 14

QTLs (Yano et al. 2001). In maize, it was shown that flowering time variation was the

94

result of cumulative effect of several small QTLs and by using 5,000 RILs, a total of 36

QTLs were identified for days to anthesis (Dell’Acqua et al. 2015). Flowering time QTLs

in brassicas mapped to similar regions in homologous chromosomes within and

between species (Lagercrantz et al. 1996; Osborn et al. 1997). Flowering related genes

in Brassica showed a close sequence similarity to one end of chromosome 5 of

Arabidopsis, which contained many flowering related genes (Bohuon et al. 1998). This

type of syntenic results showed that the same QTLs may exist in other related species.

The availability of the reference genome of pearl millet (Varshney et al. 2017) largely

facilitates the comparative genomic approaches for napiergrass as these two species

share one genome (Jauhar and Hanna 1998).

In this study, we phenotyped a biparental F1 or pseudo-F2 population that

segregated for flowering time to identify QTLs related to flowering. The traits studied

showed a continuous variation indicating that the traits are quantitative in nature. We

observed a negative phenotypic correlation between the number of flowers and

flowering time in the F1 population. Days to flowering and number of flowers were

consistent in different years and locations for the population. The heritability for

flowering date ranged from 0.60-0.87 for napiergrass. It was higher than the heritability

reported for heading date in rice which ranged from 0.37 to 0.55 (L. Zhou et al. 2016)

and lower than that reported for maize (0.82 to 0.93) (C. Wang et al. 2010). For linkage

analysis, we employed a pseudo-testcross strategy. Linkage analysis of quantitative

traits in outcrossing polyploid species by using single dose markers (1:1) is a common

practice due to the limitation of the linkage analysis software (K. K. Wu et al. 1992). In

this approach, a pseudo-testcross strategy using heterozygous markers for one parent

95

and recessively homozygous markers in other parent is employed for linkage map

construction (S. Wu et al. 2010; W. Zhou et al. 2015). This strategy has been

successfully applied in tree plants such as Pinus elliottii and P. caribaea (Shepherd et

al. 2003), legumes such as alfalfa (Adhikari et al. 2018), as well as grasses such as

orchard grass (W. Xie et al. 2011), and sugarcane (Yang, Islam, et al. 2018).

In this study, three stable QTLs for number of flowers were identified on the

linkage map constructed for N122. An initial study done in pearl millet identified QTLs

related to grain number and panicle number in LG1 of pearl millet (Bidinger et al. 2007;

Yadav et al. 2003). However, the lack of shared markers did not allow a definitive

conclusion if these two QTLs represent the same locus.

Several QTL regions related to flowering time and number of flowers were

located at the end of the LG 1 of napiergrass. Across all years, the early flowering QTLs

with the greatest effect was found on LG 1 that correspond to pseudomolecule 1 of

pearl millet (Varshney et al. 2017). This highlighted the importance of this chromosome

for flowering and it might harbor genes that control flowering in napiergrass. QTL

identification has enabled us to link variations at the trait level to variations at the

sequence level. Since a QTL can harbor tens to hundreds of genes (Gelli et al. 2016),

the identification of genes responsible for phenotypic variation poses a major challenge.

Accurate identification of the underlying genes responsible for trait variation in QTL

regions is difficult in napiergrass due to the lack of genomic and transcriptomic

resources in napiergrass. Nevertheless, comparative genomic approaches utilizing

reference genome of pearl millet has helped us to identify putative candidate genes

related to flowering in napiergrass. Several candidate genes affecting flowering time in

96

model plants have been discussed previously (Kuittinen, Sillanpa, and Savolainen 1997;

Bouché et al. 2016). In rice, QTL analysis of flowering time identified genes such as

Hd1 (CO orthologue in Arabidopsis), Hd6 (CK2 orthologue in Arabidopsis), Hd3a (FT

orthologues in Arabidopsis), and SE5 (HY1 orthologue in Arabidopsis) (Yano et al.

2001).

In our study, we located important flowering time related genes including

AGAMOUS, DELLA, DEFICIENS, PPM1 (MIKCC MADS-domain protein), WRKY, and

SERK1 (Table 3-4) in the QTL area of napiergrass that showed sequence similarity to

the pearl millet genome. The AGAMOUS gene encodes a transcription factor that

regulates genes to determine stamen and carpel development (Yanofsky et al. 1990).

DELLA protein plays a key role in negative regulation of gibberellin biosynthesis that

regulates many cellular and developmental events include flowering, pollen maturation,

and the transition from vegetative growth to flowering (Yoshida et al. 2014). DEFICIENS

is an ortholog of APETALA3 in Arabidopsis that functions in petal and stamen organ

identity (Zahn et al. 2005). PPM1 is ubiquitously expressed throughout vegetative and

reproduction tissues and may have diverse functions (Singer, Krogan, and Ashton

2007). WRKY has been related to several abiotic responses and accelerates flowering

by regulating FLOWERING LOCUS T and LEAFY (Phukan, Jeena, and Shukla 2016).

SERK1 (SOMATIC EMBRYOGENESIS RECEPTOR-LIKE KINASE1) acts as a

negative regulator of abscission metabolism and is required for anther development

(Lewis et al. 2010). Further investigations of these genes involved in flowering in wild

accessions of napiergrass are necessary to elucidate the basis for intraspecific variation

which is of great relevance to breeders. Future fine-mapping analysis by identifying

97

additional genetic variants in the QTL regions or by targeted resequencing of the QTL

intervals in a subset of individuals can lead towards accurate identification of targets for

improving flowering time in napiergrass and related crops. To verify the napiergrass

QTLs detected in our analysis, phenotyping should be repeated in multiple locations.

Breeders are more interested in QTLs that have large effect (Kearsey and Farquhar

1998) and the QTLs identified in this study can be used in the future for MAS in order to

breed late flowering napiergrass.

Conclusion

In summary, this study reports for the first time QTL analysis in napiergrass.

Three stable QTLs controlling number of flowers in napiergrass were identified, which

can explain 15%-20% of the phenotypic variation. We also identified three potential

QTLs controlling flowering time in napiergrass, which can explain 11%-14% of the

phenotypic variation. Gene models in pearl millet that mapped to two stable QTLs

harbored MADS-box transcription factors such as AGAMOUS and DEFICIENS, along

with other proteins such as DELLA, WRKY, and SERK1 that are involved in flowering

time regulation in plants. The QTLs detected in this study will be valuable information for

napiergrass breeding programs and help to understand the genetic basis of flowering.

This study confirms that flowering time and flower number are highly heritable traits in

napiergrass. Therefore, the late flowering napiergrass genotypes developed by Sinche

(2013) constitute a valuable germplasm resource to develop late flowering napiergrass

cultivars. In addition, validation of the putative candidate genes identified in this study

should lead to targets for genome editing to manipulate flowering time in napiergrass.

98

Figure 3-1. Histogram of number of flowers in the mapping population in 2012 in Citra, FL. X-axis represents the number of flowers and y-axis represents the count of plants. Average number of flowers of two parental lines, N122 and N190 are plotted. Data is adapted from Sinche, 2013.

99

Figure 3-2. Histogram of number of flowers in mapping population for 2013 in Citra, FL. X-axis represents the number of flowers and y-axis represents the count of plants. Average number of flowers of two parental lines, N122 and N190 are plotted. Data is adapted from Sinche, 2013.

100

Figure 3-3. Scatterplot of number of flowers between 2012 and 2013 in Citra, FL. X-axis represents the number of flowers in 2012 and y-axis represents the number of flowers in 2013. Data is adapted from Sinche, 2013.

0

100

200

0 10 20 30 40

Number of Flowers 2012

Nu

mb

er

of F

low

ers

20

13

101

Figure 3-4. Histogram of number of days to first flowering in 2012 in Citra, FL. X-axis represents the number of days to flowering (FT) and y-axis represents the number of accessions. Days to first flowering of two parental lines, N122 and N190 are plotted. Data is adapted from Sinche, 2013.

102

Figure 3-5. Histogram of number of days to first flowering in 2013 in Citra, FL. X-axis represents the number of days to flowering (FT) and y-axis represents the number of accessions. Days to first flowering of two parental lines, N122 and N190 are plotted. Data is adapted from Sinche, 2013.

103

Figure 3-6. Histogram of first flowering date in the mapping population in 2016 in EREC, Belle Glade, FL. X-axis represents the number of days to flowering (FT) and y-axis represents the number of accessions. Days to first flowering of two parental lines, N122 and N190 are plotted.

104

Figure 3-7. Scatterplot of first date of flowering between different years and locations.

X-axis and Y-axis represent the number of days to flowering (FT) in different years. Data for 2012 and 2013 are from Citra, FL and adapted from Sinche, 2013. Data for 2016 are from Belle Glade, FL.

R2

0.11

220

230

240

250

220 230 240 250 260 270

Days to Flowering 2012

Da

ys to

Flo

we

rin

g 2

01

3

R2

0.04

220

230

240

250

220 230 240 250 260 270


Da

ys to

Flo

we

rin

g 2

01

6

R2

0.06

220

230

240

250

220 230 240 250


Da

ys to

Flo

we

rin

g 2

01

6

105

Figure 3-8. Scatterplot between number of flowers and days to flowering in 2012 in Citra, FL. X-axis represents the number of days to flowering (FT) in 2012 and y-axis represents the number of flowers per line in 2012. Data is adapted from Sinche, 2013.

0

10

20

30

40

220 230 240 250 260 270


Nu

mb

er

of F

low

ers

20

12

106

Figure 3-9. Scatterplot between days to flowering and number of flowers in 2013 in Citra, FL. X-axis represents the number of days to flowering (FT) in 2013 and y-axis represents the number of flowers per line in 2013. Data is adapted from Sinche, 2013.

R2

0.41

0

100

200

220 230 240 250


Nu

mb

er

of F

low

ers

20

13

107

Figure 3-10. Linkage map of male parent N122 showing potential and stable QTLs. QTLs for number of flower (NF) are color coded in blue and that for flowering time (FT) are colored in green. A black bar means a GBS-SNP marker. Markers are labeled on the left and genetic distance (centimorgan as unit) are labeled on right. LG = Linkage group.

Linkage map of male parent N122

S3_66915283 0.0S3_66915258 0.6

dT6838 5.5dT44555 6.5dT27134 6.9

S3_66915278 7.5S1_7068544 8.9

dT3437 10.2dT46011 10.4dT38476 13.3dT35347 13.8dT33533 14.1dT28618 14.7dT39822 15.3dT28250 15.4

C5_27253496 16.0TS7_74798545 16.5

dT31263 19.7dT6267 21.1

dT30050 21.8C1_119093525 C1_119093531 22.0

dT37316 22.3dT12054 22.6

S1_71014695 24.0dT17414 24.9dT29057 25.8

S3_284635389 26.2dT2475 27.9

dT18370 dT12390 28.0dT41113 28.1dT30582 28.3

dC_618193 30.4dT44577 31.3dT34797 32.9dT51180 33.5dT28611 36.3dT35867 37.1dT23729 40.2dT34784 41.0dT37028 41.5

S1_193475897 41.8dT36388 42.3dT3762 42.5

dT38724 48.8TS1_74139971 dT11106 49.2

dT6132 49.9dT7984 50.2

dT13772 50.6dT53762 51.2dT8946 51.8

dT18605 52.3dT29128 53.6dT52531 56.0dT41949 60.2dT47752 68.8

dT27185 84.6

S4_21013305 91.5

dT8342 114.3dT52469 116.0dT37356 117.2dT35128 118.2dT53136 118.6dT39139 120.7dT36338 122.1dT19693 124.8dT44366 125.5dT28974 126.0

S1_249594415 126.5dT5535 127.0dT8792 127.2

dT42596 127.4dT48305 128.0dT32706 129.2dT49414 129.9

C1_251517594 130.2dC_259973 130.3

S1_251483789 130.4dT41404 130.8dT42403 131.3dT9863 131.6

S1_260924412 133.5dT35453 134.2dT45297 134.7dT28492 135.8dT47753 136.7dT32523 137.3dT21677 137.7

dC_2450640 138.4dT48962 139.3dT50248 139.6dT27136 140.5dT32605 141.2

S1_267892930 149.1

qN

FE

-1-1

qN

FE

-1-2

qN

FE

-1-3

qF

TE

-1-1

qF

TE

-1-2

LG1

dT11902 0.0dT17348 0.5dT46533 1.3dT53963 2.5dT38576 3.2

C3_124558897 3.3dT10674 4.2

TS6_10754586 4.7dT10631 4.8dT27467 6.0dT2555 8.6

dT45550 9.1dT11037 10.5

S5_31045825 10.8S6_35473070 12.0

dT35938 14.3C6_39611566 14.8

dT6835 15.9dT16233 20.7dT38911 21.0dT38805 23.3dT6794 24.4

dT14931 25.2dT16992 25.7dT46041 26.6dT16346 29.2dT47787 29.6dT42989 32.8dT50566 34.0dT9987 36.3

dT29726 41.8

dT23975 44.8dT31396 46.2dT43235 47.3

dT2439 49.2

dT39167 52.1

dT35142 56.8

dT5958 60.6dT8821 61.9

dT25121 69.1dT39882 69.2

dT30721 dT19708 69.3dT36233 69.5dT18641 71.2dT28669 71.7

dT15498 dC_1941dC_5486367 dC_428936

dT50064 dT4146074.3

dT5222 74.5dT43295 74.8dT31359 75.0dT44933 77.3dT33834 81.5dT46138 82.2

dT987 91.3dT17187 92.4

dT3299 95.9dT29166 96.5

dT2229 98.5

dT19151 108.9dT22571 110.8dT36501 111.1dT11509 111.6dT4155 111.9dT1086 112.6

dT20499 112.8

S6_118469241 119.5dT14954 121.1dT39428 123.1dT19912 124.3dT53574 124.7dT43321 127.6dT34936 127.9dT2297 129.9dT8340 131.0

dC_234325 132.0dT20274 134.9dT22921 135.8dT30440 139.1dT14807 140.0dT30703 141.4dT6973 142.1

dC_196800 142.9dT18761 143.2dT36412 143.8dT24998 144.5dT25533 145.0dT29143 145.7

S6_55827481 146.2

dT43184 153.6C6_200342258 154.5

dT9665 155.1dT18516 156.7dT18539 157.3dT13751 158.5dT50350 160.4

dT22426 163.1

dT17747 168.9

dT45130 171.6

dT29238 173.2

dT13151 177.6

dT31763 184.3

qN

FE

-2-1

LG2

S4_192354592 0.0

dT43261 3.2

dT3062 5.4

dT20572 7.7

dT9434 9.3

dT50501 14.2

dT13383 16.9

dT25619 20.8

dT15831 23.7dT29846 26.5

dT20879 dT3030dT27514

27.1

dT40535 dT18538 27.3dT48489 27.7

C4_200698121 28.2dT4430 28.9

dT23119 29.7dT43805 30.0dT3891 30.9

dC_13461 31.7S4_116088663 33.3

dT13255 37.0dT32416 38.2

C4_90405411 39.4dT8320 39.9

S4_29656682 41.2dT42579 44.4dT4620 44.5dT1631 47.5

dT43233 52.7dT18870 53.3dT8454 53.8dT7751 55.4

dT29230 66.0dT4496 66.5dT6466 67.8

dT16151 68.9TS4_49707946 69.2

dT11273 70.1dT43759 70.7

S4_109511769 71.8TS4_79794887 76.3

dT36845 76.6dT33420 77.6dT25340 77.7

dC_124707 79.3dT31777 84.2

S4_17921468 85.4dT39406 85.9dT45232 86.6

dT4301 C4_13275000 87.4dT24376 89.5dT36267 89.7dT26600 90.2dT52456 91.1dT10193 91.7dT34107 92.0dT22232 92.6dT9256 94.2

S4_17796331 95.7

qN

FE

-6-1

qN

FE

-6-2

LG6

S2_174467255 0.0

S1_203215691 14.4

C2_174467255 dC_58121dC_58157

18.9

dT1325 20.5dT24339 21.7dT33835 22.9dT10459 23.8dT21208 24.3dT41386 24.5dT16455 26.4dT38126 26.9dT20154 27.2dT24498 28.0

dT50610 37.5

dT44347 41.9dT6436 42.3dT4353 43.7

dT5027 48.8dT23432 49.4dT39900 51.7

dT861 52.5dT23905 53.4dT42142 55.0

S7_227510 55.3

S7_161179045 S7_161179072 57.7

dT21588 60.1

dT14775 64.4

dT34305 66.6

dT5751 69.4

dT17613 77.8

dT7877 83.4

dT7992 88.3dT22362 88.7

S7_168784407 90.3S7_168784355 91.7

dT26372 98.5

S7_146917990 109.0

S7_100737594 114.2

dT46088 118.4

dT36827 120.7dT36253 121.3

C7_38012075 S7_52134705 124.9

S7_52134761 126.6

S7_52134764 128.5

dT25902 136.4dT43823 137.3

dT19286 145.8dT43692 146.0

S5_1337832 147.8dT23441 149.3dT28159 150.8dT9688 151.5

dT11814 156.0dT16180 157.3dT26922 157.7

S7_119895303 159.3dT8614 160.3

dT29758 161.1dT10112 162.1

dT6553 173.8

dT37830 178.0S7_66630588 179.5

dT39337 180.7

dT44668 183.3dT25362 183.6dT28445 183.8dT9471 185.2dT3618 186.5

qF

TE

-7-1

qF

TE

-7-2

LG7

108

Figure 3-11. Linkage map of female parent N190 showing potential QTLs. QTLs for number of flower (NF) are color coded in blue and that for flowering time (FT) are colored in green. A black bar means a GBS-SNP marker. Markers are labeled on the left and genetic distance (centimorgan as unit) are labeled on right. LG = Linkage group.

Linkage map of male parent N122

dT20387 0.0dT41575 3.7

dT1702 dT3027 4.6dT909 4.8

C1_10467974 5.2dT38795 5.4

S3_64306230 5.8dT3328 6.1

dT49252 6.4dT30456 6.5dT16886 6.7dT53891 7.0dT16764 7.2dT35147 8.3dT49010 8.6dT2473 8.9dT1951 9.6

dT16527 9.9C1_220011116 12.0

dT50264 20.4dT29470 21.6dT32431 22.3dT9029 22.5

dT50826 22.7dT47548 23.0

dT45924 dT7985dT24786 dT732

23.5

C1_102656250 23.6dT13338 23.7dT19473 23.8dT18485 23.9dT46948 24.3dT51201 24.5dT23838 24.8dT23364 25.2dT23237 30.6

dT28498 38.0dT51779 38.5dT43458 39.3dT6993 41.3

dT29597 42.0dT5196 43.9

dT14180 44.3dT48800 44.6dT3431 45.2

dT48497 45.3S1_17257463 45.8

dT9167 46.1dT28049 46.3dT48243 46.6

S1_17257559 46.9S1_17257551 47.2

dT18424 47.8dT973 48.2

dT8459 50.7dT33654 50.8dT48112 51.1dT53672 52.0dT12272 53.7

dC_359278 58.8C1_133709193 58.9

dC_339765 59.0S1_171369178 59.2

dT40041 59.7dT18064 61.2

dT971 61.9dT22725 62.2

dT10843 dT19448 62.3dT47057 63.4

C1_234266014 65.4TS1_273230575 66.0

dT30844 66.4dT7314 67.2dT766 68.5

dT36979 69.2S1_217117736 73.5

dT30297 75.1

dT3813 78.0

S1_241274562 87.0

qN

FL-1

-1

qF

TL-1

-1qF

TL-1

-2

LG1

S1_6210777 0.0

S5_128073888 4.6

S1_6210761 8.3dT24515 11.1dT17924 15.6

TS5_114098156 16.0dT16773 16.2dT40995 16.5dT15817 16.8dT14651 18.3dT18480 18.8

TS5_128053793 dT13827C5_156998414 C5_112498054

dC_2426719.1

dT27345 dT30788 19.2dT9315 19.4

S5_40839446 19.6dT4737 20.2

dT31217 22.1dT32027 22.2dT46427 22.4dT32695 23.0

dT43636 29.3

dT31735 33.3

dT50145 38.1

dT43048 41.0C5_132196063 41.2

dT20788 41.7dT32207 41.9dT24792 43.5

dT46652 46.4

dT27717 49.6dT15391 50.1dT1524 51.2

dT14375 53.4dT31410 54.0dT31722 54.3

dT18219 59.0dT36949 59.7

dT24276 63.6

dT7856 72.9

dT5489 76.2

dT43892 77.6

dT48559 82.4

dT42534 91.8

dT49410 93.8

dT52149 94.9

dT32206 99.6

dT6664 101.1

dT8114 104.1

dT53440 105.2

dT40421 108.2dT14806 108.5dT22647 108.9

S5_31904873 109.7dT39799 110.2dT3053 110.6

S5_26740824 111.4dT4924 111.7

dT37762 113.4

dT26721 118.4dT26387 119.1dT6733 119.8

dT17794 120.1S5_16898738 120.9

dT37016 121.4dT47995 123.0dT34310 123.7dT51527 124.4

TS5_13509703 124.6dT54119 125.1dT53202 125.4dT32125 126.4dT27794 126.6dT2435 127.9

dT46426 129.8dT20384 130.1

C5_3132406 131.6dT15117 131.9dT29259 132.8

dT780 133.3dT15952 134.3dT26168 135.2dT36121 137.1dT32758 139.0

dT30470 142.4

qN

FL-4

-1qN

FL-4

-2

LG4

dT49930 0.0

S4_131830044 G4_131830067 6.6S4_131830067 6.7S4_131830066 7.2S4_131830063 7.3

dT22876 8.3dT12269 9.4

C4_131830077 9.5dC_61912 dC_55022 9.7

dT22567 10.3dT26116 10.7dT38883 11.0

dT36340 dT10040 11.3dT26884 11.6

dT10739 dT621 11.7dT17325 S7_89250999 11.9

dT17578 12.0dT25711 12.1dT5186 12.3dT8749 12.6

dC_10270 13.2dT649 13.7

dT53247 13.9dT53022 dT12194 14.0

S7_39646339 15.6S7_39646309 15.9S7_39646292 16.2

dT33677 16.7dT3510 17.3

dT19298 17.9dT1762 21.7

dT41196 22.3dT30018 22.7dT40278 23.3

S7_83881022 23.7dT7491 25.5

dT38871 28.1dT29364 35.0

dT8910 50.9

dC_223673 59.4

dT43632 63.3dT34571 64.0dT3675 64.6

S7_213184560 64.9dT22858 65.4dT32036 66.2dT29409 67.8

dT19025 71.6dT7711 72.5

dT39297 74.1

dT42995 78.6

dT47137 80.2

dT37923 90.6

qN

FL-1

1-1

qN

FL-1

1-2

LG11

109

Table 3-1. Descriptive statistics of flowering date and number of flowers for 185 F1

hybrids of a cross (N190 N122) at Citra, FL., in 2012, 2013 (Sinche 2013) and 2016.

Statistic Flowering datea Number of flowers (plot-1) b

2012 2013 2016 2012 2013

Min 10/25 10/26 10/23 0 5

Max 12/06c 12/03 12/02 46 256

Mean 11/15 (240) d 11/13 (238) d 11/07 (232) d 6 61

SE 1 d 0.5 d 0.05 d 0.76 2.86

H2 0.60 0.87 0.64 0.85 0.79

a Flowering date: mm/dd

b 1.8 m long plots.

c The harvest concluded before the first predicted frost on 12/06/2012 and several genotypes did not flower until that day.

d Value in parenthesis indicates number of days to first flowering counted from March 20 of that year.

SE Standard error H2 Broad sense-heritability estimated on entry mean basis

110

Table 3-2. Detailed information about the QTLs for number of flowers. LG = Napiergrass linkage groups, Peak markers represent the marker at the highest peak of QTL, LSI represents the LOD-1 support interval in cM.

Parent QTL code LG Year Peak markers Peak LOD PVE

(R2)

Allele

dir.

LSI (cM) Flanking Markers

N122 qNFE-1-1 1 2012, 2013 dT37356 8.13 0.15 +

116.9 cM - 118.2

cM dT52469, dT35128

N122 qNFE-1-2 1 2012, 2013 S1_249594415 10.76 0.20 +

123.7 cM - 126.8

cM dT36338, dT5535

N122 qNFE-1-3 1 2012, 2013 dT35453 9.10 0.17 -

133.5 cM - 134.6

cM

S1_260924412,

dT45297

N122 qNFE-2-1 2 2013 TS6_10754586 3.48 0.08 + 3.1 cM - 4.8 cM dT38576, dT10631

N122 qNFE-6-1 6 2012 S4_17921468 3.64 0.06 + 84.3 cM - 86 cM dT31777, dT45232

N122 qNFE-6-2 6 2012 dT10193 3.45 0.06 + 91.1 cM - 91.9 cM dT52456, dT34107

N190 qNFL-1-1 1 2013 dT3328 4.04 0.10 + 5.6 cM - 6.7 cM dT38795, dT53891

N190 qNFL-4-1 4 2012 S1_6210761 3.66 0.07 + 7.6 cM - 8.8 cM

S5_128073888,

dT24515

N190 qNFL-4-2 4 2012 dT40995 4.78 0.10 + 16.2 cM - 16.7 cM dT16773, dT15817

N190 qNFL-11-1 11 2013 dT8910 4.22 0.08 + 47.5 cM - 52.9 cM dT29364, dC_223673

N190 qNFL-11-2 11 2013 dT43632 3.18 0.06 - 57.9 cM - 64.2 cM dT8910, dT3675

111

Table 3-3. Detailed information about the QTLs for flowering time. LG = Napiergrass linkage groups, Peak markers represent the marker at the highest peak of QTL, LSI represents the LOD-1 support interval in cM.

Parent QTL

code

Linkage

Group

Year Peak

markers

Peak

LOD

PVE (R2) Allele dir. LSI (cM) Flanking Markers

N122 qFTE-1-1 1 2013 dT53136 6.77 0.14 +

118.3 cM - 120.8

cM dT35128, dT39139

N122 qFTE-1-2 1 2013 dT48305 5.40 0.11 -

126.5 cM - 128.3

cM

S1_249594415,

dT32706

N122 qFTE-7-1 7 2012 dT44347 3.91 0.09 - 31 cM - 42.1 cM dT24498, dT6436

N122 qFTE-7-2 7 2016 dT39337 5.57 0.12 -

178.3 cM - 182.1

cM dT37830, dT44668

N190 qFTL-1-1 1 2013 dT41575 3.60 0.07 - 3.3 cM - 4.6 cM dT20387, dT3027

N190 qFTL-1-2 1 2013 dT1951 3.33 0.07 - 8.9 cM - 11.2 cM

dT2473,

C1_220011116

112

Table 3-4. List of putative flowering related genes from the genome of pearl millet (Varshney et al. 2017). Gene name Description Gene ID Chr Start position End position Exon

count QTL

AGAMOUS AGAMOUS-like protein; K09264 MADS-box transcription factor

Pgl_GLEAN_10032931 chr1 268933236 268935744 7 qNFE-1-3

DELLA DELLA domain GRAS family transcription factor rga-like protein; K14494 DELLA protein


DEFICIENS Floral homeotic protein DEFICIENS, putative; K09264 MADS-box transcription factor


PPM1 PPM1; MIKCC MADS-domain protein PPM1; K09264 MADS-box transcription factor


WRKY K13424 WRKY transcription factor 33


SERK1 SOMATIC EMBRYOGENESIS RECEPTOR-LIKE KINASE 1; kinase/ transmembrane receptor protein serine/threonine kinase


113

CHAPTER 4 EVALUATE THE GENETIC BACKGROUND OF FLOWERING TIME IN A

NAPIERGRASS GERMPLASM COLLECTION

Introduction

Napiergrass (Cenchrus purpureus) is an important forage and biofuel candidate

(Singh, Singh, and Obeng 2013). However, there has been relatively little effort in

improving napiergrass as compared to cereal crops for increased yield and agronomic

characteristics (Wanjala et al. 2013). Napiergrass cultivation in the US is limited by its

potential for invasiveness (Sollenberger et al. 2014). Typically, napiergrass is very

persistent and only few diseases like napiergrass stunt disease and napiergrass head

smut disease have been reported with significant impact on biomass yield or crop

persistence (Farrell, Simons, and Hillocks 2002).

Existing genetic diversity of napiergrass needs to be assessed in order to

strengthen breeding programs for development of high yielding and late flowering

cultivars that can be considered as non-invasive (Wanjala et al. 2013; Sinche 2013).To

initiate this, there is a need to have accurate information on the available germplasm to

identify clones that form the base population for breeding late flowering lines. The first

step of breeding for flowering time is to create variability within the germplasm so as to

widen the base population of the candidate clones for selection (Juma 2014). Variability

can be due to intrinsic reasons or due to random genetic drift, natural selection,

mutation, gene flow and transfer (Y. Xu et al. 2017). This variability can be determined

by assessing the level of polymorphism between accessions and within each accession.

This information can be used for clone development. To determine the variability of the

available germplasm, stable and informative molecular markers are critical to study the

114

genetic diversity. Molecular characterization of germplasm will increase breeding

efficiency by providing important genetic information of the breeding materials and will

ultimately help in identification and introgression of desirable agronomic traits into high

yielding clones through marker assisted selection (MAS). This will lead to increased

napiergrass yield with decreased invasiveness.

Napiergrass cultivar discrimination has been mostly done based on

morphological and agronomic characteristics that lead to inconsistency in identification

of specific accessions (Struwig 2007). Consequently, many cultivars of napiergrass are

described with more than one name/identifier. Marker informed elimination of

redundancy in collections of napiergrass accessions will facilitate the maintenance of

the collection and its use in crop improvement programs. In this regards, molecular

markers have proven to be very effective to distinguish between morphologically related

individuals of the same species (Wanjala et al. 2013).

Hence, genetic assessment of napiergrass accessions with molecular markers

was initiated several decades ago. Characterization of different napiergrass

germplasms has been carried out using Restricted Fragment Length Polymorphism

(RFLP) (Smith et al. 1993), Random Amplification of Polymorphic DNA (RAPD) (Lowe

et al. 2003), Amplified Fragment Length Polymorphism (AFLP) (Harris-Shultz,

Anderson, and Malik 2010; Wanjala et al. 2013), inter-simple sequence repeats (ISSRs)

(de Lima et al. 2011), sequence-related amplified polymorphism (SRAP) (X.-M. Xie et

al. 2009), and simple sequence repeat (SSR) markers (Azevedo et al. 2012; López et

al. 2018; López et al. 2014). Currently, single nucleotide polymorphism (SNP) markers

are gaining popularity as they are ubiquitous in the genome, occur in a large number

115

with high density, relatively inexpensive, codominant, and can act as functional markers

if located in coding regions (Rafalski 2002; Ganal, Altmann, and Röder 2009). Further,

genome wide SNP markers can be used as diagnostic tools for fingerprinting

germplasm, to correct discrepancies that are inherent in taxonomic methods based on

morphological characteristics and to estimate genome composition of clones (Yang,

Song, et al. 2018).

Flowering time is very important to plant breeders because agronomic traits like

biomass yield and quality depend on the timing and intensity of flowering (Veeckman et

al. 2016). Flowering time is a key factor in plant adaptation and is linked to various

attributes like plant height, yield, and number of leaves (Durand et al. 2012). In many

species, flowering is induced in response to day length. Different flowering responses

are categorized as short-day, long-day, intermediate-day, or day-neutral based on the

day length requirement (Schlegel 2009; Bastow and Dean 2002). Napiergrass belongs

to the short-day plants (Osgood, Hanna, and Tew 1997; Singh, Singh, and Obeng

2013), in which flowering is favored by day lengths shorter than the corresponding

nights.

Early flowering in perennials limits vegetative growth, and delayed flowering will

hinder seed development to progress completely before the onset of adverse weather

(Grabowski et al. 2016). Delaying flowering could potentially lead to increases in

biomass yields as seen in other grasses like switchgrass where a 10 days delayed

flowering increased biomass by > 25% (Price and Casler 2014). Flowering time is an

important target trait in napiergrass breeding programs because different flowering time

helps to adapt cultivars in different geographic locations or seasons. It is also important

116

because flowering has been associated with invasiveness of napiergrass in Florida

(FLEPPC 2011). High number of wind-dispersed seeds has contributed to categorizing

napiergrass as an invasive species in Florida, thus limiting its utilization as an industrial

crop for forage or biofuel (López et al. 2014). For this reason, development of cultivars

that are late flowering is critical for napiergrass management in the field because late

flowering lines can be harvested before flowering, as development of viable seeds is

impaired by low temperatures in Florida during Nov-Feb. Studying flowering time

variation in napiergrass can help to identify potential genes that promote or hinder seed

formation in napiergrass. Mining genome sequences that are available for several grass

species for flowering related genes and their characterization can also identify

candidate genes. These candidate genes can be investigated in different populations for

late flowering. In addition to this, integrating genomics with conventional breeding will

shorten the breeding cycle for selection.

The substantial variations in flowering time within the population of napiergrass

can be exploited to not only select lines with desirable traits but also to identify

sequence variations that are associated with it or could be causal. By characterizing the

relationships between genetic variations and flowering time we might be able to identify

loci affecting flowering time variation in napiergrass. Flowering time can be controlled by

a large number of quantitative trait loci (QTLs) with small effects (Buckler et al. 2009).

Genetic and genomic approaches that have been used in other species to characterize

flowering time genes might not be feasible in napiergrass due to its large genome size

and high polyploid level. This challenge can be mitigated by using genome reduction

approaches like exome-capture sequencing methods (Evans et al., 2014). Exome

117

sequencing targets only genic regions of the genome with high depth and can produce

gene-level resolution of genome-wide patterns of sequence variation (Grabowski et al.

2016).

To acquire a global perspective of sequence variations and to obtain an accurate

population structure of napiergrass germplasm, we deeply sequenced the coding

regions of 94 accessions from the Tifton nursery. Then we used genome-wide SNPs

identified to evaluate the genetic relationships of the germplasm as well as performed

genome wide association studies targeting flowering time in napiergrass.


Plant Materials and Phenotyping

A total of 94 napiergrass accessions from the germplasm collection available at

the USDA-ARS, Tifton nursery were used in this study. Nodes from those accessions of

napiergrass were cut and planted in 6” pots on August 19, 2015 in Everglades

Research and Education Center (EREC), Belle Glade, FL. Transplanting of the pots

was done on October 12, 2015. Plots were established in a randomized complete block

design with three replications and spacing of 4 ft x 4 ft plant to plant as well as row to

row distance were maintained. Emergence of the first flower was noted every week

during October – December 2016. Days to flowering was calculated as the number of

days between the first appearance of the flower and vernal equinox (March 20)

(Lambert, Miller-Rushing, and Inouye 2010).

DNA Extraction

Young and healthy leaf tissues were harvested from each individual of the

germplasm collection. DNA extraction was done following the protocol described

118

previously (Dellaporta, Wood, and Hicks 1983). The extracted DNA samples were run

on a 2% agarose gel to check the quality and quantified with PicoGreen to meet the

requirements of exome sequencing.

Targeted Candidate Genes

Genes related to flowering in different monocots and other species were mined

using various publications (Table 5-1). In addition, Refseq database was searched for

“flowering” and filtered for “Green plants” to get potential flowering related genes (Pruitt,

Tatusova, and Maglott 2007). All gene models present in the pearl millet genome

(Varshney et al. 2017) were used for probe design to achieve a genome-wide coverage

of all coding sequences (CDS).

Probe Design

The protein sequences for flowering genes were clustered using CD-HIT

(Weizhong Li and Godzik 2006). Probes of 120 bp long with an overlap of 60 bp were

developed using Emboss 6.5.7 (Rice, Longden, and Bleasby 2000). These probes were

mapped back to the pearl millet genome (Varshney et al. 2017) using BLAT (Kent

2002). A hit was defined under cutoff: e-value ≤ 1e-05; alignment identity = alignment

length * percentage of identity ≥96 (120 bp * 80%=96 bp). Probes that hit the genome

less than three times were selected and probes containing repeats were removed. 2kb

promoter region and 500 bp downstream region of the flowering genes were also

included for probe design with the same criteria.

A relatively relaxed criteria was chosen to target missing flowering time related

genes. For this, 120 bp long probes were designed with 90 bp overlap and good probes

were selected with the criteria (Alignment * Identity) / 120 >= 80. The probes that hit the

119

genome less than 4 times were selected and probes containing repeated regions were

removed.

Probes were also designed to target the coding regions of the whole genome.

For this, sequences of 38,579 genes of pearl millet were used to retrieve 120 bp probe

sequence with maximum allowable overlap of 30 bp. Good probes were selected under

the criteria alignment * identity / 120 >= 80. Probes with multiple hits on the genome or

containing repeat regions were removed. A maximum of 2 probes were kept for each

CDS.

A relaxed criterion was implemented for genes, where no probes were designed

according to the standards above. For those genes, probes were designed with 90 bp

overlap and good probes were selected as alignment * identity /120 >= 80. Probes with

up to 2 hits on the genome were kept and probes containing repetitive regions were

discarded.

Probe Synthesis, Selection, and Sequencing

The designed probes were submitted to RapidGenomics LLC (FL, USA) for

quality check. Probes that passed quality criteria of RapidGenomics were then selected

using the following criteria: for flowering related genes, a total of 18 probes were

targeted per gene while for genome wide CDS, 1 probe per gene was targeted. Four

probes were targeted per gene for 2 kb promoter region and 1 probe per gene was

selected for 500 bp downstream region of the gene. Finally, a total of 37,000 probes

were submitted for synthesis. The synthesized probes were used to capture the DNA

fragments of the germplasm collection. The captured DNA fragments were sequenced

using the Illumina HiSeq 3000 platform (150 bp paired-end reads). The probe synthesis,

120

library preparation, target enrichment, and sequencing were performed by

RapidGenomics LLC (FL, USA).

Sequence Read Trimming and Mapping

Raw reads were trimmed using Trimmomatic (Bolger, Lohse, and Usadel 2014).

Reads that contained more than 50% low-quality bases (Q20) were removed and

adaptor sequences were trimmed. Trimmed reads were aligned to the pearl millet

genome v1.1 (Varshney et al. 2017) using BWA-mem (H. Li and Durbin 2010). For SNP

calling, we retained only uniquely mapped reads with a combination of mapping quality

‘0’ and ‘XA:Z’ tag (both indicates multiple alignment).

SNP Calling

SNPs were called using GATK 3.2.2 (McKenna et al. 2010) with settings (-T

UnifiedGenotyper -glm BOTH -R -ploidy 4 -mbq 20). For each sample, homozygous

SNPs required a depth of ≥11; heterozygous SNPs required a depth of ≥31; and at least

2 reads for minor allele were required to make a call. At the population level a mapping

quality of ≥30; minor allele frequency ≥0.05; and call rate of ≥95% was employed.

Population Structure

In order to infer population structure of the germplasm accessions, we used

DAPC available in the adegent package for R (Jombart 2008). Groups were clustered

using k-means clustering, where the best k minimized the Bayesian Information

Criterion (BIC). GWASpoly (Rosyara et al. 2016) was used with six different models for

GWAS using sequence variants that had a call rate ≥ 95% and MAF ≥ 0.05. Molecular

Evolutionary Genetics Analysis version 6.0 (Tamura et al. 2013) was used to infer

genetic relationships among the germplasm collection and to perform phylogenetic

121

analysis. Less than 5% alignment gaps, missing data, and ambiguous bases were

allowed at any position. Neighbor-Joining method was used to infer the evolutionary

history. The evolutionary distances were computed using the p-distance method and

are in the units of the number of base differences per site (Tamura et al. 2013).

Results

Candidate Genes Related to Flowering

A total of 3,542 flowering time related genes were curated from different

organisms including Arabidopsis, Brachypodium, maize, rice, Setaria, sorghum and

switchgrass (Table 4-1). This list included 1,285 flowering related genes retrieved from

the Refseq database (Pruitt, Tatusova, and Maglott 2007). CD-HIT clustered the

flowering genes into 1,279 non-redundant genes that were further compared to peptides

from the pearl millet genomes using BLAT. In pearl millet, 506 genes had BLAT hits to

the non-redundant flowering amino acid sequences. Thus the 506 genes were targeted

for the probe design.

Probe Design

The CDS sequences of 506 ortholog genes in pearl millet related to flowering

were subjected to probe design. A total of 8,314 probes were designed and 4,908

probes targeting 473 pearl millet flowering time related genes remained after filtering.

The number of probes designed per gene ranged from 1-91 (Figure 4-2) and the

number of probes designed per gene was proportional to the size of the gene (Figure 4-

3). To further cover the genes without probe designed, the selection criteria were

relaxed (see Materials and Methods) and an additional 9,099 probes were designed

targeting 422 additional genes.

122

We extracted 2kb region upstream of the flowering genes to design probes for

the promoter regions. A total of 3,215 probes were designed targeting promoter regions

of 243 pearl millet genes. We further extracted 500bp downstream of the flowering

genes and designed probes from the terminator regions. A total of 1,025 probes were

designed from the terminator regions targeting 239 flowering related genes.

Probes were also designed to target the coding regions of the whole genome.

Thus the sequences of 38,579 gene models annotated in the pearl millet genome v1.1

(Varshney et al. 2017) were included to design genome wide probes. The length of the

genes ranged from 100 bp to 5600 bp (Figure 4-1), with most of them in the range of

500 bp -1600 bp. A total of 53,525 probes targeting 29,015 out of 38,579 genes were

designed. For the genes that couldn’t be targeted, we relaxed selection criteria and

further design 7,306 probes that targeted additional 4,318 genes (Materials and

Methods).

Thus, a total of 78,988 probes were totally designed and were submitted for

quality check, after which 63,967 probes passed the quality test provided by Rapid

Genomics. From these probes, a total of 37,000 unique probes were further selected for

synthesis and sequencing. This final probe set targeted a total of 818 flowering related

genes with a maximum of 18 probes per gene. These genes were distributed on all

chromosomes of the pearl millet genome (Figure 4-4). The probe set also targeted a

total of 28,519 genome wide CDS with 1 probe per CDS. Promoter region of 240 genes

were targeted with a maximum of 4 probes per gene and terminator region of 231 genes

were targeted with a maximum of 1 probe per gene.

123

Sequence processing

A total of 375,135,615 paired-end reads (2×150 bp) were generated by Illumina

sequencing. The number of paired-end reads per sample ranged from 2.3 M to 5.6 M,

with an average of 3,990,804 read pairs per sample (Figure 4-5). A total of 354,016,510

(94.35%) read pairs passed the quality control. From the good quality reads, a total of

20,901,032 raw SNPs were called by GATK (McKenna et al. 2010). After filtering for

quality, depth, minor allele frequency (MAF), and call rate, a total of 78,129 high quality

SNPs remained that were subjected to genome wide association analysis using

GWASpoly (Rosyara et al. 2016).

A total of 2,893 SNPs (one SNP per 10 Kbp) were selected to perform DAPC.

Based on the Bayesian Information Criteria (BIC), the germplasm could be clustered

into 3 clusters (k=3) (Figure 4-6). Using these SNPs, the germplasm could be projected

into three linear discriminants (LD)s (Figure 4-7) that corresponded to their geographic

origin. Group 1 had accessions originating from Africa, group 2 had accessions that

originated from America (central and south America), and group 3 had a mix of lines

from Asia and America. However, three accessions whose origin were in Taiwan (Asia)

were grouped into the same group as accessions originating from Africa (group 1).

Seven accessions whose origin was listed as America were also grouped into the

African group 1. Group 2 exclusively contained accessions originating in America.

Group 3 contained accessions from the Americas and Asia, with one exception that

came from Africa.

124

Phylogenetic analysis

The phylogenetic analysis involved 94 samples in MEGA6 by using 2,872 SNPs

(one SNP every 10 Kbp). Positions that had less than 95% site coverage were

removed. Evolutionary analyses showed that most of the lines could be assigned based

on the geographic locations, however, there were some mixtures (Figure 4-8). Group 1

mostly contained lines that originated in Africa, while group 2 contained lines with origin

from America. Many accessions in group 2 had unknown origin. Group 3 included a mix

of African, Asian, and American lines. Asian lines were grouped in group 1 and group 3.

Genome wide association analysis

Average days to flowering in the germplasm collection ranged from 232 days to

279 days (Figure 4-9). The heritability for days to flowering was 0.58. All the high-quality

SNPs after filtering (78,129) were used for genome wide association analysis by using

GWASpoly with six different models (Figure 4-10, Figure 4-11). For every model, the

majority of p-values lie close to the 1:1 dashed line (Figure 4-10). For days to flowering

(DTF), the strongest association signal was observed with the 1-dom-alt model (Fig 4-

11) for a marker on Chromosome 5 at location 129,365,754 with an estimated marker

effect of -16.79. Other models didn’t give significant associations (Fig 4-11).

Subsequently, we explored the region 100kb upwards and 100kb downwards of this

marker to identify potential candidate genes. Ten gene models were located within this

region and we identified potential flowering candidate genes including calcium-binding

protein CML and enoyl-CoA hydratase in this region.

125

Discussion

Dissecting genetic loci controlling flowering time is critical for napiergrass

breeding as flowering is related to invasiveness and biomass accumulation. In this

research, we applied TES for high throughput genotyping of tetraploid napiergrass. We

were able to investigate sequence variants using TES which reduced genome

complexity by selectively sequencing only the coding regions. After applying quality

filter, we identified 78,129 high quality SNPs in the germplasm collection that serve as

valuable and novel genomic resource for napiergrass breeding programs. Moreover, we

performed GWAS in the napiergrass germplasm collection using these SNPs and

identified one SNP as well as two candidate genes that was associated with flowering

time.

We applied DAPC to analyze the population structure of the napiergrass

germplasm collection. DAPC is a multivariate method to analyze genetic structure and

does not rely on Hardy-Weinberg equilibrium assumptions or linkage disequilibrium

assumptions (Jombart 2008). We selected 2k high quality and evenly distributed SNPs

for the analysis and inferred that three groups existed in the germplasm collection. This

is in contrast with results from (Kandel et al. 2016) who inferred 5 groups in a subset of

this population. Those results were based on only 29 SSR markers in contrast with our

study that used 2k uniformly distributed SNPs. In this study, we had 24 lines that were

not used in the previous study (Kandel et al. 2016). Out of these, 14 lines were of

unknown origin. Our study grouped accessions N43 and N51 into the same group

similar to previous study that grouped biofuel type and naturalized napiergrass into

126

different groups (López et al. 2014). However, our results grouped N13 in a separate

group compared to N43 and N51.

We identified a total of 20 million raw SNPs and compiled a comprehensive

sequence variation database for napiergrass that included 78k high quality SNPs. This

dataset provides accurate genotypes for the germplasm collection and serves as the

foundation for population genomics analyses in napiergrass, like gene mapping and

GWAS. We were able to identify potential candidate variations using GWASpoly

(Rosyara et al. 2016), which is primarily designed for autopolyploid species, however, it

has also been successfully used in diploids, as well as allopolyploids like wheat (Phan

et al. 2018).

The regulation of flowering time is important for plant adaptation and biomass

production. In this study, we used genotypes derived from exome-capture sequencing

and flowering time information from a germplasm collection of napiergrass to

characterize the genetic architecture underlying flowering time regulation in napiergrass.

This is the first GWAS study performed in napiergrass. Recently, QTL mapping was

conducted in a bi-parental mapping population of 185 progenies with multiple QTLs

detected for flowering time and number of flowers (Chapter 3). We detected one

significant QTL in this study on chromosome 5 of the pearl millet genome, which

corresponded to linkage group 4 of napiergrass, where two QTLs (qNFL-4-1 and qNFL-

4-2) for number of flowers were identified in the previous study (Chapter 3). Other QTLs

identified in Chapter 3 were not identified by GWAS. This could be explained by the

type of population studied (germplasm collection vs. bi-parental population), statistical

method employed (GWAS vs. linkage mapping) and strict threshold to declare

127

significance (p<0.05 after Bonferroni correction vs. logarithm of odd <3) (Yang, Islam, et

al. 2018). In the 200kb region surrounding the most significant marker, we identified two

genes related to flowering. Calcium-binding protein CML cause alterations in flowering

time and can act as a switch in response to day length perception and the gain of CML

function results in early flowering (Tsai et al. 2007). Similarly, enoyl-CoA hydratase is

important for seed germination (Richmond and Bleecker 1999). Other potential

flowering candidates like phytochrome B and AP2-like factors were located at 150 kb

and 165 kb respectively away from the significant marker. The markers and genes

associated with flowering time could be fully utilized in napiergrass breeding programs

after they have been validated in a much larger population.

Identifying genes that are responsible for trait variation is a major challenge (Gelli

et al. 2016). Due to the unavailability of transcriptome sequences for napiergrass, it will

be difficult to pinpoint the underlying genes responsible for trait variation. Future

evaluations could lead us towards potential targets for improving flowering time in

napiergrass and related crops. Currently there is no report that has utilized TES in

napiergrass and we have made this first attempt to use exome sequencing for calling

variations in napiergrass for flowering related genes as well as performed GWAS.

GWAS in polyploids is challenging because most studies are based on diploid genetics

and software that are modeled using diploid genetics (Bourke et al. 2018). While there

is no software available specifically to perform GWAS in allotetraploids, recently, two

software were released that accept polyploid data, namely GWASpoly (Rosyara et al.

2016) and SHEsisPlus (Shen et al. 2016). In this study, we found 1 significant SNP

using GWASpoly. Modeling GWAS studies based on polyploidy as compared to diploidy

128

can result in identification of more associated SNPs (Ferrão et al. 2018). Therefore, the

selection of a proper model is critical in GWAS. In the future a software that is

particularly modeled for allopolyploidy might become available and may help to identify

more SNPs associated with the trait. Moreover, we were also limited by the size of the

population. Our population of 94 accessions might not have enough power to detect the

variations in flowering time, which is mostly controlled by minor effect QTLs (Dell’Acqua

et al. 2015). In addition, phenotyping may lack accuracy since it was only completed for

one year and at one location. These factors might have played a role in the limited

number of SNPs identified for the flowering time. Phenotyping the germplasm collection

at multiple locations and for multiple years may allow to identify significant associations

to the traits of interest. Further, genotyping more accessions of napiergrass germplasm

could also help in identifying other important variations that were not detected in this

study. In addition, this germplasm collection can also be phenotyped for other traits like

those related to biomass accumulation to associate these traits with sequence

variations and potential candidate genes.

Probe design is critical for improving the performance of TES for deep

sequencing. Probes can be designed from transcribed sequences (expressed sequence

tags or RNA-seq reads) or genomic sequences (whole genome sequencing, reduced

representation library, genotyping by sequencing) of the species of interest or related

species. We used the genome of a closely related species of napiergrass, pearl millet

(Varshney et al. 2017), to design probes for this study. Using 37k probes, we targeted

818 flowering genes, and 28,519 overall genes according to pearl millet genome.

129

Since napiergrass is allotetraploid with two similar sub-genomes, most of the

probes designed for this study might have multiple hits. With the unavailability of

reference genome of napiergrass, we are unable to separate the probes that could have

come from each sub-genome. While TES provides high coverage data, some non-

primary targets could also be captured, which is unavoidable for organisms that do not

have a reference genome (Peng et al. 2017).

Conclusion

Our study showed the feasibility to applying targeted enrichment sequencing with

probes designed from CDS of pearl millet genome to target the genomic regions in

napiergrass. We inferred the structure of the germplasm collection as well as

constructed phylogeny for the germplasm. TES allowed us to perform a GWAS on this

population, and we identified one significant SNP and two candidate genes for flowering

time variation. The success of this approach could be improved by increasing the

number of accessions studied, phenotyping for multiple years at multiple locations, as

well as using the reference genome of napiergrass once it becomes available. With the

availability of the genome reference, we can assess uniqueness of probes that will

increase unique mapping rate as well as minimize the proportion of off-target capture.

130

Table 4-1. Genes and publications related to flowering used in this research.

Organism Genes Reference

Arabidopsis 306 (Bouché, Lobet, Tocquin, & Périlleux, 2016)

Brachypodium 23 (Higgins, Bailey, & Laurie, 2010)

Maize 692 (Li et al., 2016)

Maize 148 (Xu et al., 2012)

Rice 201 Wiki pathways

Rice 39 (Lee & An, 2015)

Setaria 53 (Mauro-Herrera et al., 2013)

Sorghum 141 (Mace, Hunt, & Jordan, 2013)

Switchgrass 654 (Grabowski et al., 2016)

mRNA 1,285 Refseq db: flowering + greenplants

Grand Total 3,542

131

Figure 4-1. Histogram of length of flowering related genes.

0

10

20

30

40

50

60

70

80

100

-199

300

-399

500

-599

700

-799

900

-999

110

0-1

19

9

130

0-1

39

9

150

0-1

59

9

170

0-1

79

9

190

0-1

99

9

210

0-2

19

9

230

0-2

39

9

250

0-2

59

9

270

0-2

79

9

290

0-2

99

9

310

0-3

19

9

330

0-3

39

9

350

0-3

59

9

370

0-3

79

9

390

0-3

99

9

410

0-4

19

9

430

0-4

39

9

450

0-4

59

9

470

0-4

79

9

490

0-4

99

9

510

0-5

19

9

530

0-5

39

9

Num

be

r o

f ge

ne

s

Length of gene (bp), n=1213

132

Figure 4-2. Number of probes designed per gene.

0

10

20

30

40

50

60

0 20 40 60 80 100

Nu

mb

er

of

ge

ne

s

Number of probes designed per gene

133

Figure 4-3. Number of probes designed as a factor of the size of the gene.

134

Figure 4-4. Distribution of the targeted flowering genes in the genome of pearl millet. Each blue bar represents a flowering gene mapped on the pearl millet genome.

135

Figure 4-5. Number of paired-end reads per sample.

0

1,000,000

2,000,000

3,000,000

4,000,000

5,000,000

6,000,000

Num

be

r o

f re

ad

pa

irs

Sample

136

Figure 4-6. Bayesian Information Criteria (BIC) vs. number of clusters in k-means clustering suggests K=3 in the germplasm collection.

5 10 15 20

42

04

30

44

04

50

Value of BIC

versus number of clusters

Number of clusters

BIC

137

Figure 4-7. Projection of the napiergrass germplasm collection using first two linear discriminants (LDs) from discriminant analysis of principal components (DAPC). The shape of the points represent grouping by DAPC (circle = group 1; triangle = group 2; square = group 3) and the colors represent the origin continent: black = Americas; blue = Asia; and cyan = Africa.

-4

-2

0

2

-2.5 0.0 2.5 5.0

LD1

LD

2

138

Figure 4-8. Evolutionary relationships of taxa. G stands for group assigned by DAPC and are in different shapes (circle = group 1; triangle = group 2; square = group 3) and the fill colors represent the origin: black = Americas, blue = Asia, cyan = Africa, and white=unknown).

N43 G

2N

228 G

2

N51 G

2N

199 G

2

N210 G

2N

222 G

2N

203 G

2N

-Mer G

2N

212 G

2N

198

G2

N8

G2

N11

4 G

2N22

3 G

2

N21

4 G

2

N215 G2

N116 G2

N204 G2

N127 G2

N211 G2

N225 G2

N-6X G2

N205 G2

N128 G2

N129 G2

N12 G2

N132 G2

N74 G3N130 G3N9 G3

N14 G3N37 G3N239 G3N122 G3N168 G3N16 G3N20 G

3N172 G

3

NILR

I-16791 G1

N40 G

3

N35 G

3

N36 G

3

N34 G

3

N42 G

3

N39 G

3

N41 G

3

N32 G

3

N23 G

1

N24 G

1

N66 G

1

N56 G

1

N67 G

1

N68 G

1

N69 G

1N243 G

1

N244 G

1

N242 G

1

N240 G

1

N147 G

1

N17

9 G

1

N13

1 G

1

N15

2 G

1

N16

3 G

1

N17

1 G

1N164 G

1N181 G

1N183 G

1N170 G1N186 G1N182 G1N185 G1

N178 G1N137 G1N138 G1

N150 G1

N166 G1

N151 G1

N180 G1

N161 G1

N71 G1

N72 G1

N155 G1

N188 G1

N226 G1

N238 G1

N19 G1

N22 G1

N7 G

1

N70 G

1N

13 G1

N109 G

1N

ILRI-16786 G

1

N157 G

1N

190 G

1

NIL

RI-1

4984 G

1

139

Figure 4-9. Histogram for days to flowering trait in napiergrass germplasm collection.

0

5

10

15

230 240 250 260 270 280

Days to flowering

140

Figure 4-10. QQ plots for different models using GWASpoly using 78,129 SNP markers. DTF = Days to flowering.

141

Figure 4-11. Manhattan plots for different models using GWASpoly. P values adjusted with FDR at 0.05. DTF = Days to flowering.

142

CHAPTER 5 GENERATION OF INTERSPECIFIC HYBRIDS BETWEEN PEARL MILLET AND

NAPIERGRASS AND EVALUATION OF THEIR PERFORMANCE

Introduction

Napiergrass or elephantgrass (Cenchrus purpureus Schumach) is an important

forage crop that is widely used in Africa for dairy cows because of its high yields and

nutrient value (Singh, Singh, and Obeng 2013). Mature napiergrass reach plant heights

of 5-6 m and up to 20 nodes per stem (Boonman 1988) and out-yield other grasses by a

significant margin (Ra et al. 2012). Napiergrass is generally propagated via stem

cuttings. Clonal propagation increases propagation costs as cultivation is typically done

manually. Therefore, even cultivars of napiergrass with high biomass quality like Mott,

which produced average cattle gains of 0.97 kgd-1 compared to industry standard of

0.39 kgd-1 of ‘Pensacola’ bahiagrass (Paspalum notatum Flugge) (Sollenberger et al.

1989), have not been used commercially in the US because of the limitations

associated with the vegetative propagation (Diz and Schank 1993). Production of large

seeds that can be harvested and planted mechanically will propel the commercial

success of napiergrass.

Napiergrass is the fastest growing plant in the world (Karlsson and Vasil 1986)

with reported dry biomass yield of 45 dry t ha−1 year−1 in Florida (Woodard and

Sollenberger 2012) and as high as 80 dry t ha−1 in tropical countries (Vicente-Chandler,

Silva, and Figarella 1959). Therefore, napiergrass has a great potential as

lignocellulosic feedstock for biofuel production. Being non-edible and able to grow on

marginal land, napiergrass escapes the 'food versus fuel' debate and has competitive

advantage over tree species as it can be harvested for biomass in the first year after

143

planting. Additionally, lignin content, which is considered a hindrance to the

fermentation of biomass to ethanol, is much lower in napiergrass than woody biomass,

10% in napiergrass compared to 20-30% in woods (Tong, Smith, and Mccarty 1990;

Mckendry 2002). Most perennial biomass crops like switchgrass (Panicum virgatum L.)

have establishment issues due to small seed size, slow growth rate, dormancy, and

negative response to high planting densities (Noble Research Institute 2007). Leading

bioenergy candidate grasses like switchgrass, Miscanthus, and energycane are not

capable of both direct seeding and high production of biomass in the establishment year

and in contrast to napiergrass do not tolerate multiple harvests per year (Singh, Singh,

and Obeng 2013).

Napiergrass is a short-day plant and flowering in tropical climates occurs from

autumn through winter (Singh, Singh, and Obeng 2013). Early flowering cultivars

produce an abundant amount of small and wind dispersed seeds making seed

collection difficult and increasing its invasive potential (D’Antonio and Vitousek 1992;

Loope, Hamann, and Stone 1988; Schofield 1989). Therefore, napiergrass is listed as

an invasive species by the Florida Exotic Plant Pest Council (FLEPPC 2011).

Invasiveness in napiergrass can be effectively controlled by developing interspecific

triploid hybrids between napiergrass and pearl millet (Cenchrus americanus, 2n=2x=14)

(Hanna 1981). The chromosomes in the A’ genome of napiergrass are homologous to

the A genome of pearl millet (Jauhar 1981) and these two species hybridize naturally to

produce pearl millet napiergrass (PMN) hybrids that are triploids (AA’ B genome) and

thus are sterile (Singh, Singh, and Obeng 2013). PMN hybrids do not set seed and so,

do not pose a threat of uncontrolled establishment through dissemination of seeds

144

(Hanna and Monson 1980) and are not considered invasive (Jessup 2013). PMN

hybrids combine superior forage quality of pearl millet and high yielding ability of

napiergrass (Gupta and Mhere 1997; Osgood, Hanna, and Tew 1997). Some of these

hybrids produced higher biomass (18.9 Mg ha-1) than napiergrass (17.5 Mg ha-1) and

pearl millet (13.2 Mg ha-1) in Louisiana (Cuomo, Blouin, and Beatty 1996). Pearl millet

seeds are larger in comparison to napiergrass and no seed shattering occurs on the

pearl millet panicle (Fig 5-1). The shape and size of PMN seeds are similar to the

female pearl millet parent. These seeds can be planted using seed drills. Development

of seeded varieties that are sterile will represent a significant step in napiergrass

breeding because establishment of fields by seeds will allow automation of planting,

thus a significant cost reduction (Osgood, Hanna, and Tew 1997). Normally, the

resulting hybrid (2n=21) with AA’B genome has greater similarity to the napiergrass type

due to larger genetic contribution (66.7% chromosomes) and dominance of the

napiergrass B genome over the pearl millet A genome for genetic characters such as

earliness, inflorescence and leaf characteristics (Obok, Ova, and Iwo 2012; Gonzalez

and Hanna 1984). Most of the characteristics like resistance to pests, vigorous growth,

and high forage yield potential are on the B genome (Hanna 1987).

Both napiergrass and pearl millet are protogynous in nature, a phenomenon

where stigmas are exerted prior to anther exertion, therefore, they are predominately

cross-pollinated that results in high heterozygosity (Dowling, Burson, and Jessup 2014).

The heterozygous out-crossing nature of napiergrass and pearl millet leads to

significant segregation and lack of uniformity in progenies. Because of the

heterozygosity in napiergrass and pearl millet, PMN hybrids exhibit a high level of

145

heterosis (Dowling, Burson, and Jessup 2014). For successful commercial application, it

is important to get a high level of heterosis in the hybrids while maintaining uniformity in

biomass and persistence of the progenies. Uniformity, the ability of a cultivar to

produce a specific phenotype instead of a varying phenotype, increases the amount of

predictability on the total biomass yield (Makumburage and Stapleton 2011). Using

PMN hybrids for commercial cultivation requires seeds that produce plants with a high

biomass yield and a certain level of uniformity. Selection for stand uniformity was found

to be associated with increased tolerance to environmental stress in maize (Tollenaar

and Wu 1999). PMN hybrid combinations can produce non-germinating seeds

(Aken’ova and Chheda 1973), or produce hybrids with varying yield potential (Hanna

and Monson 1980). Phenotypic variation can be due to genetic variation caused by

alleles segregating within a population, epistasis and mutations, or environment, caused

by fluctuating external condition (Fraser and Schadt 2010).

Commercial production of PMN hybrids can be facilitated by utilizing cytoplasmic

male sterile (cms) lines of pearl millet. ‘Tift 23A’, a cms line of pearl millet paved the way

to produce seed-propagated PMN hybrids that can facilitate seed harvest (J. B. Powell

and Burton 1966). However, the cms lines of pearl millet are dwarf type forages. It is

critical to choose the right parental combination in order to maximize yield of

interspecific PMN hybrids. For this, the cms trait in the dwarf forage type pearl millet

needs to be introgressed into high biomass pearl millet lines and homogenized so that

progenies from these will be uniform. The backcross method has been commonly used

to transfer entire sets of chromosome from foreign cytoplasm in order to create

cytoplasmic male-sterile genotypes (Acquaah 2012). These lines can then be used to

146

cross with inbred napiergrass to produce progenies of male and female sterile PMN

hybrids.

Since PMN hybrids can produce forage until the season’s first frost, they have

the potential to address the fall forage deficit in the southeast USA (Cuomo, Blouin, and

Beatty 1996). PMN hybrids can be easily established using seeds and may produce

high yields in the first year itself. PMN hybrids are more resistant to most pests and

diseases than pearl millet (Hanna and Monson 1980). Expressed sequence tags –

simple short repeat (EST-SSR) markers can be used in order to confirm triploid hybrids

in a breeding program (Dowling et al. 2013). Therefore, developing seeded PMN

hybrids with high biomass yield and a certain level of uniformity should have a major

impact in the forage and biofuel industry. As such, the impact of different levels of

homozygosity/selfing of pearl millet and napiergrass parents on biomass yield and

uniformity will inform us about the best strategy to manage field breeding of PMN

hybrids. Therefore, it is necessary to identify the level of heterozygosity that is present

in the progenies from different crosses by evaluating biomass yield and uniformity of the

hybrids under field conditions. There have been no previous studies done on the impact

of different levels of selfing of pearl millet and napiergrass parents on the agronomic

performance of the PMN hybrids.

Introgression of cms into high biomass pearl millet should facilitate field

production of PMN seeds by open pollination of cms pearl millet with napiergrass. In this

study we report the biomass yield and uniformity of progenies produced from crosses

between four different parental types of female pearl millet (cms forage, high biomass,

147

cms high biomass, and homozygous high biomass) and a napiergrass male parent (25-

17).


Description of Male Sterile Line of Pearl Millet

In this research we used a forage type pearl millet, Tift 85D2A4 (Tift 85) which is a

cytoplasmic male sterile line that flowers in about 75 days. The cms line of pearl millet

Tift 85 was received from Dr. Wayne Hanna, University of Georgia. The A4 cytoplasm

was transferred from a wild subspecies of pearl millet (Wayne Hanna, personal

communication). Tift 85D2B1 is used to maintain sterility of Tift 85D2A4 and it is self-

fertile with good seed set. It was developed by selecting a rust and leaf spot resistant

plant from a selfed population of a BC5 plant developed by backcrossing Tift23D2B1 to a

wild grassy introduction [Pennisetum glaucum (L.) R. Br. Subsp. Monodii (Maire)

Brunken] (Hanna, Wells, and Burton 1987).

Production of cms Lines of Pearl Millet

Cytoplasmic male sterility present in Tift 85 was introgressed into three high

biomass yielding pearl millet lines based on vigor: PI 288787 01 SD (787), PI 215603 01

SD (603), and DLSBF. PI 288787 01 SD is a late flowering accession collected in India

with an average of 19 nodes, 0.880 gm seed weight and 370 cm plant height (USDA,

ARS, and NGRP 1963). PI 215603 01 SD is a late flowering accession collected in India

with an average of 370 cm plant height and 0.720 gm seed weight (USDA, ARS, and

NGRP 1954). DLSBF is an African introduction from Burkina Faso (Wayne Hanna,

personal communication). Pollen from the pearl millet elite lines was dusted onto stigma

of dwarf type cms Tift 85. Thus produced F1 progeny was back crossed to a selfed high

148

biomass line of pearl millet. Male sterility on individual plants were tested by visual

observation prior to crossing in each generation. Selfing of parental line and back

crossing with the most advanced line of the pearl millet continued. Finally, we were able

to introgress cms into one of the three elite pearl millet lines (line 787). We continued to

back cross the cms line 787 four times in order to get a cms version of the elite line of

pearl millet.

Production of PMN Hybrids

In order to obtain near inbred lines of napiergrass, two lines of napiergrass were

selfed, namely, Schank and 25-17. These two lines were selfed two times and flowering

was induced by controlling photoperiod in a growth chamber. However, selfed second

generation of napiergrass didn’t flower due to problems with photoperiod control in the

greenhouse. Therefore, crosses of parental napiergrass (25-17) with different types of

female pearl millet lines were performed (Table 5-1). The interspecific triploid (3x=21)

hybrids were produced in the growth chamber by pollinating pearl millet female plants

with napiergrass pollen.

These four crosses represent the cross of male napiergrass parent (25-17) with

four different types of female pearl-millet lines (cms parent – Tift85 [A], selfed 5th

generation -787S5 [B], BC4 generation – MS 787 BC4 [C], and original parent – P787

[D]).

Seedlings from these four crosses were grown and evaluated for triploidy based

on flowering. Nodes from the confirmed triploids were cut and replanted as replicated

clones into 6” pots. Five clones (nodes) for each plant were grown in the greenhouse in

UF/IFAS Plant Science Research and Education Unit (PSREU), Citra, FL.

149

The pots were filled with sandy soil from the field in PSREU to mimic the soil

conditions of the field. The nodes were planted on Jan 30, 2018 and were irrigated twice

daily. Fertilizer at the rate of 20-20-20 was supplied through irrigation using Dosatron.

After the nodes germinated, the ‘N’ fertilizer rate was increased to 46-0-0 on Feb 21,

2018 in order to promote vegetative growth of the plants. The plants were transplanted

to the field on March 28, 2018. The plants were irrigated with 12mm weekly in the

absence of rainfall after planting. On April 20, 2018 the plants were fertilized with 33.62

kg ha-1 N; 11.21 kg ha-1 and 33.62 kg ha-1 K plus micronutrient package followed by

irrigation. The plants were further fertilized on May 10, 2018 with 84.06 kg ha-1 N; 20.17

kg ha-1 P and 84.06 kg ha-1 K plus micronutrient followed by irrigation. On June 18,

2018 the plants were re-fertilized with 67.25 kg ha-1 N and 89.66 kg ha-1 K using the 6-

0-8 liquid fertilizer plus micronutrient package followed by irrigation.

Experimental Design

Completely randomized block design (RCBD) with 5 replications was used for the

field experiment. Individuals of the same cross type were grouped together and

randomized. Row to row distance was 1.22 m (4 ft) and plant to plant distance was

maintained at 0.91 m (3ft).

Traits Evaluated

We measured yield related attributes of the PMN hybrids. Plant height, number of

tillers, stem diameter, leaf length, leaf width, fresh biomass, and dry biomass were

evaluated. For fresh biomass, all the tillers arising from individual plants were cut using

a brush cutter and fresh weight measured with a hanging scale after seven months of

planting. Samples for dry biomass were taken on August 20, 2018 by cutting 1-2 tillers

150

per plant for 3 replications of each line and weighed. They were then dried in a plant

sample dryer at 48.8°C for 1 month after which the dry biomass weight of the samples

was measured. The dry biomass for replicates was averaged and its derived function

was used to extrapolate dry biomass from each line’s fresh biomass.

Data Analysis

Levene's test was done to compare the variances using 'car' package in R (Fox

et al. 2012). Analysis of variance (ANOVA) and Tukey's HSD test were carried out using

'agricolae' package in R (De Mendiburu 2014). Linear model was fit using the trait value

as dependent variable and cross type and replication as independent variable.

Results

Plant Height

Plant height of the four different crosses were significantly different to each other

(HSD, p<0.05). Variance due to block was significant which was accounted for in the

model. The interaction between plant height and block was not significant. Cross D had

significantly highest height followed by cross B, cross C, and cross A (Figure 5-2). The

coefficient of variation of height was highest in cross type A (23.98%) and the lowest CV

was on cross D (11.31%) (Table 5-2). Maximum height among all the groups was in

cross D, with a height of 426.72 cm. Lowest range of plant height was in cross B

(228.60 cm – 396.24 cm) and the highest in cross D (86.36 cm – 426.72 cm) (Figure 5-

3).

Tiller Number

Tiller number varied significantly between the four types of crosses (p<0.05).

Block effect was not significant for tiller number. Tiller number was significantly higher in

cross A followed by cross D (HSD, p<0.05) (Figure 5-4). Tiller number of cross B was at

151

par with cross C. Lowest number of tillers (1) was found in cross C and the highest

number of tillers (43) was found in cross A. The CV of number of tillers was highest in

cross A and lowest in cross D (Table 5-2). Similarly, the range of number of tillers was

much greater in cross A (2 - 43) as compared to other crosses (3 – 22 in cross B, 1 – 27

in cross C, and 5 – 15 in cross D) (Figure 5-5).

Stem Diameter

Stem diameter between the groups were significantly different (p<0.05) and the

block effect was not significant. Significantly highest (17.25 mm) stem diameter was

found in cross B followed by cross C which was at par with cross D (Figure 5-6). Cross

A showed the highest CV (24.81%) for stem diameter while cross D showed the lowest

CV (14.16%) (Table 5-2). The range of stem diameter was wider in cross A (5.92 mm –

21.95 mm) compared to other crosses (10.04 – 25.71 mm for cross B, 9.63 mm – 22.65

mm for cross C, and 10.65 mm – 22.06 mm for cross D) (Figure 5-7).

Leaf Length

Leaf length was significantly different among the four crosses (p<0.05). Block

effect was not significant. Leaf length of cross D was at par with that of cross B, while it

was significantly higher than cross C or cross A (Figure 5-8). The minimum leaf length

was found in cross A (82.25 cm) and the maximum leaf length was found in cross D

(106.84 cm). The CV for leaf length was highest in cross A (28.32%) and the lowest CV

of leaf length was on cross D (10.15%) (Table 5-2). Range of leaf length was lower in

cross D (72.39 cm – 132.08 cm) as compared to other crosses (38.10 cm – 130.81 cm

for cross A, 26.67 cm – 129.54 cm for cross B, and 29.21 cm – 146.05 cm for cross D)

(Figure 5-9).

152

Leaf Width

Leaf width was significantly different among the four crosses (p<0.05). Block

effect was not significant. Leaf width of cross D was at par with that of cross B, while it

was significantly higher than cross C or cross A (Figure 5-10). On the other hand, leaf

length of cross B was at par with cross D and cross C and significantly higher than

cross A. The minimum leaf width was found in cross C (25 mm) and the maximum leaf

width was found in cross B (69 mm). The CV for leaf width was highest in cross A

(18.80%) and the lowest CV of leaf width was on cross D (13.83%) (Table 5-2). Range

of leaf width was lower in cross D (27 mm – 62 mm) as compared to other crosses (23

mm – 59 mm for cross A, 30 mm – 69 mm for cross B, and 25 mm – 64 mm for cross C)

(Figure 5-11).

Plant Biomass

Fresh biomass per plant (kg) was significantly different (p < 0.05) among the four

crosses and the block effect was also significant. The mean response was different in

block 4 and block 5 as compared to the other blocks and the block effect was accounted

for in the model. The interaction between treatment and block was not significant.

Significantly highest biomass was found in cross D (8.54 kg/plant) followed by cross A

(6.88 kg/plant). Biomass for Cross C (4.97 kg/plant) and cross B (5.67 kg/plant) were at

par with each other (Fig 5-12). The lowest biomass was found in cross B and the

highest biomass was found in cross C (20.59 kg). The CV of biomass was lowest in

cross D (33.63%) and highest in cross A (67.40%). Similarly, the spread of biomass

values was wider in cross A (0.18 kg/plant – 19.16 kg/plant) as compared to other

153

crosses (1.56 kg/plant – 14.28 kg/plant in cross B, 0.12 kg/plant – 20.59 kg/plant in

cross C and 3.06 kg/plant – 19.70 kg/plant in cross D) (Figure 5-13).

Dry Biomass

Dry biomass data from single plant was extrapolated to a larger land area in

order to get dry biomass in tons ha-1. Projected dry biomass (tons ha-1) was significantly

different (p < 0.05) among the four crosses. Significantly highest biomass was found in

cross D (24.83 tons ha-1). Dry biomass for the remaining three cross types were at par

with each other (18.39 tons ha-1, 18.34 tons ha-1, and 16.30 tons ha-1 for cross A, cross

B, and cross C, respectively) (Fig 5-14). Highest dry biomass was found in cross A

(54.87 tons ha-1) and the lowest was found in cross C (0.27 tons ha-1). The CV of dry

biomass was lowest in cross D (33.63%) and highest in cross A (67.40%). Similarly, the

range of biomass values among different progeny plants was much wider in cross A

(0.36 tons ha-1 - 54.51 tons ha-1) as compared to other crosses (3.84 tons ha-1 – 47.56

tons ha-1 in cross B, 0.27 tons ha-1 – 53.72 tons ha-1 in cross C, and 8.73 tons ha-1 –

48.59 tons ha-1 in cross D) (Figure 5-15).

Coefficient of Variation

The coefficient of variation (CV) for all measured traits is shown in Figure 5-16.

For all the traits evaluated, it was observed that the CV was least on cross D and

highest on cross A. CV was the highest for dry biomass (tons/ha.) (30.95% - 73.42%)

(Table 5-2) and least for leaf width (13.83-18.80%) among all types of crosses. A

relatively lower CV was observed for stem diameter (14.16-24.81%), plant height

(10.16-23.98%), and leaf length (10.15-28.32%). On the other hand, higher CV was

154

observed for number of tillers (31.93-48.47%). A picture of four different crosses during

harvest is shown in Figure 5-17.

Correlation

All of the evaluated traits had significant, positive correlations with biomass

weight. Correlation coefficient was the highest for number of tillers (0.60) and lowest for

stem diameter (0.18). Correlation coefficients and p-values between biomass weight

and each trait are presented in Table 5-3.

Discussion

Both pearl millet and napiergrass are protogynous and are predominantly cross-

pollinated. This increases their heterozygosity which leads to a high level of heterosis in

PMN hybrids (Dowling, Burson, and Jessup 2014; J. B. Powell and Burton 1966).

Development of a cms pearl millet line optimized for biomass production will be the

most efficient system for commercial scale production of large seeded PMN hybrids

according to the method described by Powell and Burton (1966) (Dowling 2011). For

commercial success of PMN hybrids, phenotypic stability and robustness of the

genotype to consistently produce a specific phenotype is critical. To achieve this, we

introgressed cms into high biomass pearl millet lines and assessed the phenotypic

variation in biomass yield and related traits of PMN hybrids depending on different pearl

millet sources. It should be noted that our attempt to produce a male napiergrass parent

with higher level of homozygosity was not accomplished in time and therefore a

heterozygous napiergrass parent was used in all these crosses. Inherent heterozygosity

in this parent probably contributed to a high level of phenotypic variability in the

progenies.

155

The biomass produced after seven months of growth was the highest in the cross

involving the original pearl millet parent (cross D) followed by the cms parent, Tift 85

(cross A). The CV for biomass was also lowest in cross D (30.95%). However, CV for

biomass of cross A was the highest (73.42%) as compared to cross B (39.66%) and

cross C (56.55%). For the other traits evaluated, lowest CV was mostly found in cross D

followed by cross B. As expected the high biomass pearl millet parent P787 used in

cross D had a higher combining ability than the forage-type pearl millet parent Tift 85

resulting in significant differences in biomass yield. The backcross of Tift 85 with P787

was not able to restore the superior combining ability of P787. Interestingly, progenies

from cross C using 787-S5 as parent showed a significant lower biomass yield than

progenies from crosses involving the heterozygous P787. Testing the biomass

accumulation of selfed progenies from P787 at every generation and selecting superior

individuals may help to generate male sterile 787 with superior combining ability. The

difference among CV of cross D and cross B was low. For plant height, the CV of cross

B was actually the lowest among the different crosses. CV for the various traits involving

cross C (BC4 generation of Tift85 and 25-17) was lower than cross A but higher than

cross B or cross D. We saw that the CV decreased in BC4 but hasn't reached the level

of P787 or 787-S5. In theory, BC4 genotype will be 93.75% identical to the recurrent

parent (Acquaah 2012). Therefore, in the line MS 787 BC4, we are in theory able to

introgress approximately 94% of the genome of P787 parent. Continuing the back cross

to more than BC7 generation will give 99% similarity to the P787 parent and may result

in cms progenies that are more uniform.

156

Even though the average yield of PMN hybrids using cms parent Tift85 in cross A

was high, the high variability in yield would make these lines less acceptable for

commercial production because phenotypic uniformity contributes to a predictable yield.

Many factors such as variable genetic yield potential, uneven germination, variable

planting depth, soil clods, insect damage, and moisture might be responsible for non-

uniformity of plant stands etc. (Martin et al. 2005) which can be controlled by proper

agronomic practices, experimental setup, and statistical analysis. Also, the cultivar

resulting from back-crossing could differ from the initial cultivar beyond the transferred

gene(s) because of linkage drag from the association of undesirable traits with the

genes from the donor(Acquaah 2012).

Biomass yield of grasses is affected by different morphological traits. In sorghum,

biomass is correlated with plant height, number of tillers, leaf length, leaf width, stem

diameter, and flowering time (Hart et al. 2001; Murray et al. 2008; Xiao-ping et al. 2011).

Similarly, in Miscanthus, plant height, stem diameter, late flowering, and growth rate

showed the highest positive correlation with yield (Zub, Arnoult, and Brancourt-Hulmel

2011). In Trichloris crinite, foliage height and basal diameter were strongly correlated

with biomass yield (Cavagnaro et al. 2006). In napiergrass, plant height, number of

tillers, and stem diameter were significantly correlated with plant biomass (Sinche

2013). Our results show that in PMN hybrids, traits like plant height, leaf length, leaf

width, stem diameter, and number of tillers are significantly correlated with biomass

yield. This indicates that tall PMN hybrids with numerous thick tillers and wide and long

leaves produced more biomass than those with the opposite characteristics.

157

In this study, we have developed a cms version of elite pearl millet line P787.

The development of this non-dwarf biomass-type cms line of pearl millet provides new

resources for pearl millet and napiergrass breeding that can be exploited for commercial

production of uniform PMN hybrids.

158

Figure 5-1. Seeds of pearl millet dwarf cms line (A), cms high biomass pearl millet line (B), PMN hybrid (C), and napiergrass (D), respectively from left to right.

159

Figure 5-2. Boxplot of plant height (cm) for four different types of crosses studied. Small colored dots represent individual plant height of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

d b c a100

200

300

400

A B C D

Cross Type

Pla

nt h

eig

ht (c

m)

160

Figure 5-3. Histogram of plant height (cm) for four different crosses. X-axis represent plant height in cm and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

C D

A B

100 200 300 400 100 200 300 400

0

10

20

30

0

10

20

30

Plant height (cm)

161

Figure 5-4. Boxplot of number of tillers for four different types of crosses studied. Small colored dots represent number of tillers for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

a c c b0

10

20

30

40

A B C D

Cross Type

Nu

mb

er

of tille

rs

162

Figure 5-5. Histogram of number of tillers for four different crosses. X-axis represent the number of tillers and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

C D

A B

0 10 20 30 40 0 10 20 30 40

0

20

40

60

0

20

40

60

Number of tillers

163

Figure 5-6. Boxplot of stem diameter (mm) for four different types of crosses studied. Small colored dots represent the stem diameter (mm) for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

c a b b5

10

15

20

25

A B C D

Cross Type

Ste

m d

iam

ete

r (m

m)

164

Figure 5-7. Histogram of stem diameter (mm) for four different crosses. X-axis represent the stem diameter (mm) and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

C D

A B

5 10 15 20 25 5 10 15 20 25

0

10

20

30

40

0

10

20

30

40

Stem diameter (mm)

165

Figure 5-8. Boxplot of leaf length (cm) for four different types of crosses studied. Small colored dots represent the leaf length (cm) for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

c a b a

50

100

150

A B C D

Cross Type

Le

af le

ng

th (

cm

)

166

Figure 5-9. Histogram of leaf length (cm) for four different crosses. X-axis represent the leaf length (cm) and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

C D

A B

50 100 150 50 100 150

0

10

20

30

40

0

10

20

30

40

Leaf length (cm)

167

Figure 5-10. Boxplot of leaf width (mm) for four different types of crosses studied. Small colored dots represent the leaf width (mm) for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

c ab b a

30

40

50

60

70

A B C D

Cross Type

Le

af w

idth

(m

m)

168

Figure 5-11. Histogram of leaf width (mm) for four different crosses. X-axis represent the leaf width (mm) and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

C D

A B

30 40 50 60 70 30 40 50 60 70

0

10

20

30

0

10

20

30

Leaf width (mm)

169

Figure 5-12. Boxplot of fresh biomass per plant (kg) for four different types of crosses studied. Small colored dots represent the fresh biomass per plant (kg) for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

b c c a0

5

10

15

20

A B C D

Cross Type

Fre

sh

Bio

ma

ss (

kg

/pla

nt)

170

Figure 5-13. Histogram of fresh biomass per plant (kg) for four different crosses. X-axis represent the fresh biomass per plant (kg) and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

C D

A B

0 5 10 15 20 0 5 10 15 20

0

20

40

0

20

40

Fresh Biomass (kg/plant)

171

Figure 5-14. Boxplot of projected dry biomass (tons per ha) for four different types of crosses evaluated at seven months of growth. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

b b b a0

20

40

A B C D

Cross Type

Dry

Bio

ma

ss (

ton

s/h

a.)

172

Figure 5-15. Histogram of dry biomass (tons per ha.) for four different crosses evaluated at seven months of growth. X-axis represent the dry biomass (tons per ha.) and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

C D

A B

0 20 40 0 20 40

0

5

10

15

20

0

5

10

15

20

Dry Biomass (tons/ha)

173

Figure 5-16. Coefficient of variation (%) for the different traits evaluated. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

LeafWidth StemDiam TillerNum

BioMass height LeafLength

A B C D A B C D A B C D

0

20

40

60

0

20

40

60

Cross type

Co

effic

ien

t o

f va

ria

tio

n (

%)

174

Figure 5-17. Picture of four different crosses during harvest in Citra, FL. Cross A = Tift 85 × 25-17, Cross B = 787-S5 × 25-17, Cross C = MS 787 BC4 × 25-17, Cross D = P787 × 25-17.

175

Table 5-1. Details of the cross types used in the experiment.

Cross type Female parent ♀ (Pearl millet)

Male parent ♂ (Napiergrass)

Cross details Number of plants

A Tift85 × 25-17 cms parent 40 B 787-S5 × 25-17 Selfed 5th gen. 50 C MS 787 BC4 × 25-17 BC4 generation 27 D P787 × 25-17 Original parent 22

176

Table 5-2. Descriptive statistics for different types of crosses. P-value represents Pr>F for Levene’s test for homogeneity of variance (center = median), CV = coefficient of variation.

Trait Cross type P-value Mean Standard Deviation CV (%) Min Max

Plant height 0.0000

Plant height A 0.0778 223.84 53.67 23.98 86.36 340.36

Plant height B 0.9584 313.41 31.84 10.16 228.60 396.24

Plant height C 0.8109 285.09 50.58 17.74 111.76 375.92

Plant height D 0.6924 328.32 37.13 11.31 86.36 426.72

Number of tillers 0.0000

Number of tillers A 0.0758 17.88 8.67 48.47 2.00 43.00

Number of tillers B 0.6613 9.46 3.69 39.03 3.00 22.00

Number of tillers C 0.1015 8.98 4.29 47.73 1.00 27.00

Number of tillers D 0.6265 12.82 4.09 31.93 5.00 15.00

Stem diameter 0.0000

Stem diameter A 0.3174 13.31 3.30 24.81 5.92 21.95

Stem diameter B 0.9829 17.26 2.73 15.80 10.04 25.71

Stem diameter C 0.4913 16.42 2.61 15.90 9.63 22.65

Stem diameter D 0.8523 16.41 2.32 14.16 10.65 22.06

Leaf length 0.0000

Leaf length A 0.2382 82.26 23.29 28.32 38.10 130.81

Leaf length B 0.5612 106.01 11.75 11.08 26.67 129.54

Leaf length C 0.6267 101.17 15.05 14.88 29.21 146.05

Leaf length D 0.1045 106.84 10.84 10.15 72.39 132.08

Leaf width 0.0529

Leaf width A 0.4596 40.75 7.66 18.80 23.00 59.00

Leaf width B 0.8497 47.61 6.88 14.46 30.00 69.00

Leaf width C 0.7394 46.04 7.92 17.20 25.00 64.00

177

Table 5-2. Continued Trait Cross type P-

value

Mean Standard

Deviation

CV (%) Min Max

Leaf width D 0.1875 48.89 6.76 13.83 27.00 62.00

Fresh biomass 0.0000

Fresh biomass A 0.0214* 6.88 4.64 67.40 0.18 19.16

Fresh biomass B 0.6081 6.23 2.29 36.75 1.56 14.28

Fresh biomass C 0.0794 5.47 3.10 56.63 0.12 20.59

Fresh biomass D 0.7515 8.54 2.87 33.63 3.06 19.70

Dry biomass 0.0000

Dry biomass A 0.0207* 18.39 13.50 73.42 0.36 54.87

Dry biomass B 0.6462 18.34 7.27 39.66 3.84 47.56

Dry biomass C 0.1246 16.30 9.22 56.55 0.27 53.72

Dry biomass D 0.9255 24.83 7.68 30.95 8.73 48.59

178

Table 5-3. Correlation coefficients and p-values for biomass weight and biomass-related traits for PMN hybrids evaluated in Citra, FL.

Plant height Leaf length Leaf width Stem diameter Number of tillers

R 0.43 0.36 0.25 0.18 0.6

p-value <0.001 <0.001 <0.001 <0.001 <0.001

179

CHAPTER 6 CONCLUDING REMARKS

Summary

Napiergrass is an important forage and biofuel crop. However, its commercial

utilization has been lagging behind other crops due to limited genetic and genomic

resources. Before this study, there were no SSR markers derived from napiergrass

sequences publicly available, no genetic linkage map was available, and no QTL

studies for any traits were published. The lack of these genetic resources limits the

advances that can be achieved via breeding. In this project, we have developed genetic

and genomic resources that will be important in napiergrass breeding. In Chapter 2, we

constructed the first high-density genetic linkage map of napiergrass using NGS-derived

SNP markers. We also identified 5,339 SSRs using napiergrass genomic sequences

and successfully designed primers for 1,926 SSRs. These results will be useful for

future molecular breeding programs such as identification of QTLs for important traits

as well as MAS for the genetic improvement of napiergrass and comparative

genomics.

Early flowering cultivars of napiergrass produce abundant wind dispersed seeds,

which contribute to high potential of invasiveness. Controlling flowering or modifying

flowering time of napiergrass can help in reducing its invasiveness and boost its

potential as biofuel feedstock. Therefore, a better understanding of the genetic basis of

flowering time in napiergrass is necessary. To facilitate this, in Chapter 3, we

demonstrated that flowering time and number of flowers are highly heritable traits.

Therefore, we conducted the first QTL analysis in napiergrass and identified three

stable QTLs controlling number of flowers. We also identified three potential QTLs that

180

control flowering time in napiergrass. We were able to identify potential candidate genes

such as AGAMOUS, DEFICIENS, DELLA, WRKY, and SERK1 that were harbored by

the QTL regions. The QTLs detected in this study will be valuable tools for napiergrass

breeding and marker assisted selection. Similarly, the candidate genes identified could

be potential targets for genome editing to modify flowering time in napiergrass.

To improve our understanding of the genetic basis of flowering time in

napiergrass and to evaluate the genetic diversity of napiergrass, we used exome-

capture sequencing to study the germplasm collection of napiergrass, described in

Chapter 4. We identified 78,129 high quality SNPs in the germplasm collection that

serve as novel genomic resources for napiergrass breeding programs. We also

identified potential candidate genes like Calcium-binding protein CML, enoyl-CoA

hydratase, phytochrome B, and AP2-like factors in the germplasm collection. This study

showed the feasibility to apply targeted exome sequencing with probes designed from

CDS of pearl millet to target the genomic regions in napiergrass.

We also used traditional breeding approaches to increase biosafety of

napiergrass. For this, in Chapter 5, we introgressed cms available in forage type dwarf

pearl millet lines to high biomass yielding pearl millet lines. We then hybridized these

elite cms lines of pearl millet with napiergrass to generate PMN hybrids that are male

and female sterile and will not contribute to invasiveness with wind dispersed seeds. We

studied the uniformity of different types of parental combinations. Substantial variation

within each cross was found for biomass accumulation and yield related traits.

Uniformity of the progenies could be enhanced by using homozygous parents to make

the crosses. If efficient seed production can be accomplished under field conditions and

181

if uniformity can be improved without compromising biomass yield, PMN hybrids may

outperform alternative forage and biofuel crops in the near future.

Future work

The construction of the first genetic map of napiergrass has opened up new

avenues in napiergrass breeding. With this map, we were able to identify QTLs related

to flowering time and number of flowers in napiergrass. There are other traits like stem

diameter, tiller number, and yield, whose genetic basis needs to be further explored. It is

our hope that the genetic map developed in Chapter 2 can be utilized to explore other

traits of agronomic importance to identify QTLs and candidate genes for these traits.

Further, the sequence variations identified in the germplasm collection can be utilized in

the future to study the genetic architecture of other traits and to perform GWAS for

these traits. Moreover, introgression of cms into high biomass lines like P787 will

continue to obtain a homozygous cms line with superior combining ability with

napiergrass. Similarly, generation of homozygous or near-homozygous napiergrass

lines will have to be continued by repeated self-fertilization and selection of the best

performing lines. PMN hybrids from these homozygous or near homozygous parents

should be evaluated not only for increased uniformity but also for biomass yield which

may mainly depend on the different level of heterosis in the A and B genomes of the

hybrids and the combining ability of the parents. For commercial production of PMN

hybrid seeds under field condition different accessions need to be evaluated to

synchronize flowering time and identify parents that contribute to the highest seed yield.

Alternatively, highest yielding napiergrass lines can be evaluated in countries where

182

napiergrass is a native species, the preferred forage crop and invasiveness is not a

concern.

183

LIST OF REFERENCES

Acquaah, George. 2012. “Breeding Self Pollinated Species.” In Principles of Plant Genetics and Breeding, Second Edition, 303–36. Wiley Online Library.

Adhikari, Laxman, Orville M Lindstrom, Jonathan Markham, and Ali M Missaoui. 2018. “Dissecting Key Adaptation Traits in the Polyploid Perennial Medicago Sativa Using GBS-SNP Mapping.” Frontiers in Microbiology 9: 1–19. doi:10.3389/fpls.2018.00934.

Agarwal, Milee, Neeta Shrivastava, and Harish Padh. 2008. “Advances in Molecular Marker Techniques and Their Applications in Plant Sciences.” Plant Cell Reports 27 (4): 617–31. doi:10.1007/s00299-008-0507-z.

Aken’ova, M. E., and H. R. Chheda. 1973. “Interspecific Hybrids Pennisetum Typhoides S & H × P. Purpureum Schum. for Forage in the Hot Humid Lowland Tropics of West Africa.” The Nigerian Agricultural Journal 10 (10): 82–90.

Andersen, Jeppe R, and Thomas Lübberstedt. 2003. “Functional Markers in Plants.” Trends in Plant Science 8 (11): 554–60. doi:10.1016/j.tplants.2003.09.010.

Anderson, William F., Michael Casler, and Brian Baldwin. 2008. “Improvement of Perennial Forage Species as Feedstock for Bioenergy.” In Genetic Improvement of Bioenergy Crops, edited by Wilfred Vermerris, 309–45. Springer Science+Business Media, LLC.

Atienza, S. G., Z. Satovic, K. K. Petersen, O. Dolstra, and A. Martín. 2003. “Identification of QTLs Influencing Combustion Quality in Miscanthus Sinensis Anderss. II. Chlorine and Potassium Content.” Theoretical and Applied Genetics, no. 107: 857–63. doi:10.1007/s00122-003-1218-z.

Augustin, Eiiane, and Fernando Adami Tcacenco. 1993. “Isoenzymatic Characterization of Elephantgrass (Pennisetum Purpureum Schum.) Germplasm.” Rev. Brasil. Genet. 16 (3): 685–96.

Azevedo, Ana Luisa Sousa, Pricila Palla Costa, Juarez Campolina Machado, Marco Antonio Machado, Antônio Vander Pereira, and Francisco José da Silva Lédo. 2012. “Cross Species Amplification of Pennisetum Glaucum Microsatellite Markers in Pennisetum Purpureum and Genetic Diversity of Napier Grass Accessions.” Crop Science 52: 1776–85. doi:10.2135/cropsci2011.09.0480.

Bastow, Ruth, and Caroline Dean. 2002. “The Molecular Basis of Photoperiodism.” Biological Rhythm Research 37 (4): 353–80. doi:10.1016/S1534-5807(02)00296-4.

184

Bezant, Jeremy, David A. Laurie, Nicky Pratchett, Jan A N Chojecki, and Mike J. Kearsey. 1996. “Marker Regression Mapping of QTL Controlling Flowering Time and Plant Height in a Spring Barley (Hordeum Vulgare L .) Cross.” Heredity 77: 64–73.

Bhandari, Anitha P, D H Sukanya, and C R Ramesh. 2006. “Application of Isozyme Data in Fingerprinting Napier Grass (Pennisetum Purpureum Schum.) for Germplasm Management.” Genetic Resources and Crop Evolution 53. Springer: 253–64.

Bidinger, F. R., T. Nepolean, C. T. Hash, R. S. Yadav, and C. J. Howarth. 2007. “Quantitative Trait Loci for Grain Yield in Pearl Millet under Variable Postflowering Moisture Conditions.” Crop Science 47 (3): 969–80. doi:10.2135/cropsci2006.07.0465.

Bielenberg, Douglas Gary, Bradley Rauh, Shenghua Fan, Ksenija Gasic, Albert Glenn Abbott, Gregory Lynn Reighard, William R. Okie, and Christina Elizabeth Wells. 2015. “Genotyping by Sequencing for SNP-Based Linkage Map Construction and QTL Analysis of Chilling Requirement and Bloom Date in Peach [Prunus Persica (L.) Batsch].” PLoS ONE, 1–14. doi:10.1371/journal.pone.0139406.

Bodénès, Catherine, Emilie Chancerel, François Ehrenmann, Antoine Kremer, and Christophe Plomion. 2016. “High-Density Linkage Mapping and Distribution of Segregation Distortion Regions in the Oak Genome.” DNA Research 23 (2): 115–24. doi:10.1093/dnares/dsw001.

Bohuon, E J R, L D Ramsay, J A Craft, A E Arthur, D F Marshall, D J Lydiate, and Mike J. Kearsey. 1998. “The Association of Flowering Time Quantitative Trait Loci with Duplicated Regions and Candidate Loci in Brassica Oleracea.” Genetics 150: 393–401.

Bolger, Anthony M., Marc Lohse, and Bjoern Usadel. 2014. “Trimmomatic: A Flexible Trimmer for Illumina Sequence Data.” Bioinformatics 30 (15): 2114–20. doi:10.1093/bioinformatics/btu170.

Boonman, Joseph G. 1988. East Africa’s Grasses and Fodders: Their Ecology and Husbandry. Edited by Helmut Lieth and Harold A. Mooney. Springer-Science+Business Media, B.V.

Bouché, Frédéric, Guillaume Lobet, Pierre Tocquin, and Claire Périlleux. 2016. “FLOR-ID: An Interactive Database of Flowering-Time Gene Networks in Arabidopsis Thaliana.” Nucleic Acids Research 44: D1167–71. doi:10.1093/nar/gkv1054.

Bourke, Peter M., Roeland E. Voorrips, Richard G. F. Visser, and Chris Maliepaard. 2018. “Tools for Genetic Studies in Experimental Populations of Polyploids.” Frontiers in Plant Science 9: 1–17. doi:10.3389/fpls.2018.00513.

185

Brunken, Jere N. 1977. “A Systematic Study of Pennisetum Sect . Pennisetum (Gramineae).” American Journal of Botany 64 (2): 161–76.

Buckler, Edward S, James B Holland, Peter J Bradbury, Charlotte B Acharya, Patrick J Brown, Chris Browne, Elhan Ersoz, Sherry Flint-Garcia, Arturo Garcia, and Jeffrey C. Glaubitz. 2009. “The Genetic Architecture of Maize Flowering Time.” Science 325. American Association for the Advancement of Science: 714–18.

Burton, Glenn W. 1944. “Hybrids between Napier Grass and Cattail Millet.” The Journal of Heredity, 227–32.

———. 1958. “Cytoplasmic Male-Sterility in Pearl Millet (<i>Pennisetum glaucum)(L.) R. Br.” Agronomy Journal 50 (4). American Society of Agronomy: 230.

———. 1990. Grasses : New and Improved. Edited by J. Janick and J.E. Simon. Advances in New Crops. Portland, Oregon: Timber Press.

Burton, Glenn W., and D S Athwal. 1967. “Two Additional Sources of Cytoplasmic Male-Sterility in Pearl Millet and Their Relationship to Tift 23A.” Crop Science 7 (3). Crop Science Society of America: 209–11.

Cai, C, F Y Cheng, J Wu, Y Zhong, and G Liu. 2015. “The First High-Density Genetic Map Construction in Tree Peony (Paeonia Sect. Moutan) Using Genotyping by Specific-Locus Amplified Fragment Sequencing.” PLoS One, 1–16. doi:10.1371/journal.pone.0128584.

Catchen, Julian, A. Amores, P. Hohenlohe, W. Cresko, J. H. Postlethwait, and D.-J. De Koning. 2011. “Stacks: Building and Genotyping Loci de Novo from Short-Read Sequences.” G3.Genes|Genomes|Genetics 1: 171–82. doi:10.1534/g3.111.000240.

Catchen, Julian, Paul A. Hohenlohe, Susan Bassham, Angel Amores, and William A. Cresko. 2013. “Stacks: An Analysis Tool Set for Population Genomics.” Molecular Ecology 22: 3124–40. doi:10.1111/mec.12354.

Cavagnaro, Pablo F, Juan B Cavagnaro, José L Lemes, Ricardo W Masuelli, and Carlos B Passera. 2006. “Genetic Diversity among Varieties of the Native Forage Grass Trichloris Crinita Based on AFLP Markers, Morphological Characters, and Quantitative Agronomic Traits.” Genome 49: 906–18. doi:10.1139/g06-060.

Chemisquy, M. Amelia, Liliana M. Giussani, María A. Scataglini, Elizabeth A. Kellogg, and Osvaldo Morrone. 2010. “Phylogenetic Studies Favour the Unification of Pennisetum, Cenchrus and Odontelytrum (Poaceae): A Combined Nuclear, Plastid and Morphological Analysis, and Nomenclatural Combinations in Cenchrus.” Annals of Botany 106: 107–30. doi:10.1093/aob/mcq090.

186

Churchill, Gary A., and R. W. Doerge. 1994. “Empirical Threshold Values for Quantitative Trait Mapping.” Genetics 138: 963–71. doi:10.1534/genetics.107.080101.

Cingolani, Pablo, Adrian Platts, Le Lily Luan Wang, Melissa Coon, Tung Nguyen, Le Lily Luan Wang, Susan J. Land, Xiangyi Lu, and Douglas M. Ruden. 2012. “A Program for Annotating and Predicting the Effects of Single Nucleotide Polymorphisms, SnpEff.” Fly 6 (2): 80–92.

Clevenger, Josh, Carolina Chavarro, Stephanie A. Pearl, Peggy Ozias-Akins, and Scott A. Jackson. 2015. “Single Nucleotide Polymorphism Identification in Polyploids: A Review, Example, and Recommendations.” Molecular Plant 8 (6). Elsevier Ltd: 831–46. doi:10.1016/j.molp.2015.02.002.

Crawford, Joseph, Patrick J. Brown, Thomas Voigt, and D. K. Lee. 2016. “Linkage Mapping in Prairie Cordgrass (Spartina Pectinata Link) Using Genotyping-by-Sequencing.” Molecular Breeding 36 (5). Springer Netherlands: 1–12. doi:10.1007/s11032-016-0484-9.

Cuomo, Greg J, David C Blouin, and James E Beatty. 1996. “Forage Potential of Dwarf Napiergrass and a Pearl Millet X Napiergrass Hybrid.” Agronomy Journal 88: 434–38.

D’Antonio, Carla M, and Peter M Vitousek. 1992. “Biological Invasions by Exotic Grasses, the Grass/fire Cycle, and Global Change.” Annual Review of Ecology and Systematics 23. JSTOR: 63–87.

de Lima, R S N, R F Daher, L S a Gonçalves, D a Rossi, a T do Amaral Júnior, M G Pereira, and F J S Lédo. 2011. “RAPD and ISSR Markers in the Evaluation of Genetic Divergence among Accessions of Elephant Grass.” Genetics and Molecular Research : GMR 10 (3): 1304–13. doi:10.4238/vol10-3gmr1107.

De Mendiburu, Felipe. 2014. “Agricolae: Statistical Procedures for Agricultural Research.” R Package Version.

Dell’Acqua, Matteo, Daniel M. Gatti, Giorgio Pea, Federica Cattonaro, Frederik Coppens, Gabriele Magris, Aye L. Hlaing, et al. 2015. “Genetic Properties of the MAGIC Maize Population: A New Platform for High Definition QTL Mapping in Zea Mays.” Genome Biology 16 (167). Genome Biology: 1–23. doi:10.1186/s13059-015-0716-z.

Dellaporta, Stephen L, Jonathan Wood, and James B Hicks. 1983. “A Plant DNA Minipreparation: Version II.” Plant Molecular Biology Reporter 1 (4). Springer: 19–21.

187

Deschamps, Stéphane, Victor Llaca, and Gregory D. May. 2012. “Genotyping-by-Sequencing in Plants.” Biology 1. Molecular Diversity Preservation International: 460–83. doi:10.3390/biology1030460.

Devos, Katrien M, and Mike D Gale. 2000. “Genome Relationships: The Grass Model in Current Research.” The Plant Cell 12. Am Soc Plant Biol: 637–46.

Devos, Katrien M, T S Pittaway, A Reynolds, and M D Gale. 2000. “Comparative Mapping Reveals a Complex Relationship between the Pearl Millet Genome and Those of Foxtail Millet and Rice.” Theoretical and Applied Genetics 100. Springer: 190–98. doi:10.1007/s001220050026.

Diz, D. A., and S. C. Schank. 1993. “Characterization of Seed Producing Pearl Millet X Elephantgrass Hexaploid Hybrids.” Euphytica 67: 143–49. doi:10.1007/BF00022737.

Dowling, Charlie D. 2011. “Marker-Assisted Verification of Hybrids in Pearl Millet-Napiergrass (Pennisetum Glaucum [L.] R. BR. X Pennisetum Purpureum Schumach.).” Texas A&M University.

Dowling, Charlie D., Byron L Burson, Jamie L Foster, Lee Tarpley, and Russell W Jessup. 2013. “Confirmation of Pearl Millet-Napiergrass Hybrids Using EST-Derived Simple Sequence Repeat (SSR) Markers.” American Journal of Plant Sciences 4. Scientific Research Publishing: 1004–12.

Dowling, Charlie D., Byron L Burson, and Russell W Jessup. 2014. “Marker-Assisted Verification of Kinggrass (Pennisetum Purpureum Schumach . X Pennisetum Glaucum [ L .] R . Br .).” Plant Omics Journal 7 (2): 72–79.

Durand, Eléonore, Sophie Bouchet, Pascal Bertin, Adrienne Ressayre, Philippe Jamin, Alain Charcosset, Christine Dillmann, and Maud I Tenaillon. 2012. “Flowering Time in Maize: Linkage and Epistasis at a Major Effect Locus.” Genetics 190. Genetics Soc America: 1547–62.

Elshire, Robert J., Jeffrey C. Glaubitz, Qi Sun, Jesse A. Poland, Ken Kawamoto, Edward S. Buckler, and Sharon E. Mitchell. 2011. “A Robust, Simple Genotyping-by-Sequencing (GBS) Approach for High Diversity Species.” PLoS ONE 6 (5). doi:10.1371/journal.pone.0019379.

Erickson, John E., Arkorn Soikaew, Lynn E. Sollenberger, and Jerry M. Bennett. 2012. “Water Use and Water-Use Efficiency of Three Perennial Bioenergy Grass Crops in Florida.” Agriculture 2: 325–38. doi:10.3390/agriculture2040325.

Falconer, Douglas S, Trudy F C Mackay, and Richard Frankham. 1996. “Introduction to Quantitative Genetics.” In Trends in Genetics, 4th ed. [Amsterdam, The Netherlands: Elsevier Science Publishers (Biomedical Division)], c1985-.

188

Faris, J. D., B. Laddomada, and B. S. Gill. 1998. “Molecular Mapping of Segregation Distortion Loci in Aegilops Tauschii.” Genetics 149 (1): 319–27.

Farrell, G., S. A. Simons, and R. J. Hillocks. 2002. “Pests, Diseases and Weeds of Napier Grass, Pennisetum Purpureum: A Review.” International Journal of Pest Management 48 (1). Taylor & Francis: 39–48. doi:10.1080/09670870110065578.

Ferrão, Luís Felipe V., Juliana Benevenuto, Ivone de Bem Oliveira, Catherine Cellon, James Olmstead, Matias Kirst, Marcio F. R. Resende, and Patricio Munoz. 2018. “Insights into the Genetic Basis of Blueberry Fruit-Related Traits Using Diploid and Polyploid Models in a GWAS Context.” Frontiers in Ecology and Evolution 6. doi:10.3389/fevo.2018.00107.

FLEPPC. 2011. “List of Invasive Plant Species.” Florida Exotic Pest Plant Council. http://www.fleppc.org/list/2011plantlist.pdf.

Fox, John, Sanford Weisberg, Daniel Adler, Douglas Bates, Gabriel Baud-Bovy, Steve Ellison, David Firth, Michael Friendly, Gregor Gorjanc, and Spencer Graves. 2012. “Package ‘car.’” Vienna: R Foundation for Statistical Computing.

Fraser, Hunter B., and Eric E. Schadt. 2010. “The Quantitative Genetics of Phenotypic Robustness.” PLoS ONE 5 (1). doi:10.1371/journal.pone.0008635.

Ganal, Martin W, Thomas Altmann, and Marion S Röder. 2009. “SNP Identification in Crop Plants.” Current Opinion in Plant Biology 12: 211–17. doi:10.1016/j.pbi.2008.12.009.

Garner, Wightman Wells. 1933. “Comparative Responses of Long-Day and Short-Day Plants to Relative Length of Day and Night.” Plant Physiology 8 (3). American Society of Plant Biologists: 347–56.

Garrison, Erik, and Gabor Marth. 2012. “Haplotype-Based Variant Detection from Short-Read Sequencing.” arXiv:1207.3907.

Gelli, Malleswari, Sharon E. Mitchell, Kan Liu, Thomas E. Clemente, Donald P. Weeks, Chi Zhang, David R. Holding, and Ismail M. Dweikat. 2016. “Mapping QTLs and Association of Differentially Expressed Gene Transcripts for Multiple Agronomic Traits under Different Nitrogen Levels in Sorghum.” BMC Plant Biology 16 (16). BMC Plant Biology: 1–18. doi:10.1186/s12870-015-0696-x.

Gifford, Justin M., Won Byong Chae, Kankshita Swaminathan, Stephen P. Moose, and John A. Juvik. 2015. “Mapping the Genome of Miscanthus Sinensis for QTL Associated with Biomass Productivity.” Global Change Biology Bioenergy 7: 797–810. doi:10.1111/gcbb.12201.

189

Glaubitz, Jeffrey C., Terry M. Casstevens, Fei Lu, James Harriman, Robert J. Elshire, Qi Sun, and Edward S. Buckler. 2014. “TASSEL-GBS: A High Capacity Genotyping by Sequencing Analysis Pipeline.” PLoS ONE 9 (2). doi:10.1371/journal.pone.0090346.

Gonzalez, Boldomero, and Wayne W. Hanna. 1984. “Morphological and Fertility Responses in Isogenic Triploid and Hexaploid Pearl Millet X Napiergrass Hybrids.” The Journal of Heredity 75: 317–18.

Grabowski, Paul P., Joseph Evans, Chris Daum, Shweta Deshpande, Kerrie W. Barry, Megan Kennedy, Guillaume Ramstein, et al. 2016. “Genome-Wide Associations with Flowering Time in Switchgrass Using Exome-Capture Sequencing Data.” New Phytologist. doi:10.1111/nph.14101.

Gupta, S C, and O Mhere. 1997. “Identification of Superior Pearl Millet by Napier Hybrids and Napiers in Zimbabwe.” African Crop Science Journal 5 (3). African Crop Science Society: 229–37.

Haggard, J. Erron, Emily B. Johnson, and Dina A. St. Clair. 2015. “Multiple QTL for Horticultural Traits and Quantitative Resistance to Phytophthora Infestans Linked on Solanum Habrochaites Chromosome 11.” G3.Genes|Genomes|Genetics 5: 219–33. doi:10.1534/g3.114.014654.

Hanna, Wayne W. 1981. “Method of Reproduction in Napiergrass and in the 3X and 6X Alloploid Hybrids with Pearl Millet.” Crop Science. Crop Science Society of America, 123–26.

———. 1987. “Utilization of Wild Relatives of Pearl Millet.” In Proceedings of the International Pearl Millet Workshop, 33–42. Patancheru, India: ICRISAT.

Hanna, Wayne W., and Warren G Monson. 1980. “Yield, Quality, and Breeding

Behavior of Pearl Millet ✕ Napiergrass Interspecific Hybrids.” Agronomy Journal

72 (2). American Society of Agronomy: 358–60.

Hanna, Wayne W., H D Wells, and Glenn W. Burton. 1987. “Registration of Pearl Millet Inbred Parental Lines, Tift 85D2A1 and Tift 85D2B1.” Crop Science 27: 1324–25.

Harlan, J R, and J M J de Wet. 1971. “Toward a Rational Classification of Cultivated Plants.” Taxon 20 (4). International Association for Plant Taxonomy (IAPT): 509–17. doi:10.2307/1218252.

Harris-Shultz, Karen R., William F. Anderson, and Ravindra Malik. 2010. “Genetic Relationships among Napiergrass (Pennisetum Purpureum Schum.) Nursery Accessions Using AFLP Markers.” Plant Genetic Resources: Characterization and Utilization 8 (1). Cambridge Univ Press: 63–70.

190

Hart, G. E., K. F. Schertz, Y. Peng, and N. H. Syed. 2001. “Genetic Mapping of Sorghum Bicolor (L.) Moench QTLs That Control Variation in Tillering and Other Morphological Characters.” Theoretical and Applied Genetics 103: 1232–42. doi:10.1007/s001220100582.

Havey, Michael J. 2004. “The Use of Cytoplasmic Male Sterility for Hybrid Seed Production.” In Molecular Biology and Biotechnology of Plant Organelles, edited by H. Daniell and C. D. Chase, 623–34. doi:10.1007/978-1-4020-3166-3_23.

Huang, Xiaoen, Fangfang Wang, Ratnesh Singh, James A Reinert, M C Engelke, Anthony D Genovesi, Ambika Chandra, and Qingyi Yu. 2016. “Construction of High-Resolution Genetic Maps of Zoysia Matrella ( L .) Merrill and Applications to Comparative Genomic Analysis and QTL Mapping of Resistance to Fall Armyworm.” BMC Genomics 17 (562): 1–16. doi:10.1186/s12864-016-2969-7.

Huang, Xiaoqiu, and Anup Madan. 1999. “CAP3: A DNA Sequence Assembly Program.” Genome Research 9: 868–77. doi:10.1101/gr.9.9.868.

Hussain, Waseem, P. Stephen Baenziger, Vikas Belamkar, Mary J. Guttieri, Jorge P. Venegas, Amanda Easterly, Ahmed Sallam, and Jesse A. Poland. 2017. “Genotyping-by-Sequencing Derived High-Density Linkage Map and Its Application to QTL Mapping of Flag Leaf Traits in Bread Wheat.” Scientific Reports 7. Springer US: 1–15. doi:10.1038/s41598-017-16006-z.

ICRISAT, and FAO. 1996. The World Sorghum and Millet Economies: Facts, Trends and Outlook. Patancheru, India: International Crops Research Institute for the Semi-Arid Tropics, and Rome, Italy: Food and Agricultural Organization of the United Nations.

Jauhar, Prem P. 1981. Cytogenetics and Breeding of Pearl Millet and Related Species. Alan R. Liss, Inc. Progress and Topics in Cytogenetics, Volume 1.

Jauhar, Prem P, and Wayne W. Hanna. 1998. Cytogenetics and Genetics of Pearl Millet. Edited by Donald L. Sparks. Advances in Agronomy. Vol. 64. New York: Academic Press, 1949-. doi:10.1017/CBO9781107415324.004.

Jessup, Russell W. 2013. “‘Seeded-yet-Sterile’ Perennial Biofuel Feedstocks.” Advances in Crop Science and Technology 1 (2): 1–2.

Jiang, Guo-liang. 2013. “Molecular Markers and Marker-Assisted Breeding in Plants.” In Plant Breeding from Laboratories to Fields, 45–83. InTech.

Jombart, Thibaut. 2008. “Adegenet: A R Package for the Multivariate Analysis of Genetic Markers.” Bioinformatics 24 (11): 1403–5. doi:10.1093/bioinformatics/btn129.

191

Juma, Awalla Borniface. 2014. “Genetic Diversity and Population Structure of Napier Grass (Pennisetum Purpureum Schumach) in Western Kenya.” University of Eldoret, Kenya.

Jung, Christian, Gina Capistrano-Gossmann, Janina Braatz, Niharika Sashidhar, and Siegbert Melzer. 2018. “Recent Developments in Genome Editing and Applications in Plant Breeding.” Plant Breeding 137 (1): 1–9. doi:10.1111/pbr.12526.

Jung, Christian, and Andreas E Müller. 2009. “Flowering Time Control and Applications in Plant Breeding.” Trends in Plant Science 14 (10): 563–73. doi:10.1016/j.tplants.2009.07.005.

Kandel, Raju, Hari P. Singh, Bharat P. Singh, Karen R. Harris-Shultz, and William F. Anderson. 2016. “Assessment of Genetic Diversity in Napier Grass (Pennisetum Purpureum Schum.) Using Microsatellite, Single-Nucleotide Polymorphism and Insertion-Deletion Markers from Pearl Millet (Pennisetum Glaucum [L.] R. Br.).” Plant Molecular Biology Reporter 34: 265–72. doi:10.1007/s11105-015-0918-2.

Kannan, Baskaran, Senthilvel Senapathy, Arcot Gajaraj Bhasker Raj, Subhash Chandra, Arunachalam Muthiah, Arun Prabhu Dhanapal, and C Thomas Hash. 2014. “Association Analysis of SSR Markers with Phenology, Grain, and Stover-Yield Related Traits in Pearl Millet (Pennisetum Glaucum (L.) R. Br.).” The ScientificWorld Journal 2014: 1–14. doi:10.1155/2014/562327.

Karkkainen, Katri, Veikko Koski, and Outi Savolainen. 1996. “Geographical Variation in the Inbreeding Depression of Scots Pine.” Evolution 50 (1): 111. doi:10.2307/2410785.

Karlsson, Sylvia B., and Indra K. Vasil. 1986. “Growth, Cytology and Flow Cytometry of Embryogenic Cell Suspension Cultures of Panicum Maximum Jacq. and Pennisetum Purpureum Schum.” Journal of Plant Physiology 123. Gustav Fischer Verlag, Stuttgart: 211–27. doi:10.1016/S0176-1617(86)80072-4.

Kearsey, Mike J., and A. G. L. Farquhar. 1998. “QTL Analysis in Plants ; Where Are We Now?” Heredity 80: 137–42.

Kellis, Manolis, Bruce W Birren, and Eric S Lander. 2004. “Proof and Evolutionary Analysis of Ancient Genome Duplication in the Yeast Saccharomyces Cerevisiae.” Nature 428 (6983): 617–24. doi:10.1038/nature02424.

Kent, W James. 2002. “BLAT — The BLAST -Like Alignment Tool.” Genome Research 12: 656–64. doi:10.1101/gr.229202.

192

Khan, Z. R., J. A. Pickett, L. Wadhams, and F. Muyekho. 2001. “Habitat Management Strategies for the Control of Cereal Stemborers and Striga in Maize in Kenya.” Insect Science and Its Application 21 (4): 375–80. doi:10.1017/S1742758400008481.

Kim, Changsoo, Hui Guo, Wenqian Kong, Rahul Chandnani, Lan-Shuan Shuang, and Andrew H. Paterson. 2015. “Application of Genotyping by Sequencing Technology to a Variety of Crop Breeding Programs.” Plant Science 242. Elsevier Ireland Ltd: 14–22. doi:10.1016/j.plantsci.2015.04.016.

Kojima, Shoko, Yuji Takahashi, Yasushi Kobayashi, Lisa Monna, Takuji Sasaki, Takashi Araki, and Masahiro Yano. 2002. “Hd3a, a Rice Ortholog of the Arabidopsis FT Gene, Promotes Transition to Flowering Downstream of Hd1 under Short-Day Conditions.” Plant & Cell Physiology 43 (10): 1096–1105. doi:10.1093/pcp/pcf156.

Krzywinski, Martin, Jacqueline Schein, Inanc Birol, Joseph Connors, Randy Gascoyne, Doug Horsman, Steven J Jones, and Marco A Marra. 2009. “Circos: An Information Aesthetic for Comparative Genomics.” Genome Research 19 (9). Cold Spring Harbor Lab: 1639–45.

Kuittinen, H, M J Sillanpa, and O Savolainen. 1997. “Genetic Basis of Adaptation : Flowering Time in Arabidopsis Thaliana.” Theoretical and Applied Genetics 95: 573–83.

Kumar, Lalitha Sunil. 1999. “DNA Markers in Plant Improvement: An Overview.” Biotechnology Advances 17 (September): 143–82.

Lagercrantz, Ulf, Joanna Putterill, George Coupland, and Derek Lydiate. 1996. “Comparative Mapping in Arabidopsis and Brassica, Fine Scale Genome Collinearity and Congruence of Genes Controlling Flowering Time.” The Plant Journal 9 (1): 13–20.

Lambert, Allison M., Abraham J. Miller-Rushing, and David W. Inouye. 2010. “Changes in Snowmelt Date and Summer Precipitation Affect the Flowering Phenology of Erythronium Grandiflorum (Glacier Lily; Liliaceae).” American Journal of Botany 97 (9): 1431–37. doi:10.3732/ajb.1000095.

Langmead, Ben, and Steven L Salzberg. 2012. “Fast Gapped-Read Alignment with Bowtie 2.” Nature Methods 9 (4): 357–59. doi:10.1038/nmeth.1923.

Laurie, David A., Nicky Pratchett, Jeremy H Bezant, and John W Snapet. 1994. “Genetic Analysis of a Photoperiod Response Gene on the Short Arm of Chromosome 2(2H) of Hordeum Vulgare (Barley).” Heredity 72: 619–27.

193

Lee, Ilha, Anthony Bleecker, and Richard Amasino. 1993. “Analysis of Naturally Occurring Late Flowering in Arabidopsis Thaliana.” Molecular and General Genetics MGG 237. Springer: 171–76.

Lewis, Michael W., Michelle E. Leslie, Emilee H. Fulcher, Lalitree Darnielle, Patrick Healy, Ji-Young Youn, and Sarah J. Liljegren. 2010. “The SERK1 Receptor-like Kinase Regulates Organ Separation in Arabidopsis Flowers.” Plant Journal 62 (5): 817–28. doi:10.1111/j.1365-313X.2010.04194.x.

Li, Heng, and Richard Durbin. 2010. “Fast and Accurate Long-Read Alignment with Burrows-Wheeler Transform.” Bioinformatics (Oxford, England) 26 (5). Wellcome Trust Sanger Institute, Wellcome Genome Campus, Cambridge, CB10 1SA, UK. DOI - 10.1093/bioinformatics/btp698 SRC - Pubmed ID2 - 20080505 FG - 0: 589–95. doi:10.1093/bioinformatics/btp698.

Li, Heng, Bob Handsaker, Alec Wysoker, Tim Fennell, Jue Ruan, Nils Homer, Gabor Marth, Goncalo Abecasis, Richard Durbin, and 1000 Genome Project Data Processing Subgroup. 2009. “The Sequence Alignment/map Format and SAMtools.” Bioinformatics 25 (16). Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, Cambridge, CB10 1SA, UK, Broad Institute of MIT and Harvard, Cambridge, MA 02141, USA. DOI - 10.1093/bioinformatics/btp352 SRC - Pubmed ID2 - 19505943 FG - 0: 2078–79. doi:10.1093/bioinformatics/btp352.

Li, Weizhong, and Adam Godzik. 2006. “Cd-Hit: A Fast Program for Clustering and Comparing Large Sets of Protein or Nucleotide Sequences.” Bioinformatics 22 (13): 1658–59. doi:10.1093/bioinformatics/btl158.

Li, Wuxing, Yun Zhou, Xing Liu, Peng Yu, Jerry D Cohen, and Elliot M Meyerowitz. 2013. “LEAFY Controls Auxin Response Pathways in Floral Primordium Formation.” Science Signaling 6 (270): 1–6. doi:10.1126/scisignal.2003937.

Li, Yuan, Kai Yang, Wei Yang, Liwei Chu, Chunhai Chen, Bo Zhao, Yisong Li, et al. 2017. “Identification of QTL and Qualitative Trait Loci for Agronomic Traits Using SNP Markers in the Adzuki Bean.” Frontiers in Plant Science 8: 1–14. doi:10.3389/fpls.2017.00840.

Lindhout, Pim, Sjaak Van Heusden, Gerard Pet, Johan W. van Ooijen, Hans Sandbrink, Ruud Verkerk, Ria Vrielink, and Pim Zabel. 1994. “Perspectives of Molecular Marker Assisted Breeding for Earliness in Tomato.” Euphytica 79 (3). Springer: 279–86.

Loope, Lloyd L, Ole Hamann, and Charles P Stone. 1988. “Comparative Conservation Biology of Oceanic Archipelagoes: Hawaii and the Galapagos.” BioScience 38 (4). JSTOR: 272–82.

194

López, Yolanda, Aleksey Kurashev, Christine Chase, Maria Gallo, Lynn E. Sollenberger, Fredy Altpeter, and Jianping Wang. 2018. “Developing and Validating Microsatellite Markers in Elephant Grass (Pennisetum Purpureum S.).” Euphytica 214 (185): 1–`3. doi:10.1007/s10681-018-2256-6.

López, Yolanda, Jeffery Seib, Kenneth R. Woodard, Karen Chamusco, Lynn E. Sollenberger, Maria Gallo, S. Luke Flory, and Christine Chase. 2014. “Genetic Diversity of Biofuel and Naturalized Napiergrass (Pennisetum Purpureum).” Invasive Plant Science and Management 7 (2). Weed Science Society of America 810 East 10th Street, Lawrence, KS 66044-8897: 229–36. doi:10.1614/IPSM-D-13-00085.1.

Lowe, Andrew J, W Thorpe, A Teale, and J Hanson. 2003. “Characterisation of Germplasm Accessions of Napier Grass (Pennisetum Purpureum and P. Purpureum× P. Glaucum Hybrids) and Comparison with Farm Clones Using RAPD.” Genetic Resources and Crop Evolution 50 (2). Springer: 121–32.

Lu, Fei, Alexander E. Lipka, Jeffrey C. Glaubitz, Robert J. Elshire, Jerome H. Cherney, Michael D. Casler, Edward S. Buckler, and Denise E. Costich. 2013. “Switchgrass Genomic Diversity, Ploidy, and Evolution: Novel Insights from a Network-Based SNP Discovery Protocol.” PLoS Genetics 9 (1): 1–14. doi:10.1371/journal.pgen.1003215.

Lu, Hongfeng, Tao Lin, Joël Klein, Shenhao Wang, Jianjian Qi, Qian Zhou, Jinjing Sun, Zhonghua Zhang, Yiqun Weng, and Sanwen Huang. 2014. “QTL-Seq Identifies an Early Flowering QTL Located near Flowering Locus T in Cucumber.” Theoretical and Applied Genetics 127. Springer: 1–9. doi:10.1007/s00122-014-2313-z.

Ma, Xue Feng, Elaine Jensen, Nickolai Alexandrov, Maxim Troukhan, Liping Zhang, Sian Thomas-Jones, Kerrie Farrar, et al. 2012. “High Resolution Genetic Mapping by Genome Sequencing Reveals Genome Duplication and Tetraploid Genetic Structure of the Diploid Miscanthus Sinensis.” PLoS ONE 7 (3): 1–11. doi:10.1371/journal.pone.0033821.

Maheswaran, M, Ning Huang, S R Sreerangasamy, and Susan R McCouch. 2000. “Mapping Quantitative Trait Loci Associated with Days to Flowering and Photoperiod Sensitivity in Rice (Oryza Sativa L.).” Molecular Breeding 6. Springer: 145–55.

Makumburage, G. Buddhika, and Ann E. Stapleton. 2011. “Phenotype Uniformity in Combined-Stress Environments Has a Different Genetic Architecture than in Single-Stress Treatments.” Frontiers in Plant Science 2: 1–8. doi:10.3389/fpls.2011.00012.

195

Martel, E., D De Nay, S Siljak-Yakoviev, S. Brown, A. Sarr, D. De Nay, S. Siljak-Yakovlev, et al. 1997. “Genome Size Variation and Basic Chromosome Number in Pearl Millet and Fourteen Related Pennisetum Species.” Journal of Heredity 88. Am Genetic Assoc: 139–43.

Martel, E., V. Poncet, F. Lamy, S. Siljak-Yakovlev, B. Lejeune, and A. Sarr. 2004. “Chromosome Evolution of Pennisetum Species (Poaceae): Implications of ITS Phylogeny.” Plant Systematics and Evolution 249: 139–49. doi:10.1007/s00606-004-0191-6.

Martel, E., A Ricroch, and A Sarr. 1996. “Assessment of Genome Organization among Diploid Species (2n = 2x = 14) Belonging to Primary and Tertiary Gene Pools of Pearl Millet Using Fluorescent in Situ Hybridization with rDNA Probes.” Genome 39. NRC Research Press: 680–87.

Martin, K. L., P. J. Hodgen, K. W. Freeman, Ricardo Melchiori, D. B. Arnall, R. K. Teal, R. W. Mullen, et al. 2005. “Plant-to-Plant Variability in Corn Production.” Agronomy Journal 97 (6): 1603–11. doi:10.2134/agronj2005.0129.

Mathew, Lisa S, Manuel Spannagl, Ameena Al-malki, Binu George, Maria F Torres, Eman K Al-dous, Eman K Al-azwani, et al. 2014. “A First Genetic Map of Date Palm (Phoenix Dactylifera) Reveals Long-Range Genome Structure Conservation in the Palms.” BMC Genomics 15 (285): 1–10.

Matsubara, Kazuki, I. Kono, K. Hori, Y. Nonoue, N. Ono, A. Shomura, T. Mizubayashi, et al. 2008. “Novel QTLs for Photoperiodic Flowering Revealed by Using Reciprocal Backcross Inbred Lines from Crosses between Japonica Rice Cultivars.” Theoretical and Applied Genetics 117: 935–45. doi:10.1007/s00122-008-0833-0.

Matsubara, Kazuki, Eri Ogiso-Tanaka, Kiyosumi Hori, Kaworu Ebana, Tsuyu Ando, and Masahiro Yano. 2012. “Natural Variation in Hd17, a Homolog of Arabidopsis ELF3 That Is Involved in Rice Photoperiodic Flowering.” Plant and Cell Physiology 53 (4): 709–16. doi:10.1093/pcp/pcs028.

McCouch, Susan R, Y. G. Cho, M. Yano, E. Paul, M. Blinstrub, H. Morishima, and T. Kinoshita. 1997. “Report on QTL Nomenclature.” Rice Genetics Newsletter 14: 11–13.

Mckendry, Peter. 2002. “Energy Production from Biomass (Part 1): Overview of Biomass.” Bioresource Technology 83: 37–46.

McKenna, Aaron, Matthew Hanna, Eric Banks, Andrey Sivachenko, Kristian Cibulskis, Andrew Kernytsky, Kiran Garimella, et al. 2010. “The Genome Analysis Toolkit: A MapReduce Framework for Analyzing next-Generation DNA Sequencing Data.” Genome Research 20 (9): 1297–1303. doi:10.1101/gr.107524.110.

196

Melo, Arthur T. O., Radhika Bartaula, and Iago Hale. 2016. “GBS-SNP-CROP: A Reference-Optional Pipeline for SNP Discovery and Plant Germplasm Characterization Using Variable Length, Paired-End Genotyping-by-Sequencing Data.” BMC Bioinformatics 17 (29): 1–15. doi:10.1186/s12859-016-0879-y.

Miles, Cecelia M., and Marta Wayne. 2008. “Quantitative Trait Locus ( QTL ) Analysis.” Nature Education 1 (208): 1–7.

Morris, Karl, Sarah Thornber, Lesley Codrai, Christine Richardson, Adam Craig, Ari Sadanandom, Brian Thomas, and Stephen Jackson. 2010. “DAY NEUTRAL FLOWERING Represses CONSTANS to Prevent Arabidopsis Flowering Early in Short Days.” The Plant Cell 22: 1118–28. doi:10.1105/tpc.109.066605.

Muldoon, D K, and C J Pearson. 1979. “The Hybrid between Pennisetum Americanum and Pennisetum Purpureum.” Herbage Abstracts 49 (5): 189–99.

Murray, Seth C., William L. Rooney, Sharon E. Mitchell, Arun Sharma, Patricia E. Klein, John E. Mullet, and Stephen Kresovich. 2008. “Genetic Improvement of Sorghum as a Biofuel Feedstock: II. QTL for Stem and Leaf Structural Carbohydrates.” Crop Science 48: 2180–93.

Nagy, Istvan, Susanne Barth, Jeanne Mehenni-Ciz, Michael T Abberton, and Dan Milbourne. 2013. “A Hybrid next Generation Transcript Sequencing-Based Approach to Identify Allelic and Homeolog-Specific Single Nucleotide Polymorphisms in Allotetraploid White Clover.” BMC Genomics 14 (100): 1–19. doi:10.1186/1471-2164-14-100.

Negawo, Alemayehu Teressa, Abel Teshome, Alok Kumar, Jean Hanson, and Chris S Jones. 2017. “Opportunities for Napier Grass (Pennisetum Purpureum) Improvement Using Molecular Genetics.” Agronomy 7 (28): 1–21.

Noble Research Institute. 2007. “Switchgrass Establishment Requires Patience.” Ag News & Views. https://www.noble.org/news/publications/ag-news-and-views/2007/july/switchgrass-establishment-requires-patience/.

Obok, Ekemini Edet, Micheal E Aken Ova, and Godfrey A Iwo. 2012. “Forage Potentials of Interspecific Hybrids between Elephant Grass Selections and Cultivated Pearl Millet Genotypes of Nigerian Origin.” Journal of Plant Breeding and Crop Science 4 (9): 136–43. doi:10.5897/JPBCS12.019.

Osborn, Thomas C, C. Kole, I. A. P. Parkin, A. G. Sharpe, M. Kuiper, D. J. Lydiate, and Martin Trick. 1997. “Comparison of Flowering Time Genes in Brassica Rapa, B. Napus and Arabidopsis Thaliana.” Genetics 146: 1123–29.

Osgood, RV V, Wayne W. Hanna, and TL L Tew. 1997. “Hybrid Seed Production of Pearl Millet × Napiergrass Triploid Hybrids.” Crop Science 37. Crop Science Society of America: 998–99.

197

Paillard, S., T. Schnurbusch, M. Winzeler, M. Messmer, P. Sourdille, O. Abderhalden, B. Keller, and G. Schachermayr. 2003. “An Integrative Genetic Linkage Map of Winter Wheat (Triticum Aestivum L.).” Theoretical and Applied Genetics 107: 1235–42. doi:10.1007/s00122-003-1361-6.

Pandey, Manish K., Emmanuel Monyo, Peggy Ozias-Akins, Xuanquiang Liang, Patricia Guimares, Shyam N. Nigam, Hari D. Upadhyaya, et al. 2012. “Advances in Arachis Genomics for Peanut Improvement.” Biotechnology Advances 30. Elsevier Inc.: 639–51. doi:10.1016/j.biotechadv.2011.11.001.

Paterson, Andrew H., John E Bowers, Rémy Bruggmann, Inna Dubchak, Jane Grimwood, Heidrun Gundlach, Georg Haberer, et al. 2009. “The Sorghum Bicolor Genome and the Diversification of Grasses.” Nature 457 (7229): 551–56. doi:10.1038/nature07723.

Paudel, Dev R., Baskaran Kannan, Xiping Yang, Karen R. Harris-Shultz, Mahendar Thudi, Rajeev K. Varshney, Fredy Altpeter, and Jianping Wang. 2018. “Surveying the Genome and Constructing a High-Density Genetic Map of Napiergrass (Cenchrus Purpureus Schumach).” Scientific Reports 8 (1). Springer US: 1–11. doi:10.1038/s41598-018-32674-x.

Paux, Etienne, Pierre Sourdille, Ian Mackay, and Catherine Feuillet. 2012. “Sequence-Based Marker Development in Wheat: Advances and Applications to Breeding.” Biotechnology Advances 30. Elsevier Inc.: 1071–88. doi:10.1016/j.biotechadv.2011.09.015.

Peng, Ze, Wen Fan, Liping Wang, Dev R. Paudel, Dante Leventini, Barry L. Tillman, and Jianping Wang. 2017. “Target Enrichment Sequencing in Cultivated Peanut (Arachis Hypogaea L.) Using Probes Designed from Transcript Sequences.” Molecular Genetics and Genomics. Springer Berlin Heidelberg, 1–11. doi:10.1007/s00438-017-1327-z.

Phan, Huyen T.T., Kasia Rybak, Stefania Bertazzoni, Eiko Furuki, Eric Dinglasan, Lee T. Hickey, Richard P. Oliver, and Kar Chun Tan. 2018. “Novel Sources of Resistance to Septoria Nodorum Blotch in the Vavilov Wheat Collection Identified by Genome-Wide Association Studies.” Theoretical and Applied Genetics 131. Springer Berlin Heidelberg: 1223–38. doi:10.1007/s00122-018-3073-y.

Phukan, Ujjal J., Gajendra S. Jeena, and Rakesh K. Shukla. 2016. “WRKY Transcription Factors: Molecular Regulation and Stress Responses in Plants.” Frontiers in Plant Science 7: 1–14. doi:10.3389/fpls.2016.00760.

Poland, Jesse A., Patrick J Brown, Mark E Sorrells, and Jean-Luc Jannink. 2012. “Development of High-Density Genetic Maps for Barley and Wheat Using a Novel Two-Enzyme Genotyping-by-Sequencing Approach.” PloS One 7 (2). Public Library of Science: 1–8.

198

Poland, Jesse A., J Endelman, J Dawson, J Rutkoski, S Y Wu, Y Manes, S Dreisigacker, et al. 2012. “Genomic Selection in Wheat Breeding Using Genotyping-by-Sequencing.” Plant Genome 5 (3): 103–13. doi:Doi 10.3835/Plantgenome2012.06.0006.

Poland, Jesse A., and Trevor W Rife. 2012. “Genotyping-by-Sequencing for Plant Breeding and Genetics.” The Plant Genome 5 (3). Crop Science Society of America: 92–102.

Pootakham, Wirulda, Nukoon Jomchai, Panthita Ruang-areerate, Jeremy R. Shearman, Chutima Sonthirod, Duangjai Sangsrakru, Somvong Tragoonrung, and Sithichoke Tangphatsornruang. 2015. “Genome-Wide SNP Discovery and Identification of QTL Associated with Agronomic Traits in Oil Palm Using Genotyping-by-Sequencing (GBS).” Genomics 105. Elsevier B.V.: 288–95. doi:10.1016/j.ygeno.2015.02.002.

Powell, Jerrel B, and Glenn W. Burton. 1966. “A Suggested Commercial Method of Producing an Interspecific Hybrid Forage in Pennisetum.” Crop Science 6. Crop Science Society of America: 378–79.

Powell, Wayne, Gordon C Machray, and Jim Provan. 1996. “Polymorphism Revealed by Simple Sequence Repeats.” Trends in Plant Science 1 (7). Elsevier: 215–22.

Price, David L., and Michael D. Casler. 2014. “Divergent Selection for Secondary Traits in Upland Tetraploid Switchgrass and Effects on Sward Biomass Yield.” Bioenergy Research 7: 329–37. doi:10.1007/s12155-013-9374-8.

Pruitt, Kim D., Tatiana Tatusova, and Donna R. Maglott. 2007. “NCBI Reference Sequences (RefSeq): A Curated Non-Redundant Sequence Database of Genomes, Transcripts and Proteins.” Nucleic Acids Research 33: 501–4. doi:10.1093/nar/gkl842.

Punnuri, Somashekhar M., Jason G. Wallace, Joseph E. Knoll, Katie E. Hyma, Sharon E. Mitchell, Edward S. Buckler, Rajeev K. Varshney, Bharat P. Singh, and T H E Plant Genome. 2016. “Development of a High-Density Linkage Map and Tagging Leaf Spot Resistance in Pearl Millet Using GBS Markers.” Plant Genome 9 (2): 1–13. doi:10.3835/plantgenome2015.10.0106.

Ra, Kaien, Fumitaka Shiotsu, Jun Abe, and Shigenori Morita. 2012. “Biomass Yield and Nitrogen Use Efficiency of Cellulosic Energy Crops for Ethanol Production.” Biomass and Bioenergy 37. Elsevier: 330–34.

Rafalski, Antoni. 2002. “Applications of Single Nucleotide Polymorphisms in Crop Genetics.” Current Opinion in Plant Biology 5. Elsevier: 94–100.

199

Rajaram, Vengaldas, Thirunavukkarasu Nepolean, Senapathy Senthilvel, Rajeev K. Varshney, Vincent Vadez, Rakesh K Srivastava, Trushar M Shah, Ambawat Supriya, Sushil Kumar, and Basava Ramana Kumari. 2013. “Pearl Millet [Pennisetum Glaucum (L.) R. Br.] Consensus Linkage Map Constructed Using Four RIL Mapping Populations and Newly Developed EST-SSRs.” BMC Genomics 14 (159). BioMed Central Ltd: 1–15.

Ram, Hari Har, Hari Har Ram, and Rakesh Yadava. 2007. Genetic Resources and Seed Enterprises: Management and Policies. New India Publishing.

Reis, Gabriela Barreto dos, Amanda Teixeira Mesquita, Giovana Augusta Torres, Larissa Fonseca Andrade-Vieira, Antônio Vander Pereira, and Lisete Chamma Davide. 2014. “Genomic Homeology between Pennisetum Purpureum and Pennisetum Glaucum (Poaceae).” Comparative Cytogenetics 8 (3): 199–209. doi:10.3897/CompCytogen.v8i3.7732.

Rice, Peter, Lan Longden, and Alan Bleasby. 2000. “EMBOSS: The European Molecular Biology Open Software Suite.” Trends in Genetics 16 (6): 276–77. doi:10.1016/S0168-9525(00)02024-2.

Richmond, Todd A., and Anthony B. Bleecker. 1999. “A Defect in Beta-Oxidation Causes Abnormal Inflorescence Development in Arabidopsis.” The Plant Cell 11: 1911–24. doi:10.1105/tpc.11.10.1911.

Rosyara, Umesh R., Walter S. De Jong, David S. Douches, and Jeffrey B. Endelman. 2016. “Software for Genome-Wide Association Studies in Autopolyploids and Its Application to Potato.” The Plant Genome 9 (2): 1–10. doi:10.3835/plantgenome2015.08.0073.

Rozen, S, and H Skaletsky. 2000. “Primer3 on the WWW for General Users and for Biologist Programmers.” Methods in Molecular Biology 132. Whitehead Institute for Biomedical Research, Cambridge, MA, USA. SRC - Pubmed ID2 - 10547847 FG - 0: 365–86.

Russell, Joanne, Martin Mascher, Ian K. Dawson, Stylianos Kyriakidis, Cristiane Calixto, Fabian Freund, Micha Bayer, et al. 2016. “Exome Sequencing of Geographically Diverse Barley Landraces and Wild Relatives Gives Insights into Environmental Adaptation.” Nature Genetics 48 (9). Nature Publishing Group: 1024–30. doi:10.1038/ng.3612.

Samach, Alon. 2012. “Control of Flowering.” In Plant Biotechnology and Agriculture:, First Edit, 387–404. Elsevier Inc. doi:10.1016/B978-0-12-381466-1.00025-0.

Schlegel, Rolf H J. 2009. Dictionary of Plant Breeding. CRC Press.

Schnable, Patrick S., Doreen Ware, RS Fulton, and JC Stein. 2009. “The B73 Maize Genome: Complexity, Diversity, and Dynamics.” Science 326 (5956): 1112–15.

200

Schneider, Katharina. 2005. “Mapping Populations and Principles of Genetic Mapping.” In The Handbook of Plant Genome Mapping: Genetic and Physical Mapping, edited by Khalid Meksem and Gunter Kahl, 1–21. WILEY-VCH Verlag GmbH & Co. KGaA. doi:10.1002/3527603514.ch1.

Schofield, Eileen K. 1989. “Effects of Introduced Plants and Animals on Island Vegetation: Examples from Galápagos Archipelago.” Conservation Biology 3 (3). Wiley Online Library: 227–39.

Semagn, Kassa, Raman Babu, Sarah Hearne, and Michael Olsen. 2014. “Single Nucleotide Polymorphism Genotyping Using Kompetitive Allele Specific PCR (KASP): Overview of the Technology and Its Application in Crop Improvement.” Molecular Breeding. doi:10.1007/s11032-013-9917-x.

Shen, Jiawei, Zhiqiang Li, Jianhua Chen, Zhijian Song, Zhaowei Zhou, and Yongyong Shi. 2016. “SHEsisPlus, a Toolset for Genetic Studies on Polyploid Species.” Scientific Reports 6. Nature Publishing Group: 1–10. doi:10.1038/srep24095.

Shepherd, M., M. Cross, M. J. Dieters, and R. Henry. 2003. “Genetic Maps for Pinus Elliottii Var. Elliottii and P. Caribaea Var. Hondurensis Using AFLP and Microsatellite Markers.” Theoretical and Applied Genetics 106: 1409–19. doi:10.1007/s00122-002-1185-9.

Simpson, Jared T, Kim Wong, Shaun D Jackman, Jacqueline E Schein, Steven J M Jones, Inanç Birol, Jared T Simpson, et al. 2009. “ABySS: A Parallel Assembler for Short Read Sequence Data.” Genome Research 19 (June): 1117–23. doi:10.1101/gr.089532.108.

Sinche, Marco. 2013. “Genetic Improvement of Elephantgrass (Pennisetum Purpureum Schum.) as a Bioenergy Feedstock.” University of Florida.

Sinche, Marco, Baskaran Kannan, Carlos Corsato, Dev R. Paudel, Jianping Wang, and Fredy Altpeter. 2018. “Development and Characterization of Napiergrass (Pennisetum Purpureum) Mapping Population for Flowering Time and Biomass Related Traits.” Submitted for Publication.

Singer, S. D., N. T. Krogan, and N. W. Ashton. 2007. “Clues about the Ancestral Roles of Plant MADS-Box Genes from a Functional Analysis of Moss Homologues.” Plant Cell Reports 26: 1155–69. doi:10.1007/s00299-007-0312-0.

Singh, Bharat P., Hari P Singh, and Eric Obeng. 2013. “Elephantgrass.” In Biofuel Crops: Production, Physiology and Genetics, edited by Bharat P Singh, 271–91. CAB International.

Smith, Rex L., and M K U Chowdhury. 1983. “Mitochondrial DNA Polymorphism in Male-Sterile and Fertile Cytoplasms of Pearl Millet.” Crop Science 29: 809–14.

201

Smith, Rex L., M. E. Schweder, M. K. U. Chowdhury, J. C. Seib, and S. C. Schank. 1993. “Development and Application of RFLP and RAPD DNA Markers in Genetic Improvement of Pennisetum for Biomass and Forage Production.” Biomass and Bioenergy 5 (1): 51–62.

Sollenberger, Lynn E., G M Prine, W R Ocumpaugh, Wayne W. Hanna, C S Jones, S C Schank, and R S Kalmbacher. 1989. “Registration of ‘Mott’dwarf Elephantgrass.” Crop Science 29. Crop Science Society of America: 827–28.

Sollenberger, Lynn E., Kenneth R. Woodard, João M.B. B. Vendramini, John E. Erickson, Kenneth A. Langeland, M. Kimberly Mullenix, Chaein Na, et al. 2014. “Invasive Populations of Elephantgrass Differ in Morphological and Growth Characteristics from Clones Selected for Biomass Production.” Bioenergy Research 7 (May): 1382–91. doi:10.1007/s12155-014-9478-9.

Somerville, Chris, Heather Youngs, Caroline Taylor, Sarah C Davis, and Stephen P Long. 2010. “Feedstocks for Lignocellulosic Biofuels.” Science 329: 790–93.

Song, Xiaoming, Tingting Ge, Ying Li, and Xilin Hou. 2015. “Genome-Wide Identification of SSR and SNP Markers from the Non-Heading Chinese Cabbage for Comparative Genomic Analyses.” BMC Genomics 16 (328): 1–18. doi:10.1186/s12864-015-1534-0.

Stinchcombe, JR, and HE Hoekstra. 2008. “Combining Population Genomics and Quantitative Genetics : Finding the Genes Underlying Ecologically Important Traits.” Heredity 100: 158–70. doi:10.1038/sj.hdy.6800937.

Stone, Eric A., and Julien F. Ayroles. 2009. “Modulated Modularity Clustering as an Exploratory Tool for Functional Genomic Inference.” PLoS Genetics 5 (5): 1–13. doi:10.1371/journal.pgen.1000479.

Struwig, M. 2007. “Fingerprinting Pennisetum Purpureum Schumach. Varieties and Cultivars Using AFLP Analyses.” North West University.

Suárez-López, P, K Wheatley, F Robson, H Onouchi, F Valverde, and G Coupland. 2001. “CONSTANS Mediates between the Circadian Clock and the Control of Flowering in Arabidopsis.” Nature 410: 1116–20. doi:10.1038/35074138.

Tamura, Koichiro, Glen Stecher, Daniel Peterson, Alan Filipski, and Sudhir Kumar. 2013. “MEGA6: Molecular Evolutionary Genetics Analysis Version 6.0.” Molecular Biology and Evolution, 2725–29. doi:10.1093/molbev/mst197.

Taylor, D R, and P K Ingvarsson. 2003. “Common Features of Segregation Distortion in Plants and Animals.” Genetica 117: 27–35. doi:10.1023/A:1022308414864.

202

Taylor, M G, and I K Vasil. 1987. “Analysis of DNA Size , Content and Cell Cycle in Leaves of Napier Grass (Pennisetum Purpureum Schum .).” Theoretical and Applied Genetics 74: 681–86.

Thiel, T., W. Michalek, Rajeev K. Varshney, and A. Graner. 2003. “Exploiting EST Databases for the Development and Characterization of Gene-Derived SSR-Markers in Barley (Hordeum Vulgare L.).” Theoretical and Applied Genetics 106. Institute of Plant Genetics and Crop Plant Research (IPK), Corrensstrasse 3, D-06466 Gatersleben, Germany. DOI - 10.1007/s00122-002-1031-0 SRC - Pubmed ID2 - 12589540 FG - 0: 411–22. doi:10.1007/s00122-002-1031-0.

Thompson, J. B. 1919. Napier and Merker Grass : Two New Forage Crops for Florida. Gainesville: University of Florida, Agricultural Experiment Station.

Tollenaar, M, and J Wu. 1999. “Yield Improvement in Temperate Maize Is Attributed to Greater Stress Tolerance.” Crop Science 39: 1597–1604. doi:10.2135/cropsci1999.3961597x.

Tong, Xinggang, Laurence H Smith, and Perry L Mccarty. 1990. “Methane Fermentation of Selected Lignocellulosic Materials.” Biomass 21: 239–55.

Trick, Martin, Yan Long, Jinling Meng, and Ian Bancroft. 2009. “Single Nucleotide Polymorphism (SNP) Discovery in the Polyploid Brassica Napus Using Solexa Transcriptome Sequencing.” Plant Biotechnology Journal 7: 334–46. doi:10.1111/j.1467-7652.2008.00396.x.

Tsai, Yu Chang, Nikkí A. Delk, Naweed I. Chowdhury, and Janet Braam. 2007. “Arabidopsis Potential Calcium Sensors Regulate Nitric Oxide Levels and the Transition to Flowering.” Plant Signaling and Behavior 2 (6): 446–54. doi:10.4161/psb.2.6.4695.

Uitdewilligen, Jan G A M L, Anne Marie A Wolters, Bjorn B. D’hoop, Theo J A Borm, Richard G F Visser, and Herman J. van Eck. 2013. “A next-Generation Sequencing Method for Genotyping-by-Sequencing of Highly Heterozygous Autotetraploid Potato.” PLoS ONE 8 (5): 1–14. doi:10.1371/journal.pone.0062355.

USDA, ARS, and NGRP. 1954. “PI 215603.” Germplasm Resources Information Network - (GRIN). [Online Database] National Germplasm Resources Laboratory, Beltsville, Maryland. http://www.ars-grin.gov/cgi-bin/npgs/acc/display.pl?1176422.

———. 1963. “PI 288787.” Germplasm Resources Information Network - (GRIN). [Online Database] National Germplasm Resources Laboratory, Beltsville, Maryland. http://www.ars-grin.gov/cgi-bin/npgs/acc/display.pl?1218439.

203

van Ooijen, Johan W. 1992. “Accuracy of Mapping Quantitative Trait Loci in Autogamous Species.” Theoretical and Applied Genetics 84: 803–11. doi:10.1007/BF00227388.

———. 2006. “JoinMap 4.0: Software for the Calculation of Genetic Linkage Maps in Experimental Populations.” Wageningen, Netherlands: Kyazma BV.

Varshney, Rajeev K., Chengcheng Shi, Mahendar Thudi, Cedric Mariac, Jason Wallace, Peng Qi, He Zhang, et al. 2017. “Pearl Millet Genome Sequence Provides a Resource to Improve Agronomic Traits in Arid Environments.” Nature Biotechnology 35 (10): 969–76.

Veeckman, E, K Vandepoele, T Asp, and T Ruttink. 2016. “Genomic Variation in the FT Gene Family of Perennial Ryegrass (Lolium Perenne).” In Breeding in a World of Scarcity: Proceedings of the 2015 Meeting of the Section “Forage Crops and Amenity Grasses” of Eucarpia, 121–26. doi:10.1007/978-3-319-28932-8.

Vicente-Chandler, Jose, Servando Silva, and Jacinto Figarella. 1959. “The Effect of Nitrogen Fertilization and Frequency of Cutting on the Yield and Composition of Three Tropical Grasses.” Agronomy Journal 51. American Society of Agronomy: 202–6.

Voorrips, Roeland E. 2002. “MapChart: Software for the Graphical Presentation of Linkage Maps and QTLs.” The Journal of Heredity, 77–78. doi:10.1093/jhered/93.1.77.

Wang, Cuiling, Yanhui Chen, Lixia Ku, Tiegu Wang, Zhaohui Sun, Fangfang Cheng, and Liancheng Wu. 2010. “Mapping QTL Associated with Photoperiod Sensitivity and Assessing the Importance of QTL×environment Interaction for Flowering Time in Maize.” PLoS ONE 5 (11): 8–10. doi:10.1371/journal.pone.0014068.

Wang, Jia, Zhitong Chen, Surong Jin, Zhongli Hu, Yibing Huang, and Ying Diao. 2017. “Development and Characterization of Simple Sequence Repeat (SSR) Markers Based on a Full-Length cDNA Library of Napier Grass (Pennisetum Purpureum Schum).” Genes and Genomics 39 (12). The Genetics Society of Korea: 1297–1305. doi:10.1007/s13258-017-0536-5.

Wang, Shengchu, CJ Basten, P Gaffney, and Z-B Zeng. 2005. “Windows QTL Cartographer 2.5. User Manual.” International Immunology. Department of Statistics, North Carolina State University, Raleigh, NC. doi:10.1093/intimm/dxs150.

Wang, Yan, Silong Sun, Bo Liu, Hui Wang, Jie Deng, Yongcui Liao, Qian Wang, Feng Cheng, Xiaowu Wang, and Jian Wu. 2011. “A Sequence-Based Genetic Linkage Map as a Reference for Brassica Rapa Pseudochromosome Assembly.” BMC Genomics 12 (1): 239. doi:10.1186/1471-2164-12-239.

204

Wanjala, Bramwel W., Meshack Obonyo, Francis N. Wachira, Alice Muchugi, Margaret Mulaa, Jagger Harvey, Robert A. Skilton, Janice Proud, and Jean Hanson. 2013. “Genetic Diversity in Napier Grass (Pennisetum Purpureum) Cultivars: Implications for Breeding and Conservation.” AoB PLANTS 5 (January): 1–10. doi:10.1093/aobpla/plt022.

Woodard, Kenneth R., and Lynn E. Sollenberger. 2012. “Production of Biofuel Crops in Florida: Elephantgrass.” SS-AGR-297.

Wu, Changyin, Changjun You, Caishun Li, Tuan Long, Guoxing Chen, Mary E Byrne, and Qifa Zhang. 2008. “RID1, Encoding a Cys2/His2-Type Zinc Finger Transcription Factor, Acts as a Master Switch from Vegetative to Floral Development in Rice.” Proceedings of the National Academy of Sciences of the United States of America 105 (35): 12915–20.

Wu, K. K., W. Burnquist, M. E. Sorrells, T. L. Tew, P. H. Moore, and S. D. Tanksley. 1992. “The Detection and Estimation of Linkage in Polyploids Using Single-Dose Restriction Fragments.” Theoretical and Applied Genetics 83: 294–300. doi:10.1007/BF00224274.

Wu, S, J Yang, Y Huang, Y Li, T Yin, S D Wullschleger, G a Tuskan, and R Wu. 2010. “An Improved Approach for Mapping Quantitative Trait Loci in a Pseudo-Testcross: Revisiting a Poplar Mapping Study.” Bioinform Biol Insights 4: 1–8. doi:10.4137/BBI.S13402.TYPE.

Xiao-ping, Lu, Yun Jin-feng, Gao Cui-ping, and Surya Acharya. 2011. “Quantitative Trait Loci Analysis of Economically Important Traits in Sorghum Bicolor × S. Sudanense Hybrid.” Canadian Journal of Plant Science 91: 81–90. doi:10.4141/cjps09112.

Xie, S. Q., J. Y. Feng, and Y. M. Zhang. 2014. “Linkage Group Correction Using Epistatic Distorted Markers in F2 and Backcross Populations.” Heredity 112: 479–88. doi:10.1038/hdy.2013.127.

Xie, Wengang, Xinquan Zhang, Hongwei Cai, Linkai Huang, Yan Peng, and Xiao Ma. 2011. “Genetic Maps of SSR and SRAP Markers in Diploid Orchardgrass (Dactylis Glomerata L.) Using the Pseudo-Testcross Strategy.” Genome 54: 212–21. doi:10.1139/G10-111.

Xie, Xin-Ming, Feng Zhou, Xiang-Qian Zhang, and Ju-Ming Zhang. 2009. “Genetic Variability and Relationship between MT-1 Elephant Grass and Closely Related Cultivars Assessed by SRAP Markers.” Journal of Genetics 88 (3). Springer: 281–90.

Xu, Shizhong. 2008. “Quantitative Trait Locus Mapping Can Benefit from Segregation Distortion.” Genetics 180: 2201–8. doi:10.1534/genetics.108.090688.

205

Xu, Yang, Pengcheng Li, Zefeng Yang, and Chenwu Xu. 2017. “Genetic Mapping of Quantitative Trait Loci in Crops.” Crop Journal 5. Elsevier B.V.: 175–84. doi:10.1016/j.cj.2016.06.003.

Yadav, R. S., F. R. Bidinger, C. T. Hash, Y. P. Yadav, O. P. Yadav, S. K. Bhatnagar, and C. J. Howarth. 2003. “Mapping and Characterisation of QTL X E Interactions for Traits Determining Grain and Stover Yield in Pearl Millet.” Theoretical and Applied Genetics 106: 512–20. doi:10.1007/s00122-002-1081-3.

Yang, Xiping, Md. S. Islam, Sushma Sood, Stephanie Maya, Erik A. Hanson, Jack Comstock, and Jianping Wang. 2018. “Identifying Quantitative Trait Loci (QTLs) and Developing Diagnostic Markers Linked to Orange Rust Resistance in Sugarcane (Saccharum Spp.).” Frontiers in Plant Science 9 (350): 1–10. doi:10.3389/fpls.2018.00350.

Yang, Xiping, Jian Song, James Todd, Ze Peng, Dev R. Paudel, Ziliang Luo, Xiaokai Ma, et al. 2018. “Target Enrichment Sequencing of 307 Germplasm Accessions Identified Ancestry of Ancient and Modern Hybrids and Signatures of Adaptation and Selection in Sugarcane ( Saccharum Spp.), a ‘sweet’ Crop with ‘bitter’ Genomes.” Plant Biotechnology Journal, 1–11. doi:10.1111/pbi.12992.

Yang, Xiping, Jian Song, Qian You, Dev R. Paudel, Jisen Zhang, and Jianping Wang. 2017. “Mining Sequence Variations in Representative Polyploid Sugarcane Germplasm Accessions.” BMC Genomics 18 (594). BMC Genomics: 1–16. doi:10.1186/s12864-017-3980-3.

Yang, Xiping, Sushma Sood, Neil Glynn, Md Sariful Islam, Jack Comstock, and Jianping Wang. 2017. “Constructing High-Density Genetic Maps for Polyploid Sugarcane (Saccharum Spp.) and Identifying Quantitative Trait Loci Controlling Brown Rust Resistance.” Molecular Breeding 37 (116). Molecular Breeding: 1–12.

Yano, Masahiro, Yuichi Katayose, Motoyuki Ashikari, Utako Yamanouchi, Lisa Monna, Takuichi Fuse, Tomoya Baba, Kimiko Yamamoto, Yosuke Umehara, and Yoshiaki Nagamura. 2000. “Hd1, a Major Photoperiod Sensitivity Quantitative Trait Locus in Rice, Is Closely Related to the Arabidopsis Flowering Time Gene CONSTANS.” The Plant Cell Online 12. Am Soc Plant Biol: 2473–83.

Yano, Masahiro, Shoko Kojima, Yuji Takahashi, Hongxuan Lin, and Takuji Sasaki. 2001. “Genetic Control of Flowering Time in Rice, a Short-Day Plant.” Plant Physiology 127. Am Soc Plant Biol: 1425–29. doi:10.1104/pp.010710.1.

Yanofsky, Martin F., Hong Ma, John L. Bowman, Gary N. Drews, Kenneth A. Feldmann, and Elliot M. Meyerowitz. 1990. “The Protein Encoded by the Arabidopsis Homeotic Gene Agamous Resembles Transcription Factors.” Nature 346: 35–39. doi:10.1038/346035a0.

206

Yoshida, H., K. Hirano, T. Sato, N. Mitsuda, M. Nomoto, K. Maeo, E. Koketsu, et al. 2014. “DELLA Protein Functions as a Transcriptional Activator through the DNA Binding of the INDETERMINATE DOMAIN Family Proteins.” Proceedings of the National Academy of Sciences 111 (21): 7861–66. doi:10.1073/pnas.1321669111.

Yu, Jun, Songnian Hu, Jing Juan Jian Jun Wang, Gane Ka-shu Wong, Songgang Li, Bin Liu, Yajun Deng, et al. 2002. “A Draft Sequence of the Rice Genome (Oryza Sativa L. Ssp. Indica).” Science (New York, N.Y.) 296 (5565): 79–92. doi:10.1126/science.1068037.

Yu, Xiaoqing, and Shuying Sun. 2013. “Comparing a Few SNP Calling Algorithms Using Low-Coverage Sequencing Data.” BMC Bioinformatics 14 (274): 1–15. doi:10.1186/1471-2105-14-274.

Zahn, Laura M., J. Leebens-Mack, C. W. DePamphilis, H. Ma, and G. Theissen. 2005. “To B or Not to B a Flower: The Role of DEFICIENS and GLOBOSA Orthologs in the Evolution of the Angiosperms.” Journal of Heredity 96 (3): 225–40. doi:10.1093/jhered/esi033.

Zhang, Gengyun, Xin Liu, Zhiwu Quan, Shifeng Cheng, Xun Xu, Shengkai Pan, Min Xie, et al. 2012. “Genome Sequence of Foxtail Millet (Setaria Italica) Provides Insights into Grass Evolution and Biofuel Potential.” Nature Biotechnology 30 (6): 549–54. doi:10.1038/nbt.2195.

Zhang, Zhen, Haihong Shang, Yuzhen Shi, Long Huang, Junwen Li, Qun Ge, Juwu Gong, et al. 2016. “Construction of a High-Density Genetic Map by Specific Locus Amplified Fragment Sequencing (SLAF-Seq) and Its Application to Quantitative Trait Loci (QTL) Analysis for Boll Weight in Upland Cotton (Gossypium Hirsutum).” BMC Plant Biology 16 (79). BMC Plant Biology: 1–18. doi:10.1186/s12870-016-0741-4.

Zhao, Xinxin, Linkai Huang, Xinquan Zhang, Jianping Wang, Defei Yan, Ji Li, Lu Tang, Xiaolong Li, and Tongwei Shi. 2016. “Construction of High-Density Genetic Linkage Map and Identification of Flowering-Time QTLs in Orchardgrass Using SSRs and SLAF-Seq.” Scientific Reports 6 (29345): 1–10. doi:10.1038/srep29345.

Zhou, Liyuan, Shouye Liu, Weixun Wu, Daibo Chen, Xiaodeng Zhan, Aike Zhu, Yingxin Zhang, et al. 2016. “Dissection of Genetic Architecture of Rice Plant Height and Heading Date by Multiple-Strategy-Based Association Studies.” Scientific Reports 6 (29718). Nature Publishing Group: 1–10. doi:10.1038/srep29718.

207

Zhou, Sifan, Chengran Wang, Taylor F. Frazier, Haidong Yan, Peilin Chen, Zhihong Chen, Linkai Huang, et al. 2018. “The First Illumina-Based de Novo Transcriptome Analysis and Molecular Marker Development in Napier Grass (Pennisetum Purpureum).” Molecular Breeding 38 (95). Molecular Breeding: 1–14. doi:10.1007/s11032-018-0852-8.

Zhou, Wencai, Zaixiang Tang, Jing Hou, Nan Hu, and Tongming Yin. 2015. “Genetic Map Construction and Detection of Genetic Loci Underlying Segregation Distortion in an Intraspecific Cross of Populus Deltoides.” PLoS ONE, 1–13. doi:10.1371/journal.pone.0126077.

Zhou, Xiaojing, Youlin Xia, Xiaoping Ren, Yulin Chen, Li Huang, Shunmou Huang, Boshou Liao, Yong Lei, Liyin Yan, and Huifang Jiang. 2014. “Construction of a SNP-Based Genetic Linkage Map in Cultivated Peanut Based on Large Scale Marker Development Using next-Generation Double-Digest Restriction-Site-Associated DNA Sequencing (ddRADseq).” BMC Genomics 15 (351): 1–14. doi:10.1186/1471-2164-15-351.

Zub, H. W., S. Arnoult, and M. Brancourt-Hulmel. 2011. “Key Traits for Biomass Production Identified in Different Miscanthus Species at Two Harvest Dates.” Biomass and Bioenergy 35. Elsevier Ltd: 637–51. doi:10.1016/j.biombioe.2010.10.020.

208

BIOGRAPHICAL SKETCH

Dev Paudel received his Ph.D. in Agronomy from the University of Florida (UF) in

the fall of 2018. His research focuses on utilizing traditional plant breeding approaches

and modern bioinformatic and molecular techniques in the genetic improvement of

napiergrass (elephant grass) and its interspecific hybrids with pearl millet. This will

eventually help in sustainability of forage and biofuel feedstock.

Dev received his MS degree in Crop Science from Texas Tech University,

Lubbock, Texas, USA where his research focused on evaluating the potential of new

testing methods for cotton breeding. Plant breeders, ginners, farmers, and spinning

mills can use the information obtained from his research to make informed decisions for

increased profitability in the premium yarn market. After his MS degree, he worked at

the Texas A&M AgriLife Research, Pecos, Texas where he optimized nutrient media for

algae production. He has a BS degree in Agriculture from Tribhuvan University, Nepal.

Documents

GENOMIC AND BREEDING RESOURCES TO PRODUCE SEEDED …