A key path to drug discovery over the past decade has been the process of determining “shared bioactivity” between molecules, that is, determining the similarity of one molecule to another based on its three-dimensional shape.
There’s always been a catch, though. If you’re a large pharmaceutical company doing research in this area, you probably have a collection of more than a hundred million molecule shapes or conformations. In the rush to find new cures, how do you even begin to get through this much data in a reasonable timeframe, even in a lifetime?
This clever molecule-matcher uses Tesla GPUs to turn this arduous task into a real-time process, matching over two million conformations a second. With this breakthrough, drug research firms like Pfizer and Abbott Labs (which are already using this application today) can zip through their entire portfolios in seconds, rather than the hours or days it previously required.
FastROCS is based on the insight that molecules with similar shapes have volumes that overlay well, and any volume mismatch indicates dissimilarity. This approach is called Ligand 3D shape similarity, and it has proven to be an extremely powerful tool in the drug discovery process.
Brian Cole, the lead software developer behind FastROCS, said, “FastROCS running on a single computer equipped with four of the latest generation NVIDIA Tesla GPUs can achieve performance increases that are multiple orders of magnitude greater than one might see with traditional hardware.”
Want to hear the whole story? Bob Tolbert from OpenEye will be presenting the talk “FastROCS: Revolutionizing Drug Discovery on the GPU” on April 13 at the BioIT World Expo in Boston.