Related Experiment Video
Updated: May 14, 2026
10:52
Multiscale Sampling of a Heterogeneous Water/Metal Catalyst Interface using Density Functional Theory and Force-Field Molecular Dynamics
Published on: April 12, 2019
Message passing interface and multithreading hybrid for parallel molecular docking of large databases on petascale
Xiaohua Zhang1, Sergio E Wong, Felice C Lightstone
1Biosciences & Biotechnology Division, Physical and Life Sciences Directorate, Lawrence Livermore National Lab, Livermore, CA 94550, USA.
Journal of Computational Chemistry
|January 25, 2013
Summary
A new parallel computing approach, VinaLC, significantly accelerates molecular docking simulations using combined Message Passing Interface (MPI) and multithreading. This high-performance computing (HPC) method achieves high accuracy and efficiency for large-scale drug discovery tasks.
Area of Science:
- Computational chemistry and molecular modeling
- High-performance computing (HPC) for bioinformatics
- Drug discovery and development
Background:
- Molecular docking is crucial for identifying potential drug candidates.
- Traditional docking methods face scalability challenges with large compound libraries.
- Efficient utilization of supercomputing resources is needed for large-scale docking.
Purpose of the Study:
- To implement and evaluate a novel parallel scheme for the AutoDock Vina program.
- To enhance the computational efficiency and scalability of molecular docking.
- To validate the accuracy and enrichment performance of the new program, VinaLC.
Main Methods:
- Developed VinaLC, a mixed parallel scheme combining Message Passing Interface (MPI) and multithreading.
- Tested VinaLC on petascale High Performance Computing (HPC) systems.
- Designed input/output and data handling for large databases and massive CPU core utilization.
Main Results:
- VinaLC demonstrates excellent scalability, supporting over 15,000 CPUs with low overhead (3.94%).
- One million flexible compound docking calculations completed in 1.4 hours on ~15K CPUs.
- Docking accuracy validated: 64.4% of top poses have RMSD < 2.0 Å; good enrichment on 70% of DUD targets.
Conclusions:
- VinaLC significantly accelerates molecular docking on HPC architectures.
- The program achieves high docking accuracy and effective enrichment for virtual screening.
- VinaLC shows promise for large-scale drug discovery, enabling rapid identification of potential drug leads.
