Accelerating AutoDock4 with GPUs and Gradient-Based Local Search

Abstract

AUTODOCK4 is a widely used program for docking small molecules to macromolecular targets. It describes ligand−receptor interactions using a physicsinspired scoring function that has been proven useful in a variety of drug discovery projects. However, compared to more modern and recent software, AUTODOCK4 has longer execution times, limiting its applicability to large scale dockings. To address this problem, we describe an OpenCL implementation of AUTODOCK4, called AUTODOCK-GPU, that leverages the highly parallel architecture of GPU hardware to reduce docking runtime by up to 350-fold with respect to a single-threaded process. Moreover, we introduce the gradient-based local search method ADADELTA, as well as an improved version of the Solis-Wets random optimizer from AUTODOCK4. These efficient local search algorithms significantly reduce the number of calls to the scoring function that are needed to produce good results. The improvements reported here, both in terms of docking throughput and search efficiency, facilitate the use of the AUTODOCK4 scoring function in large scale virtual screening.

AMIDE v2: High-Throughput Screening Based on AutoDock-GPU and Improved Workflow Leading to Better Performance and Reliability

“…The local-search method (LSMET) initially embedded in AutoDock 4.2 was Solis-Wets. AMIDE also uses this version even if AutoDock-GPU is now proposed with LSMET ADADELTA [ 15 ]. The Solis-Wets method has been adapted to be run in parallel processing, leading to increased performances (especially by the reduction in calculation time).…”

“…Therefore, some projects exist to alleviate indicated counterparts, such as VirtualFlow orchestrator, which is based on AutoDock 4.2 docking engine [ 14 ]. Nevertheless, this software still does not support the newly released AutoDock-GPU [ 15 ]. Finally, even if the literature attests to the development of many frameworks, most are in-house developed and not available online.…”

“…A new version, called AMIDE v2, was thus developed to speed up the inverse docking process. This framework is now based on AutoDock-GPU [ 15 ] and has benefited from the total revision of the programming workflow.…”

Tetrahydroquinazole-based secondary sulphonamides as carbonic anhydrase inhibitors: synthesis, biological evaluation against isoforms I, II, IV, and IX, and computational studies

“…Docking calculations were executed using the latest AutoDockGPU docking software 29 along with its GUI AutoDockTools (ADT) 30 . The high-resolution X-ray crystal structure of hCA IX having (PDB 5FL4), hCA I (PDB 6EVR), hCA II (PDB 3K34), and hCA IV (PDB 5JN9) 31–33 was downloaded and superimposed on the structure of hCA IX.…”

Behavior of Chemokine Receptor 6 (CXCR6) in Complex with CXCL16 Soluble form Chemokine by Molecular Dynamic Simulations: General Protein‒Ligand Interaction Model and 3D-QSAR Studies of Synthetic Antagonists

“…A first screening to find the binding site of antagonists was performed in F0 of CXCR6 CG-MD, and a grid of 110 × 100 × 255 Å with a spacing of 0.375 Å was calculated using AutoGrid 4.2.6 [69]. Then, molecular docking was performed with AutoDock 4.2.6(cripps Institute, La Jolla, CA, USA) optimized for GPU, using a total of 50 runs and 25,000,000 evaluations with a Lamarckian genetic algorithm and a Solis-Wets local search [70]. Using the binding site of the molecule with the best IC50, a new grid was calculated, of 80 × 80 × 80 Å with a spacing of 0.375 Å, and a second molecular docking was performed using AutoDock 4.2.6 optimized for GPU, using a total of 50 runs and 25,000,000 evaluations with a Lamarckian genetic algorithm and a Solis-Wets local search [20].…”