Figures and TablesFor Your Publication

How to Present Your Data

Do you have enough data to write a paper?

• Does your data tell a story?• Does it support a main point?• Prior to completing your experiments, make

sure you anticipate how many data points you will need to support your main point – Anticipate appropriate statistical analyses– Anticipate how you will show your data (table/figure)

Do you have enough data to write a paper?

• Start with organizing your data into figures and tables.

• Most publications have 5 figures.

Deciding what data to use

• Ethically what can you do when some samples have problems?

• Issues of technique versus outlying values versus inconsistent results.

• When were samples collected?

• When were samples processed?

Figure Options

1) Pictorial presentation• Graph Data in connected series• Chart Data in separate series• Picture Must be seen (Photos)• Diagram Model to show concepts

2) Tables Data in an array

What is the most effective way to display your data?

• What type of data do you have?

– Microarray data?• Results of hundreds of genes must be separated

into tables or charts to make any sense. – A lot of data points?

• Results from 500 samples must be summarized in a way that is easy to understand.

• Try to use a variety in the types of figures. – Boring to have 6 identical type of graphs.

Time(hour)

midnight2:004:006:008:00

10:00noon2:004:006:008:00

10:00

Normal(mg/dl*)

100.393.688.2

100.5138.6102.4

93.8132.3103.8

93.6127.8109.2

Diabetic(mg/dl)

175.8165.7159.4

72.1271.0224.6161.8242.7219.4152.6227.1221.3

Table 2. Blood glucose levels

* decaliters/milligram Figure 11. Blood glucose levels over time for normal individual and diabetic subjects

Hour

12:00 6:00 am 12:00 6:00 pm 12:00

BloodGlucoseLevel(mg/dl)

300

250

200

150

100

50

0

BreakfastLunch Dinner

Normal

Diabetic

Tables or graphs?Tables or graphs?

Graphs are Graphs are often easier to often easier to

understandunderstand

Each table or figure should stand alone

• Understandable by itself– Should not have to read the text– Its title and descriptions should be enough

Ref: V. McMillan

Tables

• Use a table to present many numerical values

• Don’t pack in too much information!

• Don’t include columns that have the same value throughout. You can include that information in a caption or in the text.

Ref: V. McMillan

Not all data should be in the table

Water Sample

Date Total Coliform

(MPN/100ml)

Fecal Coliform

(MPN/100ml)

Enterovirus HAV

Del Mar1 10/02/05 1000 200 + -

Del Mar2 11/04/05 800 260 - -

Del Mar3 12/14/05 1200 300 + -

Tijuana1 10/08/05 1500 400 + -

Tijuana2 11/09/05 1300 320 + -

Tijuana3 12/12/05 1050 280 - -

Abbreviations: MPN, most probable number; HAV, Hepatitis A Virus+/- indicates the presence or absence of virus as detected by RT-PCR

Table 3. Indicator Bacteria and Viral Results

Table format

• Columns and rows– Organize a table so that the similar items

read down, not across

• Table title

• Footnotes

• Look at the format in other papers as a guideline

Ref: V. McMillan

Table footnotes

• Footnotes are BRIEF explanations about data including

- Exceptions

- Abbreviations

- Statistics

- p-values for data were >0.05.

• Do not write out information that belongs in the results!

Table example

++++++AG129

------G129

+/-+/-++++A129

------WT129

Table 1. Detection of Infectious DEN2 in Tissues by Indirect Plaque Assays

++----AG129

------G129

+/-+/-----A129

------WT129

Day 7 p.i.a

Day 3 p.i.a

Spinal cordBrainLymph nodesSpleenLiverSerum

Abbreviations: p.i.= post-infection+Virus was detectable.-Virus was undetecable.+/-Virus was detected in only some mice or some experiments.aMice were intravenously inoculated with 108 PFU of DEN2, PL046 strain. At Day 3 p.i., serum and tissues from various sites were harvested and processed for indirect plaque assay. Each group consisted of 3-5 mice per time point. This experiment was performed 3 times, and similar results were obtained in all experiments.

Mice strains

Figure format

• Different types– Graph– Chart– Picture (gels, flow cytometry)– Diagram

• All have– Figure Legend (title and a brief description

of experiment)

Four parts to figure legends

1. Title • One sentence to identify the main point of the figure.

2. Brief experimental details • Enough details so that the reader can understand figure.

3. Definitions • Symbols or bar patterns that are not explained in figure.

Antigen present Control

4. Statistical information • Number of samples, p-values, etc.

Ref: V. McMillan

Graphs

• Use to show a trend or pattern

• Generally a graph is not necessary when trends or relationships are not statistically significant

• If a cause and effect relationship: – X axis is the independent variable– Y axis is the dependent variable

Ref: V. McMillan

Graphing a small study

• Beware of presenting a small amount of data in graphs– Showing data in graphs can be a dramatic and

effective way to show an effect or trend but if your data set is too small you can’t say that you are seeing a potential trend or a real effect. So, showing it in a graph can be misleading.

• If you do use a graph with a small number of samples, clearly state how many data points you used

Response to Amoxicillin

0%

10%

20%

30%

40%

50%

60%

70%

No Response Response

Per

cent

age

of M

ice

Figure 1. Percentage of mice that responded to amoxicillin treatment.

Three mice were treated with 0.5 mg/ml amoxicillin for 7 days.

Incorrect if N= 3 mice!!

Use of a best-fit line•Direct correlation

•Assumption that the y-axis endpoint is dependent upon the x-axis variable

Fig 5. A comparison of the detection of anti-dengue virus IgG in the 12 serum samples is shown by ELISA in OD plotted against the percentage of detection by sensor chip. This percentage is the number of sensors reporting the presence of a bead as denoted with a relative signal magnitude (RSM) above the detection threshold. Dengue virus antigen was used to coat microtiter and sensor ship wells. This comparison demonstrates a strong association between each OD value and its corresponding sensor percentage, with a squared correlation coefficient (R2) of 0.966.

No best-fit line

• Plot points only

• Either variable could be on the x-axis

• No assumptions are being made about which variable is independent

Ref: V. McMillan

Other Types of Figures:

p=0.0171

0 5 10 15 20 25 300

20

40

60

80

100

WT129

G129AG129

.

A.

(n = 30)

(n = 21)(n = 22)

% S

urv

ival

Days Post-Infection

p<0.0001

Figure 1. Survival of mice lacking the IFN and/or IFN receptor genes following infection with dengue virus. A wild type mouse strain (WT129), a transgenic strain deficient for the gamma IFN receptor only (G129), as well as a transgenic strain deficient for the alpha/beta and gamma IFN receptor genes in combination (AG129) were infected with 108

PFU of dengue virus (DEN2). Survival over time was determined. The statistical significance (p-values) of the differences in survival between the transgenic and wildtype strains are provided.

Gels

66.266.2

97.497.4

45.045.0

31.031.0

21.521.5

14.414.4

MWMWKDaKDa

++

50mM50mM 100mM100mM 250mM250mM

MM

500mM500mM

5DLuc3D5DLuc3D5DLuc5DLuc ++

++ ++ ++ ++ ++ ++

**

**

Figure 1. Effect of dengue virus 5' and 3' UTRs on protein translation in the presence and absence of inhibitor. Luciferase reporter constructs containing both the dengue virus 5' and 3' UTR (5DLuc3D) or the dengue virus 5’UTR only (5DLuc) were transfected into cells. Cells were subsequently exposed to various concentrations of an RNA inhibitor.  Cell extracts were harvested and analyzed by SDS-Gel electrophoresis followed by Coomassie blue staining. Luciferase protein products of 28 Kda and 43 Kda are noted with an *.

RNA Inhibitor Concentration

Pie-chartsUP-REGULATED 30 MIN

23%

5%6%2%0%

5%7%0%

3%2%4%

2%

25%

6%

1%10%

1%

DOWN-REGULATED 30 MIN

10%2%2%11%

13%3%

3%0%

4%2%

1%

30%

3%

14%

METABOLISMENERGY

CELL CYCLE & DNA PROCESSINGTRANSCRIPTIONPROTEIN SYNTHESISPROTEIN FATE

CELLULAR TRANSPORT & TRANSPORT MECHANISMSCELLULAR COMMUNICATION / SIGNAL TRANSDUCTIONCELL RESCUE, DEFENSE AND VIRULENCE

REGULATION OF / INTERACTION WITH CELLULAR ENVIRONMENT

CELL FATE

CONTROL OF CELLULAR ORGANIZATION

SUBCELLULAR LOCALIZATIONPROTEIN ACTIVITY REGULATION

PROTEIN WITH BINDING FUNCTION OR COFACTOR REQUIREMENT

TRANSPORT FACILITATIONCLASSIFICATION NOT YET CLEAR-OUTUNCLASSIFIED PROTEIN & NOT PRESENT IN S. C.

Figure 1. Effect of cycloheximide upon gene expression in Saccharomyces Cerevisiae.  The wildtype yeast strain (YAS1180) was grown in the presence of 50 ng/ml cycloheximide for 30 minutes at 30C. Cells were harvested and total RNA was extracted. RNA expression of different classes of genes was determined using the Affymetrix GeneChip Expression System for Saccharomyces cerevisiae.

How to display the data?

Pt # Case

IFN-g SLA

IFN-g PHA

IL-2 SLA

IL-2 PHA

IL-10 SLA

IL-10 PHA IgE

pg/ml pg/ml pg/ml pg/ml pg/ml pg/ml ng/ml

1 Active ACL 1140 >2000 <125 700 50 >2000 6002 Active ACL <125 870 <125 120 <125 208 407 Active ACL <125 755 <125 <100 <125 600 1908 Active ACL 629 <125 <125 600 140 1160 220

12 Active ACL >2000 >2000 <125 1100 <125 >2000 14022 Active ACL <125 >2000 <125 720 <125 930 4023 Active ACL <125 250 <125 <100 80 <125 7524 Active ACL <125 1045 <125 200 <125 154 14025 Active ACL <125 <125 <125 <100 <125 100 32026 Active ACL 162 697 <125 <100 <125 468 25027 Active ACL <125 426 <125 <100 250 210 22628 Active ACL <125 <125 <125 <100 250 225 36729 Active ACL <125 777 <125 400 460 260 48030 Active ACL <125 313 <125 <100 320 770 13431 Active ACL <125 1018 <125 300 250 240 5736 Active ACL 882 720 411 1093 217 >2000 43237 Active ACL 1568 >2000 600 741 150 482 341

5 Asymptomatic <125 <125 <125 <100 <125 340 2106 Asymptomatic 320 196 <125 <100 <125 690 3809 Asymptomatic 960 >2000 <125 250 <125 720 150

15 Asymptomatic 460 >2000 <125 1500 <125 >2000 10018 Asymptomatic 210 >2000 <125 700 <125 1450 <1519 Asymptomatic <125 >2000 <125 350 <125 >2000 9020 Asymptomatic 1440 >2000 <125 1000 <125 >2000 14021 Asymptomatic 500 >2000 <125 1850 <125 >2000 6534 Asymptomatic 488 >2000 <125 <125 <125 >2000 615

3 Neg ctrl <125 >2000 <125 <100 <125 538 4804 Neg ctrl <125 >2000 <125 <100 <125 300 330

10 Neg ctrl <125 106 <125 <100 <125 235 14011 Neg ctrl <125 >2000 <125 >2000 130 >2000 25013 Neg Ctrl 980 >2000 <125 550 <125 >2000 4014 Neg ctrl <125 >2000 <125 >2000 <125 >2000 9016 Neg ctrl <125 >2000 <125 1800 <125 >2000 18017 Neg ctrl 220 >2000 <125 >2000 <125 >2000 10032 Neg Ctrl 553 1709 <125 <125 <125 <125 51033 Neg Ctrl 337 1909 <125 <125 <125 >2000 55435 Neg ctrl <125 300 <125 294 <125 >2000 348

Raw Data fromLeishmania paper

A bit confusing in a table.

How to present?

0

100

200

300

400

500

0 0.5 1 1.5 2 2.5 3 3.5Active ACL Asymptomatic Control

Figure 1. IL-10 produced by PBMCs in response to stimulation with the Leishmania antigen. Peripheral blood mononuclear cells (PBMCs) collected from people with active atypical cutaneous leishmaniasis (ACL) infection, people with asymptomatic ACL, and uninfected people (control) were stimulated with 2 mg of soluble Leishmania antigen (SLA).  IL-10 levels were measured by ELISA.

IL-1

0 L

evels

(p

g/m

l)

PBMC Samples

Get Started!!!

• Try different ways of organizing your data• Ask advice from colleagues• Look at other articles and see how other

researchers present information that is similar to your data