
This post is an attempt to capture some thoughts I have about ML models for predicting protein–ligand binding affinity, sequence- and structure-based approaches to protein modeling, and what the interplay between generative models and simulation may look like in the future.