Protein structure is dynamic, and predicting multiple conformations is key to understanding biological function. Deep learning models like AlphaFold and OpenFold have transformed structural biology, but how can we push their capabilities further?

In this replay, we will discuss a new strategy for subsampling multiple sequence alignments. This strategy leverages direct coupling analysis to form clusters of evolutionarily related sequences that bias models like AlphaFold towards distinct conformational states. We will also show how we implemented and applied this approach to a variety of biological systems in Discovery Studio Simulation, where predicted conformations were compared to known functional conformations. 

Webinar highlights:

• Learn about the new Protein Conformation Search protocol in Discovery Studio Simulation
• Explore how AlphaFold2 can be used to predict multiple conformations of proteins
• Discover synergies between AlphaFold2 and molecular dynamics simulations
• Observe how Protein Conformation Search can be applied to metamorphic proteins

Watch the Replay Now！