Sign In

Communications of the ACM

ACM News

Computers Could Make Better Drugs


View as: Print Mobile App Share: Send by email Share on reddit Share on StumbleUpon Share on Hacker News Share on Tweeter Share on Facebook
computer prediction of drug compound

Advanced computing can simulate the binding of virtual proteins and ligands which allow chemists to screen vast pools of possible compounds faster than would be possible in the laboratory. Shown is a computer prediction of a novel inhibitor binding to the JNK substrate docking site on a protein.

Credit: Pengyu Ren / Unveristy of Texas at Austin

Most drugs enter the body as small molecules, ligands that bind to the surface of target proteins, inhibiting their function and protecting a person's health. For a drug to tame a headache or reduce a swollen knee, the drug needs to be effective at small doses, and selective enough to limit side effects.

The many medicines on the shelves of the local pharmacy suggest that finding a new drug is a simple, straightforward process. But in reality, discovering a new medicine can be a Herculean effort. On average, it takes 15 years and more than $800 million in research and development for a drug to come to market, according to experts. This drives up the price of blockbuster drugs, while limiting research into less profitable medications.

Due to the time and costs involved, advanced computing is crucial to drug discovery efforts. By simulating the binding of virtual proteins and ligands, chemists can screen vast pools of possible compounds faster than would ever be possible in the laboratory.

This process trims the number of possible cures from millions to hundreds, at which point the drug candidates can be studied in the lab, thereby making drug discovery cheaper and faster. Several important HIV protease inhibitors were discovered using this method.

However, virtual "enrichment" is only helpful if the most effective molecules end up in the top 10 percent of the prediction. More often than not, they don't, leading to frustration and skepticism in the field.

Pengyu Ren, assistant professor of biomedical engineering at the University of Texas at Austin, is trying to solve this problem. Using the NSF-funded Ranger supercomputer and a large pool of known protein-ligand matches, he is fashioning a robust way to search for new drugs.

"We're testing and developing computational approaches that can best reproduce the experimental data of protein-ligand binding that has been reported in the [academic] literature," Ren explains.

The effort is one of the most comprehensive studies ever undertaken of protein-ligand interactions using all-atom simulations. By representing physical reality with far greater fidelity, Ren expects the new method to significantly improve the effectiveness of drug discovery.

"In the old days, shortcuts were necessary to achieve speed. Researchers made approximations of physical models because the computations were too expensive," Ren says. "We're adding that further layer of physics to get more accurate predictions."

The physical reality of these simulations is enabled by supercomputers at the Texas Advanced Computing Center (TACC). "Virtual drug simulations require massive computing power," Ren says, "and by having access to TACC, we're in a position to actually try out these methods."

Ren is evaluating the accuracy and efficiency of different methods by applying them to more than 200 complexes across 10 different protein families. The method with the closest correlation to the real results is deemed the most effective. If the most successful method is proven to work consistently, Ren believes chemists will adopt it.

"The promise of rapid, inexpensive computational drug discovery has thus far eluded scientists," says Michael Gonzales, life sciences program director at TACC. "Pengyu's work is an excellent example of how current advances in computing power are enabling scientists to take a fundamentally different approach to virtual drug discovery."

It's not just about methods and protocols for Ren. He is also involved in a number of collaborations that put his algorithms to the test, exploring the relationship between rigidity and protein-ligand binding, and searching for inhibitors to proteins that are involved in cancer and other diseases.

"If this works, it will improve our ability to design drug candidates that are more potent with fewer side-effects," Ren says. "But to make robust, accurate predictions, it's time to invest in the next generation of computational technologies for drug discovery."
 


Comments


Xiyang Liu

A very promising and interesting work. The expensive supercomputer at TACC is not available for most researchers in most developing countries. So it is my best wish to develop less expensive algorithms running on cloud computing platforms or GPU clusters.


Displaying 1 comment