Back to Winter 2011 index

FEATURED WINTER 2011 ARTICLE

Remote Control The Institute for Computational Molecular Science makes formerly unfathomable experimentation possible. By Brian M. Schleter

Throughout history, the greatest scientific and engineering advances have resulted from theories proven through costly, sometimes dangerous and often repetitive forms of experimentation.

Advances in computer technology are changing that dynamic—rapidly.
Every day, researchers use high-performance computers, capable of
performing quadrillions of calculations per second, to design safer, more
effective medicines; predict the effects of climate change; and search for
new treatments for HIV, influenza and other diseases.

Increasingly, researchers in the natural and life sciences are turning to
computational science to observe video simulations of molecular activity.
Now, researchers at Temple can utilize the Institute for Computational
Molecular Science (ICMS) in the College of Science and Technology.

ICMS was founded in 2009 by Laura H. Carnell Professor of Science Michael
Klein, member of the esteemed National Academy of Sciences—one of the
highest honors bestowed on a scientist—author of more than 600
publications, editor of four books and winner of the American Physical
Society’s Aneesur Rahman prize for outstanding computational research.

A leader in his field, Klein aims to bring Temple’s research to the forefront of
science and to foster cross-disciplinary team-building through ICMS, a new
valuable tool for generating new approaches to discovery of all kinds.

From increasing research capabilities for students to allowing scientists to
witness molecular interaction visually, ICMS’s computing will push the
boundaries of what can be accomplished —and what can be understood—
at Temple.

“Computational science is commonly described as the third pillar of scientific
discovery, beside theory and experimentation,” says Axel Kohlmeyer,
associate director of ICMS and associate research professor at Temple.
“We use computers to do experiments that cannot be done in reality.”

A C_60 molecule (or “buckyball”) is swallowed by a bilayer of lipid molecules, a model for a cell
membrane. Researchers study if the otherwise inert buckyballs can have negative effects on cell
membranes. Image courtesy of ICMS.

Building Bridges
In today’s research landscape, anyone who needs a quick answer or has a
large block of data to process needs high-performance computing. Within ICMS,
a dozen Temple chemists, biochemists, physicists and computer scientists use state-of-the-art computer simulations to model molecular behaviors. Similar to
the way video games simulate experiences based on data —but on a much
more expansive and complex level—computer models designed by the ICMS
group utilize scientific data to verify the results of experiments put forth by collaborating researchers.

“Providing high-performance computing tools is a bit like being a pipefitter,” Kohlmeyer says. “There is a certain infrastructure that has to be available and maintained. If you pull a certain handle, you’ll get hot water. People without
that knowledge, working by themselves, wouldn’t be able to get hot water.”

Klein, a 16-year member of the chemistry faculty and director of the Center for Molecular Modeling at the University of Pennsylvania, came to Temple to found ICMS. That decision played a role in Robert Kulathinal’s decision to join the
faculty last year.

“Temple’s successful recruitment of the ICMS group was exemplary of the university’s commitment to both innovation and excellence,” says Kulathinal,
an evolutionary geneticist and assistant professor of biology. “It confirmed that this is a place I wanted to be.”

His lab plans to use ICMS’s extraordinary computing resources to construct
three-dimensional protein models from genome sequence data, to understand how proteins have evolved over millions of years. Kulathinal says the modeling might lead to a new understanding of how proteins interact, resulting in more effective drug treatments.

Part of Klein’s strategy is to seek seed funding for researchers to collaborate across disciplines. An example, he says, would be a pharmacologist working
to develop a new compound who needs a chemist to synthesize it. The role
of the ICMS would be to use computers to help in the screening of millions of possible molecules to find potential targets and thus reduce the workload of
the synthetic chemist. By facilitating these conversations, the ICMS is creating
an environment where “random, unexpected collisions” of this sort can occur.“

We’re trying to build bridges,” Klein says.

Providing the Tools
Though the work of ICMS encompasses an array of different sciences and researchers, the common theme in all of the institute’s projects is the smart
and effective use of high-performance computing clusters, or HPC clusters.

Just as bridges need sturdy foundations, the first order of business for ICMS
is to build a state-of-the-art cyberinfrastructure at Temple. The university
now has a hybrid GPU (graphics processing unit)/CPU (central processing unit) cluster computer.

In Temple’s case, the HPC cluster features more than 100 nodes containing
more than 1,200 processor cores, 48 Nvidia Tesla GPUs (hardware made specifically for high-perfomance computing) and 120 terabytes of storage.

In other words, the HPC cluster is the equivalent of hooking up approximately 2,000 laptop computers set to process the same information. It provides more hard-drive space than 240 computers that each contain a 500-gigabyte hard
drive.

Beyond the astounding capacity of the HPC cluster at Temple, ICMS also is part
of TeraGrid, a cyberinfrastructure of even more powerful supercomputers used
for scientific research, information and data sharing, which not only encourages cross-disciplinary research, but also fosters collaboration between Temple and other TeraGrid researchers across the country.

Targeting the Impossible
Most of the ICMS group’s research is literally “trying to do the impossible,” Kohlmeyer notes, by putting theoretical models to the test. Because
simulations are numbers-based, they know the exact state of the whole
system at any point, allowing for perfectly controlled experiments.

The institute’s ultimate goal—Kohlmeyer calls it the “holy grail” of
computational science—is to arrive at a juncture where the molecular
computer simulations built are so accurate and detailed, they render many chemical and biological experiments obsolete. Imagine a scenario in which
a new drug is proven safe and effective outside the laboratory—where it
currently takes years of high-cost testing—and on a computer instead,
with every possible molecular interaction taken into account.

Though the technology is not there yet, much of the current research at
ICMS is helping researchers improve drug treatments.

For example, one team is looking at the cellular binding abilities of certain anesthesia agents.“We know that anesthetics work, but we don’t know
how they work on the molecular level,” Kohlmeyer explains. Understanding
the molecular activity of anesthetic agents, he says, will result in safer
anesthetics that can better target the area of the body they are meant to
act upon.

Polymer chemists in the group are using computer simulations to unlock
the secrets of nano-scale materials. Mixing multiple polymers (chemical
compounds) and studying the new mixture that arises on the nano level
can improve drug delivery. “The purpose is to ‘package’ a drug so it can
get where it needs to be, instead of ‘flooding’ the body and thus risking
side effects,” Kohlmeyer explains.

Another area of focus at ICMS is software development. Last summer,
Kohlmeyer supervised two undergraduate students who helped improve a simulation software package. The effect enabled a postdoctoral researcher
to run his calculations up to 15 times faster than before, greatly improving
the researcher’s efficiency.

“As I see it, the mission of ICMS, in this context, is to share our experiences
with other Temple researchers and, in return, learn from their experiences,” Kohlmeyer says. “Ultimately, we want to establish collaborative efforts that
can solve even more complex problems than the ones being worked on now.”

As ICMS grows, so too will opportunities for groundbreaking research at
Temple.

“The vision for this institute is to foster collaborative, interdisciplinary,
frontier research,” Klein says. “It will take time, but it will be driven largely
by recruitment. We need to recruit people aggressively at all levels to staff
centers and institutes, especially young people. This new generation of researchers is more adept at using computers to solve complex problems.”

Brian M. Schleter is a research freelance writer based in Baltimore City, Md.