Advertisement

< Back to Article

Identification, Replication, and Functional Fine-Mapping of Expression Quantitative Trait Loci in Primary Human Liver Tissue

Figure 3

eQTL replication.

(A) Number of gene expression traits (y-axis, log10 scaled) with best associated cis-eQTLs (black) and trans-eQTLs (blue) as a function of BF (x-axis). Counts at each threshold are plotted separately for all probes (solid) and for probes without known polymorphisms (dashed). (B) cis-eQTL associations were tested in two replication sample sets, UW and Merck. Replication sample linear regression t-test p-values were binned (x-axis) and the number of genes per bin (y-axis) is displayed separately for the UW (red) and Merck (blue) replication sets. Data are plotted for all eQTLs (thin lines) and for significant eQTLs (heavy lines). (C) cis-eQTL replication rate (y-axis) is depicted as a function of UC BF (x-axis) for UW (red), Merck (blue), either replication study (green), and for the n = 60 resampled data (purple). Replication rates are plotted separately for probes overlapping known polymorphisms (solid lines) and for probes not overlapping an annotated polymorphism (dashed). (D) Cis-eQTL replication rate (y-axis) plotted as a function of quantile binned SNP to TSS distance. Per bin mean (points) and standard errors (lines) are plotted separately for associations with UC BF>0 (grey) with UC BF>5 (black), and with UC BF>10 (blue). (E) Trans-eQTL replication p-values were binned (x-axis) and the number of genes per bin (y-axis) is displayed separately for the UW (red) and Merck (blue) replication sets. Data are plotted for all eQTLs (thin lines) and for significant eQTLs (heavy lines). (F) trans-eQTL replication rate (y-axis) is depicted as a function of UC BF (x-axis) for UW (red), Merck (blue), either replication study (green), and for the resampled data (purple).

Figure 3

doi: https://doi.org/10.1371/journal.pgen.1002078.g003