Home > iSGTW 31 October 2007 > iSGTW Feature - Grids work like a CHARMM for molecular dynamics

Feature - Grids work like a CHARMM for molecular dynamics

X-ray crystallography at room temperature revealed only one water molecule (shown as a red sphere) near the protein residue of interest (shown in the box). The protein is a variant of staphylococcal nuclease.
Images courtesy of Johns Hopkins University

In 1963, Richard Feynman said that “everything that living things do can be understood in terms of the jiggling and wiggling of atoms.” ⁽¹⁾

Molecular dynamics aims to better understand these jiggles and wiggles, using numerical methods to simulate the ways in which atoms and molecules move.

One tool facilitating this work is CHARMM—or Chemistry at HARvard Macromolecular Mechanics—a software application developed at Harvard University for modeling the structure and behavior of molecular systems.

CHARMM has multiple applications, but for Ana Damjanovic of the National Institutes of Health and Johns Hopkins University in Baltimore, Maryland, U.S., molecular dynamics simulations are the name of the game.

CHARMMing simulations

Damjanovic is using CHARMM to help her learn more about the interactions between proteins and water, an understanding of which can ultimately aid the design of medicinal drugs.

“I’m running many different simulations to determine how much water exists inside proteins and whether these water molecules can influence the proteins,” Damjanovic says.

“Water in protein interiors can play important functional roles—such as in enzymatic catalysis or charge transfer reactions—and when a protein moves, it can shift the position of these water molecules.”

“One of the bottlenecks in simulating how proteins function is that many functionally important motions occur on timescales of microseconds and longer, but molecular dynamics simulations can only sample real-time in nanoseconds,” Damjanovic says. “Sometimes even the processes that occur on nanosecond timescales, such as water penetration in some proteins, are not sampled properly with only a few simulations. To get better statistics, one needs to run many, many different simulations.”

This is where it pays to have access to grid computing, which provides the large number of processors necessary to obtain meaningful sampling.

Molecular dynamics simulations reveal that the staphylococcal nuclease protein residue can twist into another conformation, usually forcing water molecules to leave. But which conformation is adopted and for what proportion of the time? And what happens to the water structure around these conformations? Many more simulations are required to find answers.
Images courtesy of Johns Hopkins University

Using your CHARMM on grids

Tim Miller of the National Institutes of Health is working with Damjanovic to optimize the CHARMM application for running on the Open Science Grid.

“We have developed workflow management software for submitting and babysitting the jobs,” Miller explains. “Since molecular dynamics jobs usually take days to complete, and Open Science Grid sites are optimized for shorter jobs, each long job is split into many, many little jobs.”

The workflow uses customized PanDA software developed at Brookhaven National Laboratory as part of the ATLAS high energy physics experiment.

“Now our workflow software can keep track of the rather large number of jobs—what’s submitted and what’s left to be done—and it can then automatically submit the work that’s to be done,” Miller says. “If a job fails, this is detected and the job is then resubmitted. There are multiple threads running as well, and they can branch out and run two different types of analysis on the same structure.”

Damjanovic says her access to Open Science Grid with the newly developed workflow software has worked like a charm.

“All I had to do is initially submit my simulations,” she says. “A month later, I had statistical analysis of lots and lots of simulation. No pain, and no babysitting.”

- Jen Nahn, Open Science Grid

This story also appeared as an OSG Research Highlight.

_{(1) Feynman, Richard (1963). “Lectures on Physics” 1: 3-6.}