mmi genomics/ucd & msu visits – may 2003 design and analysis of cdna microarray experiments at...

MMI Genomics/UCD & MSU Visits – May 2003

Design and Analysis of cDNAMicroarray Experiments atCSIRO Livestock Industries

Toni Reverter ([email protected])CLI BioinformaticsQueensland Bioscience PrecinctBrisbane, 4067 Australia

Introduction: Analysis possibilities Challenges Process for microarray

Technical Concerns: Design Image (data) quality Data analysis

Contents


Design and Analysis of cDNA Microarray Experiments at CSIRO Livestock Industries (Toni Reverter)

Determine genes which are differentially expressed (DE).

Connect DE genes to sequence databases to search for common upstream regions.

Overlay DE genes on pathway diagrams.

Relate expression levels to other information on cells, e.g. tumor types.

Identify temporal and spatial trends in gene expression.

Seek roles of genes based on patterns of co-regulation.

…many more!!!

Analysis Possibilities(adapted from Hongzhe Li, 2002)



Challenges

Time Dependent

Chronology

Logical

1800s – DATA30-60s – METHODS50-70s – SOFTWARE1980s – COMPUTER

cDNA

Human Dependent

Skill Integration

QuantitativeComputer Sci.StatisticiansMathematicians …….

Non-QBiochemistsPhysiologistsPathologists …….

BANANA EGG

“banana omelette”

Historical Excitement Balance Interdisciplinary

Data Dependent

Paradigm

Distribution

Source Size




Array Process

cDNA “A” Cy5 cDNA “B” Cy3

Tissue Samples

Treat A Treat B

mRNA Extraction & Amplification

Hybridization

Laser 1 Laser 2

Optical Scanner

+

Image Capture

Analysis



What you see isWhat you get



Egg Level (Biochemist):1. Preparation (Printing) of the Chip2. RNA Extraction, Amplification and Hybridisation3. Optical Scanner (Reading)

Banana Level (Quantitative):1. Design2. Image (data) Quality3. Data Analysis

Replication:

1. Animal2. Sample3. Array4. Spot

Technical Concerns



############################################################## # GP3xCLI # # GenePix Processing Program by CSIRO Livestock Industries # # # # Enquiries: [email protected] # # Copyright (c) 2003 CSIRO-LI # ############################################################## GPR Input: F12.gpr Processed on: Tue Apr 8 13:40:01 EST 2003 =-=-=-=-=-=-= IMAGE QUALITY =-=-=-=-=-=-=-= Total No. of Spots ------------------------> 19200 Spots with Flag = -50 --------------------> 4720 Spots with Flag = -100 --------------------> 12 Red dye with Background >= Foreground ---> 892 Green dye with Background >= Foreground ---> 915 Median to Mean Correlation Analysis: DATA LEFT RED GREEN Corr Raw Log2 Raw Log2 ______________________________________ > 0.00 19200 19200 19200 19200 > 0.20 19199 19200 19199 19200 > 0.40 19183 19200 19192 19200 > 0.60 19008 19200 19102 19200 > 0.80 17061 19199 18541 19198 > 0.85 14466 19193 17872 19196 > 0.90 10491 19137 15786 19181 =-=-=-=-=-=-= VALID SPOTS* =-=-=-=-=-=-=-= Total No. of Valid Spots -----------------> 14433 Percentage of Valid Spots -----------------> 75.2 Total No. of Genes ------------------------> 7220 Mean No. Repetitions -----> 2 for 6600 Genes Min. No. Repetitions -----> 1 for 580 Genes Max. No. Repetitions -----> 24 for 8 Genes Log(R/G) vs 0.5*Log(R*G) ________ ____________ N 14433 14433 Mean -0.017 10.327 Std 0.617 2.079 Min -8.711 3.246 Max 4.030 15.994 Correlation 0.362 Log(R/G) across Intensity Values Intensity Spots % <0 % >0 __________________________________ ( 0 , 4) 4 100.0 0.0 ( 4 , 8) 1499 74.1 25.9 ( 8 , 12) 9847 40.4 59.6 (12 , 16) 3083 17.3 82.7 __________________________________ *NB: Valid Spot defined as spots with Background < Foreground for both Red and Green channels and with a Quality Flag of 0.

Technical Concerns: Image (Data) Quality GP3xCLI



Clever Programming Tailored to your needsN=1

for filename in R16T0S1.gpr R16T0S2.gpr R16T24S1.gpr R16T24S2.gpr S32T0S1.gpr S32T0S2.gpr S32T24S1.gpr S32T24S2.gprdo

# Get valid readings, compute log ratios

awk 'NR>30 && $NF>=0 && $4!="no_spot" && substr($4,1,5)!="score" && \ substr($4,1,5)!="custo" && substr($4,1,6)!="spotre" && \ $9>$12 && $18>$21 {print $4, $9-$12, $18-$21, log($9-$12)/log(2.0), \ log($18-$21)/log(2.0)}' $filename | sort > junk1awk '$2!=$3 {print $0, $4-$5, 0.5*($4+$5)}' junk1 > junk2

# get the median of log ratios

REC=`wc -l junk2 | awk '{print int($1/2)}'`MED=`sort -n +5 junk2 | awk -v rec=$REC 'NR==rec {print $6}'`echo "Median of file" $filename " = " $MED

# Global normalization: substract the median to each log ratio

awk -v median=$MED -v slide=$N '{print "Slide_"slide, int(slide/2+.5), \ $1, $6-median}' junk2 | sort +2 > dat.$N

N=`expr $N + 1`done

cat dat.* > total.dat



Technical Concerns: Experimental Design

O

B

A

AB

O

B

A

AB

O

B

A

AB

Reference Loop All-Pairs

Variance of Estimated Effects(Relative to the All-Pairs)

Reference

1132

Loop

4/31

8/31

All-Pairs

1121

Main effect of AMain effect of BInteraction ABContrast A-B



Technical Concerns: Experimental Design

Glonek & Solomon Factorial and Time Course Designs for cDNA Microarray Experiments

• Read pp 1-2

• DefinitionA design with a total of n slides and design matrix X is said to be admissibleif there exists no other design with n slides and design matrix X* such that

ci* ciFor all i with strict inequality for at least one i. Where ci* and ci are respectivelythe diagonal elements of (X*’X*)-1 and (X’X)-1.

• Read pp 24



What is the No. of Possible Configurations?

No. of Arrays: (S-1) to S·(S-1)

S = 3 2 6

S = 4 3 12

S = 12 11 132



SA-1



Pie-Bald black Non-Pie-Bald black

Normal

White

Recessive SA-1 = 53 = 125



x5

x5

x5

x5

x5

x5

x5

x5

x5

x5

x5

x5

x5

x5

x5

x5

x5

x5

x5

x5

x5

x5

x5

x5

x5



0 hr 24 hr


SA-1 = 1210 = 62 Billion!


0 hr 24 hr

R

R

R R

R

R

R

R

RR

R

R

G

G

G G

G

G

G

G

G

G

G

G



F HS

M TM

F HS

M HS

F TM

M HS

F HS

M HS

R

R

R

R

R

R

R

R

R

R

R

R

R

R

G

GG

G

G

G

G

G

G

G

G

G

G

G


24: 23 To 552

14: 13 To 182

pooling


Slide RES SUS 0 3 24 M F HS TM

1 0 0 1 0 -1 0.066 -0.066 0.266 -0.266

2 0 0 -1 1 0 0.600 -0.600 -0.600 0.600

3 1 -1 1 -1 0 -0.600 0.600 -0.400 0.400

4 -1 1 -1 0 1 0.600 -0.600 0.400 -0.400

5 -1 1 1 -1 0 -0.600 0.600 1 -1

6 1 -1 1 -1 0 0.666 -0.666 -0.400 0.400

7 1 -1 -1 0 1 -0.666 0.666 0 0

8 0 0 1 0 -1 -0.333 0.333 0 0

9 0 0 0 -1 1 0.333 -0.333 -0.666 0.666

10 0 0 0 1 -1 -1.000 1.000 0 0

11 -1 1 0 -1 1 -0.500 0.500 0 0

12 1 -1 0 1 -1 -1.000 1.000 -1 1

13 -1 1 0 1 -1 0.666 -0.666 0.666 -0.666

14 0 0 0 1 -1 -1.000 1.000 0 0



RES SUS 0 3 24 M F HS TM

RES 8 -8 1 0 -1 -1.766 1.766 -3.866 3.866

SUS 8 -1 0 1 1.766 -1.766 3.866 -3.866

0 8 -4 -4 -1.335 1.335 0.666 -0.666

3 10 -6 -1.033 1.033 -0.468 0.468

24 10 2.368 -2.368 -0.198 0.198

M 6.247 -6.247 0.493 -0.493

F 6.247 -0.493 0.493

HS 3.798 -3.798

TM 3.798

Sum(ABS) 29.3 29.3 22.0 23.0 27.1 21.7 21.7 17.6 17.6

Sum(ABS) 26.8 26.8 39.1 23.1 17.3 7.1 7.1 14.3 14.3

Reference Design



Technical Concerns: Statistical Analysis

Technique Choice AimReal Ideal

1. Transformation Base-2 Log Numerically tractable Gaussian

2. Normalisation Location: M - c Systematic effects Gaussian2.i. Global: - Mean

- Median - Regr. Coeff (LOWESS)

2.ii. Local: - LOWESS within print-tip-group

3. Standarisation Scale Parameter Stabilise variance Gaussian

Data Beautifying Techniques


Pin group (sub-array) effects

Boxplots of log ratios by pin groupLowess lines through points from pin groups


Technical Concerns: Statistical AnalysisAssumption: The proportion of genes that are DE is minimal

Q: Which genes to use?A: Only the ones (housekeeping) that we know are not DENB: “Boutique” arrays become a nuisance

Adapted from T Speed 2002



Except Log2, everything else applies only to Ratios:

M = log2(R/G)

Except Log2, everything else applies only within slide

Everything is beautified to identify DE genes straight from “M vs A” plot (A = Average) from a single slide or from a function of M’s (t-stat) across slides


Data Beautifying Techniques



Include the “possible systematic sources of variation” into a model-based (eg. ANOVA) analysis and the data will be implicitly normalised. Then check the residuals.

Whenever possible, avoid ratios.


Log2(Intensity) = Array +Dye +A*D +Treatment +Sample +Gene +Gene*Treatment +Gene*Sample +Residuals

Normalisation Model

Gene Model



Rockhampton Model

Log2(Intensity) = Design +Array +

Dye + Array*Dye + (Diet +) Gene + Gene*Diet + Residuals

N Levels 2

14 228 3

7,34722,041

300,936

RockhamptonMEDIUM

(4 Animals)LOW

(3 Ani, 1 Rep)

(pooled 3 Anim)HIGH

(pooled 2 Anim)

MEDIUM(Pooled & Ampl)

LOW(Pooled & Ampl)

Reference

All Pairs



Japanese Model

Log2(Intensity) = Array +Dye +

Array*Dye + Breed + Time + Breed*Time + Gene +

Gene*Br*Ti + Residuals

N Levels 12

2 24

2 36

590035,400

259,080

Japanese

JANUARY 01

JAPANESE

JUNE 01

JAPANESE

OCTOBER 01

JAPANESE

HOLSTEIN HOLSTEIN HOLSTEIN



Rockhampton Japanese

ResidualGene

Gene*Treatment

% Total VarianceExplained by

GeneGene*Treatment

0.7203.6640.137

81.13.0

0.8892.8830.133

73.83.4

REML



Clever Programming Tailored to your needs

T24 - T0

-4

-2

0

2

4

-4 -2 0 2 4

Resistant

Dis

ease

Interaction Solutions

Your Needs: “Important values are…”1. Away from (0,0)2. In quadrants 1 and 4.

Generate a new variable:

+1.0*[(R24-R0)+(S0-S24)] if R0<R24 & S0>S24

+0.5*[(R24-R0)+(S24-S0)] if R0<R24 & S0<S24

-0.5*[(R0-R24)+(S0-S24)] if R0>R24 & S0>S24

-1.0*[(R0-R24)+(S24-S0)] if R0>R24 & S0<S24

…then apply model-based clustering.



Clever Programming BAYESMIX



Human Dependent

Challenges

Interdisciplinary Skills

Minimal knowledge of the application discipline is needed

…..failing that, the Statisticians will win, ..…but with the wrong weapons.

1. Amount of Expression = Amount of Response2. Same cut-off point to judge all genes3. Over-emphasis in normalization (hence, despise “Boutique Arrays”)4. Over-emphasis in variance stabilization5. Over-emphasis in controlling false-positives



The Statistical Analysis of cDNA Microarray Data:

General:1. Still in its infancy (…possibly even embryonic stage)2. Many decisions have a heuristic rather than a

theoretical foundation3. No hope for a “One size fits all” software (even method)4. Safer to aim towards “Tailor to one’s needs”5. Integration of interdisciplinary skills is a must

Conclusion

Livestock Species:1. Tailing humans (…at the moment)2. Strong background knowledge of genetics accumulated3. Journals will soon be inundated4. We have the opportunity to participate



Thank You!


mmi genomics/ucd & msu visits – may 2003 design and analysis of cdna microarray experiments at...

Documents

correlation analysis

analysis possibilitiesadapted

getmmi genomicsucd msu

microarray technical

roles of genes

percentage of valid

image quality

green dye