Gypsum-DL

Gypsum-DL generates 3D molecular models and enumerates ionization states, tautomeric variants, stereochemical (chiral) configurations, cis/trans isomers, and ring conformations to prepare small-molecule libraries for structure-based virtual screening (SBVS).

Key Features:

• Input formats: Accepts SMILES or flat SDF input formats for virtual compound libraries.
• Comprehensive enumeration: Systematically enumerates ionization states, tautomeric variants, chiral configurations, cis/trans isomers, and ring conformations for each molecule.
• Output generation: Produces an SDF file containing each enumerated molecular form with assigned 3D coordinates.

Scientific Applications:

• Structure-based virtual screening: Provides chemically relevant 3D ligand variants to support SBVS workflows.
• Ligand library preparation for drug discovery: Captures chemical diversity and alternative molecular forms to improve identification of bioactive compounds.
• Pose prediction benchmarking: Supports evaluation of virtual screening pose prediction accuracy using curated protein–ligand datasets.

Methodology:

Demonstrated by processing the NCI Diversity Set VI (1,558 molecules), the Distributed Drug Discovery (D3) combinatorial virtual library (56,608 molecules), and the PDBBind database (4,463 protein–ligand complexes), with reported improvements in pose prediction accuracy.