3D searching at 2D speed: Ligand-based lead discovery on the GPU presented by Paul Hawkins, Ph.D., Head of Scientific Solutions from OpenEye.
This webinar was originally held on Wednesday, September 21, 2016. Access the on-demand recording below.
Navigating the space of available molecules is daunting - currently there are over 100 million known substances indexed in Chemical Abstracts and around 40 million compounds available for purchase. In virtual compound space the problem is even more challenging as virtual libraries can easily contain billions of molecules. 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 period of time and with reasonable computational resources. More accurate 3D ligand-based or structure-based methods required such a commitment of computational resources that they were infrequently used. However, the advent of inexpensive graphics processing units (GPUs) has transformed the field of ligand-based lead discovery. OpenEye’s development of a GPU version of our class-leading 3D ligand similarity tool ROCS, known as FastROCS, has enabled the routine searching of compound libraries at speeds hitherto only achievable with 2D methods (Greater than 10 million molecules per GPU per minute).
This webinar will introduce the science underlying shape similarity searching at OpenEye, the different methods of deployment of the FastROCS technology and provide case studies of the impact that this qualitative change in the speed of 3D searching has had on real-world problems.