Pool-based association scan for type 2 diabetes using 49,419 SNPs in a 92.2 Mb region on chromosome 6

L.J. Scott, A.B. Sparks, K.L. Mohlke, D.A. Hinds, M. Li, D.G. Ballinger, P.S. Chines, J. Tuomilehto, R.N. Bergman, K.A. Frazer, R.M. Watanabe, F.S. Collins, D.R. Cox, M. Boehnke

U. Michigan, Ann Arbor, MI; Perlegen Sciences, Mountain View, CA; NHGRI, Bethesda, MD; Nat’l Public Health Inst., Helsinki, Finland; U. Southern California, Los Angeles, CA

The aim of the Finland-United States Investigation of NIDDM Genetics (FUSION) Study is to identify genetic variants that contribute to development of type 2 diabetes (T2D). Based on 735 FUSION families, we identified overlapping linkage signals in a 90 cM (92.2 Mb) region of 6q. These included signals for T2D (LOD=2.66 at 95.5 cM), for the subset of T2D families with highest mean HDL/total cholesterol ratio (LOD=7.93 at 78.0 cM), and for fasting insulin QTL (LOD=2.65 at 139.0 cM). To scan for diabetes-associated variants, we assayed 49,419 SNPs in the 92.2 Mb interval (1.8 kb average density) using Perlegen’s high-density oligonucleotide array genotyping platform. Eight replicate measurements were collected for each SNP from each of 3 pools: 464 T2D cases, 152 HDL ratio cases, and 205 elderly non-diabetic controls. Indirect estimates of case-control allele frequency differences were calculated from the array data for each of the 47,269 SNPs (96% of the total) that passed QC criteria (low replicate variability, n>5). Perlegen’s reference haplotype map was used to identify SNPs that occur in the same haplotype block, and are therefore expected to provide concordant information. We selected 624 SNPs for individual genotyping by defining thresholds for estimated allele frequency differences, with lower thresholds for SNPs closer to a linkage peak, within relevant portions of genes, or with concordant haplotype block information. For redundant SNPs in the same haplotype, only the SNP with the largest estimated allele frequency difference was selected. Genotyping of the individuals in the pools is underway to evaluate the selection criteria and to identify SNPs with significant allele frequency differences for follow-up genotyping in this and other populations.