Ranger's parallel computing power enables researchers to model the largest biomolecular apparatus to date

In the 300 years since Dutch scientist Antonie van Leeuwenhoek first discovered living cells with his homemade lenses in 1674, microscopes have grown hundreds of thousands of times more powerful. Employing new methods and techniques, from electron beams and atomic probes to x-rays, the frontier of magnification has moved from the cell to the molecule, trillions of which work in tandem to create life.

"All life forms are actually a society of molecules, a very hierarchical society," Klaus Schulten, preeminent molecular biologist and professor of physics at The University of Illinois at Urbana-Champaign, explained. "But we're also more than molecules. Water is made of molecules, but it cannot repair or duplicate itself. The point about molecules in living systems is they form teams and work together."

To truly understand the human body, and to design effective medicines and treatments, it is necessary to grasp the operations of cellular molecules from the atomic level up. While the functions of biomolecules, like proteins and DNA, are well known, certain aspects of the proteins' actions elude researchers -- even when using the most powerful microscopes.

Schulten has spent his career extending the limits of microscopy by applying the immense power of supercomputers to molecular imagery. His "computational microscope" takes information from laboratory tests and turns it into dynamic, three-dimensional images with a powerful program Schulten created called NAMD (NAnoscale Molecular Dynamics, pronounced "NAM-dee"). Joining electron microscopy, x-ray crystallography, quantum chemistry and multi-scale molecular dynamics, with the massive parallel processing power of Ranger, the most powerful supercomputer in the world for open science research, Schulten's molecular simulations are opening new realms of research that help us understand fundamental aspects of how life exists on earth.

Schulten uses a football game as an analogy to explain how the computational microscope combines diverse microscopy techniques. "Crystallography," Schulten said, "is like football players listening to the national anthem before the game. They stand there, and if you take a good photograph, you can see them all precisely."

But during the game, the players (molecules) are in motion, interacting, bumping into one another, which is where electron microscopy plays a role. "Here, you can capture the biomolecules in action, but not with the same resolution as in the crystal," Schulten explained. "You don't see every detail, but you see enough that you can match the straight standing players to the actors on the field and learn what are they doing -- where are their legs? where are their heads? do they have the football?"

Combining these two methods tells you what you're looking at from the outside. But to see the molecule from the inside out and to understand how it forms and what it does, you need an all-atom representation of the protein. "Only when you know the chemical detail can you make sense of what is actually happening," Schulten said. "In football language, who has the ball? who kicks the ball? who throws the ball? You can reconstruct this detail through the application of the computational microscope."

Ranger, the newly launched supercomputer at the Texas Advanced Computing Center (TACC), will integrate the data from these varied methods on thousands of parallel processors, and output movies of the molecular machinery in motion. These information-rich visualizations, in turn, will help fuel the next round of molecular dynamics breakthroughs.

Coming from Schulten, it sounds simple, but in reality, this process is the product of more than two decades of coding and refinement, and $20 million in funding from the National Institute of Health (NIH). Today, Schulten's parallel molecular dynamics program, NAMD, is the leader for large-scale simulations of biomolecular systems (more than 100,000 atoms) and one of the most capable parallel scientific codes ever run on a supercomputer.

Schulten has used NAMD to do some of the most intensive molecular dynamics simulations ever attempted. His 2006 simulation of the satellite tobacco mosaic virus was the first to capture a whole biological organism in intricate detail. It showed the million-atom virus pulsing in a solution of water as if it were breathing. In 2007, he simulated the actions of the ribosome, modeling three million atoms and discovering new facets of this essential protein factory.

In the coming months, working with experimentalists, computational scientists, and other theorists, Schulten will use Ranger to model the largest and most complex biomolecular machinery to date: the 100 million atom chromatophores of purple bacteria.

"Organelles like the chromatophore are like the organs in the body -- small but important parts of the cell that have distinct functions," Schulten said. "In the case of the chromatophore, the function is absorbing sunlight and turning it into chemical fuel that the cell needs for many of its processes."

For decades, Schulten had studied the parts of the chromatophore one-by-one. "What Ranger will permit us to do, on a much larger scale than ever before, is look at the concerted activity of the molecules in the chromatophore, not just one at a time, but at many of them."

One of the first questions that Schulten's research explores is what gives the chromatophore its spherical shape? Or, as Schulten puts it: "How do cells build their own houses?" Using Ranger, Schulten simulated the most common chromatophore protein interacting with the cell membrane. He found that if the proteins are put on a flat membrane, they dome the membrane, form a spherical bubble, and then cut themselves off.

"We did test calculations to get a sample. Then, to be sure of how these proteins are arranged in the membrane -- how tightly packed they are -- we did all kinds of varieties of simulations, just as in the lab you do all kinds of experiments," Schulten said. "In the old days, you were happy if you could do one sample calculation. But with Ranger, we can do several of them to be sure that we're not being led astray."

The day Schulten and his colleagues submitted their computational study[1] of domed chromatophore proteins, researchers from Harvard University announced that they had shown experimentally that other proteins were curving a different cellular membrane, the endoplasmic reticulum, in a similar manner. "They are showing experimental views of this phenomenon at a much lower resolution than what the computer can do," Schulten said. "With their research, you recognize only the rough shape of the membrane, whereas we see every one of its atoms."

The formation of protein domes is just one process in living cells that will be explored by Schulten and his team on Ranger in the next few years. Additional simulations will show the functioning of all the proteins active in the chromatophore. And Schulten is only one of thousands of researchers who are using NAMD to uncover insights about the molecular machinery, providing an incredible multiplier for the field.

Perhaps as importantly, Schulten's experience using his highly-scalable NAMD code on Ranger points the way to future algorithms and scalable codes capable of modeling larger and larger molecules on next-generation high-performance computing systems.

"We can already see on the horizon that what we are doing on Ranger today will be done in the office tomorrow," Schulten said. "Someone has to begin to use these kinds of machines with many processors to teach the world how to use them effectively, to develop new algorithms that work with parallel processors, so they could be used effectively tomorrow by everybody."

Unraveling the mysteries of life with the help of a computational microscope, Schulten's research is one more reason to stay tuned to TACC and Ranger for world-changing scientific discoveries.

For more information about Klaus Schulten's research, visit http://www.ks.uiuc.edu.

Footnote:

[1] "Chandler, Hsin,Harrison, Gumbart, and Schulten; Intrinsic Curvature Properties of Photosynthetic Proteins in Chromatophores," submitted to Biophysical Journal.

-----

Source: Texas Advanced Computing Center. For original article with complete graphics, visit http://www.tacc.utexas.edu/research/users/features/sult.php.