Protein evolution with RL
A reinforcement learning framework for modeling SARS-CoV-2 protein evolution using large-scale genomic sequence data.
The research question
Can reinforcement learning model plausible mutation trajectories and predict what comes next in protein evolution?
My contribution
I developed the reinforcement learning framework and built preprocessing, training, and evaluation pipelines for mutation-trajectory generation and next-step mutation prediction.
I compared the generated trajectories with observed mutation patterns in the genomic data.
RSI and UROP
I began this work through the Research Science Institute in Bonnie Berger’s lab at MIT CSAIL and was invited to continue through UROP.
Email me about this project ↗