Teslovich et al. [15] reported common variants associated with plasma concentrations of low-density lipoprotein cholesterol (LDL), high-density lipoprotein cholesterol (HDL) and triglyceride (TG) levels in more than 100,000 individuals of European ancestry. They then reported the correlations between the lead SNPs at the loci they found and the expression levels of transcripts in liver. For the lipid dataset we have access only to summary statistics. The liver expression dataset used in this analysis is the same as the one used in [15]. In Teslovich et al., regions are defined within 500 kilobases of the lead SNPs, and the threshold for significance is . At this threshold, they found 38 SNP-to-gene eQTLs in liver (Supplementary Table 8 of [15]). Table S1 shows our results for these 38 previously reported colocalisations. A complete list of all our identified colocalisations (independently of previous reports) is provided in Tables S2, S3, S4, S5 (broken down by lipid traits). Using the coloc web server for this analysis with a PP4 >75, it took 1 minute to complete chromosome 1 and approximately 7 minutes to analyse the entire imputed genome-wide data on a laptop.
