Increasing the acceptance rate of medical implants is just one of the benefits of the discoveries Professor Mark Biggs and his research group have made with the help of eResearch SA’s high performance Corvus cluster computer.
Focusing on the application of molecular simulation methods in studying the phenomenon of protein adsorption at solid surfaces, postdoc Milan Mijajlovic says the potential applications of their work are endless.
“Our improved understanding of protein adsorption opens up the possibility to design new materials that can increase the acceptance rate of medical implants, as well as the development of new techniques for the self-assembly of nanotechnology devices,” Milan explains.
The efforts of Mark’s group in this field are split into two broad objectives: understanding the mechanisms of different stages of protein adsorption at an atomistic level, and the prediction of structures, thermodynamics and transport of adsorbed proteins.
Having discovered the importance of the interaction between protein functional groups and water molecules adsorbed at the solid surface in initiating the adsorption process, their most recent work focuses on the development of a new method for predicting free energy changes during protein adsorption.
“Proteins are large molecules that may consist of hundreds or even thousands of atoms – in order to simulate systems of practical interest, we must first solvate them with thousands of water molecules,” Milan explains.
Adding atoms of a solid surface to this mix, Mark’s group routinely models systems with several thousand atoms, a situation that is further complicated by the need to continuously evaluate interactions between pairs of atoms within these simulated systems.
With such a large number of atoms, an even greater number of mutual interactions must take place – a calculation that is very difficult to deal with on single desktop computers.
This is where eResearch SA’s high-performance computing (HPC) facilities fit in.
By using multiple processors in a single simulation, Mark and his research team are able to achieve simulation times that would otherwise be inaccessible on standard office computers.
“Using eResearch SA’s Corvus cluster computer has assisted greatly in the development of our new method for calculating the free energy of protein adsorption,” says Milan.
“The method is based on calculating an average force between a protein and a solid surface. These averages are calculated for multiple protein-surface distances and for relatively lengthy time scales – simulations like this would take years on single desktops, but with Corvus it only takes a few weeks to months to complete them.”
“Our experiences with eResearch SA have been very good, we have been particularly happy with the IT staff who maintain the Corvus cluster and other HPC facilities,” Milan continues. “They are very pleasant to work with and always happy to help with any problems we might have.”