High Performance Computing Cluster for Public Health Driven Molecular Science - Core Facility

University Of California Berkeley, Berkeley CA

Investigators

Linked publications & trials

Abstract

The College of Chemistry at the University of California, Berkeley seeks funds to purchase a high performance computing cluster with integrated data storage dedicated to the current and future needs of a broad community of NIH-supported research groups who are engaged in public health driven molecular science. This cluster will be integrated into a College of Chemistry core computing facility which itself is part of a larger suite of College of Chemistry and Interdepartmental core analytical facilities supporting NIH investigators. The goal of this proposal is to enable efficient research innovation by putting advanced computational resources into the hands of all NIH investigators, not just computational specialists. Computer modeling of chemical structure and reactivity is now a key experiment design and analysis tool. Many synthetic chemists collaborate with computational chemists to obtain insights made possible by modern software coupled with computing power. This is an important partnership for specialized cases but experimentalists can and should be enabled to do much of this every day computational work themselves, shortening the turnaround time to obtain potentially transformative insights that impact human health. Experimental scientists can explore critical aspects of structure and electronics, protein-ligand interactions, viral capsid formation and disruption, bioinformatics and many other properties of molecular science. Desktop based software is an entry point to computational approaches but true insight into complex molecular science still requires supercomputing level power. Unfortunately, most supercomputing comes with a bare command line and a very high barrier to entry due to the learning curve of arcane queuing and operating systems. The College of Chemistry at UCB has a proven track record of lowering the barrier to entry by training experimental scientists in computational techniques in our core facility and empowering them with software, hardware and consulting infrastructure tailored to their research needs. Right now our NIH funded researchers are limited to use of spare cpu cycles on an older computing cluster which itself is primarily dedicated to other projects. This proposal will fund the computing power required to serve a large group of NIH supported investigators and will leverage an existing core facility staffed by professional scientists with more than 30 years of experience in computation. Expanding the access to computational tools for experimental scientists is potentially transformative in terms of public health outcomes driven by molecular science.