Deconvoluting complex tissues for expression quantitative trait locus-based analyses

Abstract

Breast cancer genome-wide association studies have pinpointed dozens of variants associated with breast cancer pathogenesis. The majority of risk variants, however, are located outside of known protein-coding regions. Therefore, identifying which genes the risk variants are acting through presents an important challenge. Variants that are associated with mRNA transcript levels are referred to as expression quantitative trait loci (eQTLs). Many studies have demonstrated that eQTL-based strategies provide a direct way to connect a trait-associated locus with its candidate target gene. Performing eQTL-based analyses in human samples is complicated because of the heterogeneous nature of human tissue. We addressed this issue by devising a method to computationally infer the fraction of cell types in normal human breast tissues. We then applied this method to 13 known breast cancer risk loci, which we hypothesized were eQTLs. For each risk locus, we took all known transcripts within a 2 Mb interval and performed an eQTL analysis in 100 reduction mammoplasty cases. A total of 18 significant associations were discovered (eight in the epithelial compartment and 10 in the stromal compartment). This study highlights the ability to perform large-scale eQTL studies in heterogeneous tissues.

FGFR2 risk SNPs confer breast cancer risk by augmenting oestrogen responsiveness

“…This study examined a small number of patients (21 matched tumour-normal pairs and 10 tumour samples without corresponding normal tissue), but samples were microdissected to maximize any detectable differences. In contrast, a recent eQTL study of normal breast tissue failed to see an association between the risk alleles and FGFR2 expression ( 45 ), but in this study, tissue was not microdissected. It is not clear how this might fit with experimental data that suggest FGFR2 is important in increased branching during mammary gland development and in the generation of breast tumour-initiating cells ( 13–17 ): it is likely that FGFR2 may play distinct roles at different stages of breast development.…”

Single Nucleotide Polymorphisms of Caudal Type Homeobox 1 and 2 Are Associated with Barrett’s Esophagus

“…Nevertheless, these novel findings are preliminary and must be confirmed in subsequent studies with larger cohorts. We are currently recruiting a confirmation cohort for an expression quantitative trait loci (eQTL) study to associate these SNPs and haplotypes with gene expression [27]. North Carolina has a relatively high African-American population, and recruitment of more African-American patients will allow us to further understand the potential contribution of genetic variants in racial disparity of BE [28].…”

Functional annotation of breast cancer risk loci: current progress and future directions

“…Alternative in silico approaches to deconvolute cell-type-specific expression profiles have also been developed [ 81 – 83 ]. Whilst these have mainly been used to test for the association between clinical covariates and breast cancer prognosis [ 83 , 84 ], Seo et al used a deconvolution approach to examine gene expression in normal breast tissue [ 61 ]. Specifically, they modelled breast tissue as comprising four different cell types (adipocytes, epithelial, inflammatory and stromal), and identified eQTL associations at published breast cancer GWAS loci in two of these cell types—epithelial and stromal cells [ 61 ].…”