paperKB
coga / coga-kb
Help
Sign in

Chunk #30 — Results — Assessing the impact of the intersection strategy on SNP imputation quality

Source
Imputation across genotyping arrays for genome-wide association studies: assessment of bias and a correction strategy.
Embedded
yes

Text

(HumanHap300-Duo version 2): average R2 < 0.9 for each MAF bin and average R2 = 0.71 across the MAF spectrum. Imputation quality resulting from the intersection of the Illumina 1M and Affymetrix 6.0 arrays was comparable to the imputation quality resulting from the smallest input genotyped SNP set (that is, the largest number of arrays combined among the Illumina arrays), with average R2 = 0.73 across the MAF spectrum. In African Americans (Fig. 3b), poorer imputation quality was observed for all input genotyped SNP sets, as expected, but the relative patterns were similar to those observed in European Americans. The lowest imputation quality in African Americans resulted from the intersection of all five Illumina arrays, with average R2 = 0.56 across the MAF spectrum.