Recently, advances in virtual chemical library (<i>de-novo</i>) generation, efficient sampling of chemical space, virtual assessments of synthetic feasibility, and advances in GPU hardware, combined with machine learning approaches, have enabled easier access to extensive compound libraries than ever before. The developments described above, therefore, justify new approaches in virtual screening and provide the opportunity to screen compound sets several log units larger in a biological context. Therefore, we have started the development of the open-source software package CmDock. This is molecular docking software with accompanying tools for analyzing docking results, based on rDock, with the goal of optimizing docking protocols and introducing support for modern GPU hardware. To be GPU agnostic, we chose the OpenCL platform and implemented the HTVS highly efficient (Chem)PLP scoring function. Moreover, we have implemented a GPU-side genetic algorithm and can successfully perform large-scale global search experiments GPU-side in a timely manner. We have successfully validated the approach with small molecule binding site identification experiments and binding pose studies. We anticipate that these developments will be useful for future GPU-docking developments and provide a much-needed boost to <i>in silico</i> novel drug design and research.

Strides towards efficient GPU-supported molecular docking using CmDock software

Fall 2022 abstracts, posters, and presentations will launch in mid-September.

ACS Fall 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.