Rendering a VR for Viruses
In what is certainly a novel – and undoubtedly beneficial – use of supercomputer resources, University of Melbourne researchers have been utilising the full computing power of Australia's fastest supercomputer to date, in order to run one very specific virtual reality system. This is not an interactive VR, not meant for real-time human use, but that does not make it in any way less important.
What this is, is a dedicated VR system for the human rhinovirus – the most common cause of the common cold. Every known aspect of the virus is modelled in full 3D, as it folds, unfolds, and bonds with various structures inside a human cell. The goal? To work out the most efficient ways to stop it replicating, and by so doing, cure the common cold.
The rhinovirus is a nasty piece of work. Fully 70% of all severe asthma attacks are linked to an infection by this virus playing havoc with the victim's respiratory tract. Half of these triggered attacks are so severe that the patient has to be hospitalised, putting significant strain on healthcare resources throughout the world, year on year.
That's not even counting the other respiratory tract problems including pneumonia, bronchitis, acute chronic obstructive pulmonary disease, et al. All of these are complicated further by the effects of the rhinovirus, to life-threatening proportions in many cases. An effective, swift way to treat the infection would be a literal lifesaver, and free up a great many resources to battle other conditions.
Drugs to target the rhinovirus are continually being developed, and have been developed for many decades. Yet they take up to ten years to develop, and their use is still a bit hit and miss when it comes to the effect on the virus. That's where this new VR comes in.
By using existing information on how various drugs work, how they target and affect the virus, together with studies on the virus in cell cultures, a team of researchers led by Professor Michael Parker from St Vincent’s Institute of Medical Research are simulating every known aspect of the virus as a 3D, dynamic, responsive model. Throwing different compounds at it, and observing the effects in a way that is frankly not possible outside of VR.
“Our recently published work with Biota [a leading Melbourne drug company for rhinovirus work] shows that the drug binds to the shell that surrounds the virus, called the capsid. But that work doesn’t explain in precise detail how the drug and other similar acting compounds work,” Professor Parker said.
It is no easy task to simulate cellular and virus interactions at this level of detail. The simulation is far from being real-time, lagging considerably behind the actual speed of the retrovirus' progression, turning hours of biological development into days of heavy-duty computation time. However, to do even this much requires the full, considerable capabilities of an IBM Blue Gene/Q, newly installed at the University of Melbourne for this specific purpose.
In production from 1 July 2012, the IBM Blue Gene/Q is the most powerful supercomputer dedicated to life sciences research in the Southern Hemisphere and currently ranked the fastest in Australia.
“The IBM Blue Gene/Q will provide us with extraordinary 3D computer simulations
of the whole virus in a time frame not even dreamt of before,” Professor
“This work offers exciting opportunities for speeding up the discovery and development of new antiviral treatments and hopefully save many lives around the world,” he said.
Professor Parker said that previously we have only been able to run smaller simulations on just parts of the virus.
Professor James McCluskey Deputy Vice-Chancellor (Research) at the University of Melbourne said:
“The work on rhinovirus is an example of how new approaches to treat disease will become possible with the capacity of the IBM Blue Gene Q, exactly how we hoped this extraordinary asset would be utilised by the Victorian research community in collaboration with IBM.”
“This is the way we do biology in the 21st Century,” he said.
The newly operational IBM Blue Gene/Q hosted by the University of Melbourne at the VLSCI is ranked 31st on the prestigious global TOP500 list, which as the name suggests, is a yearly-updated hall of fame of the 500 most powerful computer systems in the world.
The simulation created, is literally the next best thing to blowing the real virus up to macro scale proportions, and observing how it interacts with new compounds, and environments. Precisely because it is so complete, the belief is that any treatment tested in the VR will work exactly the same way on the actual virus. In addition, this is a two-way process, for as we confirm the drugs tested work on the actual virus as they do the simulation, the simulation can be further refined to greater advantage.
Additionally of course, if the concept works for one virus, it will work for all of them. The computing power required is certainly formidable by modern standards, but as the TOP500 list also shows, the rate of improvement in computing power, even amongst supercomputers, has to be seen to be believed. It is well within the realm of possibility that an installation like this will, in less than twenty years, be within the scope of an ordinary biology lab's systems to run such a simulation.
What remains to be done here, is prove that this kind of simulation works.