AIDRC Seminar Series: Dr. Nazim Bouatta

Jun 12, 2023
AIDRC Seminar Series banner
Nazim Bouatta

Dr. Nazim Bouatta

Senior Research Fellow, Harvard Medical School

12th June 2023, 04:00pm - 05:00pm (GST)

Title:OpenFold: Insights gained from rebuilding and retraining AlphaFold2
Abstract:AlphaFold2 represents a significant milestone in protein structure prediction. However, its implementation lacks the code and data necessary to train new models for novel tasks, such as predicting protein-ligand complex structures, a core problem in drug discovery. In this presentation, I will discuss OpenFold, a fast, memory-efficient, and trainable implementation of AlphaFold2. We trained OpenFold from scratch and discovered that it can generalize remarkably well to previously unseen regions of protein structure space, even when trained on highly reduced datasets representing only ~5% of the original AlphaFold2 training data. This has profound implications for training AlphaFold2-style models in data-sparse regimes, such as nucleic acids or small molecules. OpenFold has led to exciting applications, including the release of the ESM Metagenomic Atlas by Meta AI, which contains over 600 million predicted protein structures.
Bio:Dr. Nazim Bouatta is a Senior Research Fellow in Systems Biology and Systems Pharmacology at Harvard Medical School, where he applies his expertise in machine learning, physics, and mathematics to solve complex biological problems. He co-supervised the development of OpenFold, an open-source, trainable, and optimized version of AlphaFold2. He has also led the development of a novel approach that uses machine learning to predict protein structure through a protein language model, similar to GPT and BERT. He is currently leading multiple research efforts, including co-folding proteins and ligands, using a single sequence model for protein structure prediction, and predicting RNA 3D structure. These projects, taken collectively, aim at building a multimodal foundation model for biology.  

Dr. Bouatta earned his doctoral degree in high-energy theoretical physics in the group of Prof. M. Henneaux and completed his postdoctoral training in the Department of Applied Mathematics and Theoretical Physics (DAMTP) at Cambridge University.  

Twitter: @NazimBouatta