1. Copy-number estimation using Robust Multichip Analysis - Supplementary materials for the aroma.affymetrix lab session
Henrik Bengtsson & Terry Speed
Dept of Statistics, UC Berkeley
August 7, 2007
BioC 2007

2. Affymetrix chips

3. Generic Affymetrix chip
> 1 million identical 25bp sequences * *
1.28 cm
1.28 cm
6.5 million probes/ chip
Feature size: 100µm to 18µm to 11µm and now 5µm.
Soon: 1µm, 0.8µm, with a huge increase in number of probes.

4. Abbreviated generic assay description
Start with target gDNA (genomic DNA) or mRNA.
Obtain labeled single-stranded target DNA fragments for hybridization to the probes on the chip.
After hybridization, washing, staining and scanning we get a digital image. This is summarized across pixels to probe-level intensities before we begin. They are our raw data.

5. Affymetrix probe terminology
Target DNA: ...CGTAGCCATCGGTAAGTACTCAATGATAG...
|||||||||||||||||||||||||
Perfect match (PM): ATCGGTAGCCATTCATGAGTTACTA
Mis-match (MM): ATCGGTAGCCATACATGAGTTACTA
25 nucleotides *
other PMs
Other DNA
Other seq. * PM
Target seq. X * * * MM

6. Affymetrix SNP chips (Mapping 10K, 100K, 500K)

7. Single Nucleotide Polymorphism (SNP)
Definition:
A sequence variation such that two chromosomes may differ by a single nucleotide (A, T, C, or G).
Allele A: A
...CGTAGCCATCGGTA/GTACTCAATGATAG...
Allele B: G
A person is either AA, AB, or BB at this SNP.

8. Probes for SNPs AA AB BB * * * * PMA: ATCGGTAGCCATTCATGAGTTACTA
Allele A: ...CGTAGCCATCGGTAAGTACTCAATGATAG... Allele B: ...CGTAGCCATCGGTAGGTACTCAATGATAG... PMB: ATCGGTAGCCATCCATGAGTTACTA
(Also MMs, but not in the newer chips, so we will not use these!) * *
PMA >> PMB AA *
PMA ¼ PMB AB *
PMA << PMB BB

9. Copy-number analysis with SNP arrays

10. Copy-number estimation using Robust Multichip Analysis (CRMA)
Preprocessing (probe signals)
allelic crosstalk (or quantile)
Total CN
PM = PMA + PMB
Summarization (SNP signals  )
log-additive PM only
Post-processing
fragment-length
(GC-content)
Raw total CNs R = Reference
Mij = log2(ij /Rj) chip i, probe j

11. Copy-number estimation using Robust Multichip Analysis (CRMA)
Cross-hybridization:
Allele A: TCGGTAAGTACTC
Allele B: TCGGTATGTACTC
CRMA
Preprocessing (probe signals)
allelic crosstalk (quantile)
Total CNs
PM=PMA+PMB
Summarization (SNP signals )
log-additive (PM-only)
Post-processing
fragment-length
(GC-content)
Raw total CNs
Mij = log2(ij/Rj) AA *
PMA >> PMB *
PMA ¼ PMB AB *
PMA << PMB BB

12. Copy-number estimation using Robust Multichip Analysis (CRMA)AA TT AT
CRMA
Preprocessing (probe signals)
allelic crosstalk (quantile)
Total CNs
PM=PMA+PMB
Summarization (SNP signals )
log-additive (PM-only)
Post-processing
fragment-length
(GC-content)
Raw total CNs
Mij = log2(ij/Rj)
PMT
PMA
offset +

13. Copy-number estimation using Robust Multichip Analysis (CRMA)
Preprocessing (probe signals)
allelic crosstalk (quantile)
Total CNs
PM=PMA+PMB
Summarization (SNP signals )
log-additive (PM-only)
Post-processing
fragment-length
(GC-content)
Raw total CNs
Mij = log2(ij/Rj)
PMT
PMA

14. Copy-number estimation using Robust Multichip Analysis (CRMA)
Crosstalk calibration corrects for
differences in distributions too
CRMA
Preprocessing (probe signals)
allelic crosstalk (quantile)
Total CNs
PM=PMA+PMB
Summarization (SNP signals )
log-additive (PM-only)
Post-processing
fragment-length
(GC-content)
Raw total CNs
Mij = log2(ij/Rj)
log2 PM

15. Copy-number estimation using Robust Multichip Analysis (CRMA)
Crosstalk calibration corrects for
differences in distributions too
CRMA
Preprocessing (probe signals)
allelic crosstalk (quantile)
Total CNs
PM=PMA+PMB
Summarization (SNP signals )
log-additive (PM-only)
Post-processing
fragment-length
(GC-content)
Raw total CNs
Mij = log2(ij/Rj)
log2 PM

16. Copy-number estimation using Robust Multichip Analysis (CRMA)AA
CRMA
Preprocessing (probe signals)
allelic crosstalk (quantile)
Total CNs
PM=PMA+PMB
Summarization (SNP signals )
log-additive (PM-only)
Post-processing
fragment-length
(GC-content)
Raw total CNs
Mij = log2(ij/Rj) *
PM = PMA + PMB AB * *
PM = PMA + PMB * BB * *
PM = PMA + PMB

17. Copy-number estimation using Robust Multichip Analysis (CRMA)
Preprocessing (probe signals)
allelic crosstalk (quantile)
Total CNs
PM=PMA+PMB
Summarization (SNP signals )
log-additive (PM-only)
Post-processing
fragment-length
(GC-content)
Raw total CNs
Mij = log2(ij/Rj)
The log-additive model:
log2(PMijk) = log2ij + log2jk + ijk
sample i, SNP j, probe k.
Fit using robust linear models (rlm)

18. Copy-number estimation using Robust Multichip Analysis (CRMA)
Longer fragments )
less amplified by PCR )
weaker SNP signals 
CRMA
Preprocessing (probe signals)
allelic crosstalk (quantile)
Total CNs
PM=PMA+PMB
Summarization (SNP signals )
log-additive (PM-only)
Post-processing
fragment-length
(GC-content)
Raw total CNs
Mij = log2(ij/Rj)
100K

19. Copy-number estimation using Robust Multichip Analysis (CRMA)
Longer fragments )
less amplified by PCR )
weaker SNP signals 
CRMA
Preprocessing (probe signals)
allelic crosstalk (quantile)
Total CNs
PM=PMA+PMB
Summarization (SNP signals )
log-additive (PM-only)
Post-processing
fragment-length
(GC-content)
Raw total CNs
Mij = log2(ij/Rj)
500K

20. Copy-number estimation using Robust Multichip Analysis (CRMA)
Normalize to get same
fragment-length effect
for all hybridizations
CRMA
Preprocessing (probe signals)
allelic crosstalk (quantile)
Total CNs
PM=PMA+PMB
Summarization (SNP signals )
log-additive (PM-only)
Post-processing
fragment-length
(GC-content)
Raw total CNs
Mij = log2(ij/Rj)

21. Copy-number estimation using Robust Multichip Analysis (CRMA)
Normalize to get same
fragment-length effect
for all hybridizations
CRMA
Preprocessing (probe signals)
allelic crosstalk (quantile)
Total CNs
PM=PMA+PMB
Summarization (SNP signals )
log-additive (PM-only)
Post-processing
fragment-length
(GC-content)
Raw total CNs
Mij = log2(ij/Rj)

22. Copy-number estimation using Robust Multichip Analysis (CRMA)
Preprocessing (probe signals)
allelic crosstalk (quantile)
Total CNs
PM=PMA+PMB
Summarization (SNP signals )
log-additive (PM-only)
Post-processing
fragment-length
(GC-content)
Raw total CNs
Mij = log2(ij/Rj)