Fast binding site mapping using GPUs and CUDA

Abstract

Abstract-Binding site mapping refers to the computational prediction of the regions on a protein surface that are likely to bind a small molecule with high affinity. The process involves flexibly docking a variety of small molecule probes and finding a consensus site that binds most of those probes. Due to the computational complexity of flexible docking, the process is often split into two steps: the first performs rigid docking between the protein and the probe; the second models the side chain flexibility by energy-minimizing the (few thousand) top scoring protein-probe complexes generated by the first step. Both these steps are computationally very expensive, requiring many hours of runtime per probe on a serial CPU. In the current article, we accelerate a production mapping software program using NVIDIA GPUs. We accelerate both the rigiddocking and the energy minimization steps of the program. The result is a 30x speedup on rigid docking and 12x on energy minimization, resulting in a 13x overall speedup over the current single core implementation.

FPGA Accelerator for Machine Learning Interatomic Potential-Based Molecular Dynamics of Gold Nanoparticles

“…Even otherwise, HPC developers try to simply run their algorithms on the FPGA framework to parallelize the code using the readily available tools. Although this results in rapid development, it does not provide the complete acceleration one can get from FPGAs [13].…”

GPU optimizations for a production molecular docking code

“…As each evaluation can take many CPU-hours, huge processing capability must be applied. Docking codes have been accelerated with FPGAs [15, 18, 20, 21], Cell [14], and GPUs [13, 17, 19]. …”