Utility of the pooling approach as applied to whole genome association scans with high-density Affymetrix microarrays

Alexandra Schosser,Emma Meaburn,Sarah Cohen-Woods,Cerisse Gunasinghe,Ian W Craig,Ania Korszun,Amanda Elkin,Joanna Gray,Leonard C Schalkwyk,Katrina Pirlo,Peter Mcguffin,Darya Gaysina,Anne E Farmer,Lisa Jones

doi:10.1186/1756-0500-3-274

Abstract

BackgroundWe report an attempt to extend the previously successful approach of combining SNP (single nucleotide polymorphism) microarrays and DNA pooling (SNP-MaP) employing high-density microarrays. Whereas earlier studies employed a range of Affymetrix SNP microarrays comprising from 10 K to 500 K SNPs, this most recent investigation used the 6.0 chip which displays 906,600 SNP probes and 946,000 probes for the interrogation of CNVs (copy number variations). The genotyping assay using the Affymetrix SNP 6.0 array is highly demanding on sample quality due to the small feature size, low redundancy, and lack of mismatch probes.FindingsIn the first study published so far using this microarray on pooled DNA, we found that pooled cheek swab DNA could not accurately predict real allele frequencies of the samples that comprised the pools. In contrast, the allele frequency estimates using blood DNA pools were reasonable, although inferior compared to those obtained with previously employed Affymetrix microarrays. However, it might be possible to improve performance by developing improved analysis methods.ConclusionsDespite the decreasing costs of genome-wide individual genotyping, the pooling approach may have applications in very large-scale case-control association studies. In such cases, our study suggests that high-quality DNA preparations and lower density platforms should be preferred.

Highlights

We report an attempt to extend the previously successful approach of combining SNP microarrays and DNA pooling (SNP-MaP) employing high-density microarrays
We have performed a genome-wide association study (GWAS) using pooled DNA of a depression case-control sample and the Affymetrix Genome-Wide Human SNP Array 6.0, and followed-up the ‘top-hit’ SNPs using confirmatory individual genotyping with Sequenom MassARRAY® iPLEX Gold or TaqMan®
The top-ranked SNPs of the pooling GWAS were followed-up by individually genotyping the samples used to construct the pools, and the validity of the SNP-MaP approach was assessed by comparing allele frequency estimates from pooled DNA (RAS scores) with individual genotyping data of 110 SNPs genotyped with either Sequenom MassARRAY® iPLEX Gold (108 SNPs) or TaqMan® (2 SNPs)

Summary

Introduction

We report an attempt to extend the previously successful approach of combining SNP (single nucleotide polymorphism) microarrays and DNA pooling (SNP-MaP) employing high-density microarrays. We report an attempt to extend the previously successful approach of SNP (single nucleotide polymorphism) microarrays and DNA pooling (SNP-MaP) [1,2,3,4,5,6,7,8,9,10,11,12,13,14]. Whereas earlier studies had employed a range of Affymetrix SNP microarrays interrogating between 10 K to 500 K SNPs [15], we used the Affymetrix SNP 6.0 which displays 906,600 SNP probes and 946,000 probes for the interrogation of CNVs (copy number variations) We have performed this genome-wide association study (GWAS) using pooled DNA from a large depression case-control sample (1418 cases, 1301 controls), and the SNPs with the largest differences between cases and controls were individually genotyped in the sample used to construct the pools. In contrast to the suitability of Affymetrix microarrays (up to the 500 K) for successful analysis of DNA pools as established by other groups, the properties of the Affymetrix SNP 6.0 used in the current study were unproven at the time of investigation

Methods

Results

Conclusion