Assessment of predicted pm relationships from European versus Chinese and Japanese data.

Genes with significantly different expression between Asian and European subjects were identified by Spielman et al [20] and we divided these into two groups - European-up (Eu) and European-down (Ed), using the average expression for Europeans minus the average expression for Asian (Chinese and Japanese). These two probe-lists were used to make a pair-list of all possible combinations of Eu : Ed, and filtered to only contain the probes which appear in our final discretized data (Z = 0.4). For comparisons with the non-Affymetrix data (SAFHS and Decode) this Affymetrix probe pair-list was converted into a gene symbol pair-list. The comparisons show the number of common unique pairs between the networks and the Eu : Ed pair-list.