Speaker Professor Joe Marsh, 

Title Decoding protein variant effects and molecular mechanisms through the lens of protein structure

Host Gyorgy Abrusan

 

Abstract Most known disease-causing genetic variants occur in protein-coding regions of DNA, altering the way proteins are made. While much research has focused on how variants cause loss of function, less attention has been given to non-loss-of-function mechanisms associated with disease, such as gain-of-function and dominant-negative effects. Understanding the diverse molecular mechanisms of pathogenic variants is critical for improving diagnosis and informing potential therapeutic strategies.

First I will discuss our work using large-scale computational analyses to better understand the properties of pathogenic variants associated with different molecular mechanisms, and how we are using this to improve predictions of molecular disease mechanisms. Next, I will present our use of high-throughput deep mutational scanning (DMS) experiments to explore the mechanisms and effects of protein variants across different protein targets. Finally, I will describe our efforts to use DMS data to benchmark computational variant effect predictors, addressing training biases and circularity that currently hinder the field. While evolutionary information has historically been the most important factor in variant effect prediction, emerging evidence suggests that protein structural information is playing an increasingly significant role.