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

• DOI: 10.1371/journal.pgen.1002078
• Publication Date: 2011-05-26T17:06:20Z
• Authors: Federico Innocenti; Gregory M. Cooper; Ian B. Stanaway; Eric R. Gamazon; Joshua D. Smith; Snezana Mirkov; Jacqueline Ramirez; Wanqing Liu; Yvonne S. Lin; Cliona Moloney; Shelly Force Aldred; Nathan D. Trinklein; Erin Schuetz; Deborah A. Nickerson; Ken E. Thummel; Mark J. Rieder; Allan E. Rettie; Mark J. Ratain; Nancy J. Cox; Christopher D. Brown
• Keywords: Epistasis; Genetic Association; SNP

The discovery of expression quantitative trait loci ("eQTLs") can help to unravel genetic contributions complex traits. We identified determinants human liver gene variation using two independent collections primary tissue profiled with Agilent (n = 206) and Illumina 60) arrays SNP genotyping (550K), we also incorporated data from a published study 266). found that ∼30% SNP-expression correlations in one failed replicate either the others, even at thresholds yielding high reproducibility simulations, quantified numerous factors affecting reproducibility. Our suggest drug exposure, clinical descriptors, unknown associated ascertainment analysis have substantial effects on controlling for hidden confounding variables significantly increases replication rate. Furthermore, reproducible eQTL SNPs were heavily enriched near starts ends, subsequently resequenced promoters 3′UTRs 14 genes tested haplotypes luciferase assays. For three genes, significant haplotype-specific vitro functional differences correlated directly levels, suggesting many bona fide eQTLs result variants be mechanistically isolated high-throughput fashion. Finally, given our design, able discover validate hundreds eQTLs. Many these relate traits which liver-specific analyses are likely relevant, dozens potential connections disease-associated loci. These included previously characterized contributors diabetes, response, lipid they novel candidates such as role NOD2 leprosy risk C2orf43 prostate cancer. In general, work presented here will valuable future efforts precisely identify functionally characterize variety