"M3Dock: A Collaborative Evolutionary\u2013Gradient Optimization Framework for Multi-Objective Molecular Docking"

"All protein-ligand complexes were originally obtained from the Protein Data Bank (PDB) [49]. The experiments mainly employ three widely used benchmark datasets, Astex, PoseBusters and PDBbind, which are commonly adopted for evaluating molecular docking and binding pose prediction methods.These datasets contain diverse protein\u2013ligand complexes and have been extensively adopted for benchmarking docking methods.  Astex, PoseBusters and PDBbind datasets were downloaded in MMTF format [52] from the PDB. Before docking experiments, each complex was preprocessed using PyMOL [53], which included: (1) removing water molecules from the crystal structure; (2) removing metal ions and adding hydrogen atoms to all protein and ligand atoms; (3) assigning partial charges to each atom and determining protonation states; and (4) saving the processed proteins and ligands as PDBQT files. In addition, AutoGrid4 [19] was used to generate the energy grid files (.map) required for docking calculations, which were then used in subsequent molecular docking experiments."