Machine Learning in Structural Biology

Bridget Carragher is the founding technical director of the Chan-Zuckerberg Imaging Institute. From 2002–2022, she served as the Director of the National Resource for Automated Molecular Microscopy (NRAMM), an NIH-funded national biotechnology research resource. Throughout her career, she has been involved in developing automated methods for EM aimed at improving both the quality of EM data as well as the accessibility of these techniques to the wider biological community.

Kyunghyun Cho is an Associate Professor of Computer Science and Data Science at New York University and a senior director of frontier research at the Prescient Design team within Genentech Research & Early Development (gRED). He is a leading researcher in sequence modeling and NLP and, more recently, in his role at Prescient Design he has been at the forefront of the field of antibody sequence and structure-based design.

Rhiju Das is a Howard Hughes Medical Institute investigator and Associate Professor of Biochemistry at Stanford University. His research focuses on using cryo-electron microscopy, high-throughput structure mapping experiments, and deep learning to gain a predictive understanding of RNA structure. He led the first RNA structure prediction challenge in the 2022 CASP competition, and is interested in the design of new metrics for evaluation of conformational ensembles.

Polly Fordyce is an Associate Professor of Genetics and Bioengineering and fellow of the ChEM-H Institute at Stanford University, where her laboratory focuses on developing and applying new microfluidic platforms for quantitative, high-throughput biophysics and biochemistry and single-cell genomics. She is the recipient of numerous awards including an NIH New Innovator Award, an NSF CAREER Award, the 2023 Eli Lilly Award in Biological Chemistry, and is a Chan Zuckerberg Biohub Investigator.

Tanja Kortemme is a Professor at University of California, San Francisco and member of the RosettaCommons. Tanja is interested in understanding how biological molecules communicate with each other, and how this communication encodes the processing of information. Computationally, her lab has been developing both physical energy function and machine learning models for the prediction and design of protein-protein interactions and conformational changes at the atomic level. Experimentally, her group has applied these models for the redesign of a protein interface and to create artificial DNA-binding proteins with new specificity.

Gevorg Grigoryan, PhD, is a co-founder and chief technical officer at Generate Biomedicines. He is also a research associate professor of computer science, biological sciences, and chemistry at Dartmouth College. He received two bachelor’s degrees, in computer science and biochemistry, and went on to obtain a PhD from the Massachusetts Institute of Technology, where he studied computational protein design and modeling of protein interactions. After completing his postdoctoral work at the University of Pennsylvania in 2011, Gevorg was appointed to the faculty at Dartmouth, where he obtained tenure in 2017.

Prior to co-founding Generate Biomedicines, Gevorg’s research at Dartmouth focused on understanding the relationship between amino-acid sequence and structure, seeking rigorous ways of inferring principles governing protein folding, interaction, and function from data. The powerful ideas that were discovered suggested a fundamentally new and promising paradigm for the design of macro-molecules that was based on general principles inferred from data. Armed with this new and powerful understanding, Gevorg envisioned the possibility of a company that uses the breakthrough of data-driven protein design, powered by Machine Learning, to transform therapeutic discovery towards a faster, more precise, and less expensive endeavor. Collaborating with colleagues at Flagship Pioneering brought this vision into reality, leading to the founding of Generate Biomedicines. The most enjoyable part of Gevorg’s role as the founding Chief Technology Officer at Generate has been to work with incredibly talented scientist on some of the world’s most challenging problems in molecular and therapeutic science.

Gevorg has authored over 50 peer-reviewed papers, including in Nature, Science, PNAS, and other top journals and conferences, and his work has received awards and recognition from the Alfred P. Sloan Foundation, the National Institutes of Health, the National Science Foundation, and the American Cancer Society.

The RFdiffusion team, comprised of computational biologists from UW Medicine, Columbia University, and MIT, have developed a diffusion model for protein design. Their tool, RFdiffusion, is able to generate novel proteins, addresses challenges related to de novo binder design, symmetric architecture, enzyme scaffolding, and more. Now open source, RFdiffusion is poised to accelerate scientific discovery and potentially transform medicine, vaccine development, and advanced materials development.