We sought information of the function of the eQTLs identified in this study by taking the pooled Affymetrix and Illumina expression results and grouping them into Gene Ontology (GO) categories using annotation information downloaded from the manufacturer's website. We counted the percentage of probes mapped by at least one SNP (probes representing one of the 14,177 genes in Table 1) for each GO category and tested whether this percentage was significantly higher than a random category of probes of the same size. Both Z-score and permutation-based P-values were used to assess significance. Accounting for 461 GO terms and a 5% family-wise false-positive rate, the Bonferroni correction gave a significant P-value threshold of 1.08 × 10−4. We found that transcripts with the strongest eQTLs were associated with the regulation of transcription (Supplemental Table 1). Highly significant enrichment was also found for categories concerning protein, lipid, and carbohydrate metabolism. While the evolutionary value of genetic variation in metabolism is clear, it was unexpected that the most variably expressed genes in the human genome may be regulators of transcription. Immune response genes were significantly overrepresented among eQTLs, although to a lesser extent, reflecting the genome response to a different kind of environmental pressure.
