All traits were tested for associations to cis- or trans-acting SNPs individually. Study-wise individual cutoff levels were calculated to test for cis- and trans-associations. To derive a study-wise cis-cutoff level at significance level α=0.05, we first determined all cis-acting SNPs within the data set. To allow for detailed comparisons of our results with those of Schadt et al.,16 we used the same definitions, that is, cis-acting SNPs are those that occur within 1 Mb upstream or downstream of the known or predicted 5′- and 3′-ends of any trait gene. Thus, the study-wise Bonferroni-adjusted P-value threshold equals 0.05/2 607 664=1.92 × 10−8, where 2 607 664 is the total number of possible cis-associations based on the above definition. Accordingly, the study-wise trans-specific Bonferroni's cutoff level was computed as the total number of tests for trans-associations, that is, 0.05/4.6 billion=1.08 × 10−11.
