How Deep Learning is Changing Protein Structure Prediction and Design

Pre-user, User, Pre-developer, Developer, Design, Modeling, RosettaC++, Machine Learning

This talk is aimed at graduate students, postdocs and faculty in biophysics. With the backdrop of the recent breakthrough in protein folding using deep learning, demonstrated by AlphaFold2 by DeepMind in CASP 14, I present some of the core building blocks and philosophies for the new, deep-learning approaches to protein structure prediction and design. I share some results in the context of my lab's research on using deep learning to predict antibody loop structures. https://youtu.be/6lDbMpkkGYo (opens in a new tab)