Expression Quantitative Trait Loci (eQTL)

eQTLs are classified by their genomic distance from the target gene. Cis-eQTLs are located near the gene they regulate, typically within a 1 Mb window of the transcription start site, often affecting promoter or enhancer regions. Trans-eQTLs act over long genomic distances, often on different chromosomes, and frequently encode transcription factors or signaling molecules that regulate entire gene networks. Cis effects tend to have larger individual effect sizes, while trans effects reveal higher-order regulatory architecture.

eQTL discovery uses linear regression or Spearman rank correlation to test millions of SNP-gene pairs for association. The standard model tests: Expression ~ SNP + Covariates + ε. Key covariates include:

• Principal components to correct for population stratification
• PEER factors to account for hidden batch effects
• Age, sex, and technical variables Multiple testing correction via Benjamini-Hochberg FDR is essential, with typical thresholds at FDR < 0.05.

A single genetic variant can have divergent or even opposite effects on gene expression depending on the tissue or cellular context. The GTEx Consortium has cataloged eQTLs across 54 human tissues, revealing that:

• 48% of eQTLs are shared across most tissues
• 52% are tissue-specific, enriched in distal regulatory elements
• Dynamic eQTLs respond to stimuli like drug treatment, immune activation, or hypoxia, capturing gene-by-environment interactions critical for disease modeling.

eQTLs are a primary tool for mechanistically interpreting GWAS hits. Colocalization analysis tests whether a GWAS disease signal and an eQTL signal share the same causal variant. Methods like COLOC and FINEMAP compute posterior probabilities for competing hypotheses:

• H3: Two distinct causal variants (no colocalization)
• H4: A single shared causal variant (colocalization) A posterior probability > 0.75 for H4 provides strong evidence that a GWAS locus acts through gene expression regulation.

Bulk tissue eQTLs confound cell-type-specific effects with compositional changes. Single-cell eQTL (sc-eQTL) mapping resolves regulation at cellular resolution, revealing:

• Cell-type-specific eQTLs masked in bulk RNA-seq
• Dynamic eQTLs varying along differentiation pseudotime
• Context-dependent effects only active in rare cell populations Pipelines like scQTLtools and SAIGE-QTL use mixed models with cell-level random effects to handle the sparsity and zero-inflation of single-cell data.

eQTL data enables TWAS, which imputes gene expression into large GWAS cohorts lacking expression data. The workflow:

• Train a genetic predictor of expression using eQTL reference panels
• Apply the predictor to GWAS genotypes to impute expression
• Test imputed expression for disease association Tools like PrediXcan, FUSION, and S-PrediXcan integrate eQTL weights from GTEx and other resources to prioritize causal genes at GWAS loci.

A statistical method for exploring relationships between two sets of high-dimensional variables by finding linear combinations that maximize their cross-correlation. In eQTL studies, sparse CCA is frequently applied to identify coordinated patterns between genetic variants and transcriptomic profiles, selecting only the most relevant features from each domain for improved biological interpretability.

A statistical process for removing the influence of extraneous variables that affect both the genetic exposure and the transcriptomic outcome. In eQTL mapping, failure to adjust for population stratification, batch effects, or hidden covariates can generate spurious associations. Methods like PEER factors or principal component analysis are standard for estimating and regressing out these latent confounders.

An unsupervised statistical framework that integrates multiple omics data types by decomposing their variation into a sparse set of latent factors. When eQTL data is combined with proteomics or epigenomics, MOFA can reveal latent axes of variation that capture the principal sources of coordinated biological and technical variability across molecular layers.

A multi-omics integration strategy that maps individual molecular features to predefined biological pathways before performing statistical analysis. Instead of analyzing single eQTL-gene pairs in isolation, this approach aggregates signals to the pathway level, dramatically reducing the multiple-testing burden and enhancing biological interpretability by revealing coordinated regulatory programs.

The computational correction of non-biological experimental variation introduced by processing samples in different batches, on different days, or in different laboratories. In multi-tissue eQTL consortia like GTEx, rigorous batch correction using methods like ComBat is essential to prevent technical artifacts from being misinterpreted as genuine genetic regulatory effects.