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

Research Title: Using Deep Mutational Scanning and Deep Learning Models to Understand, Predict, and Design Transcriptional Repressors

Research Description: Thousands of human proteins have been identified as transcriptional regulators. However, it remains difficult to predict the ability of any given protein sequence to control gene expression. To decrypt the relationship between sequence and repressor function, Raeline and the team have used high-throughput genetics screens to measure the repressive activity of over 115K protein domains and deep mutational scanning variants at amino-acid resolution. With these data, they are now identifying the key molecular features important for repressor function and are training integrative deep learning models to predict the repressive activity of untested sequences. Successful completion of this work will facilitate the development of new tools for precise gene regulation, mechanistic hypotheses for pathogenic variants, and the inverse design of de novo synthetic effectors.