Latest News

 

Q&A with Apratim Bhattarcharya, TACC User

Published on June 14, 2018 by Faith Singer-Villalobos

Fig. Future systems that we would like to simulate using XSEDE resources. Peptides near physiologically relevant lipid bilayers.

What is the crux of the science you're researching at this time?

Protein folding and protein-protein interactions underlie our conceptual understanding of how living systems work. The problem of protein folding refers to the process in which a chain of amino acids fold into a 3-D structure which is necessary to perform a relevant function either in isolation or through the interaction with other molecules inside the cell. Protein folding and protein-protein interactions have been studied extensively in the past. However, only recently, attempts are being made to understand these processes in physiologically relevant cellular environments. Cellular environments are highly crowded owing to the presence of other cellular machineries such as DNA, RNA, and assemblies of proteins such as the ribosomes, chaperones, etc., and hence a comprehensive understanding posits an extremely challenging problem.

The focus of our research is to develop computational statistical mechanical models using enhanced sampling algorithms to study protein folding and protein-protein interactions in crowded cellular environments. Specifically, our research involves simulating the cellular environment modeling the effects of confinement, crowding and physiologically relevant surfaces which affect protein folding and protein-protein interactions. Using a combination of Monte-Carlo and Molecular Dynamics simulations we have shown that crowding and confinement significantly alter the process of protein folding, highlighting the importance of studying these systems in a context of physiological relevance.

Why are protein folding and protein-protein interactions important to society?

Apratim Bhattacharya, Ph.D.

A comprehensive understanding of protein folding and protein-protein interactions in physiologically relevant contexts allows for the development of novel therapeutics for aggregation induced diseases, most notably the Alzheimer's disease. Alzheimer's disease affects about six million people in the U.S. alone. As an example, protein-based drugs can be discovered based on findings of this research that are expected to have similar functions of confining newly formed proteins inside the cell, preventing them from aggregating.

Why is the computational aspect of doing the science challenging or difficult?

The computational models that we study typically involve more than 25,000 atoms. A typical workflow involves development of an initial molecular configuration, description of the atomic interactions and model physics, performing the simulation using enhanced sampling methodologies achieved through performing multiple simulations in parallel, and analysis of the trajectories. These simulations evaluate millions of interactions of particles for millions of time steps which require extraordinary amounts of computational hardware and time.

How did XSEDE specifically help you overcome the challenges to improve the research?

Most of our research relies on the GROMACS software, which is one of the most widely used open-source and free software code for dynamical simulations of biomolecules and is available pre-compiled in XSEDE resources. GROMACS uses multithreading, heterogeneous CPU-GPU acceleration, state-of-the-art three-dimensional domain decomposition and ensemble-level parallelization through built-in replica-exchange capability. We have been using Stampede for most of our research. Stampede (Stampede2 currently) is a state-of-the-art system with all the necessary compilers. Since some of our research also requires several advanced techniques for accurate free-energy calculations, we often customize the source-code and re-compile in XSEDE systems. To that end, XSEDE has an exceptional resource management team where almost all of our issues are addressed in less than 24 hours.

What is the next step for your research?

We are still a long way from developing comprehensive models of these complex systems. We are taking a gradual approach to understand these systems, beginning with modeling the fundamental interactions. To that end, we have been studying the behavior of peptide backbones near hydrophobic interfaces. The obvious next step is to vary the degrees of hydrophobicity of the surface and understand how it modulates the conformations of the peptide backbone. From simplistic models of molecular interfaces, we intend to use more physiologically relevant surfaces such as different types of lipid bilayers. These systems are expected to be extremely complex and we look forward to studying these systems using state-of-the-art XSEDE resources.

###


Contact

Faith Singer-Villalobos

Communications Manager
faith@tacc.utexas.edu | 512-232-5771

Aaron Dubrow

Science And Technology Writer
aarondubrow@tacc.utexas.edu

Jorge Salazar

Technical Writer/Editor
jorge@tacc.utexas.edu | 512-475-9411