Applying the RPM to the dichotomous nicotine dependence phenotype among smokers in COGEND allowed the development of hypotheses that joint effects of specific genetic loci (tagged by SNPs genotyped in COGEND) increased the variance explained. To minimize the chance of reporting the same signal multiple times, we report two pair-wise results as independent only if at least one SNP from the first pair is not in linkage disequilibrium (r2 < 0.4) with either of the SNPs in the second pair.
