Cell Type-Specific Expression Quantitative Trait Loci

• Activity Number: 127 - Statistical Methods for Multi-Omic Data Analysis
• Type: Topic-Contributed
• Date/Time: Monday, August 9, 2021 : 1:30 PM to 3:20 PM
• Sponsor: Section on Statistics in Genomics and Genetics
• Author(s): Paul Little* and Yun Li and Danyu Lin and Wei Sun
• Companies: Fred Hutchinson Cancer Research Center and UNC-Chapel Hill and University of North Carolina at Chapel Hill and Fred Hutchinson Cancer Research Center
• Keywords: RNAseq; cell types; eQTL; gene expression; composition

Abstract

Gene expression varies across cell types, which partially explains the diverse morphologies and phenotypes across cell types despite almost identical DNA sequences. DNA regulation of gene expression also vary across cell types and traditional gene expression quantitative trait locus (eQTL) analyses often ignore such heterogeneity and map eQTLs of bulk samples as if they are composed of homogeneous cells. We have developed a method to perform cell type-specific (CTS) eQTL mapping for bulk RNA-seq data. Our method utilizes total and allele-specific expression, which further improves the power of eQTL mapping. We evaluated it using various simulations. Then we applied it to study GTEx whole blood and evaluated our results using gene expression from purified samples of a few cell types. We also applied it to study human brain from GTEx frontal cortex samples and CommonMind Consortium (CMC). With CMC data, we found that CTS eQTLs can differ between schizophrenia patients and controls. The least squares (LS) model served as the benchmark for our method. The benefits of our model over LS included higher power and appropriate control of Type I error/ false discovery rate.