To provide complementary analyses, we began by identifying clusters of positive SNPs that fall within genes in each sample. We identify the genes that are tagged by at least one cluster of nominally positive SNPs from each of the two samples (Table 1). We focus on a set of criteria for approach (2) that produce about the same ratio of “true” genes to total genes (“true”+“chance”) (0.43 vs 0.47) as those likely to be observed using the criteria on which we focused in approach (1). The SNPs that we identify in this secondary analysis thus a) display t values with p<0.05 significance in one abuser vs control comparison, b) cluster, so that at least four of these positive SNPs lie within 10kb of each other in this abuser vs control comparison, c) identify the same genes as clustered positive SNPs from the other sample (Table 2, legend, and Table S1). Three hundred forty one genes are identified in this way (Table S1). Fifty-two of the 104 genes listed in Table 2 based on approach (1) are also identified after application
