Applying Bioinformatic Tools to Better Understand Eye Diseases

Applying Bioinformatic Tools to Better Understand Eye Diseases

by

Vikrant Singh

M.Sc. Bioinformatics (Jamia Millia Islamia- Central University, India)

M. Engg Bio-Medical Engineering (Anhalt University of Applied Sciences and Martin

Luther University, Germany)

Menzies Institute for Medical Research

Submitted in fulfilment of the requirements for the Doctor of Philosophy

University of Tasmania

October 2019

ii

Declarations

“The thesis contains no material which has been accepted for the a degree or diploma by the

University or any other institution, except by way of background information and duly

acknowledged in the thesis, and to the best of my knowledge and belief no material previously

published or written by another person except where due acknowledgement is made in the text

of the thesis, nor does the thesis contain any material that infringes copyright.”

This thesis may be made available for loan and limited copying in accordance with the

Copyright Act 1968.

Student Signature: Vikrant Singh

Primary Supervisor Signature:

Alex Hewitt

iii

Acknowledgements

“No one who achieves success does so without acknowledging the help of others. The wise

and confident acknowledge this help with gratitude.” – Alfred North Whitehead

One of the pleasant aspects of writing an acknowledgement is the opportunity to thank all those

who have contributed to it. Unfortunately, the list of expression of gratitude- no matter how

extensive is always incomplete and inadequate. This acknowledgement is no exception.

First of all, I wish to express my sincere gratitude to my supervisors Prof. Alex Hewitt and

Assoc/Prof. Anthony Cook who gave me the chance to complete my PhD under their esteemed

supervision. Without your unconditional support and guidance, I am unable to complete my

thesis on time. Because of your inspiring guidance, motivation, positive criticism, continuous

encouragement and constant supervision, this work could be brought to its present shape.

Prof. Alex Hewitt provided excellent scientific guidance that helped me to expand my

knowledge about the use of bioinformatic applications to analyse the omics datasets. I want to

extend my special thanks for being so patience and understanding. You helped me to think and

work beyond my capabilities. You gave me opportunity to work on different projects that

helped me to develop my scientific aptitude. My words are enough to express my gratitude and

respect towards you. I will always be grateful to you. I want to express my special thanks to

my co-supervisor Assoc/Prof. Anthony Cook. Your support, guidance, positive feedback and

short-term discussion kept me motivated and helped me a lot during my PhD candidature.

My special thanks to Dr Guei-Sheung Liu for allowing me to work with you on a project that

helped me a lot to understand the biological concept of retinal neovascularisation along with

the computational analysis. I am very thankful for your motivation and long discussion we had

during my PhD. Other special thanks to my graduate research co-ordinators Prof. Heinrich

Korner and Prof. Wendy Oddy for your help and support in and out during my PhD

candidature. Other Special thanks to Prof. Alice Pebay and Dr. Duncan Crombie from the

University of Melbourne for Providing the differentiated cell lines for virtual karyotyping.

iv

“Oldies are the goldies”: This goes for not-so-long-gone lab mates – Fan Li, Sloan Wang whose

presence is missed for sure. I appreciate every moment of rising and fall we spent during my

PhD. Science can’t be carried out by a single man. I would like to thank Qi Wang, Peter Lu,

Peter Tran and Mohd Khairul Nizam for engaging into the short and long discussions and

always keeping the spirit high. I also appreciate the time spent in discussing captivating advises

about the future. Canteen mates are always reminded you to have lunch on time. I can’t thank

enough to Ankit Gupta, always look after me when I was not around. Your support and

encouragement can’t be replaced by anything.

Someone honestly said – “Where you go, whatever you do, your family is always there to

support and love you”. I never realised that I had come that far but, this would not be possible

without the immense love from my family. I know my Dadu (Grandfather) and Dadi

(Grandmother) must be very proud of me in heaven. My journey wouldn’t be possible with the

unconditional love and support from my Maa (Mother) and Paa (Father). Your continuous

motivation and encouragement helped me a lot to reach my goal. I am very thankful for you to

teaching me a useful lesson of my life, both socially and economically. I also thank my brother

Yug for always guarding my back in every situation.

I would also like to express my not so formal thankfulness to my friend, Mr Karan Pater for

being there and make me laugh in the weird situations, making it worse but the best time of my

life. Bharat Sir, special thanks to you for continuous motivation and helping me in a tough

situation. Prasoon Thakur, my bioinformatics buddy, thank you for having a long discussion

with me about problem-solving.

In the end, I would like to acknowledge my wonderful wife, Soniya Sharma who is there with

me since always. She is an eternal part of my past, present and future. I don’t want to thank

you but congratulate her for completing this together with hand in hand. Being a PhD student,

it is very tough sometimes to manage things, but you always backed me and helped me

throughout our PhD journey. Also, thank you for the beautiful gift I ever had as our babies

(Sovik Singh and Nivika Singh). Thank you for being there always!! Last but not least, I would

like to thank the almighty GOD for being there somewhere around me with all positive energy.

Vikrant Singh

v

Abstract

The highly specialized cells of the eye function in concert to produce clear vision, one of the

most valued senses. Visual impairment or blindness can occur as a result of disease or trauma.

Age-related macular degeneration, glaucoma, cataracts and diabetic retinopathy are common

causes of vision loss in older individuals. This thesis explores the bioinformatics approaches

based on the central dogma as a model for exploring the experimental models for human eye

diseases including quality control of stem cells to detect chromosomal abnormalities;

epigenetic age prediction of ocular tissues; identification of novel genes in Clustered Regularly

Interspaced Short Palindromic Repeats (CRISPR) knockout screening in uveal melanoma; and

RNA-Seq analysis to understand the molecular mechanisms in oxygen-induced retinopathy

and differentiation of dental pulp mesenchymal stem cells into trabecular meshwork cells.

In a virtual karyotyping study, I investigated the utility of a low-density genome-wide SNP

array for the karyotypic assessment of human pluripotent stem cells (hPSCs) using the Illumina

Infinium HumanCore BeadChip. Specifically, the resolution of these arrays in detecting

chromosomal aberrations and their ability to identify clonal variations was determined. It was

shown that the SNP array can detect chromosomal abnormalities when at least 25% of the cell

population is aberrant. Our data demonstrate that an array-based karyotyping offers an

economical and robust sampling method, in the genomic resolution, compared with standard

cytogenetic karyotyping of hPSCs. This approach could provide a reliable, rapid and cost-

effective assessment of hPSCs clonality and virtual karyotype for large-scale generation and

maintenance of hPSCs.

To investigate the effects of aging among different tissues, we calculated the DNA methylation

age of whole peripheral blood and ocular tissue from the same individual and compared it with

the person’s chronological age. We found significant differences between chronological and

epigenetic ages (p<0.048). Our study showed that there is a significant difference (mean = 44.4

years) between chronological and epigenetic age in neurosensory retinal tissue and the same

pattern identified by various tools.

Through a CRISPR knockout screening study, we used the Human GeCKOv2 pooled library

in OCM1 cell lines (uveal melanoma) to identify novel genes that are associated with

vi

tumorigenesis using the CRISPRAnalyzeR tool on next-generation sequencing data. Overall,

from our analysis, we found 15 genes that have relatively low expression in the passage 12 as

compared to the passage 0 and are associated with the metabolic process, cellular process,

primary metabolic process, cellular metabolic process, biological process and organic

substance metabolic process. Our study shows that these genes are crucial for cell proliferation;

however, further in-vitro and in-vivo validation is required.

In RNA-Seq studies, we investigated the miRNA expression in oxygen-induced retinopathy

(OIR) rat models through next-generation sequencing (RNA-Seq) data. Our RNA-Seq data

suggested that the expression of miR-143 dysregulated and mediated the regulatory networks,

leading to retinal neovascularization. Furthermore, miR-143 can influence cellular motion and

cell-matrix interaction during vascular formation. We also found that miR-126, miR-150 is

significantly down-regulated and directly involved in retinal neovascularization. We also

investigated the propensity of mesenchymal stem cells (MSC) populations (dental pulp-derived

MSCs) to differentiate into trabecular meshwork (TM) cells. We performed RNA sequencing

under two conditions: control (DPMSCs) and treated (with growth factors to differentiate into

the TM cells). Overall, we found over 8000 genes that are statistically significant and have an

association with tissue development, extracellular matrix development and related pathways,

and demonstrated the ability of DPMSCs to differentiate into TM cells. Using bioinformatics

techniques to develop models of blinding eye diseases can lead to new and better treatments to

preserve vision.

vii

Publications

• Maciej Daniszewski, Quan Nguyen, Hun S. Chy, Vikrant Singh, Duncan E. Crombie,

Tejal Kulkarni, Helena H. Liang, Priyadharshini Sivakumaran, Grace E. Lidgerwood,

Damián Hernández, Alison Conquest, Louise A. Rooney, Sophie Chevalier, Stacey B.

Andersen, Anne Senabouth, James C. Vickers, David A. Mackey, Jamie E. Craig,

Andrew L. Laslett, Alex W. Hewitt, Joseph E. Powell, Alice Pébay,"Single-Cell

Profiling Identifies Key Pathways Expressed by iPSCs Cultured in Different

Commercial Media", iScience, Volume 7, 2018, Pages 30-39, ISSN 2589-

0042,https://doi.org/10.1016/j.isci.2018.08.016. (PMID: 30267684)

• Raymond C.B. Wong, Sandy S. Hung, Stacey Jackson, Vikrant Singh, Shahnaz Khan,

Helena H. Liang, Lisa S. Kearns, Tu Nguyen, Alison Conquest, Maciej Daniszewski,

Alex W. Hewitt, Alice Pébay, Generation of a human induced pluripotent stem cell

line CERAi001-A-6 using episomal vectors, Stem Cell Research, Volume 22, July

2017, Pages 13-15, ISSN 1873-5061, https://doi.org/10.1016/j.scr.2017.05.007.

(PMID:28952926).

• Fan Li1, Sandy S.C. Hung, Mohd Khairul Nizam Mohd Khalid, Jiang-Hui Wang, Vicki

Chrysostomou, Vickie H.Y. Wong, Vikrant Singh, Kristof Wing, Leilei Tu, James A.

Bender, Alice Pébay, Anna E. King, Anthony L. Cook, Raymond C.B. Wong, Bang V.

Bui, Alex W. Hewitt, Guei-Sheung Liu, “Utility of self-destructing CRISPR/Cas

constructs for targeted gene editing in the retina”, Hum Gene Ther. 2019 Aug 2.

doi: 10.1089/hum.2019.021. (PMID: 31373227)

• Jiang-Hui Wang, Damien Linga, Leilei Tu, Vikrant Singh, Moeen Riaz, Fan Li, Alex

W. Hewitt, Gregory J. Dusting, Peter van Wijngaarden, Guei-Sheung Liu

“MicroRNA-143 plays a protective role in ischemia-induced retinal

neovascularization”, biorXiv, doi: https://doi.org/10.1101/548297.

viii

Abstract Publications

• Alex W Hewitt, Vikrant Singh, Vania Januar, Alexandra Sexton-Oates, Jamie E Craig,

Richard Saffery; Differences in epigenetic age of ocular tissue and the implications

for eye disease. Invest. Ophthalmol. Vis. Sci. 2017;58(8):1845. (ARVO Abstract

2017).

• Jiang-Hui Wang, Damien Linga, Leilei Tu, Vikrant Singh, Moeen Riaz, Fan Li, Alex

W. Hewitt, Gregory J. Dusting, Peter van Wijngaarden, Guei-Sheung Liu

“MicroRNA-143 plays a protective role in ischemia-induced retinal

neovascularization”. Invest. Ophthalmol. Vis. Sci. 2019; 60(9):3261 (ARVO Abstract

2019).

• Fan Li, Vikrant Singh, Vivienne Lu, Jinying Chen, Sandy S.C. Hung, Joanne L.

Dickinson, Phillippa Taberlay, Anthony L. Cook, Alex W. Hewitt, Guei-Sheung Liu

"A genome-wide crispr/cas9 screen to identify novel therapeutic targets for uveal

melanoma", Clinical and Experimental Ophthalmology2018;46 (Suppl 1) Pages 21-

35

ix

Table of Contents Declarations ...................................................................................................................... ii

Abstract .............................................................................................................................. v

Publications ..................................................................................................................... vii

Abstract Publications ..................................................................................................... viii

Chapter 1 Introduction ..................................................................................................... 18

1.1 Structure of the eye .................................................................................................... 18

1.2 Eye diseases ............................................................................................................... 19

1.2.1 Blinding Mendelian Disease of the RPE ................................................................ 19

1.2.2 Glaucoma ................................................................................................................ 20

1.2.3 Age-Related Macular Degeneration ....................................................................... 21

1.2.4 Diabetic Retinopathy .............................................................................................. 22

1.2.5 Cataract ................................................................................................................... 23

1.2.6 Uveal Melanoma ..................................................................................................... 23

1.2.3 Bioinformatics application in eye disease ............................................................... 25

1.2.3.1 Quality Control in Stem Cells Genomic Integrity ............................................... 25

1.2.3.1.1 Traditional Karyotyping .................................................................................... 26

1.2.3.1.2 Virtual karyotyping ........................................................................................... 27

1.2.3.1.3 Tools for detecting Copy number variations .................................................... 29

1.2.3.2 DNA Methylation ................................................................................................ 30

1.2.3.2.1 Background ....................................................................................................... 30

1.2.3.2.2 DNA methylation .............................................................................................. 31

1.2.3.2.3 DNA methylation age ....................................................................................... 31

1.2.4.1 CRISPR Cas 9 ...................................................................................................... 32

1.2.4.1.1 Background ....................................................................................................... 32

1.2.4.1.2 CRISPR Cas9 genome editing .......................................................................... 33

1.2.4.1.3 CRISPR Screening ............................................................................................ 35

1.2.4.1.4 CRISPR/Cas9 technology in eye disease .......................................................... 36

1.2.5 RNA-Seq Analysis .................................................................................................. 36

1.2.5.1 Background .......................................................................................................... 36

1.2.5.2 RNA-Seq Experiment Design .............................................................................. 37

1.2.5.3 Analysis of RNA-Seq Data .................................................................................. 39

1.2.5.3.1 Transcript Identification .................................................................................... 39

x

1.2.5.3.2 Alignment ......................................................................................................... 39

1.2.5.3.3 Transcript discovery and De novo transcript reconstruction ............................ 39

1.2.5.4 Quantification ...................................................................................................... 40

1.2.5.5 Differential Gene Expression Analysis ................................................................ 41

1.2.5.4 RNA-Seq in Eye diseases .................................................................................... 41

1.3 Outline of the Thesis .................................................................................................. 42

Chapter 2 Virtual Karyotyping ........................................................................................ 43

2.1 Introduction ................................................................................................................ 43

2.2 Methods...................................................................................................................... 44

2.2.1 Cytogenetic Karyotyping ........................................................................................ 44

2.2.2 Illumina and Axiom™ Biobank Plus Genotyping .................................................. 44

2.2.3 Array Processing and Virtual Karyotype Analysis ................................................. 45

2.3 Results ........................................................................................................................ 45

Chapter 3 DNA methylation Age .................................................................................... 52

3.1 Introduction ................................................................................................................ 52

3.2 Methods...................................................................................................................... 52

3.2.1 DNAm data collection and processing ................................................................... 52

3.2.2 DNAm age estimation by using epigenetic clock ................................................... 53

3.3 Results ........................................................................................................................ 54

3.4 Discussion .................................................................................................................. 58

Chapter 4 CRISPR Screening for uveal melanoma ......................................................... 59

4.1 Introduction ................................................................................................................ 59

4.2 Methods...................................................................................................................... 60

4.2.1 Sequencing .............................................................................................................. 60

4.2.2 Quality Control ....................................................................................................... 61

4.2.3 Adapter/Quality Trimming ..................................................................................... 61

4.2.4 CRISPR Screening Analysis ................................................................................... 61

4.2.5 Functional Enrichment Analysis ............................................................................. 62

4.3 Results ........................................................................................................................ 62

4.4 Discussion .................................................................................................................. 67

Chapter 5a RNA-Seq Analysis for Retinal neovascularization in miRNAs .................... 70

5a.1 Introduction .............................................................................................................. 70

5a.2 methods .................................................................................................................... 71

5a.2.1 Sequencing ............................................................................................................ 71

xi

5a.2.4 RNA-Sequencing Analysis ................................................................................... 71

5a.2.5 Network and Pathway Analysis ............................................................................ 72

5a.3 Results ...................................................................................................................... 72

5a.4 Discussion ................................................................................................................ 77

Chapter 5b RNA-Seq Analysis for differentiating DPMSCs into TM cells .................... 79

5b.1 Introduction .............................................................................................................. 79

5b.2 Methods.................................................................................................................... 80

5b.2.1 Sequencing ............................................................................................................ 80

5b.2.2 Dataset ................................................................................................................... 80

5b.2.4 RNA-Seq Analysis ................................................................................................ 80

5b.2.5 Functional Enrichment Analysis ........................................................................... 80

5b.4 Discussion ................................................................................................................ 82

Chapter 6 Conclusion ....................................................................................................... 84

Appendix ........................................................................................................................ 109

Method 1 iPSCs Differentiation ..................................................................................... 109

Method 2 Genome-wide CRISRP/Cas9 screen for uveal melanoma OCM-1 cells ....... 111

Method 3 miR143 Animal body experiment methods ................................................... 114

Method 4 DPMSC to TM cells differentiation protocol ................................................ 120

xii

List of Tables

Table 1.1: Various computational tools for CNV detection in whole-genome genotyping

and next-generation sequencing data. .............................................................................. 29

Table 1.2: Tools available for CRISPR screening data analysis. .................................... 36

Table 3.1: chronological age correlation with epigenetic age acceleration ..................... 55

Table 4.1: Summary of comparisons ............................................................................... 63

Table 4.2 List of genes with missing 3 or more sgRNA in UMP12 as compared to UMP0.

.......................................................................................................................................... 64

Table 4.3: List of final gene subset select based on MAGeCK and DESeq2 algorithms.

.......................................................................................................................................... 66

Table 5a.1: Statistics of mapped reads from samples OIR Scrambled, OIR-miR143,

miR126, miR143, miR150 and control summary table. .................................................. 72

Table 5a.2: List of top pathways from IPA in miR126, miR143, miR150 and OIR miR143.

.......................................................................................................................................... 76

Table 5b.1: Statistics of mapped reads from Control and treated sample with all replicates.

.......................................................................................................................................... 81

All supplementary Figures and Tables are available from the following link: .............. 125

xiii

List of Figures

Figure 1.1 Structure of the eye. Outer layer (sclera and cornea), middle layer (choroid, iris

and ciliary body) and the inner layer (macula fovea and optic disc). .............................. 18

Figure 1.2 G-banding showing metaphase arranged in the standard karyotype format. . 26

Figure 1.3: Fluorescence in-situ hybridisation (FISH) -Normal fibroblast cell line showing

normal chromosomes 11 positive for WCP11 (spectrum orange) and normal

chromosomes 15 positive for WCP 15 (spectrum green). ............................................... 27

Figure 1.4: DNA methylation (Martin and Fry, 2018). ................................................... 30

Figure 1.5: Cas9 protein domains and schematics and Crystal structure from PDB (PDB

id - 4CMP (Jinek et al., 2014). ......................................................................................... 33

Figure 1.6: Schematic picture of genome editing via the CRISPR/Cas9 system (Ran et al.,

2013). ............................................................................................................................... 34

Figure 2.1: G-banding and FISH analysis: A) Normal male fibroblast cell line BCL6 with

karyotype 46, XY. B) Karyotype of cell line BCL6 7+5 with karyotype 45,X,-

Y,der(1)t(1;15)(p34.1;q15),t(4;8)(p14;q24),der(11)t(1;11)(p34;q23),t(15;22)(p11.2;q13).

C) Normal Cell line whole chromosome painting (WCP) showing Chr11 positive with

spectrum orange and Chr15 positive with spectrum green. D) Chromosome painting,

WCP#11 (orange) and WCP#15 (green) showing duplication at Chromosome 15 and

deletion at chromosome 11. ............................................................................................. 46

Figure 2.2: Comparison table of CNV events detected by various tools with aneuploid

genome proportion (AGP). Pie chart represent the percentage of aberrations in the sample

(0,10,15,20,25,30,35,40,45,50 and 100). ......................................................................... 48

Figure 2.3: BAF/LRR Plots A) Cell line with 0% genomic aberration (no CNV events).

B) Cell line with 100% genomic aberrations (including CNV events - duplication at

chromosome 1, deletion at chromosome 11 and duplication at chromosome 15). .......... 49

Figure 3.1. (A)Chronological age versus DNAm age across all sample data set (r = 0.37,

p< 0.048 and standard error = 18 years). Mean epigenetic age difference (y- axis) versus

different tissues in (B) respectively. The bar plots also report a p-value from a non-

parametric group comparison test (Kruskal Wallis test). Each bar plot depicts the mean

value and 1 standard error. (C) Post mortem interval versus DNAm age across all sample

data set (r = -0.15, p<0.44). .............................................................................................. 55

Figure 3.2: Hannum Method (A) Chronological age versus DNAm age across all sample

data set (r = 0.32, p< 0.091 and standard error = 26 years). Mean epigenetic age difference

xiv

(y- axis) versus different tissues in (B) respectively. The bar plots also report a p-value

from a non-parametric group comparison test (Kruskal Wallis test). Each bar plot depicts

the mean value and 1 standard error. (C) Post mortem interval versus DNAm age across

all sample data set (r = 0.031, p<0.87). ............................................................................ 56

Figure 3.3: Qzhang Method (a) Chronological age versus DNAm age across all sample

data set by using Enpred Algorithm and mean epigenetic age difference (y- axis) versus

different tissues in respectively. The bar plots also report a p-value from a non-parametric

group comparison. Each bar plot depicts the mean value and 1 standard error. (b)

Chronological age versus DNAm age across all sample data set by using BLUpred and

mean epigenetic age difference (y- axis) versus different tissues in respectively. The bar

plots also report a p-value from a non-parametric group comparison. Each bar plot depicts

the mean value and 1 standard error. ............................................................................... 57

Figure 4.1: a) Scatterplot showing distribution of sgRNA read count in UMP0 and

UMP12. b) Boxplot with outliers for all samples. c) Distribution plot of all samples with

normalised read count. ..................................................................................................... 63

Figure4.2: a) Scatterplot showing significant gene using MAGeCK, b) Scatterplot

showing significant gene using sgRSEA, c) Scatterplot showing significant gene using

DESeq2, d) Scatterplot showing significant gene using EdgeR. ..................................... 65

Figure 4.3 a) Venn diagram showing the overlap genes between the different algorithms.

b) overlapping genes with final selected algorithms MAGeCK and DESeq2. ................ 66

Figure 4.4: Pathway analysis showing statistically significant pathways by using gProfiler

and visualisation using enrichment map. ......................................................................... 67

Figure 4.5: Expression level (transcripts per million- tpm) of final selected gene on TCGA

datasets. ............................................................................................................................ 69

Figure 5a.1: a) Correlation plot between the OIR Scrambled and OIR-miR143 sample. b)

Heatmap showing the expression (tpm) of statistically significant (adjusted pValue<0.05)

gene found by using Sleuth. c) Volcano plot highlighting statistically significant (adjusted

pValue<0.05) between OIR Scrambled and OIR-miR143 sample. ................................. 73

Figure 5a.2: a) Correlation plot between Control and miR126 sample. b) Volcano plot

highlighting statistically significant (adjusted pValue<0.05) between Control and miR126

sample along with highlighted genes that are related to angiogenesis. c) Heatmap showing

the expression (tpm) of statistically significant (adjusted pValue<0.05) gene found by

using Sleuth. ..................................................................................................................... 74

xv





using Sleuth. ..................................................................................................................... 74





using Sleuth. ..................................................................................................................... 75

Figure 5b.1: a) Heatmap showing the expression (tpm) of statistically significant (adjusted

pValue<0.01) gene found by using Sleuth. b) Volcano plot highlighting statistically

significant (adjusted pValue<0.01) between control and treated condition. c) Read

distribution plot across all samples. d) Statistically significant pathways by using

gProfiler. .......................................................................................................................... 81

xvi

List of Abbreviations

AAV Adeno-associated virus

AMD Age-related macular degeneration

ASO Allele-specific oligonucleotides

BAF B allele frequency

CGH Comparative genomic hybridization

CHAT Clonal Heterogeneity Analysis Tool

CNV Copy number variations

CpG 5'—C—phosphate—G—3'

CRISPR Clustered regularly interspaced short palindromic repeats

CRVO Central retinal vein occlusion

DME Diabetic macular edema

DNA Deoxyribonucleic acid

DNAm DNA methylation

DPMSC dental pulp-derived mesenchymal stem cells

DR Diabetic retinopathy

DSB Double-stranded break

FISH Fluorescence in-situ hybridisation

G-Banding Giemsa banding

GC GenCall

GEO Gene Expression Omnibus

HDR Homology-directed repair

hESCs Human embryonic stem cells

hg38 Homosapiens genome assembly

hiPSC Human induced pluripotent stem cells

HMEC Human dermal microvascular endothelial cells

hPSC Human pluripotent stem cells

HUVEC Human umbilical vein endothelial cell

IOP Intraocular pressure

iPSC Induced pluripotent stem cells

LRR Log R ratio

MAGeCK Model-based Analysis of Genome-wide CRISPR/Cas9 Knockout method

miRNA MicroRNA

xvii

MOI Multiplicity of infection

mRNA messenger ribonucleic acid

MSC Mesenchymal stem cells

NGS Next-generation sequencing

NHEJ Non-homologous end joining

OIR Oxygen induced retinopathy

PAM Protospacer adjacent motif

PDR Proliferative diabetic retinopathy

PDR Proliferative retinopathy

POAG Primary open-angle glaucoma

QC Quality control

RGC Retinal ganglion cells

rn6 Rattus norvegicus genome assembly

RNA Ribonucleic acid

RNA-Seq RNA-sequencing

ROP Retinopathy of prematurity

RPE Retinal pigment epithelium

rRNA ribosomal RNA

SD Standard deviation

sgRNA Single guided ribonucleic acid

SNP Single nucleotide polymorphism

TALEN Transcription activator-like effector nuclease

TCGA The Cancer Genomic Atlas

TGFβ Transforming growth factor-β

TIME Telomerase-immortalized microvascular endothelial cell

TM Trabecular meshwork

tpm Transcript per million

UM Uveal melanoma

UTR Untranslated region

UV Ultraviolet

VEGF Vascular endothelial growth factor

WCP Whole chromosome paints

ZFN Zinc finger nuclease

Chapter 1 Introduction

1.1 Structure of the eye

The eyes are one of the most complex organs in the human body. To enable optimal vision, all

structures within the eye must function properly in order to refract and focus light onto the

retina. The eye has three layers: outer, middle and inner (Figure 1.1). The outer layer, also

known as the fibrous coat, consists of the sclera and the cornea. The cornea is the transparent

layer in front of the iris, pupil and lens and maintains the shape of the eye. It transmits light to

the inner eye and also protects it from infection and structural damage. The middle layer, or

uveal tract, consists of the choroid, iris and ciliary body. The iris controls the size of the pupil

and therefore the amount of light entering the eye and reaching the retina.

Figure 1.1 Structure of the eye. Outer layer (sclera and cornea), middle layer (choroid, iris and ciliary body) and the inner layer (macula fovea and optic disc)1.

The ciliary body controls the shape of the lens and the production of aqueous humour whereas

the choroid provides nutrients and oxygen to the retinal layer. The inner layer of the retina is a

complex layer of neurons that capture and process light (Willoughby et al., 2010).

1 Image Source: Rhcastilhos. And Jmarchn

19

1.2 Eye diseases

Normal eye function can be affected by minor issues that resolve quickly, such as

conjunctivitis, or more serious conditions, e.g. glaucoma, that can lead to visual impairment.

With many eye diseases, there may be no early symptoms and they can only be diagnosed after

the disease reaches an advanced stage. Unfortunately, advanced blinding diseases are typically

less amenable to the treatment.

1.2.1 Blinding Mendelian Disease of the RPE

Many blinding diseases occur due to inherited degeneration of the retinal pigment epithelium

(RPE). Although specific mutations responsible for the majority of inherited retinal dystrophies

have been identified, all these conditions are currently untreatable (De Roach et al., 2013). Best

Disease, Doyne Honeycomb Retinal Dystrophies, and Sorsby Fundus Dystrophy are clinically

well-defined but remain poorly characterized in terms of their cellular pathophysiology and the

underlying mechanisms that lead to RPE degeneration. Currently, there are no treatments for

these blinding dystrophies, and as such, they are excellent candidates for disease modelling

and genetic correction.

Best disease (OMIM #153700) is an autosomal disorder of RPE that causes secondary

photoreceptor degeneration within the macula and concomitant loss of central vision, usually

beginning in the second to third decade of life (Blodi and Stone, 1990, Nordstrom and

Barkman, 1977). One of the earliest clinical manifestations of Best Disease is the appearance

of a yellowish vitelliform lesion located in the subretinal space of macula (Boon et al., 2009,

Mohler and Fine, 1981, Xiao et al., 2010). Best Disease is caused by mutations in the RPE

gene BESTROPHIN1 (BEST1), which, through mechanism(s) that remain unclear, leads to the

accumulation of subretinal fluid and auto fluorescent waste products from shed photoreceptor

outer segments (POS) (Singh et al., 2013).

Doyne Honeycomb Retinal Dystrophies (or Malattia Leventinese; OMIM #126600) is an

inherited autosomal dominant maculopathy that results in progressive and irreversible visual

loss (Zhang et al., 2014a). It is characterized by a radial pattern of many laminar drusen (a lipid

waste deposition between the RPE and Bruch’s membrane in the outer retina) that accumulate

beneath the RPE at the posterior pole (Zhang et al., 2014a). It is caused by mutations in the

20

epidermal growth factor-containing fibulin-like extracellular matrix protein 1 (EFEMP1) gene.

In this disease, drusen develop during early adulthood. The build-up of lipofuscin and

coalescence of large drusen form a honeycomb-like pattern within the macula, causing disease

progression (Stone et al., 1999).

Sorsby Fundus Dystrophy (OMIM #136900) is a rare, autosomal-dominant retinal dystrophy

caused by a missense mutation in the tissue inhibitor of metalloproteinases-3 (TIMP3) gene

(Szental et al., 2009). ). Clinically it is characterized by early-onset (<45 years of age) bilateral

loss of central vision loss due to RPE atrophy and subretinal neovascularization at the macula

(Szental et al., 2009).

1.2.2 Glaucoma

Glaucoma involves chronic neurodegeneration of the optic nerve and is a leading cause of

blindness (Blomdahl et al., 1997, Munier et al., 1998, McKinnon, 2003). Glaucoma is a

heterogeneous group of ocular disorders, with primary open-angle glaucoma (POAG) the most

common subset (Hewitt et al., 2006). POAG is defined as progressive excavation of the optic

disc with corresponding visual fields defects and an open iridotrabecular angle(Quigley,

2011a). POAG is challenging to diagnose in the early stages because the clinical features can

be subtle. The disease is characterized by the death of retinal ganglion cells (RGC) due to

axonal degeneration, resulting in visual impairment (Yasuda et al., 2014). However, despite an

active genetic component, the precise molecular drivers of POAG pathology are not known.

Currently there is no definitive treatment (Quigley, 2011a), although intraocular pressure

(IOP)-lowering medications and/or surgery can delay progression.

Globally, it is estimated that 60 million people have glaucomatous optic neuropathy and 8.4

million people are blind as a result of glaucoma (Cook and Foster, 2012). By 2020 these

numbers are set to increase to 80 million and 11.2 million, respectively (Cook and Foster,

2012). Extensive population-based epidemiology studies revealed that the prevalence of

POAG in Australia is approximately 3% in people aged over 50 years, rising to 10% by age 90

(Mitchell et al., 1996). Various comprehensive examinations are used to detect glaucoma:

• Tonometry

• Visual field test

• Pachymetry

21

• Gonioscopy

1.2.3 Age-Related Macular Degeneration

Age-related macular degeneration (AMD) is a neurodegenerative disease that affects the

central part of retina — the macula. The disease is categorized into early, intermediate and

advanced stages based on the severity and underlying pathology, including choroidal

neovascularization (Pennington and DeAngelis, 2016). AMD is further classified as dry AMD

or wet AMD. Dry AMD is more common and is related to the drusen (yellow deposition) in

the macula; this leads to loss of macula function and blurred vision. The signs of dry AMD

cannot be observed easily and it is particularly difficult to detect if it affects only one eye. Wet

AMD accounts for about 15% of cases. In wet AMD, blood vessels located under the macula

start leaking, which causes the retina to deform. Wet AMD can progress rapidly and can be

very severe, degrading vision quickly (Parmet et al., 2006). Clinical features of AMD include:

• Aging: increasing age is a consistent risk factor for both dry and wet AMD. In the

United States, the incidence of AMD was reported as 0.3% in people aged between 50-

59 years and 16% in those aged 80 years and older (Friedman et al., 2004).

• Genetics: The most common genetic variation for AMD is the long arm of

chromosome 1 (1q31). It is a common amino acid variation (tyrosine with histidine at

402bp amino acid) reported at CFH (Complement factor H) gene (Hageman et al.,

2005, Klein et al., 2005, Haines et al., 2005). The CFH gene is highly implicated in the

development of AMD, with an odds ratio of 2.4–4.6 for heterozygous patients and 3.3–

7.4 for homozygous patients (Losonczy et al., 2011).

Nominally, 5% of blindness worldwide is due to AMD (Pascolini and Mariotti, 2012). By 2020

it is estimated that 196 million people will have AMD, and this will increase up to 288 million

by 2040 (Wong et al., 2014). The global cost for AMD is around $343 billion, including $255

billion in health care (Gordois et al., 2010). According to a report published by the Australian

Institute of Health and Welfare in 2005, nearly 3.1% of the population aged 55 or more have

AMD and 10.1% of Australians aged >80 have drusen in one of their eyes, with men and

women equally affected. The early and intermediate stages of AMD are asymptomatic. AMD

can only be detected by specific eye exams:

• Amsler grid

• Dilated eye exam

22

• Fluorescein angiography

• Optical coherence tomography

• Visual acuity test

There is no definitive treatment for dry AMD, but several combinations of nutritional

supplements have been shown to impact drusen (Krishnadev et al., 2010).

• Vitamin C (500 mg)

• Vitamin E (400 IU)

• Lutein (10 mg)

• Zeaxanthin (2 mg)

• Zinc (80 mg)

• Copper (2 mg)

For treating wet AMD, anti-vascular endothelial growth factor (VEGF) drugs can be used to

reduce abnormal blood pressure in the retina (Yorston, 2014).

1.2.4 Diabetic Retinopathy

Individuals with diabetes often experience diabetic retinopathy (DR); indeed, DR is the leading

cause of vision loss among working adults (Cheung et al., 2010). Ongoing hyperglycaemia

causes the blood vessels in the retina to bleed, distorting vision (Mustafa et al., 2018). DR

classified into four stages, which are as follows:

● Mild nonproliferative retinopathy: This first stage of DR is also known as

background retinopathy. Microaneurysms (tiny bulges) on the blood vessels in the

retina occur.

● Moderate nonproliferative retinopathy: This is the second stage of DR, also known

as pre-proliferative retinopathy. At this stage, the retina swells and is unable to carry

the blood, resulting in a change in the physical appearance of the eye. Diabetic macular

edema (DME) due to deposition of blood and fluid in the macula occurs. This affects

sight as the macula is directly involved in direct vision.

● Severe nonproliferative retinopathy: At this stage, the blood vessels are blocked,

(macular ischemia). Floaters can be visible and vision is blurry.

23

● Proliferative diabetic retinopathy (PDR): This advanced stage is caused by

neovascularization. Thin and weak vessels are formed and frequently bleed because of

scar tissue, which can be responsible for retinal detachment and permanent vision loss.

The main risk factors associated with the DR are the duration of diabetes and degree of

glycaemic control, with other conditions such as hypertension, pregnancy and elevated serum

lipid levels also playing a role. It is estimated that, in 2015, nearly 2.6 million people were

visually impaired as a result of DR and it is expected that this number will increase to 3.2

million by 2020 (Leasher et al., 2016, Flaxman et al., 2017).

In Australia, the prevalence of DR is 29.1% (McKay et al., 2000). Ophthalmoscopy, with or

without dilating the eye, can be used to screen for DR by detecting microaneurysms (de Carlo

et al., 2015). Fluorescein angiography detects the vascular changes in the inner and outer

retinal blood barrier and thus is a more sensitive, although more costly, option than

ophthalmoscopy for diagnosing DR (Kwiterovich et al., 1991). It is crucial that individuals

with DR maintain blood glucose levels within target range. Treatments for later stage disease

include laser therapy, eye injections (ranibizumab and aflibercept) and eye surgery (removal

of vitreous humour).

1.2.5 Cataract

Cataracts are caused by aging of the crystalline lens; the lens becomes cloudy or even opaque

so light does not pass through and vision is blurry. Lifestyle factors such as smoking and

exposure to the sunlight and long-term oral corticosteroid therapy also contribute to cataracts

(Allen and Vasavada, 2006). Cataract can be diagnosed using various eye examinations such

as a slit-lamp exam, retinal exam, refraction and visual acuity test.

Just over half (51%) of the global blindness burden — 20 million people afflicted — is due to

cataract (World Health Organization report -(report)). Although cataract can be treated by

surgery, surgical services are inadequate in developing countries.

1.2.6 Uveal Melanoma

Uveal melanoma (UM) is the most common intraocular malignant tumour in adults and it has

a high rate of fatal metastasis. UM emerges from the melanocytes in the choroid, ciliary body

24

or the iris (also known as uveal tract) (Chang et al., 1998). The cause of UM is not entirely

understood. Sun exposure is considered the most critical risk factor for cutaneous melanoma;

however, ultraviolet (UV) radiation also plays a role in the development of UM. The iris is

unprotected from visible light and UV radiation, whereas the ciliary body and choroid are less

directly exposed. Nonetheless, around 85% of UM emerges in the choroid, 10% in the ciliary

body, and the remaining 5% in the iris, demonstrating an inverse association between exposure

to sunlight and the manifestation of UM (Weis et al., 2006). Notably, welding as an occupation

(artificial source of UV radiation) is associated with UM (Vajdic et al., 2002). Additionally,

light iris colour and skin pigmentation and the inability to tan are other factors that correlate

with increased risk for UM (Wakamatsu et al., 2008).

Genetic variants such as rs12913832, rs1129038 and rs916977 on chromosome 15q13.1, which

contribute to eye and skin pigmentation, are correlated with UM (Ferguson et al., 2016) and

mutation at BAP1 (BRAC1-associated protein 1) gene is also found to be an inclined element

for UM (Harbour et al., 2010). The most common symptoms of UM are:

• blurred vision

• visual field defect

• ocular irritation

• photopsia (flashing of lights)

• floaters

• Metamorphopsia

The incidence of UM varies depending upon sex, race and geographical location, with the rate

in men 30% higher than in women (McLaughlin et al., 2005). In the United States, around 5.1

people in 1 million are diagnosed with UM (Singh et al., 2011) whereas, in Europe, especially

in Denmark and Norway, the rate is more than 8 people in 1 million. In Spain and Italy, 2

people in 1 million are diagnosed with UM (Virgili et al., 2007). In contrast, the incidence

among the Asian population is 0.39 per million (Hu et al., 2005).

UM is diagnosed primarily by clinical examination, including slit-lamp biomicroscopy,

indirect ophthalmoscopy and ancillary diagnostic testing such as ocular ultrasonography

(Jovanovic et al., 2013b). Patients with posterior UM are mostly treated by plaque radiation

therapy or enucleation. Local radiotherapy such as 106 ruthenium, 125 iodine brachytherapy,

25

stereotactic radiosurgery or photo beam therapy are other treatment options for UM (Nathan et

al., 2015).

1.2.3 Bioinformatics application in eye disease

Bioinformatics is an interdisciplinary field that uses mathematics, computer science and

statistics to solve complex biology problems. Various bioinformatics approaches can be

applied in order to discern the underlying genetics of the eye:

● Quality Control in stem cells for genomic aberrations.

● Epigenetic age analysis.

● CRISPR screening to identify essential knockout genes.

● RNA-Seq analysis.

1.2.3.1 Quality Control in Stem Cells Genomic Integrity

iPSCs are equivalent to a human embryonic stem cell in that they have the ability to self-renew

and to differentiate into all types of embryonic germ layers (Takahashi and Yamanaka, 2006,

Takahashi et al., 2007, Yu et al., 2007). Induced pluripotent stem cells (iPSCs) show great

promise in clinical applications because of their potential use in personalised cell therapy or

disease modelling (Simara et al., 2013). To develop the best disease model, it is essential to

have a stable and intact genome. For reasons not entirely understood, iPSCs tend to accumulate

genomic aberrations and lack the functionality of repair mechanisms (Weissbein et al., 2014).

Genomic instability in human pluripotent stem cells (hPSC) was first identified in the early

2000s when karyotyping abnormalities such as trisomy of the chromosome were noted (Cowan

et al., 2004, Draper et al., 2004). Other chromosomal abnormalities observed in hPSCs include

duplication and deletions in genes or point mutations (International Stem Cell et al., 2011,

Taapken et al., 2011). Single nucleotide polymorphism (SNP) genotyping arrays can detect

these abnormalities. It is reported that iPSCs have a higher number of copy number variations

(CNV) than hPSCs (Laurent et al., 2011, Martins-Taylor et al., 2011). These variations in

iPSCs seem less evident, but recently it was found that somatic mosaicism in fibroblast cultures

causes CNVs among human and murine iPSCs (Bock et al., 2011, Hussein et al., 2011).

Because each member of a colony of iPSCs is a clonal offspring of single programmed somatic

cells, the variation present in the parent’s cells is amplified (Peterson and Loring, 2014). The

genomic aberration in stem cell lines can be identified using traditional methods such as G-

26

Banding or Fluorescence in-situ hybridisation (FISH) or using computational analysis on

genotyping arrays (Illumina and Axiom). Using comparative genomic hybridization (CGH)

array, (Chin et al., 2009) analysed the CNVs of three fibroblast lines and the hiPSCs that were

derived from the fibroblasts and unique CNVs were identified in each Human induced

pluripotent stem cells (hiPSC) (Chin et al., 2009).

1.2.3.1.1 Traditional Karyotyping

Karyotyping is a cytogenetic process of putting together all the chromosomes of an organism

to provide a genome-wide print of that individual. It is prepared by the standard procedure

following graded staining to reveal the structure of each chromosome and can detect anomalies

up to several megabase pairs or more (O'Connor, 2008). Several methods are used to detect

chromosomal abnormalities:

G-banding (Giemsa Banding)

G-banding is also known as Giemsa banding. It is a classic staining method that employs

various chemical and enzymatic treatments for chromosome preparations (Figure 1.2). The

main advantage of using this method is that permanent slides can be prepared and studied under

a standard light microscope. In this method, Giemsa stain correlates with the intense banding

fluorescent regions. Trypsin is used for the pre-treatment of a chromosome before Giemsa

staining; other stains, including the Wright stain and the Leishman stain, provide the same

pattern as Giemsa stain (Charleen M Moore and Robert G Best, 2001).

Figure 1.2 G-banding showing metaphase arranged in the standard karyotype format.

27

Fluorescence in-situ hybridisation (FISH)

FISH is the technique used to determine the chromosomal position of a deoxyribonucleic acid

(DNA) or Ribonucleic acid (RNA) probe. This cytogenetic technique was developed and first

used in the early 1980s (Pinkel et al., 1988).

Figure 1.3: Fluorescence in-situ hybridisation (FISH) -Normal fibroblast cell line showing normal chromosomes 11 positive for WCP11 (spectrum orange) and normal chromosomes 15 positive for WCP 15 (spectrum green).

In this technique, fluorescent DNA probes are attached to the complementary region of the

chromosome and emit signals that are later visualized by using probes to detect abnormalities

in the chromosome (Amann and Fuchs, 2008) (Figure 1.3).

FISH is much faster than other traditional methods for metaphase karyotyping analysis

(Sujobert et al., 2013). Specific chromosomal abnormalities such as amplification,

translocation or deletion can be reported and sketched (Ratan et al., 2017).

1.2.3.1.2 Virtual karyotyping

Virtual karyotyping is a sophisticated technique that determines karyotype from analysing a

DNA probe, which is isolated and recorded (Wang et al., 2002). It detects genomic aberrations

at high resolution compared to the traditional method that identifies abnormalities at a low

28

resolution based on Comparative genomic hybridization (CGH) (Shinawi and Cheung, 2008).

The main methods used for virtual karyotyping are as follows:

• CGH-arrays

• SNP genotyping arrays

CGH-arrays: Array-based comparative genomic hybridisation is also known as molecular

karyotyping. This method is developed to detect CNV that is undetectable using traditional

methods. CNV occurs due to the loss of function or amplification of a genomic sequence. In

CGH array, genomic DNA is extracted from peripheral blood lymphocytes or skin fibroblasts

and labelled with fluorescent dye (generally Cy3-labelled dCTP). The patient DNA is labelled

with another fluorescent dye (Cy5-labelled dCTP) with the control genomic DNA co-

hybridised for a selected set of pre-determined genomic fragments (Vermeesch et al., 2005).

The amount of Cy3 and Cy5 fluorescent emission is interpreted as normal or not. For example,

if the fluorescent intensity of Cy3 is high in the sample, then there is duplication in the sample

whereas if Cy5 fluorescent is high, a deletion is present (de Ravel et al., 2007). CGH provides

resolution more than 100-1000 times higher than traditional methods and also detects small

genomic aberrations of up to a few hundred base pairs and thus is useful for detecting

syndromes involving microdeletions or microduplications. Moreover, CGH arrays allow us to

understand the relationship between the disease and its underlying genetic defect.

SNP arrays: Single nucleotide polymorphism (SNP) array is used to detect polymorphisms

within the population. SNP is the typical type of genetic variant and presents as a different

single nucleotide within the genomic sequence. The basic principle of SNP arrays is the same

as DNA microarrays. SNP arrays work mainly on three components: immobilizing allele-

specific oligonucleotides (ASO), labelling a genomic sequence with fluorescent dyes and a

system used for recording hybridization system (LaFramboise, 2009). The ASO probes are

selected based on the sequencing representation of an individual. In the SNP arrays, two probes

are used for the SNP positions for both alleles (Rapley et al., 2004 ). SNP microarrays have

come to play an essential role since the human genome map was published. SNP array is a

potent tool to study genomic variation and for the virtual karyotyping using various

computational tools by determining the CNV from each SNP from the array and also to align

the chromosome order (Sato-Otsubo et al., 2012).Various commercial technologies such as

Affymetrix SNP genotyping and Illumina genotyping platforms are used to study CNVs.

29

1.2.3.1.3 Tools for detecting Copy number variations

CNV is discrepancy in the number of copies of a gene between individuals in the form of gain

or loss of genetic material; CNV can lead to the genetic imbalance. SNP is the standard variant

observed in the population (>1%). CNV can be described as amplification, insertion or deletion

at the specific genomic location (Thapar and Cooper, 2013). Various computational methods

used to determine the CNV in genomic samples are shown in Table 1.1.

Table 1.1: Various computational tools for CNV detection in whole-genome genotyping and next-generation sequencing data.

Software Algorithm Published URL Programming

language

Citations

PennCNV

(Wang et

al., 2007)

Hidden

Markov Model

(HMM)

2007 http://penncnv.open

bioinformatics.org/

en/latest/

C++ 1402

QuantiSNP

(Colella et

al., 2007)

Objective

Bayes Hidden-

Markov Model

2007 https://sites.google.

com/site/quantisnp/

MATLAB 586

BCFtools

CNV

(Danecek et

al., 2016)

Hidden

Markov Model

(HMM)

2016 https://samtools.git

hub.io/bcftools/how

tos/cnv-calling.html

Perl 7

CHAT (Li

and Li,

2014)

Markov Chain

Monte Carlo

(MCMC)

2014 https://sourceforge.

net/p/clonalhetanal

ysistool/wiki/CHA

T/

R 39

CNVcaller

(Wang et

al., 2017b)

Read Depth

Algorithm

2017 https://github.com/J

iangYuLab/CNVca

ller

Perl/Python 9

CNVhap

(Coin et al.,

2010)

chromosome-

wide

2010 https://bioinformati

cshome.com/tools/c

nv/descriptions/cnv

Hap.html

Java 49

30


language

Citations

haplotype

model

ADTEx

(Amarasing

he et al.,

2014)

Hidden

Markov Model

(HMM)

2014 https://sourceforge.

net/projects/adtex/

R/Python 61

PICNIC

(Greenman

et al., 2010)

Hidden

Markov Model

(HMM)

2010 https://www.sanger

.ac.uk/science/tools

/picnic

MATLAB 169

1.2.3.2 DNA Methylation

1.2.3.2.1 Background

Epigenetics is the study of genomic modification that affects gene expression without altering

the genomic sequence. In the 1970s, DNA methylation was considered to be a transcriptional

regulatory mechanism that was maintained during cell division (Riggs, 1975). Over time, the

definition of epigenetics has evolved from the cellular level to the molecular level. With recent

advancements in next-generation sequencing technologies, epigenetics has expanded to

include the study of genome-wide chromatin modification that does not change the genomic

sequence including DNA methylation, histone modification, chromatin accessibility and

microRNA regulations (Sarda and Hannenhalli, 2014).

Figure 1.4: DNA methylation (Martin and Fry, 2018).

31

1.2.3.2.2 DNA methylation

DNA methylation occurs when a methyl group is added to the 5th carbon of a cytosine residue

that is linked by a phosphate to a 5CpG dinucleotide by DNA methyltransferase (DNMT1,

DNMT3A and DNMT3B), thus resulting in 5-methylcytosine as shown in Figure 1.4

(Ballestar, 2011). DNA methylation is involved in transcriptional regulation. Most of the

genome does not have CpG sites, but a collection of CpG sites (known as a CpG island) occurs

commonly on the promoter region of housekeeping genes. However, these CpG islands do not

undergo methylation, resulting in gene activation Methylated CpG promoters are related to low

gene expressions, and CpG islands shores usually have low density, located 2kb from the CpG

islands. These CpG shores demonstrate tissue-specific methylation with gene silencing

whereas exonic methylation have active expression. Outside the shore region, the genomic

region has a lower frequency of CpG sites, which are methylated. These CpG promoters and

enhancers are regulated by tissue-specific genes (Irizarry et al., 2009). At present, DNA

methylation (CpG sites) have been identified as a useful probe in age-associated studies.

1.2.3.2.3 DNA methylation age

Aging is an irreversible biological process associated with a decline in physical function and

an increase in the rate of neurodegenerative and cardiovascular disease and cancer (Sen et al.,

2016). The aging process involves changes at a molecular and cellular level, including

epigenetic alterations, telomere abrasion or cellular agedness (Lopez-Otin et al., 2013).

Epigenetic modifications such as DNA methylation, histone modification and non-coding

RNA with the change in DNA methylation are related to the aging process (Jones et al., 2015,

Bruce Richardson, 2003). DNA methylation age (DNAm age) can be used as an epigenetic

biomarker for the aging process. DNAm age of various human tissue and cell type can be

predicted by using several tools as shown in Table 2 (Horvath, 2013) (Hannum et al., 2013).

Horvath Method for DNAm Age prediction

• Data from Illumina 27K or Illumina 450K array platform used for analysis.

• Elastic net regression model used for selecting 353 CpGs for DNAm age analysis with

the glmnet R package ((Friedman et al., 2010).

● WGCNA R package (Langfelder et al., 2013) used for meta-analysis to measure pure

age effects.

● The tool is implemented in R.

32

Hannum Method for DNAm Age prediction

• Illumina Infinium HumanMethylation450 array was used to extract the methylation

fraction values.

• 71 CpGs for DNAm age analysis.

• Elastic Net algorithm used in aging model implemented in a glmnet R package

(Friedman et al., 2010).

• Missing values claimed by the ten nearest marker approach using impute R package

(Troyanskaya et al., 2001).

• The tool is implemented in R.

1.2.4.1 CRISPR Cas 9

1.2.4.1.1 Background

Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) is a naturally occurring

anti-viral defence mechanism identified in many prokaryotic species such as bacteria and

archaea. The CRISPR loci appear on prokaryotic DNA and have been characterised as having

identical palindromic repeats of 21-40 base pairs in length (Jansen et al., 2002, Mojica et al.,

2000). The CRISPR-associated protein 9 (Cas9 protein) derived from type II CRISPR emerged

as a powerful tool for genomic engineering. An RNA-guided DNA endonuclease, Cas9 can

easily be programmed to alter a specific site with a slight change in the guided RNA sequence

(Wang et al., 2016). The Cas9 protein is able to search out and cleave the target DNA in both

natural and artificial CRISPR/Cas systems. It has six domains: REC I, REC II, Bridge Helix,

PAM Interacting, HNH and RuvC (Figure 1.5) (Jinek et al., 2014, Nishimasu et al., 2014).

33

Figure 1.5: Cas9 protein domains and schematics and Crystal structure from PDB2. The Rec I domain is the most prominent domain and binds to the guide RNA, whereas the

HNH and RuvC domains are responsible for the cleavage activity and cut single-strand DNA.

These domains are highly homologous and are also found in other proteins as well (Jinek et

al., 2014, Nishimasu et al., 2014). However, it is reported that Cas9 protein remains inactive

in the absence of guide RNA (Jinek et al., 2014).

Cas9 protein searches for the target DNA with the help of protospacer adjacent motif (PAM)

sequence (two or three base sequence located in the complementary region to the guide RNA)

(Sternberg et al., 2014).

1.2.4.1.2 CRISPR Cas9 genome editing

The initial step of CRISPR/Cas9 genome editing is to modify the genomic locus by a double-

stranded break (DSB). In mammalian cells, two types of methods are used for detection and

repair: homology-directed repair (HDR) and non-homologous end joining (NHEJ) (Vartak and

Raghavan, 2015). The HDR functions only in the replicative phase of the cell cycle by using a

homologous template to repair the missing genetic information. NHEJ can occur anywhere in

the cell cycle and works by ligating the broken DNA using chemicals based on an independent

template with the introduction of indels (small insertions and deletions mutations) of various

lengths (Vartak and Raghavan, 2015). The position of indels in the NHEJ is very crucial; if it

occurs in the coding region, this could result in a frameshift and lead to a premature stop codon

with a change in the protein function (Figure 1.6).

2 Image Source: https://sites.tufts.edu/crispr/crispr-mechanism/

34

Figure 1.6: Schematic picture of genome editing via the CRISPR/Cas9 system3. The CRISPR/Cas type II system from Streptococcus pyogenes has been widely used for

genome editing. It is employed with the aid of Cas9 nuclease and a short guide RNA (sgRNA),

which is an integration of crRNA (contains genome-targeting sequence) and tracrRNA in the

cell. Various strategies, such as lentiviral transduction (Kabadi et al., 2014) and cas9-guide

RNA ribonucleoprotein complex, can be used for delivering sgRNA and cas9 into the cells

(Lin et al., 2014). The sgRNA library can be delivered by viral vectors such as lentivirus,

retrovirus or adeno-associated virus (AAV). The AAV vector does not integrate into the host

genome, whereas lentivirus and retrovirus integrate into the host genome. Thus, the retrovirus

delivery method works for dividing cells whereas AAV works non-dividing cells (Joung et al.,

2017a). In terms of capacity, the insert size of AAV is much smaller compared to retrovirus

or lentivirus (Joung et al., 2017a). Zinc finger nuclease (ZFN) and transcription activator-like

effector nuclease (TALEN) lentiviral-based delivery methods are widely used in genomic-

editing technologies. TALENs can effectively deliver as mRNA by using lentiviral delivery

method containing inactivated reverse transcriptase, and this method was adopted by the

CRISPR/Cas9 system (El-Aneed, 2004). Type II CRISPR/Cas9 genome editing has been used

edit various cells and organisms in vivo and in vitro including human (Mali et al., 2013),

bacteria (Pyne et al., 2015), zebrafish (Hwang et al., 2013), mice (Hall et al., 2018), rat (Back

et al., 2019), rabbits (Kawano and Honda, 2017) and others.

3 Source: AcceGen R&D Team

35

1.2.4.1.3 CRISPR Screening

CRISPR Screening is a technique used for uncovering the genes that are responsible for

specific biological functions. With the advent of next-generation screening technologies,

functional phenotypes can be identified and essential genes determined using loss-of-function

and gain-of-function inquiries. Generally, genetic pooled screens are used for targeting a large

number of genes under the suitable condition for one gene agitated per cell. A cellular

population can be interrogated to uncover unanswered biological questions for a particular

disease, with enrichment or depletion of cells as per the environment. Genetic screening not

only facilitates identification of genetic variants but also helps discover target for therapeutic

intervention (Bartha et al., 2018). Genetic screening can be carried out by two methods:

● Array: siRNAs, shRNAs and sgRNA separately brought into cells.

● Pooled: a mixture of shRNA/sgRNA introduced into cells simultaneously with specific

barcoding.

In array format screening, targeting reagent is synthesised separately and is delivered by

transfection after analysis has been done on the readout of each well (Shalem et al., 2015,

Agrotis and Ketteler, 2015). In pooled array, targeting reagents are sourced efficiently through

Insilco oligonucleotide libraries and transfection has been done on plasmid libraries that cloned

for pooled screening. In pooled screening, the number of transgenes is limited during

transfection to ensure that each cell has only one sgRNA, resulting in low multiplicity of

infection (MOI) and the readout can be analysed by next-generation sequencing (Shalem et al.,

2015). Pooled screening is less expensive but more time-consuming compared to array screens.

However, it is relatively easier to apply long cultures in pooled screening compared to the array

format because eventually a smaller culture volume is required; a pooled approach is

commonly used with the CRISPR/Cas9 system (Wang et al., 2014, Koike-Yusa et al., 2014,

Zhou et al., 2014). Pooled screening array can be further divided into two categories:

● Negative Selection: the primary goal of negative selection is to find the sgRNA that is

depleted during selection, which corresponds to cell apoptosis.

● Positive Selection: the primary goal of positive selection is to find the sgRNA that

enriched during the selection resulting in cell proliferation.

Several tools have been developed for analysing CRISPR knockout screening data or other

screening data (Table 1.2).

36

Table 1.2: Tools available for CRISPR screening data analysis.


Language

Citation

MAGeCK

(Li et al.,

2014)

Mean/variance modelling and robust rank aggregation

2014 https://sourceforge.net/p/mageck/wik

i/Home/

Python 189

CRISPRAnalyzeR (Winter et al., 2017)

Workflow 2017 http://crispr-analyzer.dkfz.de/

Web-based 8

caRpools

(Winter et

al., 2016)

Workflow 2016 https://github.com/boutroslab/caRpools

R 27

1.2.4.1.4 CRISPR/Cas9 technology in eye disease

CRISPR/Cas 9 emerged as a pivotal technology for understanding the enrichment and

depletion of genes related to several diseases, including those affecting the eye.

• Recently, CRISPR/Cas9 technology was used to knockout the vegfa in mouse retina

using the lentiviral delivery method. SpCas9 and single sgRNA were delivered to erode

the vegfa gene in the mouse retina; this may contribute to developing new treatments

for ocular diseases such as AMD (Holmgaard et al., 2019).

• Another recent study reported the effect of the CRISPR/Cas9 knockout method on the

eye pigmentation gene (White) in the fruit fly, which altered the colour of head spots in

(Bai et al., 2019). Similar work has also been done on Helicoverpa zea (Boddie), with

knockout of the eye colour gene vermillion (Perera et al., 2018).

1.2.5 RNA-Seq Analysis

1.2.5.1 Background

Over the past decade, genomes, including DNA modification and RNA qualitative and

quantitative changes, have been characterised for species ranging from simple organisms to

37

humans (Consortium et al., 2007). This advance has occurred through the use of the various

genomic measurements, including comprehensive transcriptomic studies (Consortium et al.,

2007). The transcriptome is the complete set of cellular RNA transcripts, and their quantity,

for a specific developmental stage or physiological condition. The main aim of establishing

transcriptomes is to catalogue all species of transcript including mRNAs, non-coding RNA and

sRNA to determine the transcriptional structure of genes in terms of their start sites, 5' and 3'

ends, splicing pattern, other post-translational modifications and to quantify the change in

expression at each level of transcript during development and under different conditions (Wang

et al., 2009).

Initially, transcriptomics studies primarily relied on hybridisation-based microarray-based

technology with limited ability to quantify the RNA molecules expressed from the genome

over a wide range of levels (Ozsolak and Milos, 2011). The introduction of high throughput

next-generation sequencing (NGS) allowed RNA analysis from the cDNA on a massive scale

(RNA-Seq) (Metzker, 2010). Identifying and quantifying transcripts for gene expression has

been a distant goal in molecular biology since the discovery of RNA's position as the key

intermediate between the genome and the proteome. Single high-throughput sequence assay is

called RNA-Sequencing (RNA-Seq). Various RNA-Seq protocols and analysis have been

published. Currently, different applications and workflows/pipelines for RNA-Seq analysis can

be used. A researcher can adopt different strategies depending upon the organism being studied

and the research goals. RNA-Seq can be used for transcriptome profiling with other functional

genomic methods for analysing gene expression (Conesa et al., 2016).

Every RNA-Seq experiment possibly has different methods for transcript quantification,

normalization and differential expression. Our focus is to outline current resources for the

bioinformatic pipelines for analysis of RNA-Seq data.

1.2.5.2 RNA-Seq Experiment Design

One of the most important aspects of experimental design is having an appropriate RNA-

extraction protocol to remove the highly abundant ribosomal RNA (rRNA), which constitutes

90% of total RNA in the cell, leaving the 1-2% of mRNA that a researcher is usually interested

in. When studying eukaryotes, this involves choosing whether to enrich for mRNA using

poly(A) selection or to deplete rRNA. Poly(A) selection required a high proportion of mRNA

38

with minimum degradation, yielding a high fraction of reads typically onto known exons. Many

biological samples (such as tissue biopsies) are insufficient for producing suitable poly(A)

RNA-Seq libraries. For bacterial samples, mRNAs cannot be polyadenylated; the only viable

alternative is ribosomal depletion. Another consideration is to generate strand-preserving

libraries (Conesa et al., 2016).

The first generation of Illumina-based RNA-Seq used random hexamer priming to reverse-

transcribe poly(A)- selected mRNA. This methodology did not retain the information contained

on the DNA strand that is expressed and complicated the analysis and quantification of the

overlapping transcript (Mortazavi et al., 2008). Several strand-specific protocols (Levin et al.,

2010), such as the widely used dUTP method, extend the original protocol by incorporating

UTP nucleotide during the second cDNA synthesis step, before ligation adapter, followed by

the digestion of dUTP containing strand (Parkhomchuk et al., 2009). The final fragment size

(usually less than 500 base pairs for Illumina) is crucial for proper sequencing and subsequent

analysis. Sequencing can be single-end (SE), or paired-end (PE) reads, although the latter is

preferable for isoform expression analysis (Katz et al., 2010, Garber et al., 2011).

Similarly, longer reads improve mapping and transcript identification (Garber et al., 2011,

Labaj et al., 2011). The best sequencing method to use depends upon the goal. Cheaper and

short single-end reads usually are sufficient for gene expression level in well-annotated

organisms, whereas longer and paired-end reads are preferable for annotated transcriptomes.

Another critical factor is sequencing depth or library size, which is the number of sequenced-

reads for each sample. More transcripts can be quantified more precisely the deeper the sample

is sequenced (Mortazavi et al., 2008). Optimal sequencing depth depends upon the aim of the

experiment. While some researchers argue that as few 5 million mapped reads are sufficient to

accurately quantify highly expressed genes in most eukaryotic transcriptomes, up to 100

million reads are required to quantify precisely gene and transcripts that have low expression

level (Sims et al., 2014). Studying single cells with limited sample complexity is often carried

out with one million reads but may be done with as few as 50,000 reads for highly expressed

genes (Pollen et al., 2014). The complexity of the targeted transcriptome depends on the size

of the library. An experimental result showed that deep sequencing improves identification and

quantification but also detects transcriptional noise and off-target transcript. Transcriptome

coverage can be improved by assessing the saturation curve at a given sequencing depth

(Tarazona et al., 2011).

39

Finally, the number of replicates is always a crucial factor and depends on the amount of

technical variability in RNA-Seq procedure and the biological variability of the system under

study, as well as on the capacity for detecting statistically significant differences in the gene

expression between the experimental groups.

1.2.5.3 Analysis of RNA-Seq Data

A typical RNA-Seq experiment involves initial processing for sequence reads, assembling

those reads to reference sequences, annotating the transcripts, quantifying the reads per

transcript and statistically comparing transcript abundance across samples or treatments.

1.2.5.3.1 Transcript Identification

Discovering new, unannotated transcripts and the associated quantification analysis depends

upon mapping solely to the reference transcriptome and mapping the reads onto the reference

transcriptome or genome to which transcripts expressed. If the organism does not have its

genome sequenced, then reads must first be assembled into contigs and treated as an expressed

transcriptome to which reads are mapped, and coverage can be used as transcript expression

level.

1.2.5.3.2 Alignment

Alignment can be done by: mapping to the genome or mapping to the transcriptome.

Regardless of the reference sequence, whether a genome or transcriptome, reads may be

aligned uniquely or be multi-aligned reads (multireads) to the reference sequence. Genomic

multireads are mainly due to repetitive sequences or paralogous genes that have a shared

domain.

1.2.5.3.3 Transcript discovery and De novo transcript reconstruction

Identifying novel transcripts by using short reads is one of the most challenging tasks in RNA-

Seq. Short reads span across several splice junctions and make it challenging to infer all full-

length transcripts directly. Identifying the transcription start and end site (Steijger et al., 2013)

is very challenging; a tool like GRIT (Boley et al., 2014) that incorporates other data such as

5' ends improves the likelihood of annotating the expressed isoform correctly. Several methods,

40

such as Cufflinks (Roberts et al., 2011), iReckon (Mezlini et al., 2013), SLIDE (Li et al., 2011)

and StingTie (Pertea et al., 2015), add annotations to the possible list of isoforms.

In case of incomplete or unavailability of a reference genome, RNA-Seq reads can be

assembled de novo into a transcriptome by using various packages such as SOAPdenovo-Trans

(Xie et al., 2014), Oases (Schulz et al., 2012), Trans-ABySS (Grabherr et al., 2011) or Trinity

(Haas et al., 2013). In pair-end, strand-specific sequencing, long reads are preferable because

they can be more informative. It is impossible to assemble low-expressed transcripts because

they lack coverage for reliable assembly, and too many reads are problematic because this may

lead to misassembly and increase the run-time. Therefore, computationally reduction of the

number of reads is recommended for deeply sequenced samples. For comparative analyses

across the samples, it is suggested to combine all the reads from multiple samples into a single

output to obtain a set of transcripts for expression estimation (Haas et al., 2013).

1.2.5.4 Quantification

RNA-Seq is commonly used to estimate gene and transcript expression. This application is

based on the number of reads that map to each transcript. There are a few algorithms such as

Sailfish that rely on k-mer counting in reads without the need of mapping (Patro et al., 2014).

HTSeq-count (Anders et al., 2015) and featureCounts (Liao et al., 2014) are the programs used

for quantifying raw counts of mapped reads.

Several algorithms have been developed to estimate transcript-level expression by handling the

problem related to the transcript sharing most of their reads. Cufflinks (Trapnell et al., 2010)

estimates transcript expression from a mapping genome obtained by TopHat using an

expectation-maximisation approach that estimates transcript abundances. Algorithms that

quantify expression from transcriptome includes RSEM (RNA-Seq by Expectations

Maximization) (Li and Dewey, 2011), eXpress (Roberts and Pachter, 2013), Sailfish (Anders

et al., 2015) and Kallisto (Bray N, 2015). These are multi-mapping read methods among the

transcript and sequences biases are corrected by sample-normalised values within the output

(Anders et al., 2015, Roberts and Pachter, 2013). The RSEM algorithm uses an expectation

maximisation approach that returns transcripts per million (TPM) values (Li and Dewey, 2011).

The NURD (Ma and Zhang, 2013) algorithm provides an efficient way of estimating transcript

expression for single-end reads with low memory and computing cost.

41

1.2.5.5 Differential Gene Expression Analysis

Differential gene expression identifies features of the transcriptome that are differentially

expressed between or among different groups or conditions. The normalization of RNA-Seq

provides an accurate inference of expression level. This approach is mostly used for

determining sequencing depth and fails to correct for library preparation and other technical

biases (Risso et al., 2014). The methods that avoid this pitfall are TMM (Zhang et al., 2014b),

DESeq (Anders and Huber, 2010), PoissionSeq (Li et al., 2012) and UpperQuartile (Bullard et

al., 2010), ), which avoids highly variable and expressed features. Differential gene expression

can also be analysed on transcript abundance by using Sleuth (Pimentel et al., 2017).

RNA-Seq is based on the read counts that are probabilistically assigned to transcripts; the first

approach is to compute differential expression using negative binomial distribution (Anders

and Huber, 2010). A generalisation of Poisson distribution allows overdispersion (additional

variance) beyond the expected variance from a random sampling of RNA-Seq data. The data

might lose their discrete nature by using extensive normalization. Some methods, such as

edgeR (Robinson et al., 2010), take raw read counts as input and perform integrated differential

expression analysis along with normalization. In other methods, normalization is performed

before the differential expression to remove all possible biases. Recent studies have

demonstrated that the choice of the method can affect the outcome of the analysis and no single

method is preferred for all data sets (Soneson and Delorenzi, 2013).

1.2.5.4 RNA-Seq in Eye diseases

Obtaining high-quality ocular tissue and developing eye disease models has been a challenge

in ophthalmology research. RNA-Seq analysis (Yang et al., 2015). has brought tremendous

advances:

• Retinal dystrophies: Currently, 75 mutated genes have been identified as involved in

affecting retinal photoreceptor cells (RPE) in the retinal dystrophies, which leads to

progressive vision loss. Stem cell-based therapy provides us with a potential avenue for

treatment, either replacing retinal cells via differentiation or by using neuronal cell

lineage (Lu et al., 2015). Stem cell transplantation with human neural progenitor cells

(hNPCs) in a rodent model for retinal degeneration was interpreted using RNA-Seq

analysis (Jones et al., 2016). Gene expression analysis between treated and untreated

42

rat models identified 68 genes with a change in the expression due to the treatment in

hNPCs. Later enrichment of signalling (such as an increase in the survival rate in

photoreceptor cells) was noted using pathway and enrichment analysis (Jones et al.,

2016).

• Age-related macular degeneration (AMD): AMD is characterized by degeneration

in RPE cells, resulting in photoreceptor cell death. A transcriptomic study using a rat

model was conducted to estimate the change in the retinal transcriptome. Differential

gene expression analysis found the gene cluster that is down-regulated compared to the

control condition (Telegina et al., 2015).

• Glaucoma: The level of fibrotic encapsulation post-trabeculectomy surgery affects

how successful this treatment is for glaucoma (Addicks et al., 1983). RNA-Seq has

been used to determine the responses in fibrotic and non-fibrotic fibroblast cell lines

isolated from glaucoma patients (Yu-Wai-Man et al., 2017). Several genes such as

RELB, PPP1R13L and MYCOD were identified and were found to be significantly

upregulated and associated with inflammation and apoptosis (Yu-Wai-Man et al.,

2017).

1.3 Outline of the Thesis

Overall, the thesis work for applying bioinformatic tools to better understand eye diseases

comprised four research projects:

1. Resolution of a low-density array for virtual karyotyping.

2. Variation in epigenetic age in ocular tissues.

3. A genome-wide CRISPR/Cas 9 screen to identify a novel therapeutic target for uveal

melanoma.

4. RNA-Seq Analysis:

a. To identify novel targets for retinal neovascularization in OIR rat models.

b. To identify the genes that are involved in the differentiation of dental pulp-

derived mesenchymal stem cells to trabecular meshwork cells.

43

Chapter 2 Virtual Karyotyping

2.1 Introduction

Human embryonic stem cells (hESCs), derived by isolating the inner cell mass from

preimplantation embryos, have the ability to differentiate into all cell-types found in the body

(Reubinoff et al., 2000, Thomson et al., 1998). These pluripotent stem cells can be maintained

indefinitely without changing their characteristic profile. Improved understanding of key hESC

pluripotency regulatory networks has allowed the generation of induced pluripotent stem cells

(iPSCs) by transfecting adult somatic cells with specific transcription factors known to be

involved in pluripotency (Yu et al., 2007, Park et al., 2008, Takahashi et al., 2007). These

iPSCs share many of the same characteristics of hESCs and the ability to generate iPSCs

directly from well characterised or genetically defined patients has opened a number of novel

avenues for disease modelling and therapy (Kiskinis and Eggan, 2010).

Prior to using hESCs or iPSCs for in vitro or in vivo applications, each cell line must undergo

rigorous characterization to ensure they are truly representative of pluripotent stem cells. A

series of immunological and molecular techniques can be used to confirm expression of

markers indicative of undifferentiated stem cells and that the cells have the capacity to

differentiate into each embryonic germ layer, thereby ensuring pluripotency potential (Muller

et al., 2010, Bock et al., 2011). Furthermore, given that hESCs and iPSCs have exhibited

chromosomal instability, each cell line must be inspected for the incorporation of genomic

aberrations (Hamada et al., 2012).

As the stem cell field moves from small-scale disease modelling — studying known mutations

with readily detectable phenotypes — towards larger-scale studies using hundreds of cell lines

to model complex heterogeneous diseases, improvements in pluripotency screening and

detection of karyotypic aberrations is required (Merkle and Eggan, 2013). To this end, a

quantitative polymerase chain reaction panel (TaqMan® hPSC Scorecard TM Panel; Life

Technologies) has been established as a high-throughput method to quantitatively confirm

pluripotency (Bock et al., 2011, Fergus et al., 2016). However, traditionally, the genomic

stability quality control process has been performed using low-throughput, low resolution and

relatively expensive cytogenetic karyotyping techniques (Henson et al., 2005). Virtual

karyotyping using data from comparative genomic hybridisation (CGH) or single nucleotide

44

polymorphism (SNP) arrays have been applied to hESCs and iPSCs to detect chromosomal

changes at a much higher resolution to traditional karyotyping methods (International Stem

Cell et al., 2011). Currently, the costs of these published methods are similar to performing a

standard cytogenetic karyotyping with microscopic analysis of chromosomes with G-banding

(Alkan et al., 2011).

The aim of this study was to investigate the utility of a relatively inexpensive, low-density

genome-wide SNP array for the karyotypic assessment of human induced pluripotent stem cells

(hiPSCs). Karyotypically normal and abnormal iPSC clones were pooled at specific ratios and

then genotyped on the Illumina Infinium HumanCore BeadChip and UK Biobank Axiom array.

Specifically, we sought to determine the resolution of these arrays in detecting chromosomal

aberrations and assess the ability of these arrays to identify clonal variations.

2.2 Methods

Differentiation Protocol was performed “Duncan Crombie” and full details

are in Supplementary Method 1.

2.2.1 Cytogenetic Karyotyping

Cytogenetic karyotyping was performed by the Cytogenetics Laboratory at Monash Health

(Clayton, Victoria, Australia). G-banding was performed using a standard protocol and a

minimum of 20 metaphase spreads with no overlapping chromosomes were examined. A total

of 50 metaphase spreads were examined in our MT1 line that was used for clonal mixing.

Following cytogenetic karyotyping, normal and abnormal iPSC clones were generated at

specific ratios.

2.2.2 Illumina and Axiom™ Biobank Plus Genotyping

Genotyping was also performed on two platforms by using Illumina HumanCore Arrays and

UK Biobank Axiom array (Catalogue number: 902186). For Illumina Genotyping, total DNA

was extracted using the QIAamp DNA Mini Kit (Qiagen) according to the manufacturer’s

45

instructions. Cell pellets were resuspended in 200 µL PBS (Life Technologies). Samples were

treated with Proteinase K for 10 min at 56oC. DNA was eluted in nuclease-free water. DNA

concentration and quality were initially determined by spectrophotometry prior to picogreen

assessment. For UK Biobank Axiom genotyping, 200ng genomic DNA in 20 ul (10 ng/ul) were

used to run on the SNP 6.0 axiom chip.

2.2.3 Array Processing and Virtual Karyotype Analysis

Genomic DNA was hybridized to Infinium HumanCore Arrays (Illumina). The GenCall (GC)

scores as calculated by Illumina’s GenomeStudio were used as a preliminary quality control

metric. GC scores range between 0 and 1 and reflect the reliability of each genotype call (scores

>0.7 indicate high quality genotypes and SNPs above this level were used for further analysis).

Probes passing the initial quality control (QC) were used for copy number variant (CNV)

assessment. The B allele frequency (BAF) and the log R ratio (LRR) were extracted from

GenomeStudio. As explained by Illumina, “the BAF represents a proportion of hybridised

sample that carries the B allele as designated by the Infinium Assay”. This thus means that in

a normal cell line, each locus AA, AB or BB would have a respective BAF of 0.0, 0.5 and 1.

Other values would indicate an abnormal copy number. Similarly, the LRR represents a logged

ratio of “observed probe intensity to expected intensity”. Hence, a deviation from zero

corresponds to a change in copy number. Matching fibroblast and iPSC genotypes were

assessed for discordant heterozygote and discordant homozygote SNP call rates. CNV analysis

was performed using PennCNV (Wang et al., 2007) , BCFtools CNV (Danecek et al., 2016)

and QuantiSNP (Colella et al., 2007). Assessment of clonality was performed using the Clonal

Heterogeneity Analysis Tool (CHAT) (Li and Li, 2014) in the R statistical environment. In

CHAT, fibroblast and corresponding iPSC-derived samples were used as “tumour-normal”

pairs.

2.3 Results

Traditional cytogenetic methods reported a normal male fibroblast cell line with karyotype 46,

XY as shown in Figure 2.1A. The karyotype of cell line BCL6 7+5 is

45,X,Y,der(1)t(1;15)(p34.1;q15),t(4;8)(p14;q24),der(11)t(1;11)(p34;q23),t(15;22)(p11.2;q13)

(Figure 2.1B). The chromosome abnormalities observed included a derivative chromosome 1

that resulted from a translocation between the long arms of chromosomes 1 and 15, which

46

resulted in gain of 15q from band 15q15 to 15qter. Two translocations were also observed: one

between the short arm of chromosome 4 and the long arm of chromosome 8 and the other

between the short arm of chromosome 15 and the long arm of chromosome 22. These appeared

to be balanced cytogenetically; however, small losses or gains of material are outside the limit

of detection of conventional chromosome analysis. There was a derivative chromosome 11 that

resulted from a translocation between the short arm of chromosome 1 and the long arm of

chromosome 11. This resulted in loss of 11q from band 11q23 to 11qter.

Figure 2.1: G-banding and FISH analysis: A) Normal male fibroblast cell line BCL6 with karyotype 46, XY. B) Karyotype of cell line BCL6 7+5 with karyotype 45,X,-Y,der(1)t(1;15)(p34.1;q15),t(4;8)(p14;q24),der(11)t(1;11)(p34;q23),t(15;22)(p11.2;q13). C) Normal Cell line whole chromosome painting (WCP) showing Chr11 positive with spectrum orange and Chr15 positive with spectrum green. D) Chromosome painting, WCP#11 (orange) and WCP#15 (green) showing duplication at Chromosome 15 and deletion at chromosome 11.

FISH analysis using whole chromosome paints (WCP) showed normal chromosome 11

showing positive WCP11 in spectrum orange, whereas the derivative of chromosome 11

showed a portion of the chromosome that is positive for WCP11, with the remaining material

A B

C D

47

negative. Chromosome 15 showed positive WCP15 in spectrum green and the derivative

chromosome 15 involved in the translocation with chromosome 22 with a portion of the

chromosome being negative for the WCP15 and at last derivative chromosome 1 showing

positivity for the WCP15. Normal cell line WCP showing chromosomes 11 positive for WCP

11 in spectrum orange and chromosome 15 positive for WCP 15 in spectrum green as shown

in Figure 2.1C. Whole chromosome paints (WCP) 11 (spectrum orange) and 15 (spectrum

green) shown in Figure 2.1D. The derivative chromosome 1 is positive for WCP 15, consistent

with the conventional chromosome analysis and virtual karyotyping which showed a gain of

15q from band 15q15. The derivative chromosome 11 showed partial hybridization with WCP

11, with the distal q arm negative. This is also consistent with the conventional cytogenetics

and virtual karyotyping, which showed deletion of 11q from 11q23.

Further CNVs were assessed by using three programs: PennCNV, QuantiSNP, BCFtools CNV

and aneuploid genome proportion was calculated using CHAT. In these CNV readings, the

regions of the genome that are aberrant in terms of copy numbers were identified by using two

different measurements: BAF and LRR. Briefly, normal BAF values are 0, 0.5 or 1 and any

variation from these numbers will indicate abnormalities in copy numbers. Similarly, the

normal LRR value is 0 and any variation from 0 indicates an abnormality in copy numbers. To

notify the aberration, we selected the CNV across the sample by 10 contiguous SNPs or 1mb

spanning length. The CNV events recorded in all the samples are shown in Figure 2.2 (where

(-) no aberrations and (+) aberrations) with the change in genomic proportion and CNV event

in all samples shown in Supplementary Table 2.1 with genomic proportion shown in

Supplementary Table 2.2.

48

Figure 2.2: Comparison table of CNV events detected by various tools with aneuploid genome proportion (AGP). Pie chart represent the percentage of aberrations in the sample (0,10,15,20,25,30,35,40,45,50 and 100).

The results showed both variations of BAF and LRR in chromosome 1, 11and 15. The same

genomic aberrations were detected by both traditional methods and with the genotyping arrays.

Further BAF/LRR plots were drawn to show the higher resolution of CNV events at which

position these events were reported as shown in Figure 2.3(All samples BAF/LRR Plots are in

Supplementary Figures 2.1). From our workflow, we are able to detect the aberrations in cell

lines with in at least 25% abnormalities.

MBE-01878

MBE-01878-CL6-7+5

QuantiSNP PennCNV BCFtools CHAT (AGP)

---- ---- ----- 0.05

---- ---- ----- 0.75

-+-- ---- ----- 0.23

---- ---- ----- 0.99

+-+- ---+ -+--- 0.28

--+- ---+ -+--- 0.3

+-++ --++ ++-++ 0.37

---- --++ -+--- 0.41

---- --++ -+-+- 0.49

++++ -+++ +++++ 0.74

++++ ++++ +++++ 0.97

49

Figure 2.3: BAF/LRR Plots A) Cell line with 0% genomic aberration (no CNV events). B) Cell line with 100% genomic aberrations (including CNV events - duplication at chromosome 1, deletion at chromosome 11 and duplication at chromosome 15).

This data suggest that the assay is at least as sensitive than a G-banding to detect single cell

variability within samples. It thus establishes that this assay is reliable for testing of

chromosomal abnormalities in cells. Hence, we used CNV analysis for the assessment of

chromosomal stability of hPSCs using established lines.

UK Biobank genotype data were analysed by using Axiom Analysis suite (Inc;, 2019) where

raw CEL format file were used to generate the PennCNV formatted output files for all samples

by using Axion Long Format Export Tools (Inc, 2019) and CNV events are detected by using

PennCNV. Furthermore, we have also compared the CNV events on the same samples which

are genotyped by using different platforms (Illumina and UK Biobank) for detecting genomic

aberrations are shown in Supplementary Table 2.3.

A B

50

2.4 Discussion We demonstrated that a low-resolution CNV array-based karyotyping offers an economical

and robust sampling method, both in the number of cells studied and genomic resolution,

compared with standard cytogenetic karyotyping genomic resolution for aberration detection.

Our data also showed that this method provides sufficient resolution to detect changes not

found through G-banding, which is regularly used for assessing genetic stability in hPSCs. In

Australia, these arrays are approximately one-third to one-quarter of the price of a standard

karyotype analysis. Together with the higher resolution than G-banding assay, this method

provides a clear advantage as it allows for a more economical and more efficient screening of

genetic rearrangements in hPSCs. This could be beneficial for allocating financial resources

for large-scale generation and maintenance of large repositories of iPSCs, as currently initiated

by consortia worldwide.

The virtual karyotype described has the advantage of generating more in-depth data than the

current mainstream assay for checking hPSC genetic stability (G-banding); it allows for DNA

fingerprinting of samples, thus permitting direct coding and tracing of samples; and it allows

population structure/outlier detection based on genome-wide profile.

There are, however, some important caveats to our findings. First, balanced translocations are

not readily identifiable using SNP-based virtual karyotyping. Second, despite DNA being

extracted from many more cells than could be visualized by cytogenetic karyotyping, we

hypothesise that virtual karyotyping would only be robust after a set number of passages,

assuming that most chromosomal changes/aberrations occur during reprogramming.

hPSCs must be characterized prior to using them in vitro or in vivo to understand fundamental

mechanisms or model disease. The field is moving from boutique experiments towards larger

scale studies, involving hundreds of cell lines. In this context, it is obvious that standard

characterization of hPSCs will have to evolve from time-consuming assessment of cells by

PCR, immunochemistry or even teratoma formation, to more high-throughput economical

assays. Such an approach already exists for some aspects of characterizing hPSCs. In particular,

pluripotency can now be assessed using high-throughput embryoid body assay TaqMan

scorecard (TaqMan hPSC Scorecard™ assay, Life Tech) in place of standard

immunochemistry for markers of the three germ layers. It is clear from our data that virtual

karyotyping with low-density SNP array is a natural complement to this approach. With

51

ongoing reductions in costs, the eventual incorporation of next-generation sequencing into the

large-scale profiling of iPSCs will only improve the resolution of CNV assessment. In

summary, we have confirmed that virtual karyotyping on a low-density SNP array is a good

addition to the evolving panels of assays for the economical and high-throughput assessment

of iPSCs.

52

Chapter 3 DNA methylation Age

3.1 Introduction

Many eye diseases manifest due to inherited degeneration of the retinal pigment epithelium;

despite the identification of specific mutations causing the majority of inherited retinal

dystrophies, all these conditions are currently untreatable (De Roach et al., 2013). Genome-

wide association studies (Sheffield and Stone, 2011) have identified genetic causes of ocular

diseases but genetic variants remain poorly understood.

Epigenetic variation refers to genetic modifications that are not due to a change in the DNA

sequence (Petronis, 2010) this can be influenced by environment interactions including DNA

methylation, post-translational histone modifications and chromatin remodelling (Gibney and

Nolan, 2010). DNA methylation (cytosine-5 methylation within CpG dinucleotide) is a

common and well-studied epigenetic mechanism because it is highly variable across different

cell types. All tissues of an organism are affected by aging and the accumulated cellular

changes (Garinis et al., 2008). Recent studies showed that age can predict DNA methylation

levels in specific tissues (Bocklandt et al., 2011, Hannum et al., 2013). A biomarker of aging,

also referred to as the epigenetic clock, DNA methylation levels can be used to calculate the

DNA methylation age (DNAm age) (Horvath, 2013).

The aim of this study was to determine the DNAm age of the optic nerve, RPE/choroid and

neurosensory retina tissue from the eye and peripheral whole blood from the same individual.

In addition, we also compared chronological age and DNAm age between ocular tissue and

blood.

3.2 Methods

3.2.1 DNAm data collection and processing

Whole blood from the subclavian vein and whole eyes were obtained post-mortem. Specimens

were taken from eight donors (all male) with no known ophthalmic disease. Blood samples

were collected in anticoagulant (EDTA)-containing tubes. All ocular tissue was collected and

53

stored within 12 hours’ post-mortem. The vitreous body was discarded by circumferential pars

plana incision. The neurosensory retina was removed and the RPE and choroid were dissected

from the scleral wall. Dura matter was separated from the anterior segment of the optic nerve

before storage. Dissected ocular tissue was stored in QIAGEN Allprotect Tissue Reagent

(QIAGEN, Hiden, Germany) and DNA extraction was subsequently performed using the

QIAGEN DNeasy Blood & Tissue Kit (QIAGEN). Following bisulfite conversion using the

Methyl Easy bisulphite modification kit (Human Genetic Signatures, Sydney, NSW,

Australia), samples were hybridized to Illumina Infinium HumanMethylation450 (Illumina

Inc, San Diego, CA, USA) BeadChips (HM450K) according to the manufacturer's protocols.

Array processing and Beadstudio analysis was conducted through the Australian Genome

Research Facility (Melbourne, VIC, Australia). This study was approved by the human

research ethics committee of the University of Western Australia (RA/4/1/4805).

3.2.2 DNAm age estimation by using epigenetic clock

The epigenetic clock is a prediction method used for calculating age based on DNA

methylation levels of 353 CpGs dinucleotides. DNAm (Predicted age) can be correlated with

chronological age in various tissue and cell types (Horvath et al., 2015) using a DNAm age

calculator tool (Horvath, 2013). It defines DNAm age using a penalized regression model on

Illumina 450K and 27K platform. The beta values were obtained from the raw data in IDAT

format from 450k array using minfi (R package) (Aryee et al., 2014). Normalization and pre-

processing were included in the tool to improve the accuracy, specifically the median error

(Horvath, 2013). Meta-Analysis R function in the WGCNA R package (Langfelder et al., 2013)

was used to measure the pure age effects irrespective of the tissue type. We also predicted the

epigenetic age by another method developed by Hannum et al (2013) (Hannum et al., 2013).

This method is based on the 71 CpGs dinucleotides and uses an elastic net regression model

on the DNA methylation levels from the whole blood (Hannum et al., 2013). We used the

online version of the DNA methylation age calculator (Horvath, 2017) for calculating

epigenetic age based on Hannum et al (2013) using advanced analysis for blood data. An online

version of the DNA methylation age calculator is also available at (Horvath, 2017). The DNAm

age was also calculated by using Enpred and BLUpred predictor developed by Qzhang (Zhang

et al., 2019). Source code for reproducing results and figures of the paper is available on GitHub

(https://github.com/hewittlab/Ocular_mDNA_age).

54

3.3 Results

The mean (SD) age of the eight male donors was 60.6 (11.3) years. In total, we analysed 29

samples. We studied 485,512 CpGs that were present on the Illumina 450K platform. The

Horvath method found a linear relationship between the DNAm age and chronological age in

blood and ocular tissues, with a correlation of 0.37 (p<0.048) and standard error of 18 years

(Figure 3.1A) whereas the Hannum method predicted with a correlation of 0.32 (p<0.091) and

standard error of 26 years (Figure 3.2). The correlations between the chronological and DNA

methylation age are reported in Table 3.1. There was a significant association between the

measures. The correlation coefficient was strongest in neurosensory retina (r2 = 0.93,

p<0.00081).

55

Figure 3.1. (A)Chronological age versus DNAm age across all sample data set (r = 0.37, p< 0.048 and standard error = 18 years). Mean epigenetic age difference (y- axis) versus different tissues in (B) respectively. The bar plots also report a p-value from a non-parametric group comparison test (Kruskal Wallis test). Each bar plot depicts the mean value and 1 standard error. (C) Post mortem interval versus DNAm age across all sample data set (r = -0.15, p<0.44). Table 3.1: chronological age correlation with epigenetic age acceleration

Tissue Type Correlation P-value

Peripheral Blood 0.86 0.013

Optic Nerve 0.87 0.011

RPE/Choroid 0.83 0.021

Neurosensory Retina 0.93 0.00081

10 20 30 40 50 60 70 80

1020

3040

5060

7080

Horvath cor=0.37, p=0.048

Chronological Age

DNAm

Age H

orvath

Blood

Neurosensory Retina

Optic Nerve

RPE/Choriod

Blood Optic Nerve RPE/Choroid Neurosensory Retina

p = 3e−05

Tissues

Differe

nce in

epigen

etic an

d chro

nologi

cal ag

e(year

s)−60

−40−20

020

2 4 6 8 10 12

−60−40

−200

20

C cor=−0.15, p=0.44

Post Mortem Interval

Age A

ccelera

tion Di

fferenc

e

b

a

c

56

Figure 3.2: Hannum Method (A) Chronological age versus DNAm age across all sample data set (r = 0.32, p< 0.091 and standard error = 26 years). Mean epigenetic age difference (y- axis) versus different tissues in (B) respectively. The bar plots also report a p-value from a non-parametric group comparison test (Kruskal Wallis test). Each bar plot depicts the mean value and 1 standard error. (C) Post mortem interval versus DNAm age across all sample data set (r = 0.031, p<0.87).

Surprisingly, in neurosensory retinal tissue we found a significant age difference of 44.4 years

between the DNAm age (mean = 16.1 years) and chronological age (mean = 60.6 years) (Figure

3.1B). The mean DNAm age of the RPE/Choroid tissue was 31.8 years, which was also

significantly different from the mean chronological age of 60.4 years, and the age difference

of optic nerve tissue was 11.3 years. We did not find any significant age difference in the blood

tissue as compared to the DNAm age (mean = 58.1 years) and chronological age (mean = 59.8

years). We tested the DNAm with all causes of mortality across the samples. Respiratory failure

had the maximum age acceleration difference (mean = 38 years with overall p<0.47). In

10 20 30 40 50 60 70 80

2040

6080

100120

140

Hannum cor=0.32, p=0.091

Chronological Age

DNAm

Age H

annum

Blood

Neurosensory Retina

Optic Nerve

RPE/Choriod

Blood Optic Nerve RPE/Choroid Neurosensory Retina

p = 2.5e−05

Tissues

Differe

nce in e

pigeneti

c and ch

ronolog

ical age

(years)

−60−40

−200

2040

60

2 4 6 8 10 12

−60−40

−200

2040

60

C cor=0.031, p=0.87

Post Mortem Interval

Age Ac

celerati

on Diff

erence

b

a

c

57

addition, we found no significant relationship between the measure of age acceleration

difference and post-mortem interval (p<0.44) (Figure 3.1C).

Figure 3.3: Qzhang Method (a) Chronological age versus DNAm age across all sample data set by using Enpred Algorithm and mean epigenetic age difference (y- axis) versus different tissues in respectively. The bar plots also report a p-value from a non-parametric group comparison. Each bar plot depicts the mean value and 1 standard error. (b) Chronological age versus DNAm age across all sample data set by using BLUpred and mean epigenetic age difference (y- axis) versus different tissues in respectively. The bar plots also report a p-value from a non-parametric group comparison. Each bar plot depicts the mean value and 1 standard error. The DNAm age was also predicted by Qzhang method. Best linear unbiased prediction and

Elastic net prediction model were used, and a similar pattern was also recognised in our dataset.

By the following predictor, we have found the mean age difference of 63 years (r= 0.25, p<

0.0000001667) (Figure 3.3a) in the neurosensory retina by Enpred predictor whereas 51 years

(r=72.9, p< 0.0000009404) (Figure 3.3b) mean difference were found by BLUpred method.

58

3.4 Discussion

Tissue aging is a complex and dynamic process that compromises biological function and

increases susceptibility to disease and death. The aging process is still poorly understood

(Berdasco and Esteller, 2012) and some body tissues may be more prone to the effects of aging

than others. It is therefore valuable to understand the epigenetic contribution to age-dependent

changes in gene expression. During aging and the pathogenesis of ocular disease, changes in

chromatin lead to the dysregulation of gene expression. Retinal diseases such as glaucoma

(McDonnell et al., 2014) , an optic neuropathy that occurs due to loss of RGC, are influenced

by genetic mutation and aging (Leske et al., 1995). Multiple epigenome-wide studies (EWASs)

have identified chronological aging-associated changes in the DNA methylome (Rakyan et al.,

2010, Marttila et al., 2015, Yuan et al., 2015). Age-related alteration in the DNAm has been

reported previously (Hernandez et al., 2011, Maegawa et al., 2010, Gronniger et al., 2010).

DNAm age has been validated across different complex tissues and organs including brain,

breast, kidney, liver and lung (Horvath, 2013) and can be compared to chronological age.

Individual cells acquire a specialized gene expression profile specific to their tissue type and

function and further changes in the chromatin landscape are required in various tissue types

such as the lens and retina (Otteson, 2011). Understanding the tissue-specific differential

methylation pattern can help tease out underlying mechanisms of tissue-specific processes

(Ollikainen et al., 2010).

Age-related macular degeneration (AMD) causes irreversible central vision loss and its

pathogenesis involves a complex interaction of both genetic and environmental risk factors

(Ding et al., 2009). The role of epigenetics in AMD is poorly understood. Alterations in

epigenetic pattern may be the link between environmental factors and gene expression for the

development of disease. Our data shows that DNAm age is highly correlated (>0.80) with

chronological age in ocular tissue but that the DNAm age of peripheral blood differs to that of

ocular tissue. The main limitation of this work is that it is not yet completely understood what

biomarker of aging is measured and what represents noise and epigenetic drift (Teschendorff

et al., 2013).

59

Chapter 4 CRISPR Screening for uveal melanoma

4.1 Introduction

Uveal melanoma (UM) (OMIM #155720) is a rare form of a cancer of the eyes and is the most

common intraocular malignancy in adults. It arises primarily from melanocytes in the choroid

(80%), iris (5%) or the ciliary body (15%) (which is collectively known as uveal tract, one of

the inner layers of eye) (Chang et al., 1998). In the Unites States, the incidence of UM is 5.1

people in 1 million (Singh et al., 2011) compared to more than 8 people in 1 million in Norway

and Denmark and 2 people in 1 million in Spain and Italy (Virgili et al., 2007). The most

common symptoms of UM are blurred vision, floaters, photopsia, metamorphopsia, a dark spot

on the iris, change in the shape and size of pupil and change in the position of the eyeball in

the eye socket. Sun exposure is a risk factor (Singh et al., 2004). UM can be detected using slit

lamp biomicroscopy and indirect ophthalmoscopy and ancillary diagnostic testing such as

ocular ultrasonography (Jovanovic et al., 2013a). Most UM (small- or medium-sized tumours)

are treated by brachytherapy, teletherapy and dosimetry (Stannard et al., 2013). Regardless of

the treatment modality, nearly 45% of patients die from metastases (Kujala et al., 2003), with

90% of metastases involving the liver (Bedikian et al., 1995). Metastases also occur

sporadically in lung and bone (Gragoudas, 2006, Collaborative Ocular Melanoma Study, 2001,

Diener-West et al., 2004, Landreville et al., 2008).

The molecular profile of UM comprises various genomic aberrations and somatic gene

alterations. The most common genomic aberrations are monosomy 3, 1p loss, 1q gain, 6q loss,

6p gain, 8p loss and 8q gain (Damato et al., 2010). Monosomy 3 is observed in nearly 50% of

tumours and is associated with metastatic disease (Thomas et al., 2012). There is no proven

standard care for the patients who developed metastatic disease. Dacarbazine, a

chemotherapeutic option for cutaneous melanoma but has limited activity in UM (Dummer et

al., 2012, Carvajal et al., 2014, Robert et al., 2013). Other chemotherapy including

temozolomide, cisplatin, treosulfan, fotemustine have been investigated in UM but results have

been discouraging (Pereira et al., 2013, Spagnolo et al., 2013, Augsburger et al., 2009).

Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) is a feature of the

bacterial immune system. The CRISPR-associated protein 9 (Cas9) derived from the CRISPR

60

type II system has emerged as an effective genomic engineering tool. As an RNA-directed

DNA endonuclease, Cas9 can be readily programmed to target and modify a specific site with

a slight shift in the directed RNA sequence, and this technique can be used for genome editing,

control, biochemical and structural studies (Wang et al., 2016). Genomic screening for a

phenotype of interest can be undertaken using custom- or ready-made guided RNA libraries

(Joung et al., 2017b). We used a Genome-Scale CRISPR Knockout GeCKOv2 (lentiCRISPR

v2.0 and lentiGuide-Puro) pooled library (provided by Feng Zhang, Catalogue# 1000000048)

that targets around 19050 genes throughout the whole human genome, with a total of 1233411

sgRNAs pooled together (6 gRNAs per targeted gene, control sgRNAs designed not to target

the genome, and sgRNAs targeting miRNA). This pooled knockout screen experiment is akin

to performing tens of thousands of small experiments simultaneously, creating thousands of

knockouts at once. The cells with key genes that are knocked out will die with these related

guide RNAs depleted from the population. By identifying the missing guide RNAs, we can

determine which genes are crucial for the initiation and proliferation of UM.

The lentiCRISPRv2 library was amplified and packaged with lentivirus. The UM cell line

OCM-1 was transduced with lentiCRISPRv2 library and puromycin was added to maintain a

low Multiplicity of infection (MOI) to ensure each cell received only one sgRNA. A large

volume of transfected OCM-1 cells was maintained and cells were collected at two time points:

once immediately after transduction, and finally after 12 passages. Genomic DNA was

extracted, and the presence of the guide RNA library was identified using next-generation deep

sequencing. The web-based tools CRISPRAnalyzeR and gProfiler were then employed for

bioinformatics and functional analysis. The aim of the following study is to identify the

essential genes essential for the proliferation and survival of uveal melanoma via a forward

genetic genome‐wide CRISPR/Cas9 screening approach.

4.2 Methods

CRISPR Screening experiment was performed by “Fan Li” and full details


4.2.1 Sequencing

The UM (OCM1) cell line was used in three biological replicates and library preparation and

sequencing on NOVOSEQTM 6000 Sequencing system were performed at the Ramaciotti

61

Centre for Genomics - University of New South Wales, Australia. The S4 flow cell type with

2X150 bp output per flow cell configuration was used. Later, bcl2fastq2 Conversion Software

v2.20 was used to generate the Fastq files.

4.2.2 Quality Control

The Fastq files were obtained using the NovoSeq platform. Initially the quality of the Fastq

files was assessed by the FASTQC and it was confirmed that all the reads had phred score >

25. FASTQC is a quality control tool used to check the quality of raw sequences generated by

the various sequencing platforms and to provide an interactive report including basic statistics

(reads quality score), adapter contents and many more (Andrews, 2010).

4.2.3 Adapter/Quality Trimming

The adapter content was removed using the BBDUK module from BBTOOLS

(https://jgi.doe.gov/data-and-tools/bbtools/). In BBDUK, DUK stands for decontamination

using k-mers. It was developed to perform quality filter trimming and remove adapter content

via different parameters including GC filtering, k-mers matching and length filtering. In total,

not more than 9.78% of the bad reads were removed along with the adapter contents from the

Fastq files.

4.2.4 CRISPR Screening Analysis

CRISPR/Cas9 Knockout Screening analysis was done using CRISPRAnalyzeR (Winter et al.,

2017). CRISPRAnalyzeR is a web-based application featuring eight different algorithms to

identify the hit calling including DESeq2 (Love et al., 2014), MAGeCK (Li et al., 2014), EdgeR

(Robinson et al., 2010), sgRSEA (Noh, 2015), Mann-Whitney (Hart, 2001),ScreenBEAM (Yu

et al., 2016) and BAGEL (Hart and Moffat, 2016). In the following analysis, five different

programs were used with default parameters to identify the ranked-based possible hits:

● Wilcox - the Wilcox method, a test for comparing two different groups on the fold

change for each gene, was implemented on the two-sided Mann-Whitney-U test (Hart,

2001). Later, to correct for multiple testing the corrected P-values were calculated using

the Benjamini-Hochberg method.

• MAGeCK - MAGeCK is a model-based CRISPR/Cas9 knockout method to identify

the negative and positive selected sgRNAs and genes across different experimental

62

conditions. The program uses median-normalised method to adjust the effective size of

libraries and read-count distribution and then uses the mean variance estimation model

and negative binomial to determine the significant sgRNAs and genes across the

different conditions. Afterwards, a robust ranking aggregation (RRA) algorithm(Kolde

et al., 2012) was used to rank the significant genes in negative and positive selection,

respectively (Li et al., 2014).

• DESeq2 - DESeq2 is a differential expression analysis method implemented in R

programming language as a package. It uses read count of all sgRNAs/genes for given

different experimental conditions to estimate the differential effects based on a negative

binomial model. Enrichment and depletion of the sgRNAs/genes were estimated on

log2 fold change between the two different groups using the Wald test (Love et al.,

2014).

• EdgeR - EdgeR is another method used for differential expression analysis and it is

also implemented in R programming language. In this program, read count data were

modelled using an over-dispersed Poisson model. Gene dispersion was estimated by

conditional maximum likelihood method and an adapted Fisher’s exact test was used

for assessing the differential expression for sgRNAs/genes (Robinson et al., 2010).

• sgRSEA - single-guide RNA Set Enrichment Analysis (sgRSEA is another program

that is also implemented in the R programming language for enrichment analysis of

CRISPR/Cas9 Knockout Screen Data. It ranks the sgRNA on the signal-to-noise ratio

and calculates sgRNA ranking by enrichment score (which is one-sided Kolmogorov

Smirnov statistic method) and reflects the enriched and depleted sgRNA, respectively.

4.2.5 Functional Enrichment Analysis

The functional analysis was done using gProfiler (Reimand et al., 2007) a web-based tool for

functional profiling significant genes from high throughput data. A list of significant genes was

supplied and statistically significant gene ontology pathways were identified. Cytoscape

(Shannon et al., 2003) and Enrichment map (Isserlin et al., 2014) were used for visualization

of statistically significant pathways as mentioned in the following paper (Reimand et al., 2019).

4.3 Results

The quality-trimmed Fastq files from two different conditions i.e., Passage 0 (P0) and Passage

12 (P12) from UM (OCM1) cell lines were mapped to the Human_GeCKOv2_Library

63

(Sanjana et al., 2014) using MAGeCK, and the mapping summary for all samples are given

below in Table 4.1 with Gini index and the distribution for all the samples shown in Figure 4b

and 4c. Table 4.1: Summary of comparisons

Sample id Reads Mapped Percentage Total

sgRNA

Zero

Counts

Gini

Index4

1 UMP0_1 387529989 306261122 79% 119461 286 0.06

2 UMP0_2 427247327 311118165 73% 119461 276 0.06

3 UMP0_3 359356334 286941293 80% 119461 69 0.06

4 UMP12_1 416965737 322517953 77% 119461 6309 0.17

5 UMP12_2 679106595 524070831 77% 119461 3320 0.12

6 UMP12_3 417905201 309265060 74% 119461 8800 0.21

Figure 4.1: a) Scatterplot showing distribution of sgRNA read count in UMP0 and UMP12. b) Boxplot with outliers for all samples. c) Distribution plot of all samples with normalised read count.

4 The Gini Index of the read count distribution indicates more evenness of the count distribution.

a b

c

64

The read count data were further accessed and sgRNAs with less than 100 read counts were

calculated in P0 and P12. In our data, 0.45% of P0 reads had less than 100 read counts whereas

3.9% of P12 reads had less than 100 read counts. The number of sgRNAs with less than 10

read counts in P12 compared to P0 and only 0.16% sgRNAs were identified (Figure 4.1a).

From the initial screening, the sgRNAs with read counts less than 100 in P12 compared to P0

with read counts more than 100 were used and 56 genes were identified with low expression

in 3 or more sgRNAs in P12 and considered as possible hits (Table 4.2).

Table 4.2 List of genes with missing 3 or more sgRNA in UMP12 as compared to UMP0.

Gene Number of Missing sgRNA

Gene Number of Missing sgRNA

CDIPT 4 PHAX 3 DUX2 4 POLG2 3

GOLGA8B 4 POLR2I 3 hsa-mir-3929 4 PRPF40A 3

OR5AR1 4 PSMA3 3 RPL35A 4 RAD9A 3

CIT 3 RBM8A 3 CLPB 3 RPL13 3 COPA 3 RPL21 3

CWC22 3 RPL23 3 DDX56 3 RPL8 3 DHX33 3 RPS27 3 EIF2S1 3 S100A10 3

EIF6 3 SART3 3 GAPDH 3 SCAF11 3 H2AFZ 3 SLC25A3 3

hsa-mir-4459 3 SNRPD3 3 hsa-mir-566 3 SNRPF 3 hsa-mir-619 3 SON 3 hsa-mir-663a 3 SOX10 3 hsa-mir-8078 3 SRA1 3

LONP1 3 SRSF1 3 MAD2L1 3 SYNCRIP 3 MED14 3 UBL5 3 MEF2C 3 VARS2 3 MESP1 3 WDR43 3

MRPS10 3 WDR70 3 OGDH 3 WDR74 3

65

The read count data were used to identify significant hits through different algorithm

implemented on a web-based platform CRISPRAnalyzeR. From our analysis in negative

selection, we found 48 significant genes from MAGeCK (depleted P-value<0.05)

(Supplementary Table 4.1) (Figure 4.2a), 25 significant genes from sgRSEA(adjusted P-

value<0.05) (Supplementary Table 4.2) (Figure 4.2b), 77 significant genes from DESeq2

(adjusted P-value<0.001) (Supplementary Table 4.3) (Figure 4.2c) and 2164 significant genes

from EdgeR (adjusted P-value< 0.05) (Supplementary Table 4.4) (Figure 4.2d), respectively.

Figure4.2: a) Scatterplot showing significant gene using MAGeCK, b) Scatterplot showing significant gene using sgRSEA, c) Scatterplot showing significant gene using DESeq2, d) Scatterplot showing significant gene using EdgeR.

a b

dc

66

A number of overlapping genes were identified (Figure 4.3a). The final selection for

identifying hits was done using MAGeCK and DESeq2 algorithms and 15 genes were

identified as hit candidates (Figure 4.3b). Almost all sgRNA associated with genes had low

expression in Passage 12 compared to the Passage 0 (Table 4.3) (Supplementary Figure 4.1).

Figure 4.3 a) Venn diagram showing the overlap genes between the different algorithms. b) overlapping genes with final selected algorithms MAGeCK and DESeq2. Table 4.3: List of final gene subset select based on MAGeCK and DESeq2 algorithms.

Genes Total sgRNA in P0 and P12 Number of sgRNA with low

expression in P12

WASH1 6 6

SLC3A2 6 6

ABT1 6 6

NDUFB10 6 6

RPL35 6 5

COQ2 6 6

LSM11 6 6

KATNB1 6 6

UBL5 6 6

MRPL22 6 4

HIST2H4A 6 6

HTRA2 6 6

SPDYE5 4 4

CCNA2 6 6

POLR3K 6 6

a b

67

The final selected subset genes were used for the functional analysis and significant gene

ontology pathways were identified. The detailed pathways interactions are shown in Figure

4.4.

Figure 4.4: Pathway analysis showing statistically significant pathways by using gProfiler and visualisation using enrichment map.

Functional analysis suggested the major pathways with the most intersects were: metabolic

process, cellular process, primary metabolic process, cellular metabolic process, biological

process and organic substance metabolic process (Supplementary Table 4.5).

4.4 Discussion

Cancer is a malignant disease generally caused by genetic alterations from activation of

innate/acquired mutations and epigenetic alterations that lead to functional abnormality and

initiate the process of tumorigenesis (Hanahan and Weinberg, 2000). High-throughput

sequencing platforms can provide useful insight into understanding various mutated genes.

However, genes that are involved in the initial step of tumorigenesis remain elusive (Hanahan

and Weinberg, 2000). The CRISPR/Cas9 screening method for annotation of gene function has

uncovered various mutations that are involved in the origin of cancer or in the progression into

biological process

metabolism

cellular component organization

cellular metabolic process

rna metabolic process

cellular protein complex

ribonucleoprotein complex biogenesis

protein targeting

ribosome assembly

protein complex assembly

organic cyclic compound

primary metabolic process

small subunit processome

metabolic process

organic substance biosynthetic

organic substance metabolic

cellular process cellular nitrogen compound

cellular biosynthetic processgene expression

biosynthetic process

nitrogen compound transport

intracellular protein transport

cellular component biogenesis

cellular macromolecule metabolic

organelle organization

protein complex subunit

rna processingcellular component organization

ribosome biogenesis

68

metastasis (Luo, 2016). Several oncogenes have already been uncovered, e.g. KRAS and

PIK3CA are fatal hits in the colorectal cancer cell line DLD-1 and HCT116, BCR and ABL are

fatal hits in the chronic myelogenous leukaemia cell line KBM7 that harbours BCR-ABL

translocation (Wang et al., 2014). UM is an intraocular malignant tumour with a high potential

for lethal metastasis and a low mutational load compared to cutaneous melanoma (Luo, 2016).

However, currently there are no effective therapies for treating UM and most clinical trials are

only Phase I or II (Yang et al., 2018). Hence, it is crucial to set up genome-wide CRISPR/Cas9

loss-of-function screening to identifying the novel genes that underpin the development,

growth, and survival of the cancer cell lines (Shalem et al., 2014).

In screening loss-of-function analysis, we found 15 genes as possible hits. One of the possible

target genes SLC3A2 (Solute carrier family 3 member 2; also known as CD98hc) is a

transmembrane protein and exists as a heavy chain of heterodimer with large neutral amino

acid transporter LAT1 in cells (Estrach et al., 2014). SLC3A2 overexpression widely occurs in

different cancer cells such as skin squamous cell carcinoma (Estrach et al., 2014), Gastric

cancer (Wang et al., 2017a) and osteosarcoma (Zhu et al., 2017). The consumption of L- type

transporters (LATs) including L-leucine into the cells, signalling and protein synthesis were

stimulating in mammalian targets. Different L-type amino acid transporters are active in

different stage of prostate cancer and are responsible for the cell growth (Wang et al., 2013).

Another possible target gene COQ2 (Coenzyme Q2, Polyprenyl Transferase) encodes for the

biosynthesis of COQ (ubiquinone), an electron and proton carrier in the mitochondrial

respiratory chain and a lipid-soluble antioxidant [RefSeq 2009]. A deficiency in coenzyme Q10

related to the COQ2 gene has been associated with retinopathy in some patients (Desbats et al.,

2016) and a homozygous mutation visual dysfunction trait related to rod-cone retinopathy has

also been observed (Quinzii et al., 2014). RPL35 and POLR3K are differentially expressed in

the breast cancer MCF-7 and MDA-MB-231 breast tumour cell lines and are associated with

the translation- and transcription-related molecular pathways (Satih et al., 2010). CCNA2

(CyclinA2) is a highly conserved member of the cyclin family and functions as a regulator in

the cell cycle [RefSeq 2016]. CCNA2 is identified as a prognosis biomarker for estrogen-

positive breast cancer and tamoxifen resistance (Gao et al., 2014).

69

Figure 4.5: Expression level (transcripts per million- tpm) of final selected gene on TCGA datasets.

We further assessed the tumour subgroup expression on our hit candidates using the online

portal ULCAN (Chandrashekar et al., 2017) to check the expression level on different stages

based on genomic data from The Cancer Genomic Atlas (TCGA) (Cancer Genome Atlas

Research et al., 2013) and found that all genes had relative fair expression (tpm) on the UM

genomic dataset as shown in Figure 4.5 whereas WASH1 gene expression was not found.

Overall, our selected candidate genes in the CRISPR/cas9 knockout screening analysis showed

lower expression in UM OCM_P12 as compared to the P0, suggesting these genes have a

crucial role in the cell growth and development of UM cell lines. However, further invitro and

in vivo analysis are required to validate our findings.

70

Chapter 5a RNA-Seq Analysis for Retinal neovascularization in miRNAs

5a.1 Introduction

Retinal neovascularization is a sight-threatening condition that occurs in ischemic retinopathies

including retinopathy of prematurity (ROP), central retinal vein occlusion (CRVO) and

proliferative retinopathy (PDR) (Vinekar et al., 2016, Jonas et al., 2017, Smith and Steel, 2015).

Angiogenesis influences the neovascularization in various conditions such as hypoxia,

ischemia or inflammation (Campochiaro et al., 2016). Pathological angiogenesis in the retina

has to be monitored carefully as it may lead to visual impairment (Bressler, 2009). Vascular

endothelial growth factor (VEGF) is a crucial regulator of retinal angiogenesis and impacts

endothelial cell migration and proliferation after hypoxia (Bressler, 2009). VEGF inhibitors

are effective in some but not all patients, indicating that other pathways might play a crucial

role as well. Moreover, VEGF inhibitors such as ranibizumab and aflibercept are very

expensive and are associated with some adverse effects (Ross et al., 2016). Therefore, new

treatments for retinal neovascularization are needed.

MicroRNAs (miRNAs) are a small class of endogenous non-coding RNA 21-25 nucleotide

sequences in length that act by complementary base pairing at the 3’ untranslated region (3

UTR) of targeted mRNA, resulting in translation repression or degradation of targeted mRNA

(McClelland and Kantharidis, 2014). Major biological processes such as cell growth, cell death,

development and cell differentiation are regulated by the miRNAs (Raghunath and Perumal,

2015). miRNAs are an important modulator of blood vessel formation and function. Various

retinal miRNAs such as miRNA-150 (miR-150), miRNA-126 (miR-126) and miRNA-155

(miR-155) have been shown to be dysregulated in animal models of retinal neovascularization,

although study results have been inconsistent (Liu et al., 2015, Bai et al., 2011, Yan et al.,

2015). miR-143 have been shown involve in regulating actin actin-related protein subunits

smooth muscle and TM cells (Li et al., 2017).

We performed next-generation sequencing (NGS) RNA-Seq in vivo and in vitro in an OIR

(Oxygen-induced retinopathy) Rat model to identify novel genes associated with retinal

neovascularization. miR-143, miR-126 and miR-150 were evaluated in human endothelial cells

(telomerase-immortalized microvascular endothelial cell [TIME], human dermal

microvascular endothelial cells [HMEC] and human umbilical vein endothelial cell [HUVEC])

to assess for any impact of dysregulation involved in retinal neovascularization.

71

5a.2 methods

In-vivo and In-vitro experiment for OIR models was performed “Jiang-Hui

Wang” and “Jinying Chen” and full details are in Supplementary Method 3.

5a.2.1 Sequencing

The libraries of cDNA from the biological samples were sequenced using an Illumina HiSeq-

2500 RNA-Seq platform as 100 bp single end at the Australian Genome Research Facility

(AGRF, Melbourne, VIC, Australia). HiSeq Control software (HCS) v2.268 and Real time

analysis (RTA) v1.18.66.3 were used. The Fastq data were generated using the Illumina

bcl2fastq 2.20.0.422 pipeline.

5a.2.2 Dataset We used two datasets:

● Two groups from rat retina including OIR scrambled RNA and OIR-miR143 with two

biological replicates were sequenced.

● Four groups of mRNA samples from human endothelial cells (telomerase-

immortalized microvascular endothelial cell [TIME], human dermal microvascular

endothelial cell [HMEC] and human umbilical vein endothelial cell [HUVEC])

including Control, miR126, miR143, miR150 with three biological replicates were

sequenced.

The raw data were deposited in NCBI's Gene Expression Omnibus (GEO) and are accessible

through the GEO accession number XXXXXXXX (Data submission in progress).

5a.2.3 Quality Control The raw Fastq files were obtained by the Illumina HiSeq 2500. Initially the quality of the Fastq

files were assessed by the FASTQC and it was confirmed that all the reads had phred score >

28. FASTQC is a quality control tool that generates interactive reports including basic

statistics, adapter content information and base quality score for reads in a Fastq file (Andrews,

2010).

5a.2.4 RNA-Sequencing Analysis

Kallisto (v0.46.0) is a new approach to quantify the gene expression abundance from an RNA-

Seq data. It employs the novel concept of pseudoalignment for determining the compatibility

of reads with target, with no requirement for alignment. The Kallisto pseudoalignment

72

approach finds potential transcripts using K-mers in the reads. In addition to pseudoalignment,

the expectation-maximum (EM) algorithm is performed on a likelihood based on equivalence

class. This allows the EM algorithm to scale in complexity to the number of equivalence classes

rather than number of reads. It builds the indices for quantification that are useful for assessing

the reliability of abundance estimates produced by the bootstrap samples (Bray et al., 2016).

Sleuth (Pimentel et al., 2017) is a program that implements a statistical algorithm for

differential expression analysis on transcript abundance which is quantified using Kallisto. It

has been designed for exploring differentially expressed data using the shiny application in

RStudio (Team, 2015) and ggplots (Wickham, 2016) an R package were used for generating

plots. Heatmaps were generated using ClustVis (Metsalu and Vilo, 2015).

5a.2.5 Network and Pathway Analysis

Ingenuity Pathway Analysis Software (IPA: (Ingenuity Systems, Redwood City, CA, USA,

http://www.ingenuity.com/) is a web-based application used for solving various biological

problems in system biology. Functional analysis of genes and their networks were used in the

analysis and the significance of any gene function in a network was denoted by p-value less

than 0.05. The right-tailed Fisher Exact Test was used to calculate the p-value.

5a.3 Results

The statistics of mapped reads from the RNA-Seq data are shown in Table 5a.1 for all the

samples and more than 70% of reads were mapped to the reference genome. OIR Scrambled

and OIR miR143 were mapped to the Rattus norvegicus genome assembly (rn6) whereas

Control and other miRNA samples were mapped to the Homosapiens genome assembly (hg38)

and transcript abundance was estimated with bootstrapping the samples 100 times. The

quantified transcript abundance was used for calculating differential gene expression across all

the conditions using Sleuth. Table 5a.1: Statistics of mapped reads from samples OIR Scrambled, OIR-miR143, miR126, miR143, miR150 and control summary table.

Sample Mapped Reads Total Reads Mapping Percentage Bootstraps

OIR Scrambled_1 12931393 18041628 71.68% 100

OIR Scrambled_2 13747121 18864406 72.87% 100

OIR miR143_1 13477391 17694969 76.17% 100

73

OIR miR143_2 12070325 16723655 72.18% 100

Control_2 19376283 20703074 93.59% 100

Control_3 18464997 19807112 93.22% 100

Control_4 19125654 20485604 93.36% 100

miR126_2 17182109 18561094 92.57% 100

miR126_3 16252186 17647104 92.1% 100

miR126_4 16390376 17607238 93.09% 100

miR143_2 17676654 18985807 93.1% 100

miR143_3 17802910 19086343 93.28% 100

miR143_4 16881020 17908706 94.26% 100

miR150_2 19324683 20806718 92.88% 100

miR150_3 18433277 19883336 92.71% 100

miR150_4 17616278 19099779 92.23% 100

Figure 5a.1: a) Correlation plot between the OIR Scrambled and OIR-miR143 sample. b) Heatmap showing the expression (tpm) of statistically significant (adjusted pValue<0.05) gene found by using Sleuth. c) Volcano plot highlighting statistically significant (adjusted pValue<0.05) between OIR Scrambled and OIR-miR143 sample.

74

Figure 5a.2: a) Correlation plot between Control and miR126 sample. b) Volcano plot highlighting statistically significant (adjusted pValue<0.05) between Control and miR126 sample along with highlighted genes that are related to angiogenesis. c) Heatmap showing the expression (tpm) of statistically significant (adjusted pValue<0.05) gene found by using Sleuth.


75


Overall, from our analysis we found 122 genes were differentially expressed (upregulated and

downregulated genes) in OIR Scrambled and OIR-miR143 based on adjusted p<0.05 as shown

in Figure 5a.1 (Supplementary table 5a.1), 1163 genes were differentially expressed

(upregulated and downregulated genes) in Control and miR126 based on adjusted p<0.01 as

shown in Figure 5a.2 (Supplementary table 5a.2), 445 genes were differentially expressed


shown in Figure 5a.3 (Supplementary table 5a.3) and 192 genes were differentially expressed


shown in Figure 5a.4 (Supplementary Table 5a.4).

76

Table 5a.2: List of top pathways from IPA in miR126, miR143, miR150 and OIR miR143. miRNA Diseases or Functions Annotation p-value # Molecules miR126 Development of vasculature 7.27E-17 149 miR126 Endothelial cell development 2.86E-09 60 miR126 Proliferation of endothelial cells 3.34E-07 50 miR126 Migration of endothelial cells 4.6E-14 68 miR126 Cell movement of endothelial cells 1.18E-13 71 miR126 Cell movement of endothelial cell lines 1.45E-07 24 miR126 Migration of endothelial cell lines 4.56E-07 20 miR126 Migration of vascular endothelial cells 0.00000138 30 miR126 Movement of vascular endothelial cells 0.00000581 31 miR126 Angiogenesis 9.7E-17 145 miR126 Vasculogenesis 7.4E-13 114 miR126 Morphogenesis of blood vessel 0.0000118 14 miR126 Formation of blood vessel 0.0000324 26 miR126 Development of endothelial tissue 9.86E-10 62 miR126 Formation of endothelial tube 0.0000166 14 miR126 Permeability of vascular system 0.000016 25 miR143 Development of vasculature 0.00000016 58 miR143 Endothelial cell development 0.000255 23 miR143 Cell proliferation of vascular endothelial cells 0.000863 13 miR143 Proliferation of endothelial cells 0.0018 19 miR143 Migration of vascular endothelial cells 0.00000968 16 miR143 Migration of endothelial cells 0.000277 22 miR143 Migration of microvascular endothelial cells 0.000451 6 miR143 Angiogenesis 2.63E-07 56 miR143 Vasculogenesis 0.00127 37 miR143 Permeability of vascular system 0.00011 13 miR150 Development of vasculature 0.00000972 29 miR150 Vascularization 0.00185 8 miR150 Binding of endothelial cells 0.000251 9 miR150 Adhesion of endothelial cells 0.00426 6 miR150 Endothelial cell development 0.000432 13 miR150 Proliferation of endothelial cells 0.00514 10 miR150 Migration of vascular endothelial cells 0.0000198 10 miR150 Cell movement of endothelial cells 0.0000432 15 miR150 Migration of endothelial cells 0.0000641 14 miR150 Vasculogenesis 0.00000652 25 miR150 Angiogenesis 0.000014 28 miR150 Vascularization of absolute anatomical region 0.00144 6 miR150 Formation of blood vessel 0.00207 7 miR150 Permeability of vascular system 0.00553 6

OIR miR143 Tubulation of vascular endothelial cells 0.0000264 6

77

miRNA Diseases or Functions Annotation p-value # Molecules OIR miR143 Interaction of endothelial cells 0.000415 7 OIR miR143 Migration of vascular endothelial cells 0.0000251 8 OIR miR143 Migration of endothelial cells 0.00026 10 OIR miR143 Vasculogenesis 0.0000738 17 OIR miR143 Angiogenesis 0.000121 19

Ingenuity Pathway Analysis Software (IPA) was employed to identify the statistically

significant biological pathways associated with the differential expressed genes across all the

samples. All the significant genes from the groups were used as input for pathway analysis by

IPA and major pathways are shown in Table 5a.2. Angiogenesis and vascularization were

associated with all groups.

5a.4 Discussion

Ischemic retinopathies are the leading cause of visual impairment in working adults and are

characterized by microvascular degeneration and preretinal neovascularization (Desjarlais et

al., 2019). Neovascularization is implicated in tractional retinal detachment and vitreous

haemorrhage in patients with diabetes. Retinal hypoxia is the underlying cause of retinal

neovascularization and fundus biomicroscopy has revealed that large vessels create new

irregular vascular networks on the retinal surface or on the vitreous(Group, 1991) (B.A., 2008).

miRNAs expression profile in oxygen-induced retinopathy (OIR) was used to identify

dysregulation that might contribute to the pathological processes of retinal neovascularization

(Wang et al., 2017c, Liu et al., 2016, Shen et al., 2008). Several in vivo and in vitro studies

suggested that miRNA is a crucial regulator of organ development and homeostasis (Vidigal

and Ventura, 2015). Human retina is difficult to sample for analysis. Therefore, animal models

are commonly used to investigate the role of miRNA expression in retinal neovascularization.

miRNA profiling expression in human specimens is emerging as a tool for disease

characterisation and identifying novel biomarkers or therapeutic targets for treatment and

diagnosis (Lu et al., 2005).

Various studies have already identified miRNAs that are downregulated in the retinas of OIR

in mice models including miR-126 (Bai et al., 2011), miR-218 (Han S, 2016), miR-410 (Chen

et al., 2014), miR-150 (Liu et al., 2015) and miR-129-5p (Liu et al., 2016). In the present study,

miR-143, miR-126 and miR-150 were down-regulated in the OIR rats at Passage 14 compared

78

with those in the normoxic rats. Studies have shown that mending endothelial-specific miR-

126 and miR-150 in retina can reduce the neovascularization in OIR mice models through

directly targeting pro-angiogenesis genes (Liu et al., 2015, Bai et al., 2011). miR-143 is down-

regulated in diseases with angiogenesis, notably cancers such as colon and lung cancer (Calin

and Croce, 2006).

Overall, we found novel genes that are downregulated in the OIR rat model and associated with

angiogenesis and vascularization in miR-143. Moreover, we observed the same pattern with

miR-126 and miR150. Identifying these novel genes will allow us to develop new therapeutic

targets and approaches for addressing retinal neovascularization.

79

Chapter 5b RNA-Seq Analysis for differentiating DPMSCs into TM cells

5b.1 Introduction

Glaucoma is a leading cause of irreversible blindness (Blomdahl et al., 1997, Munier et al.,

1998). The disease is characterized by the death of retinal ganglion cells (RGC) due to axonal

degeneration, resulting in visual impairment (Yasuda et al., 2014). Globally, 60 million people

are estimated to have glaucomatous optic neuropathy, and 8.4 million people are blind as a

result of glaucoma (Cook and Foster, 2012). Elevated intraocular pressure (IOP) is associated

with the development of glaucoma and IOP is dependent upon the flow of aqueous humour

(Langenberg et al., 2008, Quigley, 2011b). One of the functions of the specialized trabecular

meshwork (TM) tissue located at the anterior chamber angle of the eye is to regulate aqueous

humour drainage through the trabecular outflow pathway and hence balance IOP within the

eye (Gasiorowski and Russell, 2009, Du et al., 2013, Paulavičiūtė-Baikštienė et al., 2013,

Langenberg et al., 2008). Imbalance in the outflow of aqueous humour promotes pathological

IOP and later leads to the death of RGC (Paulavičiūtė-Baikštienė et al., 2013, Kwon et al.,

2009). It is not completely understood how dysfunction of TM is involved in glaucoma (Stamer

and Acott, 2012). Various animal models of glaucoma have been developed (Gasiorowski and

Russell, 2009, Weinreb and Lindsey, 2005, Kokotas et al., 2012, Gelatt, 1977, Zhang et al.,

2003, Tripathi et al., 1994, Flugel-Koch et al., 2004, Caballero et al., 2003, Bradley et al.,

2001). however, patient-specific disease cells provide more information than animal models

(Stamer et al., 1995, Gasiorowski and Russell, 2009). Human disease models can be derived

from pluripotent stem cells (HSCs), which could differentiate into TM cells.

Currently, mesenchymal stem cells (MSCs) can be directly differentiated from PSCs

(Pluripotent Stem Cells) (Menendez et al., 2013), but the conditions required to differentiate

MSCs into TM cells are still unknown. We differentiated dental pulp-derived MSCs

(DPMSCs) into TM cells using various growth factors such as retinoic acid (RA),

(transforming growth factor-β) (TGF-β2), BMP4 and examined the process under two

conditions (control and treated). The aim of this study is to examine the ability of dental pulp-

derived MSCs to differentiate into TM like cells by using Next-generation sequencing

bioinformatics analysis.

80

5b.2 Methods

Differentiation Protocol was performed “Nicola McDonald” and full details


5b.2.1 Sequencing

The libraries of cDNA from the biological samples were sequenced using an Illumina HiSeq-

2500 RNA-Seq platform as 150 bp single end at the Australian Genome Research Facility

(AGRF, Melbourne, VIC, Australia). HiSeq Control Software (HCS) v2.268 and Real-time

analysis (RTA) v1.18.66.3 were used to perform the image in real-time. The Fastq data were

generated using the Illumina bcl2fastq 2.20.0.422 pipeline.

5b.2.2 Dataset

Two groups of DPMSCs (control and treated) with two replicates were sequenced. The raw

data have been deposited in NCBI's Gene Expression Omnibus (GEO) and are accessible

through the GEO accession number XXXXXXXX (Data submission in progress).

5b.2.3 Quality Control Described in Chapter 5a. 5b.2.4 RNA-Seq Analysis

Described in Chapter 5a. 5b.2.5 Functional Enrichment Analysis

gProfiler (Reimand et al., 2016) was used for the functional enrichment analysis to identify the

statistically significant pathways for the selected genes, including gene ontology and other

gene-related functions.

5b.3 Results The statistics of mapped reads from the RNA-Seq data are shown in Table 5b.1 for all the

samples. More than 82% of reads were mapped to the Homosapiens (hg38) reference genome,

and transcript abundance was estimated with bootstrapping the samples 100 times. The

81

quantified transcript abundance was used for calculating differential gene expression across all

the conditions using Sleuth. Table 5b.1: Statistics of mapped reads from Control and treated sample with all replicates.

Figure 5b.1: a) Heatmap showing the expression (tpm) of statistically significant (adjusted pValue<0.01) gene found by using Sleuth. b) Volcano plot highlighting statistically significant (adjusted pValue<0.01) between control and treated condition. c) Read distribution plot across all samples. d) Statistically significant pathways by using gProfiler.

Condition Sample Id Mapped Reads Total Reads Mapping Percentage Bootstraps Control_1 S1_L1 13596851 16190560 83.98% 100 Control_1 S1_L2 13302861 15844733 83.96% 100 Control_1 S1_L3 13966207 16621921 84.02% 100 Control_1 S1_L4 13813545 16426765 84.09% 100 Control_2 S3_L1 13824442 16260986 85.02% 100 Control_2 S3_L2 13462380 15838698 85% 100 Control_2 S3_L3 14230329 16727478 85.07% 100 Control_2 S3_L4 13994383 16440843 85.12% 100 Treated_1 S4_L1 13167883 16008766 82.25% 100 Treated_1 S4_L2 12839258 15605535 82.27% 100 Treated_1 S4_L3 13552777 16458102 82.35% 100 Treated_1 S4_L4 13348810 16203611 82.38% 100 Treated_2 S6_L1 13943373 16891372 82.55% 100 Treated_2 S6_L2 13664856 16559655 82.52% 100 Treated_2 S6_L3 14334411 17346326 82.64% 100 Treated_2 S6_L4 14181421 17155166 82.67% 100

82

Overall, we found 8224 genes were differentially expressed (upregulated and downregulated

genes) under control and treated conditions based on adjusted p<0.01 (Figure 5b.1 a and Figure

5b.1 b) (Supplementary Table 5b.1). All the samples had three replicates, and the average of

each sample was used in the heatmap. The distribution of all samples’ read count (tpm) of

control and treated condition is plotted in Figure 5b.1 c. gProfiler were employed to identify

the statistically significant biological pathways associated with the differential expressed gene

across all the samples (Figure 5b.1 d). All the significant genes from the groups were used as

input for pathway analysis using gProfiler, and all statistically significant pathways are shown

in Supplementary Table 5b.2. The majority of the pathways were associated with tissue

development, extracellular matrix development, cell differentiation, cell morphogenesis

differentiation and cellular component development.

5b.4 Discussion

PSCs are self-renewing and have the capacity to differentiate into any cell type. MSCs are

multipotent cells that can be derived from mesenchymal or non-mesenchymal tissue by in vivo

and in vitro techniques (da Silva Meirelles et al., 2006). DPMSCs thoroughly expressed MSCs

marker and are able to differentiate into neural cells, adipocytes and osteoblasts and are

appropriate for tissue regeneration (Kawashima et al., 2017). Dental pulp stem cells (DPSCs),

also known as postnatal dental pulp stem cells, were isolated initially by Gronthos et al.

(Gronthos et al., 2002) from the third molar and, and their identity was confirmed by the ability

to differentiate into neural ectoderm cells and adipocytes (Zhang et al., 2008). We hypothesized

that we could differentiated DPMSCs into TM cells using RA, TGF-β2 and BMP4 growth

factors. RA, TGF-β2 and BMP4 growth factors are found in the aqueous humour and related

to the anterior segment development and are required for development of functional TM cells

(Fuchshofer et al., 2009, Furuta and Hogan, 1998). It is also known that BMP7 is involved in

eye development and highly expressed in the adult TM cells (Fuchshofer et al., 2009).

We used next-generation sequencing data to identify the genes and the associated molecular

pathways which included cell differentiation (GO:0030154), regulation of neuronal

differentiation (GO:0045664), neuronal differentiation (GO:0030182), mesenchymal cell

differentiation (GO:0048762) and endodermal cell differentiation (GO:0035987). The TGF-β-

related pathways are linked with TM cells, and are associated with IOP and therefore POAG.

TGF-β is well known for its expression of ocular hypertensive mediators, including cellular

83

contractility (Pervan, 2017). The TGF-β related pathways such as Signaling by TGF-beta

family members (REAC:R-HSA-9006936), Signaling by TGF-beta Receptor Complex

(REAC:R-HSA-170834) and Downregulation of TGF-beta receptor signaling (REAC:R-HSA-

2173788) are found in our analysis and thus confirm the involvement of TGF-β in

differentiation process in our study.

Overall, next-generation sequencing and pathway analysis results showed that the

differentiation protocol using growth factor (RA, TGF-β and BMP4) was an essential element

for differentiating DPMSCs into cells that share some characteristics with TM cells. The results

from this study will contribute to developing patient-specific TM cell models to better

understand the cellular and molecular mechanism of TM cell dysfunction in glaucoma.

84

Chapter 6 Conclusion

The eyes are a very complex organ consisting of three layers: the outer (sclera and cornea),

middle (uvea) and the inner (retina). For optimal vision, all eye components must function

properly in order to refract and focus light onto the retina. This thesis comprised four research

projects to understand the various eye-disease models by using bioinformatics applications.

iPSCs are pluripotent cells that are self-renewing and have the potential to differentiate into

any cell. However, during the differentiation process, the cell lines tend to accumulate genomic

aberrations. Applying virtual karyotyping, we used the computational tools PennCNV,

QuantiSNP, BCFtoolsCNV and CHAT to investigate copy number variation in control and

differentiated cell lines with a selection criterion of 10 contagious SNPs or 1mb spanning

length using Genotyping arrays. From our analysis, we demonstrated higher resolution of

clonal variations (amplification and deletion) at the chromosomal level with virtual

karyotyping and showed that this technique is equivalent to traditional karyotyping methods.

We confirmed that virtual karyotyping on a low-density SNP array is a practical option for

cost-effective, high-throughput assessment of iPSCs.

Epigenetics refers to genetic modifications, including DNA methylation, post-translational

histone modifications and chromatin remodelling, that do not affect genomic sequence and are

influenced by environment interactions. In our second study, we determined the DNAm age of

the optic nerve, RPE/choroid and neurosensory retina tissue from the eye and peripheral whole

blood from the same individual. We found a linear relationship between the methylation age

and biological age in blood and ocular tissues (r2= 0.37; p=0.048). The methylation age of the

blood sample was closest to the biological age, whereas the methylation age of neurosensory

retina was statistically significantly different to mean chronological age by 16.1 years (r2=0.93,

p<0.001). The methylation age of RPE/Choroid and optic nerve also differed from the

biological age. However, the cause of the donor’s death and the preservation interval time did

not have any effect on the DNAm and biological age across the tissues. We identified

differences between the methylation and biological age in blood and ocular tissues. This

observation could have implications for treating eye disease due to epigenetic aging.

CRISPR Screening is another method for identifying the genes that are responsible for cell

proliferation and cell apoptosis by characterizing genomic data with the phenotype of interest.

85

In our CRISPR Screening study, the lentiCRISPRv2 library was amplified and packaged with

lentivirus on UM cell line OCM-1. The OCM-1 cell lines transduced with the lentiCRISPRv2

library and Multiplicity of infection was maintained by puromycin to ensure each cell had a

sgRNA. Later, cells were collected at two different passages (P0 and P12). NovoSeq (next-

generation sequencing) was performed to identify the candidate hits using CRISPRAnalyzeR.

Overall, we identified 15 genes that had lower expression in P12 as compared to the P0 as

possible candidates. From the functional analysis, our selected genes associated with the

metabolic process, cellular process and biological process.

The selected gene expression was checked across the Cancer Genomic Atlas and 14 out of 15

genes and expression were found on the UM dataset. Our analysis shows that selected genes

may play a crucial role in cell proliferation in UM cell lines. However, further in vivo and in

vitro validation is required.

RNA-Sequence analysis is another way to identify gene expression under control and disease

conditions. We used the RNA-Seq method to analyse two different datasets. In the first part of

the analysis, we implemented the RNA-Seq computational tool to find the genes that are

responsible for neovascularization in OIR rat models in vitro (rat retina including OIR

scrambled RNA and OIR-miR143) and also in the in vivo model human endothelial cells

(telomerase-immortalized microvascular endothelial cell [TIME], human dermal

microvascular endothelial cell [HMEC] and human umbilical vein endothelial cell [HUVEC]).

We perform pathway analysis to identify the pathways responsible for the neovascularization

from our analysis in the in vitro sequence data where we treated the miR143 in OIR rat model.

We found significant genes indirectly related to angiogenesis, a step in the development of

neovascularization. Other miRNAs such as miR126 and miR150 had already been reported to

contribute to the development of neovascularization. However, we are the first to report the

role of miR143 in the development of retinal neovascularization and also confirm the

expression by analysing in vivo sequencing data.

In the second RNA-seq study, we differentiated dental pulp-derived MSCs (DPMSCs) into TM

cells using the growth factors such as RA, TGF-β2 and BMP4 and examined them under

control and treated conditions. We performed next generation sequencing (NGS) RNA-Seq to

identify the differential expressed genes and pathways that might be involved in the

differentiation of DPMSCs into TM cells. The molecular pathways identified from the analysis

86

included cell differentiation (GO:0030154), regulation of neuronal differentiation

(GO:0045664), neuronal differentiation (GO:0030182), mesenchymal cell differentiation

(GO:0048762) and endodermal cell differentiation (GO:0035987). Other pathways such as

signaling by TGF-beta family members (REAC:R-HSA-9006936), signaling by TGF-beta

Receptor Complex (REAC:R-HSA-170834) and downregulation of TGF-beta receptor

signaling (REAC:R-HSA-2173788) were also found, which further confirms that the DPMSCs

differentiated into the TM cells. Overall, our results indicate that the growth factor is essential

for the differentiation of DPMSCs into TM cells. Thus, this method of developing models of

eye disease, such as glaucoma, can help reveal the underlying cellular and molecular

mechanisms.

Overall, I have implemented various Bioinformatics approaches to understand the different

aspect of eye disease models including quality control of iPSCs to detect the genomic

aberrations in high-resolution as compared to the traditional method, Predicting DNA

methylation age of ocular tissue compared to biological age, knock-out gene identification by

CRISPR screening method and RNA-Seq analysis to find the role of miR-143 in the OIR rat

model and Differentiation of DPMSCs into TM like cells. In future it would be interesting to

compare the DNA methylation age prediction in knowing ocular disease patients and to

compare with the existing result to unfold how epigenetic age plays a crucial role in ocular

tissues in healthy and disease conditions. Also, it would be interesting to perform the single

cell analysis to better understand differentiated cell lines and how closely related they are

related to TM like cells. However, the sensitivity and efficacy of the pipeline can be tested for

qualitative analysis.

87

References

ADDICKS, E. M., QUIGLEY, H. A., GREEN, W. R. & ROBIN, A. L. 1983. Histologic characteristics of filtering blebs in glaucomatous eyes. Arch Ophthalmol, 101, 795-8.

AGROTIS, A. & KETTELER, R. 2015. A new age in functional genomics using CRISPR/Cas9 in arrayed library screening. Front Genet, 6, 300.

ALKAN, C., COE, B. P. & EICHLER, E. E. 2011. Genome structural variation discovery and genotyping. Nat Rev Genet, 12, 363-76.

ALLEN, D. & VASAVADA, A. 2006. Cataract and surgery for cataract. BMJ, 333, 128-32. AMANN, R. & FUCHS, B. M. 2008. Single-cell identification in microbial communities by

improved fluorescence in situ hybridization techniques. Nat Rev Microbiol, 6, 339-48. AMARASINGHE, K. C., LI, J., HUNTER, S. M., RYLAND, G. L., COWIN, P. A., CAMPBELL, I. G. &

HALGAMUGE, S. K. 2014. Inferring copy number and genotype in tumour exome data. BMC Genomics, 15, 732.

ANDERS, S. & HUBER, W. 2010. Differential expression analysis for sequence count data. Genome Biol, 11, R106.

ANDERS, S., PYL, P. T. & HUBER, W. 2015. HTSeq--a Python framework to work with high-throughput sequencing data. Bioinformatics, 31, 166-9.

ANDERSON, A. L., STANGER, S. J., MIHALAS, B. P., TYAGI, S., HOLT, J. E., MCLAUGHLIN, E. A. & NIXON, B. 2015. Assessment of microRNA expression in mouse epididymal epithelial cells and spermatozoa by next generation sequencing. Genom Data, 6, 208-11.

ANDREWS, S. 2010. FastQC: a quality control tool for high throughput sequence data. ARYEE, M. J., JAFFE, A. E., CORRADA-BRAVO, H., LADD-ACOSTA, C., FEINBERG, A. P.,

HANSEN, K. D. & IRIZARRY, R. A. 2014. Minfi: a flexible and comprehensive Bioconductor package for the analysis of Infinium DNA methylation microarrays. Bioinformatics, 30, 1363-9.

AUGSBURGER, J. J., CORREA, Z. M. & SHAIKH, A. H. 2009. Effectiveness of treatments for metastatic uveal melanoma. Am J Ophthalmol, 148, 119-27.

B.A., B. 2008. Hypoxia and Retinal Neovascularization. Penn J. (eds) Retinal and Choroidal Angiogenesis. Springer, Dordrecht.

BACK, S., NECARSULMER, J., WHITAKER, L. R., COKE, L. M., KOIVULA, P., HEATHWARD, E. J., FORTUNO, L. V., ZHANG, Y., YEH, C. G., BALDWIN, H. A., SPENCER, M. D., MEJIAS-APONTE, C. A., PICKEL, J., HOFFMAN, A. F., SPIVAK, C. E., LUPICA, C. R., UNDERHILL, S. M., AMARA, S. G., DOMANSKYI, A., ANTTILA, J. E., AIRAVAARA, M., HOPE, B. T., HAMRA, F. K., RICHIE, C. T. & HARVEY, B. K. 2019. Neuron-Specific Genome Modification in the Adult Rat Brain Using CRISPR-Cas9 Transgenic Rats. Neuron, 102, 105-119 e8.

BAI, X., ZENG, T., NI, X. Y., SU, H. A., HUANG, J., YE, G. Y., LU, Y. Y. & QI, Y. X. 2019. CRISPR/Cas9-mediated knockout of the eye pigmentation gene white leads to alterations in colour of head spots in the oriental fruit fly, Bactrocera dorsalis. Insect Mol Biol.

BAKER, M., ROBINSON, S. D., LECHERTIER, T., BARBER, P. R., TAVORA, B., D'AMICO, G., JONES, D. T., VOJNOVIC, B. & HODIVALA-DILKE, K. 2012. Use of the mouse aortic ring assay to study angiogenesis. Nat Protoc, 7, 89-104.

BALLESTAR, E. 2011. Epigenetic alterations in autoimmune rheumatic diseases. Nat Rev Rheumatol, 7, 263-71.

88

BARTHA, I., DI IULIO, J., VENTER, J. C. & TELENTI, A. 2018. Human gene essentiality. Nat Rev Genet, 19, 51-62.

BEDIKIAN, A. Y., LEGHA, S. S., MAVLIGIT, G., CARRASCO, C. H., KHORANA, S., PLAGER, C., PAPADOPOULOS, N. & BENJAMIN, R. S. 1995. Treatment of uveal melanoma metastatic to the liver: a review of the M. D. Anderson Cancer Center experience and prognostic factors. Cancer, 76, 1665-70.

BERDASCO, M. & ESTELLER, M. 2012. Hot topics in epigenetic mechanisms of aging: 2011. Aging Cell, 11, 181-6.

BIRMINGHAM, E., NIEBUR, G. & MCHUGH, P. 2012. Osteogenic differentiation of mesenchymal stem cells is regulated by osteocyte and osteoblast cells in a simplified bone niche.

BLODI, C. F. & STONE, E. M. 1990. Best's vitelliform dystrophy. Ophthalmic Paediatr Genet, 11, 49-59.

BLOMDAHL, S., CALISSENDORFF, B. M., TENGROTH, B. & WALLIN, O. 1997. Blindness in glaucoma patients. Acta Ophthalmol Scand, 75, 589-91.

BOCK, C., KISKINIS, E., VERSTAPPEN, G., GU, H., BOULTING, G., SMITH, Z. D., ZILLER, M., CROFT, G. F., AMOROSO, M. W., OAKLEY, D. H., GNIRKE, A., EGGAN, K. & MEISSNER, A. 2011. Reference Maps of human ES and iPS cell variation enable high-throughput characterization of pluripotent cell lines. Cell, 144, 439-52.

BOCKLANDT, S., LIN, W., SEHL, M. E., SANCHEZ, F. J., SINSHEIMER, J. S., HORVATH, S. & VILAIN, E. 2011. Epigenetic predictor of age. PLoS One, 6, e14821.

BOLEY, N., STOIBER, M. H., BOOTH, B. W., WAN, K. H., HOSKINS, R. A., BICKEL, P. J., CELNIKER, S. E. & BROWN, J. B. 2014. Genome-guided transcript assembly by integrative analysis of RNA sequence data. Nat Biotechnol, 32, 341-6.

BOON, C. J., KLEVERING, B. J., LEROY, B. P., HOYNG, C. B., KEUNEN, J. E. & DEN HOLLANDER, A. I. 2009. The spectrum of ocular phenotypes caused by mutations in the BEST1 gene. Prog Retin Eye Res, 28, 187-205.

BRADLEY, J. M., KELLEY, M. J., ZHU, X., ANDERSSOHN, A. M., ALEXANDER, J. P. & ACOTT, T. S. 2001. Effects of mechanical stretching on trabecular matrix metalloproteinases. Investigative ophthalmology & visual science, 42, 1505-1513.

BRAY N, P. H., MELSTED P, ET AL 2015. Near-optimal RNA-Seq quantification. arXiv:1505.02710.

BRAY, N. L., PIMENTEL, H., MELSTED, P. & PACHTER, L. 2016. Near-optimal probabilistic RNA-seq quantification. Nat Biotechnol, 34, 525-7.

BRESSLER, S. B. 2009. Introduction: Understanding the role of angiogenesis and antiangiogenic agents in age-related macular degeneration. Ophthalmology, 116, S1-7.

BRUCE RICHARDSON 2003. Impact of aging on DNA methylation. Ageing Research Reviews , 2, 245-261. BULLARD, J. H., PURDOM, E., HANSEN, K. D. & DUDOIT, S. 2010. Evaluation of statistical

methods for normalization and differential expression in mRNA-Seq experiments. BMC Bioinformatics, 11, 94.

CABALLERO, M., LITON, P. B., EPSTEIN, D. L. & GONZALEZ, P. 2003. Proteasome inhibition by chronic oxidative stress in human trabecular meshwork cells. Biochemical and biophysical research communications, 308, 346-352.

CALIN, G. A. & CROCE, C. M. 2006. MicroRNA signatures in human cancers. Nat Rev Cancer, 6, 857-66.

89

CAMPOCHIARO, P. A., AIELLO, L. P. & ROSENFELD, P. J. 2016. Anti-Vascular Endothelial Growth Factor Agents in the Treatment of Retinal Disease: From Bench to Bedside. Ophthalmology, 123, S78-S88.

CANCER GENOME ATLAS RESEARCH, N., WEINSTEIN, J. N., COLLISSON, E. A., MILLS, G. B., SHAW, K. R., OZENBERGER, B. A., ELLROTT, K., SHMULEVICH, I., SANDER, C. & STUART, J. M. 2013. The Cancer Genome Atlas Pan-Cancer analysis project. Nat Genet, 45, 1113-20.

CARVAJAL, R. D., SOSMAN, J. A., QUEVEDO, J. F., MILHEM, M. M., JOSHUA, A. M., KUDCHADKAR, R. R., LINETTE, G. P., GAJEWSKI, T. F., LUTZKY, J., LAWSON, D. H., LAO, C. D., FLYNN, P. J., ALBERTINI, M. R., SATO, T., LEWIS, K., DOYLE, A., ANCELL, K., PANAGEAS, K. S., BLUTH, M., HEDVAT, C., ERINJERI, J., AMBROSINI, G., MARR, B., ABRAMSON, D. H., DICKSON, M. A., WOLCHOK, J. D., CHAPMAN, P. B. & SCHWARTZ, G. K. 2014. Effect of selumetinib vs chemotherapy on progression-free survival in uveal melanoma: a randomized clinical trial. JAMA, 311, 2397-405.

CHANDRASHEKAR, D. S., BASHEL, B., BALASUBRAMANYA, S. A. H., CREIGHTON, C. J., PONCE-RODRIGUEZ, I., CHAKRAVARTHI, B. & VARAMBALLY, S. 2017. UALCAN: A Portal for Facilitating Tumor Subgroup Gene Expression and Survival Analyses. Neoplasia, 19, 649-658.

CHANG, A. E., KARNELL, L. H. & MENCK, H. R. 1998. The National Cancer Data Base report on cutaneous and noncutaneous melanoma: a summary of 84,836 cases from the past decade. The American College of Surgeons Commission on Cancer and the American Cancer Society. Cancer, 83, 1664-78.

CHARLEEN M MOORE & ROBERT G BEST 2001. Chromosome Preparation and Banding. In eLS, (Ed.).

CHARLES HUANG, C. Y., REUBEN, P. M., D'IPPOLITO, G., SCHILLER, P. C. & CHEUNG, H. S. 2004. Chondrogenesis of human bone marrow-derived mesenchymal stem cells in agarose culture. The Anatomical Record Part A: Discoveries in Molecular, Cellular, and Evolutionary Biology, 278, 428-436.

CHEN, J., CONNOR, K. M., ADERMAN, C. M., WILLETT, K. L., ASPEGREN, O. P. & SMITH, L. E. 2009. Suppression of retinal neovascularization by erythropoietin siRNA in a mouse model of proliferative retinopathy. Invest Ophthalmol Vis Sci, 50, 1329-35.

CHEN, N., WANG, J., HU, Y., CUI, B., LI, W., XU, G., LIU, L. & LIU, S. 2014. MicroRNA-410 reduces the expression of vascular endothelial growth factor and inhibits oxygen-induced retinal neovascularization. PLoS One, 9, e95665.

CHEN, X. & ROSBASH, M. 2016. mir-276a strengthens Drosophila circadian rhythms by regulating timeless expression. Proc Natl Acad Sci U S A, 113, E2965-72.

CHEUNG, N., MITCHELL, P. & WONG, T. Y. 2010. Diabetic retinopathy. Lancet, 376, 124-36. CHIN, M. H., MASON, M. J., XIE, W., VOLINIA, S., SINGER, M., PETERSON, C.,

AMBARTSUMYAN, G., AIMIUWU, O., RICHTER, L., ZHANG, J., KHVOROSTOV, I., OTT, V., GRUNSTEIN, M., LAVON, N., BENVENISTY, N., CROCE, C. M., CLARK, A. T., BAXTER, T., PYLE, A. D., TEITELL, M. A., PELEGRINI, M., PLATH, K. & LOWRY, W. E. 2009. Induced pluripotent stem cells and embryonic stem cells are distinguished by gene expression signatures. Cell Stem Cell, 5, 111-23.

COIN, L. J., ASHER, J. E., WALTERS, R. G., MOUSTAFA, J. S., DE SMITH, A. J., SLADEK, R., BALDING, D. J., FROGUEL, P. & BLAKEMORE, A. I. 2010. cnvHap: an integrative population and haplotype-based multiplatform model of SNPs and CNVs. Nat Methods, 7, 541-6.

90

COLLABORATIVE OCULAR MELANOMA STUDY, G. 2001. Assessment of metastatic disease status at death in 435 patients with large choroidal melanoma in the Collaborative Ocular Melanoma Study (COMS): COMS report no. 15. Arch Ophthalmol, 119, 670-6.

CONESA, A., MADRIGAL, P., TARAZONA, S., GOMEZ-CABRERO, D., CERVERA, A., MCPHERSON, A., SZCZESNIAK, M. W., GAFFNEY, D. J., ELO, L. L., ZHANG, X. & MORTAZAVI, A. 2016. A survey of best practices for RNA-seq data analysis. Genome Biol, 17, 13.

CONSORTIUM, E. P., BIRNEY, E., STAMATOYANNOPOULOS, J. A., DUTTA, A., GUIGO, R., GINGERAS, T. R., MARGULIES, E. H., WENG, Z., SNYDER, M., DERMITZAKIS, E. T., THURMAN, R. E., KUEHN, M. S., TAYLOR, C. M., NEPH, S., KOCH, C. M., ASTHANA, S., MALHOTRA, A., ADZHUBEI, I., GREENBAUM, J. A., ANDREWS, R. M., FLICEK, P., BOYLE, P. J., CAO, H., CARTER, N. P., CLELLAND, G. K., DAVIS, S., DAY, N., DHAMI, P., DILLON, S. C., DORSCHNER, M. O., FIEGLER, H., GIRESI, P. G., GOLDY, J., HAWRYLYCZ, M., HAYDOCK, A., HUMBERT, R., JAMES, K. D., JOHNSON, B. E., JOHNSON, E. M., FRUM, T. T., ROSENZWEIG, E. R., KARNANI, N., LEE, K., LEFEBVRE, G. C., NAVAS, P. A., NERI, F., PARKER, S. C., SABO, P. J., SANDSTROM, R., SHAFER, A., VETRIE, D., WEAVER, M., WILCOX, S., YU, M., COLLINS, F. S., DEKKER, J., LIEB, J. D., TULLIUS, T. D., CRAWFORD, G. E., SUNYAEV, S., NOBLE, W. S., DUNHAM, I., DENOEUD, F., REYMOND, A., KAPRANOV, P., ROZOWSKY, J., ZHENG, D., CASTELO, R., FRANKISH, A., HARROW, J., GHOSH, S., SANDELIN, A., HOFACKER, I. L., BAERTSCH, R., KEEFE, D., DIKE, S., CHENG, J., HIRSCH, H. A., SEKINGER, E. A., LAGARDE, J., ABRIL, J. F., SHAHAB, A., FLAMM, C., FRIED, C., HACKERMULLER, J., HERTEL, J., LINDEMEYER, M., MISSAL, K., TANZER, A., WASHIETL, S., KORBEL, J., EMANUELSSON, O., PEDERSEN, J. S., HOLROYD, N., TAYLOR, R., SWARBRECK, D., MATTHEWS, N., DICKSON, M. C., THOMAS, D. J., WEIRAUCH, M. T., et al. 2007. Identification and analysis of functional elements in 1% of the human genome by the ENCODE pilot project. Nature, 447, 799-816.

COOK, C. & FOSTER, P. 2012. Epidemiology of glaucoma: what's new? Can J Ophthalmol, 47, 223-6.

COWAN, C. A., KLIMANSKAYA, I., MCMAHON, J., ATIENZA, J., WITMYER, J., ZUCKER, J. P., WANG, S., MORTON, C. C., MCMAHON, A. P., POWERS, D. & MELTON, D. A. 2004. Derivation of embryonic stem-cell lines from human blastocysts. N Engl J Med, 350, 1353-6.

DA SILVA MEIRELLES, L., CHAGASTELLES, P. C. & NARDI, N. B. 2006. Mesenchymal stem cells reside in virtually all post-natal organs and tissues. J Cell Sci, 119, 2204-13.

DAMATO, B., DOPIERALA, J. A. & COUPLAND, S. E. 2010. Genotypic profiling of 452 choroidal melanomas with multiplex ligation-dependent probe amplification. Clin Cancer Res, 16, 6083-92.

DANECEK, P., MCCARTHY, S. A., HIPSCI, C. & DURBIN, R. 2016. A Method for Checking Genomic Integrity in Cultured Cell Lines from SNP Genotyping Data. PLoS One, 11, e0155014.

DE CARLO, T. E., CHIN, A. T., BONINI FILHO, M. A., ADHI, M., BRANCHINI, L., SALZ, D. A., BAUMAL, C. R., CRAWFORD, C., REICHEL, E., WITKIN, A. J., DUKER, J. S. & WAHEED, N. K. 2015. Detection of Microvascular Changes in Eyes of Patients with Diabetes but Not Clinical Diabetic Retinopathy Using Optical Coherence Tomography Angiography. Retina, 35, 2364-70.

91

DE RAVEL, T. J., DEVRIENDT, K., FRYNS, J. P. & VERMEESCH, J. R. 2007. What's new in karyotyping? The move towards array comparative genomic hybridisation (CGH). Eur J Pediatr, 166, 637-43.

DE ROACH, J. N., MCLAREN, T. L., PATERSON, R. L., O'BRIEN, E. C., HOFFMANN, L., MACKEY, D. A., HEWITT, A. W. & LAMEY, T. M. 2013. Establishment and evolution of the Australian Inherited Retinal Disease Register and DNA Bank. Clin Experiment Ophthalmol, 41, 476-83.

DESBATS, M. A., MORBIDONI, V., SILIC-BENUSSI, M., DOIMO, M., CIMINALE, V., CASSINA, M., SACCONI, S., HIRANO, M., BASSO, G., PIERREL, F., NAVAS, P., SALVIATI, L. & TREVISSON, E. 2016. The COQ2 genotype predicts the severity of coenzyme Q10 deficiency. Hum Mol Genet, 25, 4256-4265.

DESJARLAIS, M., RIVERA, J. C., LAHAIE, I., CAGNONE, G., WIRT, M., OMRI, S. & CHEMTOB, S. 2019. MicroRNA expression profile in retina and choroid in oxygen-induced retinopathy model. PLoS One, 14, e0218282.

DIENER-WEST, M., REYNOLDS, S. M., AGUGLIARO, D. J., CALDWELL, R., CUMMING, K., EARLE, J. D., GREEN, D. L., HAWKINS, B. S., HAYMAN, J., JAIYESIMI, I., KIRKWOOD, J. M., KOH, W. J., ROBERTSON, D. M., SHAW, J. M., THOMA, J. & COLLABORATIVE OCULAR MELANOMA STUDY GROUP, R. 2004. Screening for metastasis from choroidal melanoma: the Collaborative Ocular Melanoma Study Group Report 23. J Clin Oncol, 22, 2438-44.

DING, X., PATEL, M. & CHAN, C. C. 2009. Molecular pathology of age-related macular degeneration. Prog Retin Eye Res, 28, 1-18.

DONOVAN, D., BROWN, N. J., BISHOP, E. T. & LEWIS, C. E. 2001. Comparison of three in vitro human 'angiogenesis' assays with capillaries formed in vivo. Angiogenesis, 4, 113-21.

DRAPER, J. S., SMITH, K., GOKHALE, P., MOORE, H. D., MALTBY, E., JOHNSON, J., MEISNER, L., ZWAKA, T. P., THOMSON, J. A. & ANDREWS, P. W. 2004. Recurrent gain of chromosomes 17q and 12 in cultured human embryonic stem cells. Nat Biotechnol, 22, 53-4.

DU, Y., YUN, H., YANG, E. & SCHUMAN, J. S. 2013. Stem cells from trabecular meshwork home to TM tissue in vivo. Investigative ophthalmology & visual science, 54, 1450-1459.

DUMMER, R., HAUSCHILD, A., GUGGENHEIM, M., KEILHOLZ, U., PENTHEROUDAKIS, G. & GROUP, E. G. W. 2012. Cutaneous melanoma: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol, 23 Suppl 7, vii86-91.

EL-ANEED, A. 2004. Current strategies in cancer gene therapy. European journal of pharmacology, 498, 1-8.

ESTRACH, S., LEE, S. A., BOULTER, E., PISANO, S., ERRANTE, A., TISSOT, F. S., CAILLETEAU, L., PONS, C., GINSBERG, M. H. & FERAL, C. C. 2014. CD98hc (SLC3A2) loss protects against ras-driven tumorigenesis by modulating integrin-mediated mechanotransduction. Cancer Res, 74, 6878-89.

FERGUS, J., QUINTANILLA, R. & LAKSHMIPATHY, U. 2016. Characterizing Pluripotent Stem Cells Using the TaqMan(R) hPSC Scorecard(TM) Panel. Methods Mol Biol, 1307, 25-37.

FERGUSON, R., VOGELSANG, M., UCISIK-AKKAYA, E., RAI, K., PILARSKI, R., MARTINEZ, C. N., RENDLEMAN, J., KAZLOW, E., NAGDIMOV, K. & OSMAN, I. 2016. Genetic markers of pigmentation are novel risk loci for uveal melanoma. Scientific reports, 6, 31191.

92

FLAXMAN, S. R., BOURNE, R. R. A., RESNIKOFF, S., ACKLAND, P., BRAITHWAITE, T., CICINELLI, M. V., DAS, A., JONAS, J. B., KEEFFE, J., KEMPEN, J. H., LEASHER, J., LIMBURG, H., NAIDOO, K., PESUDOVS, K., SILVESTER, A., STEVENS, G. A., TAHHAN, N., WONG, T. Y., TAYLOR, H. R. & VISION LOSS EXPERT GROUP OF THE GLOBAL BURDEN OF DISEASE, S. 2017. Global causes of blindness and distance vision impairment 1990-2020: a systematic review and meta-analysis. Lancet Glob Health, 5, e1221-e1234.

FLUGEL-KOCH, C., OHLMANN, A., FUCHSHOFER, R., WELGE-LUSSEN, U. & TAMM, E. R. 2004. Thrombospondin-1 in the trabecular meshwork: localization in normal and glaucomatous eyes, and induction by TGF-beta1 and dexamethasone in vitro. Exp Eye Res, 79, 649-63.

FRIEDLANDER, M. R., CHEN, W., ADAMIDI, C., MAASKOLA, J., EINSPANIER, R., KNESPEL, S. & RAJEWSKY, N. 2008. Discovering microRNAs from deep sequencing data using miRDeep. Nat Biotechnol, 26, 407-15.

FRIEDMAN, D. S., O'COLMAIN, B. J., MUNOZ, B., TOMANY, S. C., MCCARTY, C., DE JONG, P. T., NEMESURE, B., MITCHELL, P., KEMPEN, J. & EYE DISEASES PREVALENCE RESEARCH, G. 2004. Prevalence of age-related macular degeneration in the United States. Arch Ophthalmol, 122, 564-72.

FRIEDMAN, J., HASTIE, T. & TIBSHIRANI, R. 2010. Regularization Paths for Generalized Linear Models via Coordinate Descent. J Stat Softw, 33, 1-22.

FUCHSHOFER, R., STEPHAN, D. A., RUSSELL, P. & TAMM, E. R. 2009. Gene expression profiling of TGFβ2- and/or BMP7-treated trabecular meshwork cells: Identification of Smad7 as a critical inhibitor of TGF-β2 signaling. Experimental eye research, 88, 1020-1032.

FURUTA, Y. & HOGAN, B. L. 1998. BMP4 is essential for lens induction in the mouse embryo. Genes & development, 12, 3764-3775.

GAO, T., HAN, Y., YU, L., AO, S., LI, Z. & JI, J. 2014. CCNA2 is a prognostic biomarker for ER+ breast cancer and tamoxifen resistance. PLoS One, 9, e91771.

GARBER, M., GRABHERR, M. G., GUTTMAN, M. & TRAPNELL, C. 2011. Computational methods for transcriptome annotation and quantification using RNA-seq. Nat Methods, 8, 469-77.

GARINIS, G. A., VAN DER HORST, G. T., VIJG, J. & HOEIJMAKERS, J. H. 2008. DNA damage and ageing: new-age ideas for an age-old problem. Nat Cell Biol, 10, 1241-7.

GASIOROWSKI, J. Z. & RUSSELL, P. 2009. Biological properties of trabecular meshwork cells. Experimental eye research, 88, 671-675.

GELATT, K. 1977. Animal models for glaucoma. Investigative ophthalmology & visual science, 16, 592-596.

GIBNEY, E. R. & NOLAN, C. M. 2010. Epigenetics and gene expression. Heredity (Edinb), 105, 4-13.

GORDOIS, A., PEZZULLO, L. & CUTLER, H. 2010. The Global Economic Cost of Visual Impairment Access Economics Pty Limited

GRABHERR, M. G., HAAS, B. J., YASSOUR, M., LEVIN, J. Z., THOMPSON, D. A., AMIT, I., ADICONIS, X., FAN, L., RAYCHOWDHURY, R., ZENG, Q., CHEN, Z., MAUCELI, E., HACOHEN, N., GNIRKE, A., RHIND, N., DI PALMA, F., BIRREN, B. W., NUSBAUM, C., LINDBLAD-TOH, K., FRIEDMAN, N. & REGEV, A. 2011. Full-length transcriptome assembly from RNA-Seq data without a reference genome. Nat Biotechnol, 29, 644-52.

93

GRAGOUDAS, E. S. 2006. Proton beam irradiation of uveal melanomas: the first 30 years. The Weisenfeld Lecture. Invest Ophthalmol Vis Sci, 47, 4666-73.

GREENMAN, C. D., BIGNELL, G., BUTLER, A., EDKINS, S., HINTON, J., BEARE, D., SWAMY, S., SANTARIUS, T., CHEN, L., WIDAA, S., FUTREAL, P. A. & STRATTON, M. R. 2010. PICNIC: an algorithm to predict absolute allelic copy number variation with microarray cancer data. Biostatistics, 11, 164-75.

GRONNIGER, E., WEBER, B., HEIL, O., PETERS, N., STAB, F., WENCK, H., KORN, B., WINNEFELD, M. & LYKO, F. 2010. Aging and chronic sun exposure cause distinct epigenetic changes in human skin. PLoS Genet, 6, e1000971.

GRONTHOS, S., BRAHIM, J., LI, W., FISHER, L. W., CHERMAN, N., BOYDE, A., DENBESTEN, P., ROBEY, P. G. & SHI, S. 2002. Stem cell properties of human dental pulp stem cells. J Dent Res, 81, 531-5.

GROUP, E. T. D. R. S. R. 1991. Grading diabetic retinopathy from stereoscopic color fundus photographs--an extension of the modified Airlie House classification. ETDRS report number 10. Early Treatment Diabetic Retinopathy Study Research Group. Ophthalmology, 98, 786-806.

HAAS, B. J., PAPANICOLAOU, A., YASSOUR, M., GRABHERR, M., BLOOD, P. D., BOWDEN, J., COUGER, M. B., ECCLES, D., LI, B., LIEBER, M., MACMANES, M. D., OTT, M., ORVIS, J., POCHET, N., STROZZI, F., WEEKS, N., WESTERMAN, R., WILLIAM, T., DEWEY, C. N., HENSCHEL, R., LEDUC, R. D., FRIEDMAN, N. & REGEV, A. 2013. De novo transcript sequence reconstruction from RNA-seq using the Trinity platform for reference generation and analysis. Nat Protoc, 8, 1494-512.

HAGEMAN, G. S., ANDERSON, D. H., JOHNSON, L. V., HANCOX, L. S., TAIBER, A. J., HARDISTY, L. I., HAGEMAN, J. L., STOCKMAN, H. A., BORCHARDT, J. D., GEHRS, K. M., SMITH, R. J., SILVESTRI, G., RUSSELL, S. R., KLAVER, C. C., BARBAZETTO, I., CHANG, S., YANNUZZI, L. A., BARILE, G. R., MERRIAM, J. C., SMITH, R. T., OLSH, A. K., BERGERON, J., ZERNANT, J., MERRIAM, J. E., GOLD, B., DEAN, M. & ALLIKMETS, R. 2005. A common haplotype in the complement regulatory gene factor H (HF1/CFH) predisposes individuals to age-related macular degeneration. Proc Natl Acad Sci U S A, 102, 7227-32.

HAINES, J. L., HAUSER, M. A., SCHMIDT, S., SCOTT, W. K., OLSON, L. M., GALLINS, P., SPENCER, K. L., KWAN, S. Y., NOUREDDINE, M., GILBERT, J. R., SCHNETZ-BOUTAUD, N., AGARWAL, A., POSTEL, E. A. & PERICAK-VANCE, M. A. 2005. Complement factor H variant increases the risk of age-related macular degeneration. Science, 308, 419-21.

HALL, B., CHO, A., LIMAYE, A., CHO, K., KHILLAN, J. & KULKARNI, A. B. 2018. Genome Editing in Mice Using CRISPR/Cas9 Technology. Curr Protoc Cell Biol, 81, e57.

HAMADA, M., MALUREANU, L. A., WIJSHAKE, T., ZHOU, W. & VAN DEURSEN, J. M. 2012. Reprogramming to pluripotency can conceal somatic cell chromosomal instability. PLoS Genet, 8, e1002913.

HAN S, K. Y., SUN B, HAN QH, CHEN Y, WANG YC. 2016. Expression of Concern: MicroRNA-218 Inhibits Oxygen-induced Retinal Neovascularization via Reducing the Expression of Roundabout 1. Chin Med J (Engl), 129, 2772.

HANAHAN, D. & WEINBERG, R. A. 2000. The hallmarks of cancer. Cell, 100, 57-70. HARBOUR, J. W., ONKEN, M. D., ROBERSON, E. D., DUAN, S., CAO, L., WORLEY, L. A.,

COUNCIL, M. L., MATATALL, K. A., HELMS, C. & BOWCOCK, A. M. 2010. Frequent mutation of BAP1 in metastasizing uveal melanomas. Science, 330, 1410-1413.

94

HART, A. 2001. Mann-Whitney test is not just a test of medians: differences in spread can be important. BMJ, 323, 391-3.

HART, T. & MOFFAT, J. 2016. BAGEL: a computational framework for identifying essential genes from pooled library screens. BMC Bioinformatics, 17, 164.

HENSON, N. L., HEATON, M. L., HOLLAND, B. H., HAWKINS, K. C., RAWLINGS, B. A., EANES, E. A., BOZOF, R. G., POWELL, S. W., GRAU, R. E., FORTNEY, J. A., PEEBLES, B. G., KUMAR, D., YOON, J. I., GODBY, K. N., COLLINS, J. A., SOOD, R., HIXSON, D., BOWYER, F. P., 3RD, BLACK, A. C., JR. & YOUNG, H. E. 2005. Karyotypic analysis of adult pluripotent stem cells. Histol Histopathol, 20, 769-84.

HERNANDEZ, D. G., NALLS, M. A., GIBBS, J. R., AREPALLI, S., VAN DER BRUG, M., CHONG, S., MOORE, M., LONGO, D. L., COOKSON, M. R., TRAYNOR, B. J. & SINGLETON, A. B. 2011. Distinct DNA methylation changes highly correlated with chronological age in the human brain. Hum Mol Genet, 20, 1164-72.

HEWING, N. J., WESKAMP, G., VERMAAT, J., FARAGE, E., GLOMSKI, K., SWENDEMAN, S., CHAN, R. V., CHIANG, M. F., KHOKHA, R., ANAND-APTE, B. & BLOBEL, C. P. 2013. Intravitreal injection of TIMP3 or the EGFR inhibitor erlotinib offers protection from oxygen-induced retinopathy in mice. Invest Ophthalmol Vis Sci, 54, 864-70.

HEWITT, A. W., CRAIG, J. E. & MACKEY, D. A. 2006. Complex genetics of complex traits: the case of primary open-angle glaucoma. Clin Experiment Ophthalmol, 34, 472-84.

HOLMGAARD, A., ALSING, S., ASKOU, A. L. & CORYDON, T. J. 2019. CRISPR Gene Therapy of the Eye: Targeted Knockout of Vegfa in Mouse Retina by Lentiviral Delivery. Methods Mol Biol, 1961, 307-328.

HORVATH, S. 2017. DNA methylation age calculator [Online]. Available: https://dnamage.genetics.ucla.edu/ [Accessed 30 January 2017 2017].

HORVATH, S., GARAGNANI, P., BACALINI, M. G., PIRAZZINI, C., SALVIOLI, S., GENTILINI, D., DI BLASIO, A. M., GIULIANI, C., TUNG, S., VINTERS, H. V. & FRANCESCHI, C. 2015. Accelerated epigenetic aging in Down syndrome. Aging Cell, 14, 491-5.

HSU, J. B., CHIU, C. M., HSU, S. D., HUANG, W. Y., CHIEN, C. H., LEE, T. Y. & HUANG, H. D. 2011. miRTar: an integrated system for identifying miRNA-target interactions in human. BMC Bioinformatics, 12, 300.

HU, D. N., YU, G. P., MCCORMICK, S. A., SCHNEIDER, S. & FINGER, P. T. 2005. Population-based incidence of uveal melanoma in various races and ethnic groups. Am J Ophthalmol, 140, 612-7.

HUSSEIN, S. M., BATADA, N. N., VUORISTO, S., CHING, R. W., AUTIO, R., NARVA, E., NG, S., SOUROUR, M., HAMALAINEN, R., OLSSON, C., LUNDIN, K., MIKKOLA, M., TROKOVIC, R., PEITZ, M., BRUSTLE, O., BAZETT-JONES, D. P., ALITALO, K., LAHESMAA, R., NAGY, A. & OTONKOSKI, T. 2011. Copy number variation and selection during reprogramming to pluripotency. Nature, 471, 58-62.

HWANG, W. Y., FU, Y., REYON, D., MAEDER, M. L., KAINI, P., SANDER, J. D., JOUNG, J. K., PETERSON, R. T. & YEH, J. R. 2013. Heritable and precise zebrafish genome editing using a CRISPR-Cas system. PLoS One, 8, e68708.

INC, T. F. S. 2019. Axiom™ Long Format Export Tool v1.3 USER GUIDE. . INC;, T. F. S. 2019. AxiomTMAnalysis Suite (AxAS) v4.0.1 USER GUIDE. IRIZARRY, R. A., LADD-ACOSTA, C., WEN, B., WU, Z., MONTANO, C., ONYANGO, P., CUI, H.,

GABO, K., RONGIONE, M., WEBSTER, M., JI, H., POTASH, J., SABUNCIYAN, S. & FEINBERG, A. P. 2009. The human colon cancer methylome shows similar hypo- and

95

hypermethylation at conserved tissue-specific CpG island shores. Nat Genet, 41, 178-186.

ISSERLIN, R., MERICO, D., VOISIN, V. & BADER, G. D. 2014. Enrichment Map - a Cytoscape app to visualize and explore OMICs pathway enrichment results. F1000Res, 3, 141.

JANSEN, R., EMBDEN, J. D., GAASTRA, W. & SCHOULS, L. M. 2002. Identification of genes that are associated with DNA repeats in prokaryotes. Mol Microbiol, 43, 1565-75.

JINEK, M., JIANG, F., TAYLOR, D. W., STERNBERG, S. H., KAYA, E., MA, E., ANDERS, C., HAUER, M., ZHOU, K., LIN, S., KAPLAN, M., IAVARONE, A. T., CHARPENTIER, E., NOGALES, E. & DOUDNA, J. A. 2014. Structures of Cas9 endonucleases reveal RNA-mediated conformational activation. Science, 343, 1247997.

JONAS, J. B., MONES, J., GLACET-BERNARD, A. & COSCAS, G. 2017. Retinal Vein Occlusions. Dev Ophthalmol, 58, 139-167.

JONES, M. J., GOODMAN, S. J. & KOBOR, M. S. 2015. DNA methylation and healthy human aging. Aging Cell, 14, 924-32.

JONES, M. K., LU, B., SAGHIZADEH, M. & WANG, S. 2016. Gene expression changes in the retina following subretinal injection of human neural progenitor cells into a rodent model for retinal degeneration. Mol Vis, 22, 472-90.

JOUNG, J., KONERMANN, S., GOOTENBERG, J. S., ABUDAYYEH, O. O., PLATT, R. J., BRIGHAM, M. D., SANJANA, N. E. & ZHANG, F. 2017a. Genome-scale CRISPR-Cas9 knockout and transcriptional activation screening. nature protocols, 12, 828-863.

JOUNG, J., KONERMANN, S., GOOTENBERG, J. S., ABUDAYYEH, O. O., PLATT, R. J., BRIGHAM, M. D., SANJANA, N. E. & ZHANG, F. 2017b. Genome-scale CRISPR-Cas9 knockout and transcriptional activation screening. Nat Protoc, 12, 828-863.

JOVANOVIC, P., MIHAJLOVIC, M., DJORDJEVIC-JOCIC, J., VLAJKOVIC, S., CEKIC, S. & STEFANOVIC, V. 2013a. Ocular melanoma: an overview of the current status. Int J Clin Exp Pathol, 6, 1230-44.

JOVANOVIC, P., MIHAJLOVIC, M., DJORDJEVIC-JOCIC, J., VLAJKOVIC, S., CEKIC, S. & STEFANOVIC, V. 2013b. Ocular melanoma: an overview of the current status. International journal of clinical and experimental pathology, 6, 1230.

KABADI, A. M., OUSTEROUT, D. G., HILTON, I. B. & GERSBACH, C. A. 2014. Multiplex CRISPR/Cas9-based genome engineering from a single lentiviral vector. Nucleic Acids Res, 42, e147.

KATZ, Y., WANG, E. T., AIROLDI, E. M. & BURGE, C. B. 2010. Analysis and design of RNA sequencing experiments for identifying isoform regulation. Nat Methods, 7, 1009-15.

KAWANO, Y. & HONDA, A. 2017. Gene Targeting in Rabbits: Single-Step Generation of Knock-out Rabbits by Microinjection of CRISPR/Cas9 Plasmids. Methods Mol Biol, 1630, 109-120.

KAWASHIMA, N., NODA, S., YAMAMOTO, M. & OKIJI, T. 2017. Properties of Dental Pulp-derived Mesenchymal Stem Cells and the Effects of Culture Conditions. J Endod, 43, S31-S34.

KISKINIS, E. & EGGAN, K. 2010. Progress toward the clinical application of patient-specific pluripotent stem cells. J Clin Invest, 120, 51-9.

KLEIN, R. J., ZEISS, C., CHEW, E. Y., TSAI, J. Y., SACKLER, R. S., HAYNES, C., HENNING, A. K., SANGIOVANNI, J. P., MANE, S. M., MAYNE, S. T., BRACKEN, M. B., FERRIS, F. L., OTT, J., BARNSTABLE, C. & HOH, J. 2005. Complement factor H polymorphism in age-related macular degeneration. Science, 308, 385-9.

96

KOIKE-YUSA, H., LI, Y., TAN, E.-P., VELASCO-HERRERA, M. D. C. & YUSA, K. 2014. Genome-wide recessive genetic screening in mammalian cells with a lentiviral CRISPR-guide RNA library. Nature biotechnology, 32, 267-273.

KOKOTAS, H., KROUPIS, C., CHIRAS, D., GRIGORIADOU, M., LAMNISSOU, K., PETERSEN, M. B. & KITSOS, G. 2012. Biomarkers in primary open angle glaucoma. Clinical Chemistry and Laboratory Medicine, 50, 2107-2119.

KOLDE, R., LAUR, S., ADLER, P. & VILO, J. 2012. Robust rank aggregation for gene list integration and meta-analysis. Bioinformatics, 28, 573-80.

KRISHNADEV, N., MELETH, A. D. & CHEW, E. Y. 2010. Nutritional supplements for age-related macular degeneration. Curr Opin Ophthalmol, 21, 184-9.

KUJALA, E., MAKITIE, T. & KIVELA, T. 2003. Very long-term prognosis of patients with malignant uveal melanoma. Invest Ophthalmol Vis Sci, 44, 4651-9.

KWITEROVICH, K. A., MAGUIRE, M. G., MURPHY, R. P., SCHACHAT, A. P., BRESSLER, N. M., BRESSLER, S. B. & FINE, S. L. 1991. Frequency of adverse systemic reactions after fluorescein angiography. Results of a prospective study. Ophthalmology, 98, 1139-42.

KWON, Y. H., FINGERT, J. H., KUEHN, M. H. & ALWARD, W. L. 2009. Primary open-angle glaucoma. New England Journal of Medicine, 360, 1113-1124.

LABAJ, P. P., LEPARC, G. G., LINGGI, B. E., MARKILLIE, L. M., WILEY, H. S. & KREIL, D. P. 2011. Characterization and improvement of RNA-Seq precision in quantitative transcript expression profiling. Bioinformatics, 27, i383-91.

LAFRAMBOISE, T. 2009. Single nucleotide polymorphism arrays: a decade of biological, computational and technological advances. Nucleic Acids Res, 37, 4181-93.

LANDREVILLE, S., AGAPOVA, O. A. & HARBOUR, J. W. 2008. Emerging insights into the molecular pathogenesis of uveal melanoma. Future Oncol, 4, 629-36.

LANGENBERG, T., KAHANA, A., WSZALEK, J. A. & HALLORAN, M. C. 2008. The eye organizes neural crest cell migration. Developmental Dynamics, 237, 1645-1652.

LAURENT, L. C., ULITSKY, I., SLAVIN, I., TRAN, H., SCHORK, A., MOREY, R., LYNCH, C., HARNESS, J. V., LEE, S., BARRERO, M. J., KU, S., MARTYNOVA, M., SEMECHKIN, R., GALAT, V., GOTTESFELD, J., IZPISUA BELMONTE, J. C., MURRY, C., KEIRSTEAD, H. S., PARK, H. S., SCHMIDT, U., LASLETT, A. L., MULLER, F. J., NIEVERGELT, C. M., SHAMIR, R. & LORING, J. F. 2011. Dynamic changes in the copy number of pluripotency and cell proliferation genes in human ESCs and iPSCs during reprogramming and time in culture. Cell Stem Cell, 8, 106-18.

LEASHER, J. L., BOURNE, R. R., FLAXMAN, S. R., JONAS, J. B., KEEFFE, J., NAIDOO, K., PESUDOVS, K., PRICE, H., WHITE, R. A., WONG, T. Y., RESNIKOFF, S., TAYLOR, H. R. & VISION LOSS EXPERT GROUP OF THE GLOBAL BURDEN OF DISEASE, S. 2016. Global Estimates on the Number of People Blind or Visually Impaired by Diabetic Retinopathy: A Meta-analysis From 1990 to 2010. Diabetes Care, 39, 1643-9.

LESKE, M. C., CONNELL, A. M., WU, S. Y., HYMAN, L. G. & SCHACHAT, A. P. 1995. Risk factors for open-angle glaucoma. The Barbados Eye Study. Arch Ophthalmol, 113, 918-24.

LEVIN, J. Z., YASSOUR, M., ADICONIS, X., NUSBAUM, C., THOMPSON, D. A., FRIEDMAN, N., GNIRKE, A. & REGEV, A. 2010. Comprehensive comparative analysis of strand-specific RNA sequencing methods. Nat Methods, 7, 709-15.

LI, B. & DEWEY, C. N. 2011. RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome. BMC Bioinformatics, 12, 323.

97

LI, B. & LI, J. Z. 2014. A general framework for analyzing tumor subclonality using SNP array and DNA sequencing data. Genome Biol, 15, 473.

LI, J., WITTEN, D. M., JOHNSTONE, I. M. & TIBSHIRANI, R. 2012. Normalization, testing, and false discovery rate estimation for RNA-sequencing data. Biostatistics, 13, 523-38.

LI, J. J., JIANG, C. R., BROWN, J. B., HUANG, H. & BICKEL, P. J. 2011. Sparse linear modeling of next-generation mRNA sequencing (RNA-Seq) data for isoform discovery and abundance estimation. Proc Natl Acad Sci U S A, 108, 19867-72.

LI, W., XU, H., XIAO, T., CONG, L., LOVE, M. I., ZHANG, F., IRIZARRY, R. A., LIU, J. S., BROWN, M. & LIU, X. S. 2014. MAGeCK enables robust identification of essential genes from genome-scale CRISPR/Cas9 knockout screens. Genome Biol, 15, 554.

LIAO, Y., SMYTH, G. K. & SHI, W. 2014. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics, 30, 923-30.

LIN, S., STAAHL, B. T., ALLA, R. K. & DOUDNA, J. A. 2014. Enhanced homology-directed human genome engineering by controlled timing of CRISPR/Cas9 delivery. Elife, 3, e04766.

LIU, C. H., SUN, Y., LI, J., GONG, Y., TIAN, K. T., EVANS, L. P., MORSS, P. C., FREDRICK, T. W., SABA, N. J. & CHEN, J. 2015. Endothelial microRNA-150 is an intrinsic suppressor of pathologic ocular neovascularization. Proc Natl Acad Sci U S A, 112, 12163-8.

LIVAK, K. J. & SCHMITTGEN, T. D. 2001. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods, 25, 402-8.

LOPEZ-OTIN, C., BLASCO, M. A., PARTRIDGE, L., SERRANO, M. & KROEMER, G. 2013. The hallmarks of aging. Cell, 153, 1194-217.

LOSONCZY, G., FEKETE, A., VOKO, Z., TAKACS, L., KALDI, I., AJZNER, E., KASZA, M., VAJAS, A., BERTA, A. & BALOGH, I. 2011. Analysis of complement factor H Y402H, LOC387715, HTRA1 polymorphisms and ApoE alleles with susceptibility to age-related macular degeneration in Hungarian patients. Acta Ophthalmol, 89, 255-62.

LOVE, M. I., HUBER, W. & ANDERS, S. 2014. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol, 15, 550.

LU, B., LIN, Y., TSAI, Y., GIRMAN, S., ADAMUS, G., JONES, M. K., SHELLEY, B., SVENDSEN, C. N. & WANG, S. 2015. A Subsequent Human Neural Progenitor Transplant into the Degenerate Retina Does Not Compromise Initial Graft Survival or Therapeutic Efficacy. Transl Vis Sci Technol, 4, 7.

LU, J., GETZ, G., MISKA, E. A., ALVAREZ-SAAVEDRA, E., LAMB, J., PECK, D., SWEET-CORDERO, A., EBERT, B. L., MAK, R. H., FERRANDO, A. A., DOWNING, J. R., JACKS, T., HORVITZ, H. R. & GOLUB, T. R. 2005. MicroRNA expression profiles classify human cancers. Nature, 435, 834-8.

LUO, J. 2016. CRISPR/Cas9: From Genome Engineering to Cancer Drug Discovery. Trends Cancer, 2, 313-324.

MA, X. & ZHANG, X. 2013. NURD: an implementation of a new method to estimate isoform expression from non-uniform RNA-seq data. BMC Bioinformatics, 14, 220.

MAEGAWA, S., HINKAL, G., KIM, H. S., SHEN, L., ZHANG, L., ZHANG, J., ZHANG, N., LIANG, S., DONEHOWER, L. A. & ISSA, J. P. 2010. Widespread and tissue specific age-related DNA methylation changes in mice. Genome Res, 20, 332-40.

MALI, P., YANG, L., ESVELT, K. M., AACH, J., GUELL, M., DICARLO, J. E., NORVILLE, J. E. & CHURCH, G. M. 2013. RNA-guided human genome engineering via Cas9. Science, 339, 823-6.

98

MARTIN, E. M. & FRY, R. C. 2018. Environmental Influences on the Epigenome: Exposure- Associated DNA Methylation in Human Populations. Annu Rev Public Health, 39, 309-333.

MARTINS-TAYLOR, K., NISLER, B. S., TAAPKEN, S. M., COMPTON, T., CRANDALL, L., MONTGOMERY, K. D., LALANDE, M. & XU, R. H. 2011. Recurrent copy number variations in human induced pluripotent stem cells. Nat Biotechnol, 29, 488-91.

MARTTILA, S., KANANEN, L., HAYRYNEN, S., JYLHAVA, J., NEVALAINEN, T., HERVONEN, A., JYLHA, M., NYKTER, M. & HURME, M. 2015. Ageing-associated changes in the human DNA methylome: genomic locations and effects on gene expression. BMC Genomics, 16, 179.

MAUCK, R., YUAN, X. & TUAN, R. 2006. Chondrogenic differentiation and functional maturation of bovine mesenchymal stem cells in long-term agarose culture. Osteoarthritis and cartilage, 14, 179-189.

MCCLELLAND, A. D. & KANTHARIDIS, P. 2014. microRNA in the development of diabetic complications. Clin Sci (Lond), 126, 95-110.

MCDONNELL, F., O'BRIEN, C. & WALLACE, D. 2014. The role of epigenetics in the fibrotic processes associated with glaucoma. J Ophthalmol, 2014, 750459.

MCKAY, R., MCCARTY, C. A. & TAYLOR, H. R. 2000. Diabetic retinopathy in Victoria, Australia: the Visual Impairment Project. Br J Ophthalmol, 84, 865-70.

MCKINNON, S. J. 2003. Glaucoma: ocular Alzheimer's disease? Front Biosci, 8, s1140-56. MCLAUGHLIN, C. C., WU, X. C., JEMAL, A., MARTIN, H. J., ROCHE, L. M. & CHEN, V. W. 2005.

Incidence of noncutaneous melanomas in the U.S. Cancer, 103, 1000-7. MENENDEZ, L., KULIK, M. J., PAGE, A. T., PARK, S. S., LAUDERDALE, J. D., CUNNINGHAM, M.

L. & DALTON, S. 2013. Directed differentiation of human pluripotent cells to neural crest stem cells. Nature protocols, 8, 203-212.

MERKLE, F. T. & EGGAN, K. 2013. Modeling human disease with pluripotent stem cells: from genome association to function. Cell Stem Cell, 12, 656-68.

METSALU, T. & VILO, J. 2015. ClustVis: a web tool for visualizing clustering of multivariate data using Principal Component Analysis and heatmap. Nucleic Acids Res, 43, W566-70.

METZKER, M. L. 2010. Sequencing technologies - the next generation. Nat Rev Genet, 11, 31-46.

MEZLINI, A. M., SMITH, E. J., FIUME, M., BUSKE, O., SAVICH, G. L., SHAH, S., APARICIO, S., CHIANG, D. Y., GOLDENBERG, A. & BRUDNO, M. 2013. iReckon: simultaneous isoform discovery and abundance estimation from RNA-seq data. Genome Res, 23, 519-29.

MITCHELL, P., SMITH, W., ATTEBO, K. & HEALEY, P. R. 1996. Prevalence of open-angle glaucoma in Australia. The Blue Mountains Eye Study. Ophthalmology, 103, 1661-9.

MOHLER, C. W. & FINE, S. L. 1981. Long-term evaluation of patients with Best's vitelliform dystrophy. Ophthalmology, 88, 688-92.

MOJICA, F. J., DIEZ-VILLASENOR, C., SORIA, E. & JUEZ, G. 2000. Biological significance of a family of regularly spaced repeats in the genomes of Archaea, Bacteria and mitochondria. Mol Microbiol, 36, 244-6.

MORI-AKIYAMA, Y., AKIYAMA, H., ROWITCH, D. H. & DE CROMBRUGGHE, B. 2003. Sox9 is required for determination of the chondrogenic cell lineage in the cranial neural crest. Proceedings of the National Academy of Sciences, 100, 9360-9365.

99

MORTAZAVI, A., WILLIAMS, B. A., MCCUE, K., SCHAEFFER, L. & WOLD, B. 2008. Mapping and quantifying mammalian transcriptomes by RNA-Seq. Nat Methods, 5, 621-8.

MULLER, F. J., GOLDMANN, J., LOSER, P. & LORING, J. F. 2010. A call to standardize teratoma assays used to define human pluripotent cell lines. Cell Stem Cell, 6, 412-4.

MUNIER, A., GUNNING, T., KENNY, D. & O'KEEFE, M. 1998. Causes of blindness in the adult population of the Republic of Ireland. Br J Ophthalmol, 82, 630-3.

MUSTAFA, W. A., ABDUL-NASIR, A. S. & YAZID, H. 2018. Diabetic Retinopathy (DR) on Retinal Image: A Pilot Study. Journal of Physics: Conference Series, 1019.

NATHAN, P., COHEN, V., COUPLAND, S., CURTIS, K., DAMATO, B., EVANS, J., FENWICK, S., KIRKPATRICK, L., LI, O. & MARSHALL, E. 2015. Uveal melanoma UK national guidelines. European journal of cancer, 51, 2404-2412.

NIEMELÄ, S., MIETTINEN, S., SARKANEN, J. & ASHAMMAKHI, N. 2008. Adipose tissue and adipocyte differentiation: molecular and cellular aspects and tissue engineering applications. Topics in Tissue Engineering, 4, 26.

NISHIMASU, H., RAN, F. A., HSU, P. D., KONERMANN, S., SHEHATA, S. I., DOHMAE, N., ISHITANI, R., ZHANG, F. & NUREKI, O. 2014. Crystal structure of Cas9 in complex with guide RNA and target DNA. Cell, 156, 935-49.

NIXON, B., STANGER, S. J., MIHALAS, B. P., REILLY, J. N., ANDERSON, A. L., DUN, M. D., TYAGI, S., HOLT, J. E. & MCLAUGHLIN, E. A. 2015. Next Generation Sequencing Analysis Reveals Segmental Patterns of microRNA Expression in Mouse Epididymal Epithelial Cells. PLoS One, 10, e0135605.

NOH, J. A. B., C. 2015. sgRSEA: Enrichment Analysis of CRISPR/Cas9 Knockout Screen Data. NORDSTROM, S. & BARKMAN, Y. 1977. Hereditary maculardegeneration (HMD) in 246 cases

traced to one gene-source in central Sweden. Hereditas, 84, 163-76. O'CONNOR, C. 2008. Karyotyping for chromosomal abnormalities. Nature Education OLLIKAINEN, M., SMITH, K. R., JOO, E. J., NG, H. K., ANDRONIKOS, R., NOVAKOVIC, B., ABDUL

AZIZ, N. K., CARLIN, J. B., MORLEY, R., SAFFERY, R. & CRAIG, J. M. 2010. DNA methylation analysis of multiple tissues from newborn twins reveals both genetic and intrauterine components to variation in the human neonatal epigenome. Hum Mol Genet, 19, 4176-88.

OTTESON, D. C. 2011. Eyes on DNA methylation: current evidence for DNA methylation in ocular development and disease. J Ocul Biol Dis Infor, 4, 95-103.

OZSOLAK, F. & MILOS, P. M. 2011. RNA sequencing: advances, challenges and opportunities. Nat Rev Genet, 12, 87-98.

PARK, I. H., ZHAO, R., WEST, J. A., YABUUCHI, A., HUO, H., INCE, T. A., LEROU, P. H., LENSCH, M. W. & DALEY, G. Q. 2008. Reprogramming of human somatic cells to pluripotency with defined factors. Nature, 451, 141-6.

PARKHOMCHUK, D., BORODINA, T., AMSTISLAVSKIY, V., BANARU, M., HALLEN, L., KROBITSCH, S., LEHRACH, H. & SOLDATOV, A. 2009. Transcriptome analysis by strand-specific sequencing of complementary DNA. Nucleic Acids Res, 37, e123.

PARMET, S., LYNM, C. & GLASS, R. M. 2006. Age-Related Macular Degeneration. JAMA, 295, 2438-2438.

PASCOLINI, D. & MARIOTTI, S. P. 2012. Global estimates of visual impairment: 2010. Br J Ophthalmol, 96, 614-8.

100

PATRO, R., MOUNT, S. M. & KINGSFORD, C. 2014. Sailfish enables alignment-free isoform quantification from RNA-seq reads using lightweight algorithms. Nat Biotechnol, 32, 462-4.

PAULAVIČIŪTĖ-BAIKŠTIENĖ, D., BARŠAUSKAITĖ, R. & JANULEVIČIENĖ, I. 2013. New insights into pathophysiological mechanisms regulating conventional aqueous humor outflow. Medicina (Kaunas), 49, 165-9.

PENNINGTON, K. L. & DEANGELIS, M. M. 2016. Epidemiology of age-related macular degeneration (AMD): associations with cardiovascular disease phenotypes and lipid factors. Eye Vis (Lond), 3, 34.

PEREIRA, P. R., ODASHIRO, A. N., LIM, L. A., MIYAMOTO, C., BLANCO, P. L., ODASHIRO, M., MALONEY, S., DE SOUZA, D. F. & BURNIER, M. N., JR. 2013. Current and emerging treatment options for uveal melanoma. Clin Ophthalmol, 7, 1669-82.

PERERA, O. P., LITTLE, N. S. & PIERCE, C. A., 3RD 2018. CRISPR/Cas9 mediated high efficiency knockout of the eye color gene Vermillion in Helicoverpa zea (Boddie). PLoS One, 13, e0197567.

PERTEA, M., PERTEA, G. M., ANTONESCU, C. M., CHANG, T. C., MENDELL, J. T. & SALZBERG, S. L. 2015. StringTie enables improved reconstruction of a transcriptome from RNA-seq reads. Nat Biotechnol, 33, 290-5.

PERVAN, C. L. 2017. Smad-independent TGF-beta2 signaling pathways in human trabecular meshwork cells. Exp Eye Res, 158, 137-145.

PETERSON, S. E. & LORING, J. F. 2014. Genomic instability in pluripotent stem cells: implications for clinical applications. J Biol Chem, 289, 4578-84.

PETRONIS, A. 2010. Epigenetics as a unifying principle in the aetiology of complex traits and diseases. Nature, 465, 721-7.

PIMENTEL, H., BRAY, N. L., PUENTE, S., MELSTED, P. & PACHTER, L. 2017. Differential analysis of RNA-seq incorporating quantification uncertainty. Nat Methods, 14, 687-690.

PINKEL, D., LANDEGENT, J., COLLINS, C., FUSCOE, J., SEGRAVES, R., LUCAS, J. & GRAY, J. 1988. Fluorescence in situ hybridization with human chromosome-specific libraries: detection of trisomy 21 and translocations of chromosome 4. Proc Natl Acad Sci U S A, 85, 9138-42.

POLLEN, A. A., NOWAKOWSKI, T. J., SHUGA, J., WANG, X., LEYRAT, A. A., LUI, J. H., LI, N., SZPANKOWSKI, L., FOWLER, B., CHEN, P., RAMALINGAM, N., SUN, G., THU, M., NORRIS, M., LEBOFSKY, R., TOPPANI, D., KEMP, D. W., 2ND, WONG, M., CLERKSON, B., JONES, B. N., WU, S., KNUTSSON, L., ALVARADO, B., WANG, J., WEAVER, L. S., MAY, A. P., JONES, R. C., UNGER, M. A., KRIEGSTEIN, A. R. & WEST, J. A. 2014. Low-coverage single-cell mRNA sequencing reveals cellular heterogeneity and activated signaling pathways in developing cerebral cortex. Nat Biotechnol, 32, 1053-8.

PYNE, M. E., MOO-YOUNG, M., CHUNG, D. A. & CHOU, C. P. 2015. Coupling the CRISPR/Cas9 System with Lambda Red Recombineering Enables Simplified Chromosomal Gene Replacement in Escherichia coli. Appl Environ Microbiol, 81, 5103-14.

QIAN, S.-W., LI, X., ZHANG, Y.-Y., HUANG, H.-Y., LIU, Y., SUN, X. & TANG, Q.-Q. 2010. Characterization of adipocyte differentiation from human mesenchymal stem cells in bone marrow. BMC developmental biology, 10, 47.

QUIGLEY, H. A. 2011a. Glaucoma. Lancet, 377, 1367-77. QUIGLEY, H. A. 2011b. Glaucoma. The Lancet, 377, 1367-1377.

101

QUINZII, C. M., HIRANO, M. & DIMAURO, S. 2014. Mutant COQ2 in multiple-system atrophy. N Engl J Med, 371, 81-2.

RAGHUNATH, A. & PERUMAL, E. 2015. Micro-RNAs and their roles in eye disorders. Ophthalmic Res, 53, 169-86.

RAKYAN, V. K., DOWN, T. A., MASLAU, S., ANDREW, T., YANG, T. P., BEYAN, H., WHITTAKER, P., MCCANN, O. T., FINER, S., VALDES, A. M., LESLIE, R. D., DELOUKAS, P. & SPECTOR, T. D. 2010. Human aging-associated DNA hypermethylation occurs preferentially at bivalent chromatin domains. Genome Res, 20, 434-9.

RAN, F. A., HSU, P. D., LIN, C.-Y., GOOTENBERG, J. S., KONERMANN, S., TREVINO, A. E., SCOTT, D. A., INOUE, A., MATOBA, S. & ZHANG, Y. 2013. Double nicking by RNA-guided CRISPR Cas9 for enhanced genome editing specificity. Cell, 154, 1380-1389.

RAPLEY, RALPH & HARBRON & STUART, J. M. 2004 Molecular analysis and genome discovery. J. Wiley.

RATAN, Z. A., ZAMAN, S. B., MEHTA, V., HAIDERE, M. F., RUNA, N. J. & AKTER, N. 2017. Application of Fluorescence In Situ Hybridization (FISH) Technique for the Detection of Genetic Aberration in Medical Science. Cureus, 9, e1325.

REIMAND, J., ARAK, T., ADLER, P., KOLBERG, L., REISBERG, S., PETERSON, H. & VILO, J. 2016. g:Profiler-a web server for functional interpretation of gene lists (2016 update). Nucleic Acids Res, 44, W83-9.

REIMAND, J., ISSERLIN, R., VOISIN, V., KUCERA, M., TANNUS-LOPES, C., ROSTAMIANFAR, A., WADI, L., MEYER, M., WONG, J., XU, C., MERICO, D. & BADER, G. D. 2019. Pathway enrichment analysis and visualization of omics data using g:Profiler, GSEA, Cytoscape and EnrichmentMap. Nat Protoc, 14, 482-517.

REIMAND, J., KULL, M., PETERSON, H., HANSEN, J. & VILO, J. 2007. g:Profiler--a web-based toolset for functional profiling of gene lists from large-scale experiments. Nucleic Acids Res, 35, W193-200.

REPORT, W. REUBINOFF, B. E., PERA, M. F., FONG, C. Y., TROUNSON, A. & BONGSO, A. 2000. Embryonic

stem cell lines from human blastocysts: somatic differentiation in vitro. Nat Biotechnol, 18, 399-404.

RIGGS, A. D. 1975. X inactivation, differentiation, and DNA methylation. Cytogenet Cell Genet, 14, 9-25.

RISSO, D., NGAI, J., SPEED, T. P. & DUDOIT, S. 2014. Normalization of RNA-seq data using factor analysis of control genes or samples. Nat Biotechnol, 32, 896-902.

ROBERT, C., DUMMER, R., GUTZMER, R., LORIGAN, P., KIM, K. B., NYAKAS, M., ARANCE, A., LISZKAY, G., SCHADENDORF, D., CANTARINI, M., SPENCER, S. & MIDDLETON, M. R. 2013. Selumetinib plus dacarbazine versus placebo plus dacarbazine as first-line treatment for BRAF-mutant metastatic melanoma: a phase 2 double-blind randomised study. Lancet Oncol, 14, 733-40.

ROBERTS, A. & PACHTER, L. 2013. Streaming fragment assignment for real-time analysis of sequencing experiments. Nat Methods, 10, 71-3.

ROBERTS, A., PIMENTEL, H., TRAPNELL, C. & PACHTER, L. 2011. Identification of novel transcripts in annotated genomes using RNA-Seq. Bioinformatics, 27, 2325-9.

ROBINSON, M. D., MCCARTHY, D. J. & SMYTH, G. K. 2010. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics, 26, 139-40.

102

ROSS, E. L., HUTTON, D. W., STEIN, J. D., BRESSLER, N. M., JAMPOL, L. M., GLASSMAN, A. R. & DIABETIC RETINOPATHY CLINICAL RESEARCH, N. 2016. Cost-effectiveness of Aflibercept, Bevacizumab, and Ranibizumab for Diabetic Macular Edema Treatment: Analysis From the Diabetic Retinopathy Clinical Research Network Comparative Effectiveness Trial. JAMA Ophthalmol, 134, 888-96.

SANJANA, N. E., SHALEM, O. & ZHANG, F. 2014. Improved vectors and genome-wide libraries for CRISPR screening. Nat Methods, 11, 783-784.

SARDA, S. & HANNENHALLI, S. 2014. Next-generation sequencing and epigenomics research: a hammer in search of nails. Genomics Inform, 12, 2-11.

SATIH, S., CHALABI, N., RABIAU, N., BOSVIEL, R., FONTANA, L., BIGNON, Y. J. & BERNARD-GALLON, D. J. 2010. Gene expression profiling of breast cancer cell lines in response to soy isoflavones using a pangenomic microarray approach. OMICS, 14, 231-8.

SATO-OTSUBO, A., SANADA, M. & OGAWA, S. 2012. Single-nucleotide polymorphism array karyotyping in clinical practice: where, when, and how? Semin Oncol, 39, 13-25.

SCHULZ, M. H., ZERBINO, D. R., VINGRON, M. & BIRNEY, E. 2012. Oases: robust de novo RNA-seq assembly across the dynamic range of expression levels. Bioinformatics, 28, 1086-92.

SEN, P., SHAH, P. P., NATIVIO, R. & BERGER, S. L. 2016. Epigenetic Mechanisms of Longevity and Aging. Cell, 166, 822-839.

SHALEM, O., SANJANA, N. E., HARTENIAN, E., SHI, X., SCOTT, D. A., MIKKELSON, T., HECKL, D., EBERT, B. L., ROOT, D. E., DOENCH, J. G. & ZHANG, F. 2014. Genome-scale CRISPR-Cas9 knockout screening in human cells. Science, 343, 84-87.

SHALEM, O., SANJANA, N. E. & ZHANG, F. 2015. High-throughput functional genomics using CRISPR-Cas9. Nature Reviews Genetics, 16, 299-311.

SHANNON, P., MARKIEL, A., OZIER, O., BALIGA, N. S., WANG, J. T., RAMAGE, D., AMIN, N., SCHWIKOWSKI, B. & IDEKER, T. 2003. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res, 13, 2498-504.

SHAROW, K. A., TEMKIN, B. & ANN ASSON-BATRES, M. 2012. Retinoic acid stability in stem cell cultures. International Journal of Developmental Biology, 56, 273.

SHEFFIELD, V. C. & STONE, E. M. 2011. Genomics and the eye. N Engl J Med, 364, 1932-42. SHEN, J., YANG, X., XIE, B., CHEN, Y., SWAIM, M., HACKETT, S. F. & CAMPOCHIARO, P. A.

2008. MicroRNAs regulate ocular neovascularization. Mol Ther, 16, 1208-16. SHINAWI, M. & CHEUNG, S. W. 2008. The array CGH and its clinical applications. Drug Discov

Today, 13, 760-70. SIMARA, P., MOTL, J. A. & KAUFMAN, D. S. 2013. Pluripotent stem cells and gene therapy.

Transl Res, 161, 284-92. SIMS, D., SUDBERY, I., ILOTT, N. E., HEGER, A. & PONTING, C. P. 2014. Sequencing depth and

coverage: key considerations in genomic analyses. Nat Rev Genet, 15, 121-32. SINGH, A. D., RENNIE, I. G., SEREGARD, S., GIBLIN, M. & MCKENZIE, J. 2004. Sunlight

exposure and pathogenesis of uveal melanoma. Surv Ophthalmol, 49, 419-28. SINGH, A. D., TURELL, M. E. & TOPHAM, A. K. 2011. Uveal melanoma: trends in incidence,

treatment, and survival. Ophthalmology, 118, 1881-5. SINGH, R., SHEN, W., KUAI, D., MARTIN, J. M., GUO, X., SMITH, M. A., PEREZ, E. T., PHILLIPS,

M. J., SIMONETT, J. M., WALLACE, K. A., VERHOEVEN, A. D., CAPOWSKI, E. E., ZHANG, X., YIN, Y., HALBACH, P. J., FISHMAN, G. A., WRIGHT, L. S., PATTNAIK, B. R. & GAMM, D. M. 2013. iPS cell modeling of Best disease: insights into the pathophysiology of an inherited macular degeneration. Hum Mol Genet, 22, 593-607.

103

SMITH, J. M. & STEEL, D. H. 2015. Anti-vascular endothelial growth factor for prevention of postoperative vitreous cavity haemorrhage after vitrectomy for proliferative diabetic retinopathy. Cochrane Database Syst Rev, CD008214.

SONESON, C. & DELORENZI, M. 2013. A comparison of methods for differential expression analysis of RNA-seq data. BMC Bioinformatics, 14, 91.

SPAGNOLO, F., GROSSO, M., PICASSO, V., TORNARI, E., PESCE, M. & QUEIROLO, P. 2013. Treatment of metastatic uveal melanoma with intravenous fotemustine. Melanoma Res, 23, 196-8.

STAMER, W. D. & ACOTT, T. S. 2012. Current understanding of conventional outflow dysfunction in glaucoma. Current opinion in ophthalmology, 23, 135.

STAMER, W. D., SEFTOR, R. E., WILLIAMS, S. K., SAMAHA, H. A. & SNYDER, R. W. 1995. Isolation and culture of human trabecular meshwork cells by extracellular matrix digestion. Current eye research, 14, 611-617.

STANNARD, C., SAUERWEIN, W., MAREE, G. & LECUONA, K. 2013. Radiotherapy for ocular tumours. Eye (Lond), 27, 119-27.

STEIJGER, T., ABRIL, J. F., ENGSTROM, P. G., KOKOCINSKI, F., HUBBARD, T. J., GUIGO, R., HARROW, J., BERTONE, P. & CONSORTIUM, R. 2013. Assessment of transcript reconstruction methods for RNA-seq. Nat Methods, 10, 1177-84.

STERNBERG, S. H., REDDING, S., JINEK, M., GREENE, E. C. & DOUDNA, J. A. 2014. DNA interrogation by the CRISPR RNA-guided endonuclease Cas9. Nature, 507, 62-7.

STONE, E. M., LOTERY, A. J., MUNIER, F. L., HEON, E., PIGUET, B., GUYMER, R. H., VANDENBURGH, K., COUSIN, P., NISHIMURA, D., SWIDERSKI, R. E., SILVESTRI, G., MACKEY, D. A., HAGEMAN, G. S., BIRD, A. C., SHEFFIELD, V. C. & SCHORDERET, D. F. 1999. A single EFEMP1 mutation associated with both Malattia Leventinese and Doyne honeycomb retinal dystrophy. Nat Genet, 22, 199-202.

SUJOBERT, P., CUCCUINI, W., VIGNON-PENNAMEN, D., MARTIN-GARCIA, N., ALBERTINI, A. F., UZUNOV, M., REDJOUL, R., DOMBRET, H. & RAFFOUX, E. 2013. Evidence of differentiation in myeloid malignancies associated neutrophilic dermatosis: a fluorescent in situ hybridization study of 14 patients. J Invest Dermatol, 133, 1111-4.

SZENTAL, J. A., HARPER, C. A., BAIRD, P. N., MICHALOVA, K. & GUYMER, R. H. 2009. Sorsby's Fundus Dystrophy: a case report to raise awareness of the disease and potential future treatments. Clin Experiment Ophthalmol, 37, 325-7.

TAAPKEN, S. M., NISLER, B. S., NEWTON, M. A., SAMPSELL-BARRON, T. L., LEONHARD, K. A., MCINTIRE, E. M. & MONTGOMERY, K. D. 2011. Karotypic abnormalities in human induced pluripotent stem cells and embryonic stem cells. Nat Biotechnol, 29, 313-4.

TAKÁCS, R., MATTA, C., SOMOGYI, C., JUHÁSZ, T. & ZÁKÁNY, R. 2013. Comparative analysis of osteogenic/chondrogenic differentiation potential in primary limb bud-derived and C3H10T1/2 cell line-based mouse micromass cultures. International journal of molecular sciences, 14, 16141-16167.

TAKAHASHI, K. & YAMANAKA, S. 2006. Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell, 126, 663-76.

TARAZONA, S., GARCIA-ALCALDE, F., DOPAZO, J., FERRER, A. & CONESA, A. 2011. Differential expression in RNA-seq: a matter of depth. Genome Res, 21, 2213-23.

TEAM, R. 2015. RStudio: Integrated Development for R. RStudio. TELEGINA, D. V., KORBOLINA, E. E., ERSHOV, N. I., KOLOSOVA, N. G. & KOZHEVNIKOVA, O. S.

2015. Identification of functional networks associated with cell death in the retina of OXYS rats during the development of retinopathy. Cell Cycle, 14, 3544-56.

104

TESCHENDORFF, A. E., WEST, J. & BECK, S. 2013. Age-associated epigenetic drift: implications, and a case of epigenetic thrift? Hum Mol Genet, 22, R7-R15.

THAPAR, A. & COOPER, M. 2013. Copy number variation: what is it and what has it told us about child psychiatric disorders? J Am Acad Child Adolesc Psychiatry, 52, 772-4.

THOMAS, S., PUTTER, C., WEBER, S., BORNFELD, N., LOHMANN, D. R. & ZESCHNIGK, M. 2012. Prognostic significance of chromosome 3 alterations determined by microsatellite analysis in uveal melanoma: a long-term follow-up study. Br J Cancer, 106, 1171-6.

THOMSON, J. A., ITSKOVITZ-ELDOR, J., SHAPIRO, S. S., WAKNITZ, M. A., SWIERGIEL, J. J., MARSHALL, V. S. & JONES, J. M. 1998. Embryonic stem cell lines derived from human blastocysts. Science, 282, 1145-7.

TRAPNELL, C., WILLIAMS, B. A., PERTEA, G., MORTAZAVI, A., KWAN, G., VAN BAREN, M. J., SALZBERG, S. L., WOLD, B. J. & PACHTER, L. 2010. Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat Biotechnol, 28, 511-5.

TRIPATHI, R. C., LI, J., CHAN, W. A. & TRIPATHI, B. J. 1994. Aqueous humor in glaucomatous eyes contains an increased level of TGF-β2. Experimental eye research, 59, 723-728.

TROYANSKAYA, O., CANTOR, M., SHERLOCK, G., BROWN, P., HASTIE, T., TIBSHIRANI, R., BOTSTEIN, D. & ALTMAN, R. B. 2001. Missing value estimation methods for DNA microarrays. Bioinformatics, 17, 520-5.

VAJDIC, C. M., KRICKER, A., GIBLIN, M., MCKENZIE, J., AITKEN, J., GILES, G. G. & ARMSTRONG, B. K. 2002. Sun exposure predicts risk of ocular melanoma in Australia. International journal of cancer, 101, 175-182.

VARTAK, S. V. & RAGHAVAN, S. C. 2015. Inhibition of nonhomologous end joining to increase the specificity of CRISPR/Cas9 genome editing. FEBS journal, 282, 4289-4294.

VERMEESCH, J. R., MELOTTE, C., FROYEN, G., VAN VOOREN, S., DUTTA, B., MAAS, N., VERMEULEN, S., MENTEN, B., SPELEMAN, F., DE MOOR, B., VAN HUMMELEN, P., MARYNEN, P., FRYNS, J. P. & DEVRIENDT, K. 2005. Molecular karyotyping: array CGH quality criteria for constitutional genetic diagnosis. J Histochem Cytochem, 53, 413-22.

VIDIGAL, J. A. & VENTURA, A. 2015. The biological functions of miRNAs: lessons from in vivo studies. Trends Cell Biol, 25, 137-47.

VINEKAR, A., CHIDAMBARA, L., JAYADEV, C., SIVAKUMAR, M., WEBERS, C. A. & SHETTY, B. 2016. Monitoring neovascularization in aggressive posterior retinopathy of prematurity using optical coherence tomography angiography. J AAPOS, 20, 271-4.

VIRGILI, G., GATTA, G., CICCOLALLO, L., CAPOCACCIA, R., BIGGERI, A., CROCETTI, E., LUTZ, J. M., PACI, E. & GROUP, E. W. 2007. Incidence of uveal melanoma in Europe. Ophthalmology, 114, 2309-15.

WAKAMATSU, K., HU, D.-N., MCCORMICK, S. A. & ITO, S. 2008. Characterization of melanin in human iridal and choroidal melanocytes from eyes with various colored irides. Pigment Cell & Melanoma Research, 21, 97-105.

WANG, Q., TIFFEN, J., BAILEY, C. G., LEHMAN, M. L., RITCHIE, W., FAZLI, L., METIERRE, C., FENG, Y. J., LI, E., GLEAVE, M., BUCHANAN, G., NELSON, C. C., RASKO, J. E. & HOLST, J. 2013. Targeting amino acid transport in metastatic castration-resistant prostate cancer: effects on cell cycle, cell growth, and tumor development. J Natl Cancer Inst, 105, 1463-73.

105

WANG, S., HAN, H., HU, Y., YANG, W., LV, Y., WANG, L., ZHANG, L. & JI, J. 2017a. SLC3A2, antigen of mAb 3G9, promotes migration and invasion by upregulating of mucins in gastric cancer. Oncotarget, 8, 88586-88598.

WANG, T., WEI, J. J., SABATINI, D. M. & LANDER, E. S. 2014. Genetic screens in human cells using the CRISPR-Cas9 system. Science, 343, 80-4.

WANG, T. L., MAIERHOFER, C., SPEICHER, M. R., LENGAUER, C., VOGELSTEIN, B., KINZLER, K. W. & VELCULESCU, V. E. 2002. Digital karyotyping. Proc Natl Acad Sci U S A, 99, 16156-61.

WANG, X., ZHENG, Z., CAI, Y., CHEN, T., LI, C., FU, W. & JIANG, Y. 2017b. CNVcaller: highly efficient and widely applicable software for detecting copy number variations in large populations. Gigascience, 6, 1-12.

WANG, Y., WU, S., YANG, Y., PENG, F., LI, Q., TIAN, P., XIANG, E., LIANG, H., WANG, B., ZHOU, X., HUANG, H. & ZHOU, X. 2017c. Differentially expressed miRNAs in oxygeninduced retinopathy newborn mouse models. Mol Med Rep, 15, 146-152.

WANG, Z., GERSTEIN, M. & SNYDER, M. 2009. RNA-Seq: a revolutionary tool for transcriptomics. Nat Rev Genet, 10, 57-63.

WEINREB, R. N. & LINDSEY, J. D. 2005. The importance of models in glaucoma research. Journal of glaucoma, 14, 302-304.

WEIS, E., SHAH, C. P., LAJOUS, M., SHIELDS, J. A. & SHIELDS, C. L. 2006. The association between host susceptibility factors and uveal melanoma: a meta-analysis. Archives of ophthalmology, 124, 54-60.

WEISSBEIN, U., BENVENISTY, N. & BEN-DAVID, U. 2014. Quality control: Genome maintenance in pluripotent stem cells. J Cell Biol, 204, 153-63.

WICKHAM, H. 2016. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York. WILLOUGHBY, C. E., PONZIN, D., FERRARI, S., LOBO, A., LANDAU, K. & OMIDI, Y. 2010.

Anatomy and physiology of the human eye: effects of mucopolysaccharidoses disease on structure and function – a review. Clinical & Experimental Ophthalmology, 38.

WINTER, J., BREINIG, M., HEIGWER, F., BRUGEMANN, D., LEIBLE, S., PELZ, O., ZHAN, T. & BOUTROS, M. 2016. caRpools: an R package for exploratory data analysis and documentation of pooled CRISPR/Cas9 screens. Bioinformatics, 32, 632-4.

WINTER, J., SCHWERING, M., PELZ, O., RAUSCHER, B., ZHAN, T., HEIGWER, F. & BOUTROS, M. 2017. CRISPRAnalyzeR: Interactive analysis, annotation and documentation of pooled CRISPR screens. bioRxiv, 109967.

WONG, W. L., SU, X., LI, X., CHEUNG, C. M., KLEIN, R., CHENG, C. Y. & WONG, T. Y. 2014. Global prevalence of age-related macular degeneration and disease burden projection for 2020 and 2040: a systematic review and meta-analysis. Lancet Glob Health, 2, e106-16.

WORDINGER, R. J., FLEENOR, D. L., HELLBERG, P. E., PANG, I.-H., TOVAR, T. O., ZODE, G. S., FULLER, J. A. & CLARK, A. F. 2007. Effects of TGF-β2, BMP-4, and gremlin in the trabecular meshwork: implications for glaucoma. Investigative ophthalmology & visual science, 48, 1191-1200.

XIAO, Q., HARTZELL, H. C. & YU, K. 2010. Bestrophins and retinopathies. Pflugers Arch, 460, 559-69.

XIE, Y., WU, G., TANG, J., LUO, R., PATTERSON, J., LIU, S., HUANG, W., HE, G., GU, S., LI, S., ZHOU, X., LAM, T. W., LI, Y., XU, X., WONG, G. K. & WANG, J. 2014. SOAPdenovo-

106

Trans: de novo transcriptome assembly with short RNA-Seq reads. Bioinformatics, 30, 1660-6.

YAN, L., LEE, S., LAZZARO, D. R., ARANDA, J., GRANT, M. B. & CHAQOUR, B. 2015. Single and Compound Knock-outs of MicroRNA (miRNA)-155 and Its Angiogenic Gene Target CCN1 in Mice Alter Vascular and Neovascular Growth in the Retina via Resident Microglia. J Biol Chem, 290, 23264-81.

YANG, H. J., RATNAPRIYA, R., COGLIATI, T., KIM, J. W. & SWAROOP, A. 2015. Vision from next generation sequencing: multi-dimensional genome-wide analysis for producing gene regulatory networks underlying retinal development, aging and disease. Prog Retin Eye Res, 46, 1-30.

YANG, J., MANSON, D. K., MARR, B. P. & CARVAJAL, R. D. 2018. Treatment of uveal melanoma: where are we now? Ther Adv Med Oncol, 10, 1758834018757175.

YASUDA, M., TANAKA, Y., RYU, M., TSUDA, S. & NAKAZAWA, T. 2014. RNA sequence reveals mouse retinal transcriptome changes early after axonal injury. PLoS One, 9, e93258.

YORSTON, D. 2014. Anti-VEGF drugs in the prevention of blindness. Community Eye Health, 27, 44-6.

YU, J., SILVA, J. & CALIFANO, A. 2016. ScreenBEAM: a novel meta-analysis algorithm for functional genomics screens via Bayesian hierarchical modeling. Bioinformatics, 32, 260-7.

YU, J., VODYANIK, M. A., SMUGA-OTTO, K., ANTOSIEWICZ-BOURGET, J., FRANE, J. L., TIAN, S., NIE, J., JONSDOTTIR, G. A., RUOTTI, V., STEWART, R., SLUKVIN, II & THOMSON, J. A. 2007. Induced pluripotent stem cell lines derived from human somatic cells. Science, 318, 1917-20.

YU-WAI-MAN, C., OWEN, N., LEES, J., TAGALAKIS, A. D., HART, S. L., WEBSTER, A. R., ORENGO, C. A. & KHAW, P. T. 2017. Genome-wide RNA-Sequencing analysis identifies a distinct fibrosis gene signature in the conjunctiva after glaucoma surgery. Sci Rep, 7, 5644.

YUAN, T., JIAO, Y., DE JONG, S., OPHOFF, R. A., BECK, S. & TESCHENDORFF, A. E. 2015. An integrative multi-scale analysis of the dynamic DNA methylation landscape in aging. PLoS Genet, 11, e1004996.

ZHANG, N., LI, J. H., WANG, J. A. & ZHANG, H. K. 2008. [The effect of bone marrow mesenchymal stem cell transplantation on diabetic cardiomyopathy]. Zhonghua Xin Xue Guan Bing Za Zhi, 36, 1115-9.

ZHANG, Q., VALLERGA, C. L., WALKER, R. M., LIN, T., HENDERS, A. K., MONTGOMERY, G. W., HE, J., FAN, D., FOWDAR, J., KENNEDY, M., PITCHER, T., PEARSON, J., HALLIDAY, G., KWOK, J. B., HICKIE, I., LEWIS, S., ANDERSON, T., SILBURN, P. A., MELLICK, G. D., HARRIS, S. E., REDMOND, P., MURRAY, A. D., PORTEOUS, D. J., HALEY, C. S., EVANS, K. L., MCINTOSH, A. M., YANG, J., GRATTEN, J., MARIONI, R. E., WRAY, N. R., DEARY, I. J., MCRAE, A. F. & VISSCHER, P. M. 2019. Improved precision of epigenetic clock estimates across tissues and its implication for biological ageing. Genome Med, 11, 54.

ZHANG, T., XIE, X., CAO, G., JIANG, H., WU, S., SU, Z., ZHANG, K. & LU, F. 2014a. Malattia leventinese/Doyne honeycomb retinal dystrophy in a chinese family with mutation of the EFEMP1 gene. Retina, 34, 2462-71.

ZHANG, X., CLARK, A. F. & YORIO, T. 2003. Interactions of endothelin-1 with dexamethasone in primary cultured human trabecular meshwork cells. Investigative ophthalmology & visual science, 44, 5301-5308.

107

ZHANG, Z. H., JHAVERI, D. J., MARSHALL, V. M., BAUER, D. C., EDSON, J., NARAYANAN, R. K., ROBINSON, G. J., LUNDBERG, A. E., BARTLETT, P. F., WRAY, N. R. & ZHAO, Q. Y. 2014b. A comparative study of techniques for differential expression analysis on RNA-Seq data. PLoS One, 9, e103207.

ZHAO, D., NGUYEN, C. T., WONG, V. H., LIM, J. K., HE, Z., JOBLING, A. I., FLETCHER, E. L., CHINNERY, H. R., VINGRYS, A. J. & BUI, B. V. 2017. Characterization of the Circumlimbal Suture Model of Chronic IOP Elevation in Mice and Assessment of Changes in Gene Expression of Stretch Sensitive Channels. Front Neurosci, 11, 41.

ZHOU, Y., ZHU, S., CAI, C., YUAN, P., LI, C., HUANG, Y. & WEI, W. 2014. High-throughput screening of a CRISPR/Cas9 library for functional genomics in human cells. Nature, 509, 487-491.

ZHU, B., CHENG, D., HOU, L., ZHOU, S., YING, T. & YANG, Q. 2017. SLC3A2 is upregulated in human osteosarcoma and promotes tumor growth through the PI3K/Akt signaling pathway. Oncol Rep, 37, 2575-2582.

Uncategorized References

BAI, Y., BAI, X., WANG, Z., ZHANG, X., RUAN, C. & MIAO, J. 2011. MicroRNA-126 inhibits ischemia-induced retinal neovascularization via regulating angiogenic growth factors. Exp Mol Pathol, 91, 471-7.

COLELLA, S., YAU, C., TAYLOR, J. M., MIRZA, G., BUTLER, H., CLOUSTON, P., BASSETT, A. S., SELLER, A., HOLMES, C. C. & RAGOUSSIS, J. 2007. QuantiSNP: an Objective Bayes Hidden-Markov Model to detect and accurately map copy number variation using SNP genotyping data. Nucleic Acids Res, 35, 2013-25.

HANNUM, G., GUINNEY, J., ZHAO, L., ZHANG, L., HUGHES, G., SADDA, S., KLOTZLE, B., BIBIKOVA, M., FAN, J. B., GAO, Y., DECONDE, R., CHEN, M., RAJAPAKSE, I., FRIEND, S., IDEKER, T. & ZHANG, K. 2013. Genome-wide methylation profiles reveal quantitative views of human aging rates. Mol Cell, 49, 359-67.

HORVATH, S. 2013. DNA methylation age of human tissues and cell types. Genome Biol, 14, R115.

INTERNATIONAL STEM CELL, I., AMPS, K., ANDREWS, P. W., ANYFANTIS, G., ARMSTRONG, L., AVERY, S., BAHARVAND, H., BAKER, J., BAKER, D., MUNOZ, M. B., BEIL, S., BENVENISTY, N., BEN-YOSEF, D., BIANCOTTI, J. C., BOSMAN, A., BRENA, R. M., BRISON, D., CAISANDER, G., CAMARASA, M. V., CHEN, J., CHIAO, E., CHOI, Y. M., CHOO, A. B., COLLINS, D., COLMAN, A., CROOK, J. M., DALEY, G. Q., DALTON, A., DE SOUSA, P. A., DENNING, C., DOWNIE, J., DVORAK, P., MONTGOMERY, K. D., FEKI, A., FORD, A., FOX, V., FRAGA, A. M., FRUMKIN, T., GE, L., GOKHALE, P. J., GOLAN-LEV, T., GOURABI, H., GROPP, M., LU, G., HAMPL, A., HARRON, K., HEALY, L., HERATH, W., HOLM, F., HOVATTA, O., HYLLNER, J., INAMDAR, M. S., IRWANTO, A. K., ISHII, T., JACONI, M., JIN, Y., KIMBER, S., KISELEV, S., KNOWLES, B. B., KOPPER, O., KUKHARENKO, V., KULIEV, A., LAGARKOVA, M. A., LAIRD, P. W., LAKO, M., LASLETT, A. L., LAVON, N., LEE, D. R., LEE, J. E., LI, C., LIM, L. S., LUDWIG, T. E., MA, Y., MALTBY, E., MATEIZEL, I., MAYSHAR, Y., MILEIKOVSKY, M., MINGER, S. L., MIYAZAKI, T., MOON, S. Y., MOORE, H., MUMMERY, C., NAGY, A., NAKATSUJI, N., NARWANI, K., OH, S. K., OH, S. K., OLSON, C., OTONKOSKI, T., PAN, F., PARK, I. H., PELLS, S., PERA, M. F., PEREIRA, L. V., QI, O., RAJ, G. S., REUBINOFF, B., ROBINS, A., ROBSON, P., ROSSANT, J., et al. 2011. Screening ethnically diverse human embryonic stem cells

108

identifies a chromosome 20 minimal amplicon conferring growth advantage. Nat Biotechnol, 29, 1132-44.

LANGFELDER, P., MISCHEL, P. S. & HORVATH, S. 2013. When is hub gene selection better than standard meta-analysis? PLoS One, 8, e61505.

LI, X., ZHAO, F., XIN, M., LI, G., LUNA, C., LI, G., ZHOU, Q., HE, Y., YU, B., OLSON, E., GONZALEZ, P. & WANG, S. 2017. Regulation of intraocular pressure by microRNA cluster miR-143/145. Sci Rep, 7, 915.

LIU, C. H., WANG, Z., SUN, Y., SANGIOVANNI, J. P. & CHEN, J. 2016. Retinal expression of small non-coding RNAs in a murine model of proliferative retinopathy. Sci Rep, 6, 33947.

TAKAHASHI, K., TANABE, K., OHNUKI, M., NARITA, M., ICHISAKA, T., TOMODA, K. & YAMANAKA, S. 2007. Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell, 131, 861-72.

WANG, H., LA RUSSA, M. & QI, L. S. 2016. CRISPR/Cas9 in Genome Editing and Beyond. Annu Rev Biochem, 85, 227-64.

WANG, K., LI, M., HADLEY, D., LIU, R., GLESSNER, J., GRANT, S. F., HAKONARSON, H. & BUCAN, M. 2007. PennCNV: an integrated hidden Markov model designed for high-resolution copy number variation detection in whole-genome SNP genotyping data. Genome Res, 17, 1665-74.

109

Appendix

Method 1 iPSCs Differentiation

Ethics Experimental work performed in this study was approved by the Human Research Ethics

committees of the University of Melbourne (0605017, 0829937) and Royal Victorian Eye and

Ear Hospital (11/1031H) in accordance with the requirements of the National Health &

Medical Research Council of Australia and conformed with the Declarations of Helsinki.

Informed consent was obtained from all participants prior to sample collection.

Fibroblast Isolation and culture

Primary human dermal fibroblasts were isolated from skin punch biopsy specimens using

standard techniques (15). Human fibroblasts were cultured in DMEM with high glucose, 10%

fetal bovine serum (FBS), 100 U/mL Penicillin and 100 µg/mL streptomycin (all medium

components and supplements were purchased from Life Technologies, unless otherwise

noted).

hPSC lines, maintenance and passaging

The newly generated iPSC lines MT1 and BD1, the hESC line H9 (2), the abnormal hESC line

BG01V (49, +12, +17, XXY) (17, 18) and the iPSC lines FA3, FA4 (16) and the iPSC lines

FA3, FA4 (16) were maintained undifferentiated using mitotically inactivated mouse (for MT1,

H9 and BG01V) or human foetal skin fibroblasts (WS1, ATCC CRL-1502 for FA3, FA4) in

Knockout Serum Replacement (KSR) medium containing 20 ng/mL FGF2 (Millipore) as

described in (16).

iPSC Reprogramming and Selection of Primary iPSC Clones

Fibroblasts were reprogrammed to iPSCs with non-integrating episomal vectors containing 6

reprogramming factors (OCT-4, SOX2, KLF4, L-MYC, LIN28, and shRNA for p53) as

110

described previously (Piao et al. 2014; Hernández et al. 2016). 60,000 fibroblasts were

transfected with 1 µg each of pCXLE-hOCT4-shP53, pCXLE-hSK and pCXLE-hUL plasmids

(plasmids gifts from Prof. S. Yamanaka; Addgene plasmids #27077, 27078, 27080) using the

nucleofection kit for primary fibroblasts (Lonza) and nucleofected using the Nucleofector 2b

Device (Lonza, Program T-016). Cells were cultivated in fibroblast medium for 6 more days

and then trypsinized and 1 x 105 cells were plated onto a 10 cm dish with MEF feeders. Media

was changed to iPSC culture media the following day. iPSC colonies start appearing around

Day 17-30 from transfection and were manually dissected and expanded for further

characterisation. Cell lines generated were named MT1 and BD1, followed by clone (CL)

number.

Multi-lineage Differentiation of iPSCs

Neuroectodermal differentiation was carried out as described in (21). Embryoid bodies (EBs)

at day 7 were plated on gelatinized dishes to differentiate for 7-14 days cultured in 20% KSR

media without bFGF supplementation. Cells were fixed with 4% paraformaldehyde

(Proscitech) and permeabilized and blocked with 10% normal goat serum (Life Tech.)

containing 0.1% Triton X-100 (Sigma-Aldrich).

Immunocytochemistry

Cells were immunostained using the following antibodies: mouse anti-OCT3/4 (Santa Cruz

Biotechnology), mouse anti-TRA-1-60 (Millipore), CD9 (TG30, hybridoma supernatant),

mouse anti-beta-tubulin III (TUBB3, Abcam, clone TU-20), rabbit anti-alpha-fetoprotein

(AFP, Calbiochem, clone 1G7), mouse anti-smooth muscle actin (SMA, R&D systems, clone

1A4). Cells were then immunostained with the appropriate conjugated secondary antibodies

(goat anti-mouse IgG2b-Alexa 568, goat anti-mouse IgM-Alexa 488, goat anti-mouse IgG2a-

Alexa 647, Molecular probes-Invitrogen). Nuclei were counter-stained with DAPI (1𝜇g/mL,

Sigma-Aldrich) and mounted in Vectashield (Vector Labs). Specificity of the staining was

verified by the absence of staining in negative controls consisting of the appropriate negative

control immunoglobulin fraction (Dako).

111

Method 2 Genome-wide CRISRP/Cas9 screen for uveal melanoma OCM-1

cells

Amplification of pooled sgRNA library

Pooled sgRNA library (GeCKO v2.0 DNA plasmids) for screening was amplified as previously

described (Joung et al., 2017b). Briefly, diluted half-library of LentiCRISPRv2 (Library A and

Library B) was electroporated in Endura ElectroCompetent Cells and recovered in SOC

medium. A 40000 dilution of transformation was plated for overnight growth and

transformation efficiency was calculated by multiplying the number of colonies by dilution

fold. Colonies were harvested, bacteria pellets were weighed and plasmid DNA was generated

using Maxiprep Kit.

Lentivirus generation for pooled sgRNA library (LentiCRISPRv2) Lentivirus was produced in HEK293FT cells packaging amplified GeCKO v2.0 DNA plasmids

(Library A+ Library B). One day prior to transfection: Seeding 2x107 cells per T175 flask, 2

flasks in total.

Reaction mixture was set up as below.

Tube A:

Component Per T175 flask

Library A+B (1:1 ratio) 22.5 µg

pMD2.G 7.9 µg

pCMV D8.91 14.6 µg

Opti-MEM 4.5 mL Tube B:

Component Per T175 flask

Lipofectamine 2000 130 µL

Opti-MEM 4.5 mL Tube A and Tube B were incubated at RT for 10 minutes separately, then mixed evenly and

incubated at RT for another 10 minutes. Medium was aspirated with only 1 mL left to cover

112

cells. 9 mL of the transfection mixture was added per flask. Medium was changed after 4-6 h.

Cells along with medium were collected to a 50 mL centrifuge tube 48 h later. Cell suspension

was centrifuged for 5 minutes at 300g, and the supernatant was transferred to a new sterile

centrifuge tube, followed by flowing through a 0.45µm sterile filter. 5x PEG virus precipitation

solution was added to concentrate the virus and the mixture was placed at 4 ºC overnight.

Supernatant was discarded followed by centrifuge at 4 ºC at 1500 g for 30 min. The lentivirus

particle was resuspended in 2mL Opti-MEM and aliquot to be stored at -80ºC.

Kill-curve with puromycin selection for OCM-1

2x105/well OCM-1 cells were seeded in 6-well plate to reach 80% confluency the next day.

Different puromycin concentrations (0, 0.25, 0.50, 1, 2 and 2.5 µg/mL) was added to the wells.

Cells were examined under microscope 24 h after puromycin treatment. The lowest

concentration of puromycin which killed all the cells within 2 days was selected as the

concentration for CRISPR screen for OCM-1. The selected concentration of puromycin was

then tested in OCM-1 cells seeded in similar confluency in T175 flask before applying to large-

scale cell culture.

Lentiviral transduction and screening in OCM-1 cell

Lentiviral transduction

One day before LentiCRISPRv2 transduction, 2x105/well OCM-1 cells were seeded in 6-well

plate. Before infection, the medium was replaced with 1ml RPMI with 10%FBS, and polybrene

was added (final concentration 8µg/mL). A series of different volumes of LentiCRISPRv2

virus (0, 0.1, 0.25, 0.5, 1 and 2.5 µL) were added to each well. Two days after infection, cell

culture medium was replaced with 2 mL of fresh medium supplemented with selected

concentration of puromycin. Two wells of cells without Lentivirus infection were used as

controls (positive control: with puromycin, negative control: without puromycin). Two days

after puromycin treatment, live cells were collected and counted. Multiplicity of infection

(MOI) was calculated as number of live cells in puromycin selection condition/number of live

cells in the negative control. The amount of Lentivirus with a MOI around 0.3 was selected for

113

pooled screens. Lentiviral transduction and puromycin selection were further tested in larger

cell numbers using the selected puromycin concentration and Lentiviral volume in proportion.

Knockout screening in large scale expansion of OCM-1 cells

A total of 30 T225 flasks of OCM-1 cells (1x107 cells per flask) were seeded one day before

Lentiviral transduction. Cells were transduced with LentiCRISPRv2 in medium supplemented

with polybrene (8µg/mL). Puromycin was added 2 days after lentiviral transduction. Cells were

harvested 2 days after puromycin treatment and 1x10^8 cells were maintained in puromycin

selection condition, with the remaining infected cells being collected and stored in freezer

referred as UMP0. The cells in culture were maintained in puromycin condition for 1 week in

total and then 1x108 cells were cultured and passaged in puromycin-free medium for 12

passages. OCM-1 cells were collected at P12 and stored in freezer referred as UMP12. Three

biological replicates of CRISPR knockout screening were performed in uveal melanoma cell

line OCM-1. Sample ID for P0 were listed as UMP0_1, UMP0_2, UMP0_3, and P12 as

UMP12_1, UMP12_2, UMP12_3.

gDNA preparation for NGS analysis

Frozen cell pellets of OCM-1 P0 and OCM-1 P12 were stored at -80C until use. Cells were

used to maintain a coverage of >500. Genomic DNA was extracted using Zymo Research

Quick-DNA Midiprep Plus Kit according to the manufacturer's instructions. sgRNA library

was PCR amplified for NGS. PCR reaction and conditions were as follows.

Component 50 µL reaction (µL) Final

concentration

NEBNext High Fidelity PCR master mix, 2x 25 1x

Pooled sgRNA library sample / gDNA of OCM-

1

1 0.4 ng/µL

114

NGS_Lib_Fwd primer 1.25 0.25 µM

NGS_Lib_KO_Rev primer (barcode) 1.25 0.25 µM

Nuclease-free water 21.5

Total 50

Thermal cycling condition

Cycle number Denature Anneal Extend

1 95 ºC, 5 min

2-13 98 ºC, 20 s 60 ºC, 15 s 72 ºC, 15 s

14 72 ºC, 1 min

PCR products were purified using QIAquick PCR purification kit according to the

manufacturer's protocol. NGS was performed in two stages: Stage 1, after amplification of

pooled sgRNA library to determine sgRNA distribution. Stage 2, at the end of the screen.

Method 3 miR143 Animal body experiment methods

Animal housing

Animals were housed in standard cages in a temperature-controlled environment with free

access to food and water. To minimize possible light-induced damage to the eye, 12 hours of

light (50 lux illumination) alternated with 12 hours of dark (<10 lux illumination). This study

was approved by the Animal Ethics Committees of St Vincent's Hospital (Melbourne) and the

University of Melbourne (Reference Number 004/16 and 13-044UM).

115

Extraction of total RNAs and microRNAs

The eyes of OIR and normoxic rats were enucleated at P14 and P29, respectively, and stored

in RNAlater (catalogue no. 76104; Qiagen, Chadstone, VIC, Australia), and the retinas were

dissected immediately under a dissection microscope. Total RNA and miRNA were isolated

from the dissected retina using a miRNeasy mini kit (catalogue no. 217084; Qiagen). The

quantity of RNA was measured using spectrophotometry, and the integrity of total RNA was

confirmed by a Bioanalyzer at the Australian Genome Research Facility (AGRF, Melbourne,

VIC, Australia). The quality of RNA was determined by RNA Integrity Number (RIN); a RIN

of >7 and concentration of >20 ng/µL are considered acceptable for miRNA sequencing. We

paired qualified RNA samples (n=3) from OIR and normoxic rats at P14 for miRNA-Seq.

MicroRNA next-generation sequencing

One microgram of RNAs containing miRNAs in triplicate from the normoxic rats and OIR rats

at P14 were prepared as per the manufacturers’ instructions (Illumina Inc. San Diego, CA,

USA). The libraries of cDNA from the biological samples for normoxic and OIR groups were

sequenced by using an Illumina Hiseq-2500 RNA-seq platform as 50 bp single end chemistry

at Australian Genome Research Facility (AGRF, Melbourne, VIC, Australia) (Nixon et al.,

2015, Anderson et al., 2015). Briefly, the sequence reads from all samples were analysed for

overall quality, screened for the presence of any contaminants and trimmed accordingly. The

cleaned sequence reads were then processed through the quantification modules miRDEEP2

ver2.0.0.7 pipeline for known miRNA expression profiling. Rat miRNAs from miRBASE

release21 were used for mapping and quantification (Friedlander et al., 2008). Differential

miRNA expression analysis was undertaken using the eLimma bioconductor package

(https://www.bioconductor.org/packages/release/bioc/vignettes/limma/inst/doc/usersguide.pd

f). Expression profiling comparisons were performed for mature miRNAs between the

normoxic rats and OIR rats at P14 with a data filter set to ≤ -1 Log2 (Fold Change) and false

discovery rate (FDR) of 0.05.

Quantitative Polymerase Chain Reaction (qPCR)

116

The miRNAs profiles and putative target genes selected as per the computational analysis were

validated by qPCR with Taqman miRNA assay and gene assay reagents (Life Technologies

Australia, Mulgrave, VIC, Australia) respectively according to the manufacturer’s instructions.

Real-time quantitative PCR was performed using the StepOne™ Real-Time PCR Systems

(Applied Biosystems, Foster City, CA, USA). Briefly, for miRNA reverse transcription (RT),

100 ng of total RNA was reversely transcribed to individual cDNA with individual miRNA RT

primer using a miRNA RT kit (Life Technologies Australia). For mRNA reverse transcription,

total RNA (100 ng) was reverse-transcribed to cDNA using a high-capacity RT kit (catalog no.

4374996; Life Technologies Australia). Quantitative PCR was then performed using a TaqMan

faster PCR master mix and commercially available rat miRNA and gene probe sets (TaqMan

Gene Expression Assay, Life Technologies Australia), respectively. The miRNAs selected for

validation were miR-143-3p, miR-145-5p, miR-126-3p, miR-150-5p and miR-451-5p. The U6

small nuclear RNA was employed as an endogenous control to normalise the expression levels

of miRNAs. The putative target genes selected for validation was TAK1. Rat beta-actin

(ACTB) was used as a reference gene. For analysis of miRNA and gene expression, relative

expression levels of miRNAs and target genes between normoxic and OIR rats were calculated

using DDCt method which has been previously described by Livak (Livak and Schmittgen,

2001). All qPCR assays for validation of candidate miRNAs and genes were performed in

duplicate using samples of cohorts that different from the one used for miRNA sequencing.

Animal allocation

Rats were equally randomly subjected to different treatment groups from same littermate or

cage by investigators (JHW and GSL for OIR studies; SMP and BVB for ERG/OCT) and

housing in same room at the animal facility.

Intravitreal Injection

Intravitreal injections were performed in normoxic rats and OIR rats at P14 using a surgical

microscope similar to that previously described (Hewing et al., 2013, Chen et al., 2009). In

brief, after making a guide track through the conjunctiva and sclera at the superior temporal

hemisphere behind the limbus using a 30-gauge needle, a hand-pulled glass micropipette

117

connected to a 10 μL Hamilton syringe (Bio-Strategy, Broadmeadows, VIC, Australia) was

inserted into the vitreal cavity. One microliter of low (15 ng) or high dose (90 ng) of 5Z-7-

Oxozeaenol diluted in 1% DMSO, or balanced salt solution containing 1% DMSO, was

injected into the eyes of OIR rats at P14 at a rate of 200 nL/s using a UMP3-2 Ultra Micro

Pump (World Precision Instruments, Inc. Sarasota, FL, USA). Similarly, a total of 1 μg of miR-

126, miR-143, miR-145, or miR-150 mimics (GE Healthcare Dharmacon, Parramatta, NSW,

Australia; Supplementary Table 2) was injected into the eye of the OIR rats at P14, and an

equal amount of non-targeting scrambled negative control RNAs was injected into the

contralateral eye of the same animal. At P18, the rats were sacrificed to collect the retina for

further assessment. Any issues with the injection, including large backflow upon removal of

the needle, hemorrhaging of external or internal vessel were noted and eyes were excluded

from the study. Similarly, Brown Norway rats were intravitreally injected low dose of 5Z-7-

Oxozeaenol or balanced salt solution containing 1% DMSO. The animals were then subjected

to Electroretinography and Optical coherence tomography 28 days after the injection.

Endothelial Tube Formation assay

Endothelial tube formation was assessed as the previous described (Donovan et al., 2001). BD

Matrigel™ Basement Membrane Matrix (catalog no. 356234; Becton Dickinson; Bedford,

MA) was thawed at 4°C overnight and then spread evenly over each well (50 μL) of 96-well

plate. The plate was incubated at 37°C for 30 minutes before seeding to allow Matrigel to

polymerize. HUVECs were then seeded at 2 x 104 cells per well and grown in 100 μL of

endothelial cell basal medium-2 (EBM-2) supplied with EGM™-2 BulletKit™ (catalog no.

CC-5035; Lonza). Cells were immediately treated with 1000 nM TAK1 inhibitor (5Z-7-

Oxozeaenol; catalogue no. 3604/1; Tocris Bioscience, Bristol, UK) or left untreated as controls.

Images were taken for each group 6 hours after treatment. The endothelial lumen formation

and branch point were quantified by using the Angiogenesis Analyzer Plugin for ImageJ

version 1.48 software (http://imagej.nih.gov/ij/; provided in the public domain by the National

Institutes of Health, Bethesda, MD, USA). The control samples were defined as 100% tube

formation, and the percentage change in tube formation relative to control was calculated for

the treated samples. Experiments were repeated at least twice.

Cell Proliferation assay

118

For the cell proliferation assay, a cell proliferation assay kit was used as per the manufacturer’s

instruction (CYQUANT® NF, catalog no. C35006; Life Technologies Australia). Briefly,

HUVECs were seeded at 5,000 per well in 96-well plate. The cells were then treated with 5Z-

7-Oxozeaenol at 1000 nM or left untreated as controls. The medium was removed and replaced

with the diluted binding solution with DNA dye 24 hours or 48 hours after drug treatment. To

allow the DNA being stained, the plate was incubated at 37°C for 1 hour prior to assessment

of fluorescence intensity. The fluorescence intensity of samples was measured using a

fluorescence microplate reader (Spark® 20M, Tecan, Switzerland) with excitation at ~485 nm

and emission detection at ~530 nm. The results were normalized to the percentage of control.

Experiments were repeated at least twice.

Cell Migration

HUVECs were seeded at a concentration of 2 x 105 cells/well in a 6-well tissue culture plate

prior to treatment. Cell monolayers were then wounded by scraping, followed by treatment

with 5Z-7-Oxozeaenol at 1000 nM or left untreated as controls. Cells were incubated at 37 °C

for 24 hours. Images of cells were taken immediately after scraping and 24 hours after 5Z-7-

Oxozeaenol treatment under phase contrast microscope (TE2000-U, Nikon, Japan). For each

image, the distance between the gap was quantified and normalised to the original would

distance. Four fields per well were quantified with using the ImageJ version 1.48 software, and

experiments were repeated at least twice.

Aortic Ring Assay

Aortic ring assay was performed as previously described (Baker et al., 2012). Briefly, aortae

were removed from euthanized C57BL/6 mice and immediately transferred to a culture dish

containing ice-cold EBM-2 supplied with EGM™-2 BulletKit™ (Lonza). The peripheral

fibroadipose tissue was carefully removed. The aortae were sectioned into rings ~1 mm in

width. Each aortic ring was placed in an individual well covered with 120 uL growth factor-

reduced BD Matrigel™ Basement Membrane Matrix (catalog no. 356234; Becton Dickinson),

and was cultured in 200uL EBM-2 for 9 days. For explants treated with 5Z-7-Oxozeaenol, 5Z-

7-Oxozeaenol was added to the medium at 1000 nM. The medium was replaced and the drug

119

was supplemented every 3 days for all groups (n=6 per group). Images of individual explants

were taken on day 0 to 9 after plating using Zeiss microscope. The neo-formed vascular

sprouting area was quantified with Adobe Photoshop Elements.

Electroretinography (ERG)

A low dose (18 ng) of 5Z-7-Oxozeaenol or vehicle was injected intravitreally into the eyes of

male Brown Norway rats; 28 days following the injection, the rats underwent ERG. Rats were

dark-adapted overnight prior to the ERG assessment under fully dark-adapted conditions.

Details for functional assessment are as previously described (Zhao et al., 2017). Briefly, a pair

of custom-made chloride silver active and ring-shaped reference electrodes (99.9%, A&E

Metal Merchants, Sydney, NSW, AU), connected to platinum leads (F-E-30, Grass Telefactor,

West Warwick, RI), was located in the central cornea and sclera, respectively. A stainless-steel

needle electrode (F-E2-30, Grass Telefactor) was inserted subcutaneously into the tail of the

animal to replace ground. Measurements were recorded simultaneously for both eyes. ERG

analysis as previously described consists of three major waveforms. Each waveform reflects

the function of different retinal cells, such as the photoreceptor (a-wave), bipolar cell (b-wave),

and ganglion cell (scotopic threshold response, STR).

Spectral-domain Optical coherence tomography (OCT)

Following ERG measurement, rat eyes were imaged using spectral domain-OCT (Envisu

R2200 VHR, Bioptigen, Inc., Morrisville, NC, USA). Volume scans consisting of 1000 A-

scans per 100 B-scans (Equally space across the 1.4 mm vertical dimension; 100 horizontal B-

scans evenly spaced in the vertical dimension) centred on the optic nerve head (ONH; 1.4 ×

1.4 × 1.57 mm) were obtained from both eyes. Four B-scans that crossed the ONH were

analyzed using FIJI software (https://fiji.sc/). In each B-scan the inner limiting membrane,

retinal nerve fiber layer (RNFL), inner plexiform layer and Bruch's membrane were manually

segmented by a masked observer as previously described(Zhao et al., 2017). Total retinal

thickness (TRT) was measured from the inner limiting to Bruch’s membrane. Retinal nerve

fibre layer (RNFL) thickness was measured from the inner limiting membrane to the inner

aspect of the inner plexiform layer. Outer retinal thickness was measured from Bruch’s

membrane to the outer plexiform layer.

120

Dual-Luciferase Reporter Assay

The oligonucleotides comprising the miR-143 target sequence were synthesized by Integrated

Device Technology (Singapore), then annealed and cloned into pGL3 Luciferase Reporter

Vectors (pCMV_Luciferase, Promega, Madison, WI, USA) with a reporter gene encoding

firefly luciferase (pCMV_Luciferase/miRE). The miR-143 target sequence within the 3’

untranslated regions (UTRs) of TAK1 was predicted by the miRTarBase and TargetScan as

described previously (Hsu et al., 2011, Chen and Rosbash, 2016). Constructs containing target

sequences of miR-143 (Supplementary Table 3) was confirmed by restriction enzyme digestion

and Sanger sequencing at AGRF (Melbourne, VIC, Australia). Transfection was performed in

12-well plates with 80% confluent HEK293A cells. Two hundred nanogram of firefly

luciferase plasmid and 50 ng of the Renilla luciferase plasmid (pRL-SV40; Promega) were co-

transfected into cells by using Lipofectamine 2000 (catalog no. 11668027; Life Technologies

Australia). Cells were simultaneously co-transfected with 40 nM miR-143 mimics or

scrambled negative control RNAs. For each transfection assay, the pCMV_Luciferase vector

serves as positive control. After 48 hours, cell lysates were prepared from each transfected

culture, and Dual-Luciferase assay was performed using the dual luciferase kit (catalog no.

E1910; Promega). Luminescence was measured by using a microplate reader (Spark® 20M,

Tecan). Luciferase activity was normalized to Renilla luciferase activity and to positive

controls.

Method 4 DPMSC to TM cells differentiation protocol

Cell culture

Cultures of human dental pulp MSCs (DP-MSCs; Lonza) were grown as adherent monolayers

in Dental Pulp Stem Cell Basal Media (DPSC-GM) supplemented with Human Dental Pulp

Stem Cell Growth Supplement (DPSC-GS), L-glutamine, ascorbic acid and

gentamicin/amphotericin-B (GA) (all from Lonza). Cultures were maintained at 37°C in a

humidified atmosphere with 5% CO2 in T-25 or T-75 cell culture flasks (Greiner Bio-One),

and were passaged once they reached approximately 80% confluence and the MSCs had their

medium replaced every 3-4 days.

121

For experimental procedures, DP-MSCs were allowed to reach 80-90% confluence in T-75

flasks and harvested by enzymatic digestion using Trypsin/EDTA (Lonza) respectively.

Harvested cells were washed once in complete Dental Pulp Stem Cell Growth Media (DPSC-

GM), pelleted centrifugation and counted.

For adipogenic, chondrogenic and osteogenic differentiations, DP-MSCs were seeded into 6

well plates at 1 x 105 cells/well, 8 x 106 cells/mL (as multiple 10 μL droplets) and 5 x 104

cells/well respectively. Chondrogenic cultures were differentiated 2 hours post-seeding by

addition of 2 mL (per well) of Complete Chondrogenesis Differentiation Medium (StemPRO®

Osteocyte/Chondrocyte Differentiation Basal Medium supplemented with StemPRO®

Chondrogenesis Supplement [Life Technologies]) while adipogenic and osteogenic cultures

were left to adhere overnight and were differentiated 3 days post-passage by addition of 2 mL

Complete Adipogenesis Differentiation Medium (StemPRO® Adipocyte Differentiation Basal

Medium supplemented with StemPRO® Adipogenesis Supplement [Life Technologies]) or 2

mL Complete Osteogenesis Differentiation Medium (StemPRO® Osteocyte/Chondrocyte

Differentiation Basal Medium supplemented with StemPRO® Osteogenesis Supplement [Life

Technologies]) respectively.

For growth factor experiments, cells were plated out at 5 x 103 cells/cm2 (5 x 104 cells/well of

a 6-well plate) for RNA collection and allowed to adhere overnight. For immunofluorescent

(IF) staining, cells were seeded into plastic 4 well Lab-Tek Chamber Slides (Thermo Scientific)

at 9 x 104 cells/well and allowed to attach overnight. For both RNA and IF experiments, cells

were treated the following day with different concentrations of RA, TGF-β2 and BMP4 for

different time periods.

Histochemical staining Oil Red O, Alcian Blue and Alizarin Red S were used for staining of adipogenic, chondrogenic

and osteogenic cultures respectively(Niemelä et al., 2008, Qian et al., 2010, Charles Huang et

al., 2004, Mauck et al., 2006, Takács et al., 2013, Birmingham et al., 2012, Mori-Akiyama et

al., 2003). Oil Red O was prepared as a 0.3% solution by dissolving 300 mg of Oil Red O

powder (Allied Chemical) in 100 mL of 99% isopropanol. Cells were washed using Hank’s

Balanced Salt Solution (HBSS) (Life Technologies) and fixed in 3.7% formaldehyde for 30

122

minutes. Following fixation, cells were again washed and then covered with 60% isopropanol

for five minutes. Cells were then stained using sterile filtered Oil Red O diluted 3:2 in distilled

water for 15 minutes. Following staining, cells were washed to remove Oil Red O and stained

with Haematoxylin for one minute. Cells were again washed and then emerged in HBSS for

imaging. Alcian Blue was prepared as a 1% solution by dissolving 1 g of Alcian Blue powder

(Koch-Light Limited) in 100 mL of 0.1 M hydrochloric acid (HCl). Cells were washed twice

and then fixed using 3.7% formaldehyde for 30 minutes. Following fixation, cells were washed

and then stained with Alcian Blue for 30 minutes. Cells were then washed in 0.1 M HCl three

times and emerged in distilled water to neutralise acidity and allow for imaging. Alizarin Red

S was prepared as a 2% solution by dissolving 2 g of Alizarin Red S powder (Difco

Laboratories) in 100 mL of distilled water and adjusting the pH to 4.1-4.3 by addition of

ammonium hydroxide (NH4OH). Alizarin Red S must be stored in the dark. Cells were washed

twice in HBSS and then fixed using 3.7% formaldehyde for 30 minutes. Following fixation

cells were washed twice and then stained with Alizarin Red S for 15 minutes. Cells were then

rinsed to remove stain and emerged in HBSS to allow for imaging. All cell types were imaged

using an Olympus IX71 microscope.

Treatment of MSCs with RA, TGF-β2 and BMP4 Initially, DP-MSCs were treated singularly with two concentrations of each growth factor (RA,

TGF-β2 and BMP4) in their respective growth mediums. Concentrations used were derived

from current literature(Wordinger et al., 2007, Sharow et al., 2012). RA stock (Sigma) was

made up to 1 mM in DMSO for use in subsequent experiments. Stock concentrations of

Recombinant Human Transforming Growth Factor-β2 (TGF-β2) and Recombinant Human

Bone Morphogenic Protein 4 (BMP4) (both from Life Technologies) were reconstituted to 10

μg/mL in 0.1% bovine serum albumin (BSA)/HBSS for experimental use. Cells were treated

for 48 hours with 0.1 μM and 1 μM concentrations of RA, 0.5 ng/mL and 5 ng/mL

concentrations of TGF-β2 and 1 ng/mL and 10 ng/mL concentrations of BMP4 or control

(0.1% BSA/HBSS for TGF-β2 and BMP4 or DMSO for RA). Subsequent experiments using

both AD-MSCs and DP-MSCs involved using different combinations of these growth factors

applied to cells for different time periods and are described in the results section.

RNA extraction, cDNA preparation and qRT-PCR

123

RNA was extracted using RNeasy spin column kits (Qiagen) as per the supplied manufacturer’s

protocol. Briefly, cells were washed twice using HBSS and then harvested by addition of 350

μL of Buffer RLT. Cell lysates were homogenised using a QIAshredder and transferred to a

RNeasy spin column. Lysates were treated with DNase (Qiagen) for 15 minutes at room

temperature and eluted using 30 μL of RNase-free water.

cDNA was prepared as per the manufacturer’s protocol supplied by Qiagen. Briefly, a volume

of thawed RNA corresponding to the amount of cDNA required was added to 9 μL of master

mix and RNase-free water up to a volume of 20 μL. Master mix was prepared for the number

of cDNA reactions required and, per reaction, consisted of 2 μL 10x Buffer RT, 2 μL dNTP

Mix (5 mM each dNTP), 2 μL Oligo dT Random Primer 9, 2 μL RNase inhibitor (10 units/μL),

1 μL Omniscript Reverse Transcriptase and RNase-free water up to 20 μL (all from Qiagen).

Tubes were briefly vortexed, centrifuged and then incubated at 37°C for one hour. Following

incubation, cDNA was placed on ice for qRT-PCR analysis or stored at -20°C.

For RT-qPCR reactions, cDNA (20 μL) was diluted to 2 ng/μL using nuclease-free water.

Quantitative real time PCR (qRT-PCR) was performed using a Step-One Plus Detection

System (BioRad) with C/EBPα, PPARγ2, FABP4, SOX9, ACAN, COL2A1, ALPL, RUNX2,

COL1A1, MYOC, CHI3L1, MGP, PITX2, FOXC1, FOXC2 and PAWR probe/primer mixes

(Life Technologies) (Supplementary Table 1). cDNA (5 μL) was added to 15 μL of TaqMan

PCR mix (10 μL 2X TaqMan Master Mix, 1 μL 20X probe/primer mix, 4 μL nuclease-free

water). To each well of a 96 well PCR plate, 20 μL of total reaction was added and incubated

using the following program: initial denaturation for 20 seconds at 95°C and 40 cycles of

amplification (denaturation for 1 second at 95°C, annealing and extension for 20 seconds at

60°C). Relative gene expression analysis was performed using the 2-ΔΔCt method where the

relative expression of target mRNA was normalised against the housekeeping gene HPRT1 and

presented relative to untreated control cultures.

Immunofluorescent staining of MSC cultures For IF, cells were allowed to adhere overnight and then treated using a combination of growth

factors for five days. Cells were then washed three times using 600 μL of PBS (pH 7.4) and

fixed for 10 minutes using 200 μL of 4% paraformaldehyde. Following fixation cells were

124

washed three times using PBS for five minutes each. Cells were permeabilised by the addition

of 400 μL of ice-cold 100% methanol and placed in the freezer for 10 minutes. Following

treatment with methanol, cells were washed with PBS for five minutes. From this point,

chamber slides were manipulated inside a humidified light-tight box at room temperature in

order to prevent drying and fluorochrome fading. Cells were then blocked using 400 μL of 5%

normal goat serum diluted in PBS Triton X for 60 minutes. During blocking, 1:25 and 1:75

antibody dilutions were prepared for both PITX2 (rabbit polyclonal from Thermo Scientific)

and FOXC1 (rabbit polyclonal from Bethyl Laboratories, Inc.) and a 1:50 dilution was prepared

for histone 3 (rabbit polyclonal from Cell Signaling Technology). Following blocking, 250 μL

of antibody dilution was added to each well and incubated overnight at 4°C. Following

overnight incubation, cells were washed three times using PBS for 5 minutes each. Cells were

then incubated in 250 μL of secondary antibody (anti-rabbit Alexa Fluor 594 [Life

Technologies]) diluted 1:500 in PBS Triton X for 2 hours. Cells were then washed in PBS three

times for five minutes each, incubated in a 1:500 diluted Hoechst solution (Life Technologies)

for five minutes at RT and then washed in PBS again for five minutes. The plastic wells of the

chamber slide were then removed and a coverslip was applied using ProLong Gold Antifade

mounting medium (Life Technologies) and sealed using clear nail varnish. Slides were then

examined under an Olympus IX71 microscope using the appropriate wavelength excitation

filter (CY3 and TXRED).

Phagocytosis assay pHrodo green E. coli bioparticles conjugate (Life Technologies) were prepared according to

the manufacturer’s protocol and added to bioparticle positive wells for two hours. Briefly, 1

vial of bioparticles was thawed and resuspended in 1 mL of PBS (pH 7.4). This solution was

then transferred to a large, thin-walled PCR tube and sonicated for five minutes, pulsing 30

seconds on/30 seconds off in a mixture of water and ice. Following this, 1 mL of pre-warmed

PBS was added to bring the final concentration of the bioparticles to 1 mg/mL. Culture medium

was then removed and replaced with 400 μL of bioparticles solution. Wells which did not

require bioparticles were replaced with 400 μL of PBS (pH 7.4). Chamber slides were then

wrapped in aluminium foil in order to protect the bioparticles from light and placed on a warm

plate at 37°C for two hours. After two hours, solutions were removed from wells and the cells

were washed twice using PBS. Cells were then fixed using 4% paraformaldehyde for 10

minutes at room temperature. Following this, cells were washed twice and then stained with

125

Hoechst (1:500) (Life Technologies) for five minutes. Following staining, cells were washed

twice with PBS. Wells were then removed and coverslips were added using approximately 80

μL of ProLong Gold Antifade mounting medium (Life Technologies). Slide edges were sealed

using clear nail varnish and allowed to dry. Imaging was performed using an Olympus

Fluoview FV1200 confocal microscope.

Statistical analysis Raw data generated from each experiment was collated in Microsoft Excel 2010 and statistical

analysis was performed using GraphPad Prism v6.0. DP-MSC multipotent differentiation

experiments were analysed using an independent t-test. Treatments were pooled and analysed

against the undifferentiated control. DP-MSC singular and combination treatments with RA,

TGF-b2 and BMP4 were analysed using a one-way Analysis of Variance (ANOVA) for each

cell type. Relative gene expression levels for each treatment were pooled and analysed against

the vehicle treated control. Significant results were differentiated using a Tukey’s post-hoc test.

All data are presented as the mean ± standard error of the mean (SEM), and significant results

are defined as *p<0.05, **p<0.01, ***p<0.001 and ****p<0.0001.

All supplementary Figures and Tables are available from the following link: https://drive.google.com/open?id=17i-lOVi0TgR2dAFkh5uHu4K0hc3BHGR9

Documents

Applying Bioinformatic Tools to Better Understand Eye Diseases