Abstract
Characterizing individual transcriptome variation is fundamental for deciphering human biology and disease. Demographic traits such as ancestry, sex, age and BMI, simultaneously affect gene expression and alternative splicing variation. However, how these variables mechanistically interplay to ultimately define an individual’s phenotype is not well understood. Here, we implement a statistical framework to quantify the joint contribution of these four demographic and 17 clinical traits as drivers of gene expression and alternative splicing variation across 46 human tissues. We demonstrate that demographic traits have different contributions to expression variability that strongly depend on the tissue. Whereas multiple traits can influence a gene additively in specific tissues, we find that interactions are rare. Contrary to expression, variation in tissue splicing is dominated by ancestry and a large fraction of splicing differences between populations are under genetic control. Among those, we find that alternative splicing in ribosomal proteins differs between human populations across most tissues. Furthermore, we observe that clinical traits can have important contributions to tissue transcriptome variation. Type 1 and 2 diabetes affect multiple tissues, particularly the tibial nerve, where their impact resembles that of biological aging. Overall, our study illustrates the power of multi-tissue and multi-trait transcriptome analysis and provides an extensive characterization of the main drivers of human transcriptome variation.
Zoom info
Join Zoom Meeting at https://uio.zoom.us/j/63184835161?pwd=ZWhsaHRNVTZkdDJzakFBZ1EwTDJRUT09
Meeting ID: 631 8483 5161; Passcode: 310067
Junior talk
Xiaoran (Richard) Lai, postdoctoral fellow at the department of Biostatistics, University of Oslo, will present his work on "Explaining non-linear effects and gene-gene interactions modulating the expression of a gene in single cells."