Interchromosomal LD in HapMap Genotypes

Then, using all of these tags were used to calculate intrachromosomal LD within each panel. Analysis was done using an EM algorithm based on code from Goncalo Abecasis. Marker pairs with r2 0.6 were stored for further analysis. As shown in Table 2, the number of distinct pairs of SNPs with r2 0.6 and 0.8 are shown. The numbers are shown for the original tag set. Additionally, to facilitate cross panel comparisons, all the SNPs which were captured by the original tag sets, were also analyzed for high interchromosomal LD. This extension was done because we wanted to investigate if any SNP pairs displayed high interchromosomal LD in multiple analysis panels. With initial analysis limited to tag SNPs, while specific regions might overlap, the exact SNPs used as tags might not be the same. The higher number of SNP pairs seen for the YRI panel, particularly at the lower threshold, presumably derives from the larger number of polymorphic SNPs in this panel, and is artifactual. Since it was presumed that a large source of observed interchromosomal LD might come from SNPs which were mismapped, the maximum local r2 value was determined. It was presumed that if (at least) one of a SNP pair was mismapped, then they would be much less likely to display high local LD for both SNPs.