Great Potential Tapped. Traversing Unexplored Regions of Chemical Space with the GPU" was presented by Paul Hawkins, PhD., Head of Scientific Solutions, on Wednesday, September 26, 2018 at 12pm EDT / 9am PDT (US).
The advent of the GPU (Graphics Processing Unit) offered the potential to revolutionise a wide variety of computational methods, as it provided compute power that, until recently, was only accessible through expensive, specialized high performance computing (HPC) facilities. The wide availability of inexpensive GPUs has transformed a number of areas of computational molecular design; in this webinar we will look at the effect of porting some widely used lead discovery tools to the GPU.
Until recently the sheer size of the universe of molecules has dictated the methods that can be used to search it; only very rapid 2D methods were fast enough to complete a search in a reasonable time and with reasonable computational resources. Porting of already fast 2D fingerprint search methods from the CPU to the GPU has enabled rapid searching of unprecedentedly large databases of molecules.
More accurate 3D ligand-based have required such a commitment of computational resources that they were infrequently used on large databases. Porting OpenEye’s class-leading 3D ligand similarity tool ROCS to the GPU (FastROCS) has enabled the routine searching of compound libraries at speeds hitherto only achievable with 2D methods (> 20 million molecules/GPU/minute).
This webinar will highlight the impact of the GPU on ligand-based lead discovery tools, with particular emphasis on how the use of FastROCS has enabled the searching of enormous virtual libraries (> 10^9 molecules) that were previously out of reach for any 3D method.