Vratin Srivastava just completed a successful summer internship on computational protein design within the Protein Modeling and Simulation team at BIOVIA.
During his internship, Vratin leveraged state-of-the-art generative machine learning models to create an end-to-end Python pipeline for designing novel peptide binders to protein targets. He utilized Bayesian optimization techniques and iteratively refined binder designs. This approach aimed to develop a pipeline that could generate peptide designs with minimal parameter input from users. The project explored whether Bayesian optimization could potentially improve the efficiency of the design process compared to methods that generate and filter large numbers of designs.
Everyone in the organization, especially Rohith Mohan and Reed Harrison who guided me through my internship and other scientists working in the modeling and simulation team, were extremely generous, extremely helpful. It was a very rewarding experience, and I’m really thankful for the opportunity to work on this project.
Curious to learn more about Vratin’s project? Watch the video below where Vratin shares details about his project and his experience working with the BIOVIA team.
Vratin recently graduated with a Master’s degree in Computational Biology from Carnegie Mellon University and holds an undergraduate degree in Chemical Engineering. His research interests include the application of machine learning methods to problems in drug discovery and structural biology, with his Master’s thesis revolving around the application of deep learning to determine protein binding pocket similarity. He will be working as a Research Assistant at the Department of Molecular Physiology and Biophysics at Vanderbilt University after his internship, leveraging computational methods to study protein conformational ensembles. His future goals are to advance the field of computational drug discovery, and design methods that aid in the creation of novel therapeutics.