We genotyped 14 SNPs × 35 samples in duplicate (N=490 genotype pairs); both genotypes were available for 465 (94.9%), of which 0 were discordant, estimating our overall genotyping error rate at less than 1 in 465 (<0.21%). All markers except rs1480481 (p=1.13×10−10 and not further analyzed) satisfied our HWE cut-off for inclusion. Linkage disequilibrium patterns in this sample are virtually identical to those in HapMap data (Figure S1).
