MOLS 2.0

MOLS 2.0 performs peptide modeling and induced-fit protein–ligand docking to explore conformational space and predict peptide–protein and nonpeptide small molecule binding poses with receptor flexibility.

Key Features:

• Induced-fit docking: Enables induced-fit docking of peptide ligands and nonpeptide small molecules to flexible receptor proteins using mutually orthogonal Latin squares (MOLS) sampling.
• MOLS sampling: Employs mutually orthogonal Latin squares to sample conformational spaces efficiently.
• Receptor flexibility: Permits receptor residue deviations up to 3.0 Å from starting positions to accommodate binding-induced conformational changes.
• Improved scoring system: Implements an enhanced scoring function that improves ranking effectiveness by 24%.
• Performance optimizations: Code optimizations reduce peptide docking times by two orders of magnitude and increase nonpeptide small molecule docking speed approximately 14-fold.

Scientific Applications:

• Benchmarking and comparison: Validated and benchmarked against peptide–protein complexes, nonpeptide small molecule ligands, and the Astex Diverse set, showing superior performance versus GOLD v5.2.1, AutoDock Vina, HPEPDOCK, AutoDock CrankPep v1.0, HADDOCK, PatchDock, and RosettaLigand.
• Peptide–protein docking: Achieves success rates of 6%, 37%, and 89% at the top 1, 10, and 100 ranked predictions, respectively.
• Small molecule–protein docking: Achieves success rates of 14%, 31%, and 49% at the top 1, 10, and 100 ranked predictions, respectively.

Methodology:

Samples conformational space using mutually orthogonal Latin squares (MOLS) and analyzes sampled conformations and docking poses with a variant of mean field theory.