Skip to main content

Chip type: GenomeWideSNP_5

Genome-Wide Human SNP Array 5.0

Affymetrix' press release: Affymetrix Releases Its Most Comprehensive Single Array for Genetic Studies, Feb 20, 2007.


The Genome-Wide Human SNP Array 5.0 is a PM-only chip. SNPs: It contains the exact same set of 500,568 SNPs as was available on the Mapping250K_{Nsp|Sty} chip set. They have the same rsID. The difference is that this array and the 500K arrays use different probes to interrogate the SNPs. The probes on this chip type were chosen from the ones on 500K arrays. Allele A probes were selected and put down in the array as technical replicates, and same for allele B probes. There are equal number of allele A and allele B probes, but they are may not be with the same "shift". That is, the A and B probes were not selected in pairs. The probes were selected according to their genotype performance by looking how well they separated the genotype groups. However, by "chance", approximately 350K of the SNPs turned how to have paired A and B probes. The other 150K SNPs have unpaired A and B probes. Since Affymetrix, in general, optimize the layout of probes from their sequences etc., the unpaired probes are spatially dislocated, whereas the paired probes are next to each other on the array. There are also 2748 units named "AFFX-SNP_nnnnnn"; I don't what these are. Copy-number non-polymorphic probes (CNPs): In addition to the above 500K SNPs, there are also 340,742 CN units. Almost all of these are single-probe units. Their positions were chosen to "fill" in the gaps between the SNPs.
Source: Private communication with Affymetrix, Henrik Bengtsson, 2007-04-22

Two CDFs:

Currently Affymetrix provides two different CDF files for this chip type, one named GenomeWideSNP_5.cdf that contains only the SNPs, and one named GenomeWideSNP_5.Full.cdf that contains the SNPs as well as all the non-polymorphic units.

Furthermore, this is what Affymetrix writes in their 'Data Sheet: Affymetrix: Genome-Wide Human SNP Array 5.0' (2007): "Two alternate SNP List (CDF) files are available. These files identify which SNPs are available for downstream analysis. GenomeWideSNP_5.cdf is the default set of 440,794 SNPs that are accessible via the Genotyping Console (BRLMM-P algorithm). GenomeWideSNP_5.Full.cdf is for advanced users who wish to look at all SNPs from the previous-generation Mapping 500K Array Set. This CDF file can only be accessed from the command-line tool, snp5-probeset-genotype, which is part of the Affymetrix Power Tools distribution. The full CDF file includes SNPs that may have lower per-SNP accuracy or call rates. It is expected that the performance of some of these SNPs will improve with different or future algorithms." /HB 2007-11-26

...and now also an "Alternative CDF":

"Affymetrix is providing alternative CDF files, for the Affymetrix genotyping array- Genome-Wide Human SNP Array 5.0. These CDF files provide uniform access to the Copy Number probe content by presenting "one probe per probeset" in both the full and default updated CDF files. This uniform access applies to both Affymetrix genotyping arrays with Copy Number content- Genome-Wide Human SNP Array 5.0 and Genome-Wide Human SNP Array 6.0." /HB 2007-12-02


This is the CDF file that is recommended and support in aroma.affymetrix. Note, make sure to rename it such that you replace the dots with commas.

> cdf <- AffymetrixCdfFile$byChipType("GenomeWideSNP_5", tags="Full,r2")
> cdf

Path: annotationData/chipTypes/GenomeWideSNP_5
Filename: GenomeWideSNP_5,Full,r2.cdf
Filesize: 249.46MB
File format: v4 (binary; XDA)
Chip type: GenomeWideSNP_5,Full,r2
Dimension: 2166x2166
Number of cells: 4691556
Number of units: 920928
Cells per unit: 5.09
Number of QC units: 6
RAM: 0.00MB


The early access version of this chip type was labeled BI_SNP with the below CDF. This means that some early CEL files might have this chip type in the header instead of newer GenomeWideSNP_5 label.

> cdf <- AffymetrixCdfFile$byChipType("BI_SNP")
> cdf
Path: annotationData/chipTypes/BI_SNP
Filename: BI_SNP.cdf
Filesize: 237.08MB
Chip type: BI_SNP
RAM: 0.00MB
File format: v4 (binary; XDA)
Dimension: 2166x2166
Number of cells: 4691556
Number of units: 844401
Cells per unit: 5.56
Number of QC units: 6

Q: What SNPs are on the GenomeWide Human SNP v5.0 array?

A: The SNPs on this array, are identical to the ones in the 500K chip set, they are even in the same order (but using different probes and PMs only). In fact, the ordering are even the same in the CDFs as this example illustrates:

cdfNsp <- AffymetrixCdfFile$byChipType("Mapping250K_Nsp")
cdfSty <- AffymetrixCdfFile$byChipType("Mapping250K_Sty")
cdf50 <- AffymetrixCdfFile$byChipType("GenomeWideSNP_5",
un500K <- c(getUnitNames(cdfNsp), getUnitNames(cdfSty))
snps500K <- grep("\^SNP_", un500K, value=TRUE)
un50 <- getUnitNames(cdf50)
snps50 <- grep("\^SNP_", un50, value=TRUE)
print(identical(snps50, snps500K));   # Gives TRUE



By Affymetrix: