Arvind Ramanathan: Accelerating Simulations for COVID-19 Therapies

Using ML/AI to Help Slow the Spread

Are you sick of COVID-19? So was Arvind Ramanathan, Ph.D. He was certain that AI/ML could interface with rigorous physics-based methods to address drug discovery challenges. Ramanathan presented to the Society for HPC Professionals on January 27. His thought-provoking, virtual presentation was entitled “AI-Driven Adaptive Multiresolution Molecular Simulations on Heterogeneous Computing Platforms.

Ramanathan is a computational biologist in the Data Science and Learning Division at Argonne National Laboratory and a senior scientist at the University of Chicago Consortium for Advanced Science and Engineering (CASE). His research interests are at the intersection of data science, high performance computing and biological/biomedical sciences.

The problem is that COVID-19 needs a comprehensive strategy to identify the small molecules (or other therapeutic strategies) in order to treat the virus. However, this also requires an understanding of the molecular mechanisms that the virus uses to evade immune response within its host cells. Given that these processes occur at larger length- and timescales, the question was: how do we accelerate the simulations of complex biological phenomena?

Ramanathan proposes Stream-AI-MD, a novel instance of applying deep learning methods in a streaming manner to drive adaptive molecular dynamics (MD) simulations. His explanation of how AI/ML works on existing supercomputing platforms and emerging heterogenous platforms led to a brief introduction to Argonne. The laboratory houses some of the biggest and fastest supercomputers on the planet.

DeepDiveMD was also employed to help with the ML during the course of his research. “AI-driven adaptive MD simulations are at least two orders of magnitude better than traditional sampling,” said Ramanathan.

Ramanathan’s broader collaborations have resulted in obtaining novel insights into the molecular mechanisms of how the virus replicates. Research has just now started providing ideas on how to design small molecules that can inhibit the virus. These molecules, as well as other structural biology techniques, have been validated in the lab.

 

Author

Sarah F. Hill

Source Name

HPE Data Science

News Category
Events
Institute Happenings
UH Data Science News
Research Topics