For our linear model, we extracted the posterior effect size of the top genetic variant from the mash model for each feature (gene, transcript, exon and junction). We imputed residualized expression using an individual’s genotype dosage (j) and feature (i) posterior effect size (equation (9)) using PyTorch (v.1.11.0+cu113)98:9\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{Predicted}}\,{\mathrm{expression}}_{i}={{\mathrm{effect}}\,{\mathrm{size}}\,{({\mathrm{eQTL}})}_{j} \times{\mathrm{genotype}}}_{j}$$\end{document}Predictedexpressioni=effectsize(eQTL)j×genotypej
