flâneur

Nourmohammad - statistical physics of evolving systems lab

sites.google.com · 36 words · saved by 1 readers

Proteins play a central role in all parts of biology from immune recognition to brain activity. A key challenge is to predict how protein sequence and structure determines function, such as the protein’s binding affinity to ligands or its enzymatic activity. With the growth of molecular data, machine learning has become a powerful tool in protein science. However, these techniques often generate black-box models, which are powerful but hard to interpret. In this project, we aim to characterize a structure-function map for proteins by developing equivariant neural networks that take protein structures as input and through transformations that respect the physical symmetries in the data, learn interpretable models of protein structures that could reflect the underlying biophysical function. The resulted biophysically grounded models will be used to develop generative models to design proteins with desired function. This project will involve development of novel AI techniques suitable fo

Sign in - Google Accounts Sign in Use your Google Account Email or phone Forgot email? Not your computer? Use a private browsing window to sign in. Learn more about using Guest mode Next Create account

related reading