Abstract
A central task in expression quantitative trait locus analysis is to identify cis-eGenes, i.e., genes whose expression levels are regulated by at least one local genetic variant. Existing cis-eGene identification methods are either computationally expensive, requiring thousands of permutations per gene (FastQTL), or statistically underpowered (eigenMT and TreeQTL). We propose ClipperQTL, which requires only one permutation for data sets with large sample sizes (>450; ClipperQTL works on smaller data sets too). We show that ClipperQTL performs as well as FastQTL and runs up to 500 times faster. The R package ClipperQTL is available at https://github.com/heatherjzhou/ClipperQTL .