Human complex phenotypes are shaped by multiple genetic and environmental factors and their interactions. Understanding how genetic and environmental factors interact can provide new mechanistic insights into the genetic architecture of human complex traits and facilitate the development of personalized precision medicine. However, detecting and interpreting genotype-environment (GxE) interactions is challenging due to their small effects and limited functional interpretation. We propose a statistical framework, called transcriptome-wide interaction studies (TWIS), to interpret GxE by integrating transcriptomic data. It tests whether genetic variants interact with the environment to influence complex traits and diseases through gene expression regulation. It also unifies existing methods that use transcriptomic data to interpret GxE into two main categories and shows their connection and distinction. We equip TWIS with an estimation method that requires only precomputed gene expression weights and GxE summary statistics as inputs. We applied TWIS to UK Biobank data (N = 375,791) and identify genes that have sex-differential effects on complex traits and diseases through gene expression regulation. We found that SLC2A2 (P = 5e-80) and PKD2 (P = 7e-98) have sex-differential effects on gout. Overall, TWIS enables us to uncover the molecular mechanisms underlying GxE interactions and advance personalized precision medicine.
Jiacheng Miao, University of Wisconsin-Madison
Yixuan Wu, University of Wisconsin-Madison
Lauren L. Schmitz, University of Wisconsin-Madison
Qiongshi Lu, University of Wisconsin-Madison