A beginner’s approach to deep learning applied to VS and MD techniques

D’Hondt, Stijn; Oramas, José; De Winter, Hans

doi:10.1186/s13321-025-00985-7

Journal of Cheminformatics

Table 2 Summary of the tools mentioned throughout this review used to setup VS steps, MD simulations, or DL models

From: A beginner’s approach to deep learning applied to VS and MD techniques

Name DL/VS/MD-tool	Description	Application example(s)
ML/DL software libraries
PyTorch [6]	Open-source ML/DL framework for NN development and training	[37, 141]
TensorFlow [5]	Open-source ML/DL framework for NN development and training	[137, 140, 174, 179]
VS tools
AutoDock Vina [196, 197]	Open-source program for molecular docking	[39, 40, 43, 53, 81]
SMINA [19]	Fork of AutoDock Vina focused on improving scoring functions and energy minimization	[106, 110]
GNINA [20, 198]	Fork of SMINA, employing CNNs for improved support of scoring functions and ligand optimization	[106, 110]
QuickVina 2 [199]	Fork of AutoDock Vina using heuristics to reach significant speedups at similar accuracy	[47]
QuickVina-W [21]	Update of QuickVina 2, providing the ability to dock blindly if the docking site is unknown	[106, 110]
GLIDE [22]	Schrödinger software package for ligand-receptor docking	[106, 110]
RDKit [23]	Open-source cheminformatics toolbox	[34, 47, 53, 106]
Open Babel [24]	Open-source cheminformatics toolbox	[79]
OEChem KT [25, 200, 201]	OpenEye Scientific programming library for chemistry and cheminformatics	[103]
PaDEL [38]	Open-source software for calculating molecular descriptors and fingerprints	[37, 39]
AutoClickChem [48]	Currently out-of-date open-source software for automated in silico chemical synthesis	[47]
MD tools
GROMACS [134]	Open-source software for high-performance MD simulation and output analysis	[183]
Amber [135]	Suite of biomolecular simulation programs	[179]
NAMD [136]	Open-source software for high-performance simulation of large biomolecular systems	[137, 174]
MDTraj [180]	Open-source Python library for MD trajectory manipulation and analysis	[179]
PyReweighting [181]	Open-source Python toolkit to facilitate the reweighting of accelerated MD simulations	[179]
General databases for ML/DL development
UCI ML repository [188]	Collection of accessible databases for ML/DL training and analysis
Kaggle [189]	Data science community platform
Google Dataset Search [190]	Search engine for freely available online data
Compound library databases
ChEMBL [12,13,14]	Manually curated database of bioactive molecules with drug-like properties	[36, 37, 39, 53]
PubChem [15]	World’s largest free chemical information database	[36]
ZINC [16, 17]	Curated collection of commercially available chemical compounds for VS	[47, 53, 174]
PDBbind [35]	Collection of experimentally measured binding affinity data for biomolecular complexes	[34, 79, 81, 106, 110]
Selleck [202]	Bioactive compound libraries that consist of small molecules with validated biological and pharmacological activities	[39]
TargetMol [44]	Research supplier for compound libraries of small molecule compounds	[43]
UniProt [203]	Freely accessible resource for protein sequence and functional information	[64]
Benchmarking sets
Astex Diverse Set [82]	Diverse, high-quality test set for the validation of protein–ligand docking performance	[79]
PoseBusters [114]	Python package to perform standard quality checks on DL-based protein–ligand docking methods	[112]
CASF-2016105	Open-access benchmark to assess and compare scoring functions in several metrics	[104]

This is a noncomprehensive list, meant to inspire researchers of the range of tools at their disposal. It is divided into (1) software libraries used to develop ML/DL models, (2) the various mentioned VS tools and (3) MD tools, (4) databases useful for training DL models, in general as well as (5) compound library databases, and lastly (6) mentioned tools useful for benchmarking certain DL models

Back to article page

ISSN: 1758-2946

Contact us

Submission enquiries: journalsubmissions@springernature.com