Building Better Biomolecular Models with Differentiable Programming

The Irving K. Barber Faculty of Science presents this talk as part of the Department of Biology Seminar Series

About the talk

Developing accurate, physics‑based computational models of nucleic acids, proteins and lipids requires fitting many unknown parameters across diverse experimental datasets. Conventional modelling approaches are often opaque and difficult to reproduce, creating barriers to extending or improving existing models, such as refitting parameters to incorporate new experimental data.

In this seminar, Dr. Engle will highlight how she and her collaborators are leveraging differentiable programming, a tool borrowed from machine learning, to build biomolecular models that are more extensible, reproducible and transparent. She will also demonstrate how these models enable the inverse design of synthetic biological systems, opening new possibilities for biomolecular research and engineering.

Speaker

Dr. Megan Engel
Department of Biological Sciences
University of Calgary

This is a free event. No preregistration is required.