Chembench

Chembench provides integrated cheminformatics capabilities for mining, curation, visualization, analysis, and QSAR-based modeling of chemical and biological data to support computer-assisted drug design and computational toxicology.

Key Features:

• Data mining and curation: Mining and curation of chemical and biological datasets for downstream modeling and analysis.
• Visualization and analysis: Visualization and analytical functions for exploring chemical genomics and assay data.
• QSAR modeling: Integration and application of quantitative structure–activity relationship (QSAR) models for activity and toxicity prediction.
• Virtual screening: Virtual screening of compound libraries, including DrugBank, to prioritize candidate molecules.
• Consensus prediction: Generation of consensus computational hits across models to improve candidate selection.

Scientific Applications:

• Computer-assisted drug design: Use of QSAR models and virtual screening to design or select compounds and compound libraries with improved hit rates in screening.
• Computational toxicology: Application of QSAR and modeling approaches for toxicity prediction.
• SARS-CoV-2 main protease (M^pro) inhibitor discovery: Leveraging experimental data from SARS-CoV M^pro and QSAR-based virtual screening of DrugBank to identify 42 consensus computational hits, three of which were experimentally confirmed by NCATS.
• Chemical genomics data analysis: Analysis and modeling of chemical genomics datasets to inform bioactivity and target predictions.

Methodology:

Computational methods include chemical and biological data mining and curation, visualization and analysis, QSAR modeling, virtual screening, and consensus scoring; QSAR models were built using experimental data from SARS-CoV M^pro to predict inhibitors for SARS-CoV-2 M^pro with virtual screening of DrugBank yielding 42 consensus hits.