Perfomance of WGA DNA in an Illumina GoldenGate BeadArray assay Free

38

Gene association studies require well phenotyped study participants, some of whom may not be able to provide sufficient high quality genomic DNA. Whole genome amplification (WGA) offers a means to enrich DNA quantities allowing for optimal selection of cases and controls. Methods: We recently genotyped 1536 SNPs using an Illumina GoldenGate™ BeadArray assay in an ovarian cancer study consisting of 2368 cases and controls, using both genomic (n=1086) and WGA (n=1282) DNA. To understand the performance of WGA compared to genomic DNA, we included 122 samples in duplicate, representing both genomic and WGA. WGA DNA was prepared from 100ng DNA, using the REPLI-G WGA protocol (Qiagen). We also compared the performance of replicates of WGA preparations (n=85) and replicate WGA preparations (n=15). There was extreme discordance between WGA and genomic DNA for two samples, which were excluded from the analysis below, suspected resulting from sample mix-up. SNP call rate thresholds were evaluated at two specific levels, ≥95% and ≥99%. Results: Among the 1431 SNPs with an overall call rate ≥95%, 97% of the SNPs yielded genotypes with the genomic DNA, while 93.62% did so when WGA DNA was used. 98% of the genomic DNA and 93.22% of the WGA DNA yielded successful genotypes. Among the 1368 SNPs with an overall ≥99%, 93.82% SNPs yielded genotypes with the genomic DNA while 86.39% did so with the WGA DNA. 91% of the WGA DNA and 97.5% for genomic DNA yielded successful genotypes. Using data from the 1431/1368 SNPs (at the ≥95%/≥99% levels respectively) successfully genotyped in both preparations, concordance between replicates of WGA was 99.99%/100% and between replicate WGA was 99.93%/99.99%. Comparing WGA and genomic DNA, the concordance was 99.1%/99.5%; there were 1526/838 inconsistencies, ranging from 0 mismatches (n=7/120) to 46 having >9 mismatches. Over 65%/88% of the discordances were due to a reduction towards homozygosity in the WGA DNA, reflecting potential allele drop out during the WGA process. Conclusion: The high concordance between genotype calls from genomic and WGA DNA suggests that WGA DNA may be successfully used in gene association studies, although the call rates were lower with WGA DNA. Using a SNP call rate threshold of at least 95% yielded less than 1% discordance in this study.