Awarded in 2022
Home Department: Computer Science
Faculty Advisors: Michael Bassik (Genetics) and Anshul Kundaje (Genetics and Computer Science)

Research Title: Development of Deep-Learning Guided Mutational Scans to Allow Fast Mapping of Sequence to Function

Research Description: Deep mutational scanning (DMS) measures the functional impact of thousands of protein versions in a single experiment but is limited by scale and lacks a framework for predicting how variation affects protein function. Recent work in the Bassik and Kundaje labs has enabled high-throughput DMS assays to systematically measure the effect of variation found in the sequence of the molecular building blocks (i.e., amino acids) of regulatory proteins with gene silencing and activation activities. Here, Akshatkumar proposes augmenting DMS experiments with computational deep learning models to study the function of 50 proteins and 120,000 amino acid variants, enabling more efficient experimental design and an improved understanding of protein function.