GPU grid helping to find AIDS cure

It used to be that the future of humanity’s health lay in the hands of physicians, biomedical researchers and the center for disease control. Now you can add graphics processing units (GPUs) to the mix.

Using computational techniques, researchers are now able to show how various proteins unfurl to accelerate the spread of viruses, and one university in particular is putting its substantial weight behind using those methods to find a potential cure for AIDS.

Hospital del Mar Medical Research Institute and UPF (Pompeu Fabra University) in Barcelona have been using GPUGRID.net, a voluntary distributed computing platform leveraging GPU accelerators to deliver “virtual supercomputing” performance to examine how a protein responsible for the maturation of the virus releases itself to initiate infection.

The process is done using molecular simulations to explain a specific step in the maturation of the HIV virions, i.e., how newly formed inert virus particles become infectious, which is essential in understanding how the virus replicates. It’s believed the research could be crucial to the design of future antiretrovirals.

AIDS is, of course, a devastating disease that directly attacks and weakens the human immune system, making it vulnerable to a wide range of infections, and is responsible for the death and infection of millions of people around the world. AIDS is also caused by the HIV virus.

The action of a protein called ‘HIV protease’ is responsible for the initial step of the whole HIV virus maturation process, which is what enables the virus to become infectious.

What the scientists in Barcelona discovered Using ACEMD, a GPU-accelerated molecular dynamics software, is that HIV protease acts like a pair of scissors, cutting the long chain of connected proteins that form HIV into individual proteins that will form the infectious structure of new virons. This initiates the whole HIV maturation process.

Using the thousands of Nvidia GPU accelerators in GPUGRID.net, a voluntary distributed computing platform (think, crowd-sourced supercomputing), researchers were able to run thousands of highly complex computer simulations, each for hundreds of nanoseconds (billionths of a second) for a total of almost a millisecond, giving them a very high-probability that their simulation accurately represented real-world behaviors.

Simulations of this length and complexity would have been practically unfeasible to achieve using a computing system based on CPUs alone. Also, by leveraging a distributed network of computers the scientists saved millions of dollars they would have had to spend on a fully fledged supercomputer to achieve the same result.

“GPUs have been crucial,” said Dr. Gianni De Fabritiis, lead researcher, explaining that without them it would have been very difficult to simulate such slow biological processes from an atomistic point of view.

“It allows to get us closer to biology and see how proteins work,” he said.

Dr. De Fabritiis explained that his team had coded the molecular dynamics software ACEMD in 2008 specifically for GPUs, and the payoffs seem clear.

“Running molecular dynamics simulations was prohibitive before in terms of costs, now we aggregate a simulation time of over a millisecond routinely,” he said.

Dr. De Fabritiis added that having a structural picture of this step in HIV maturation was important because it facilitated being able to perform drug design on it. 

From gaming to game changing… GPUs have certainly come a long way.