Decoding the Molecular Target Landscape of Chalcones: An Integrated Cheminformatics Analysis Using PubChem Patent Data, Protein Target Profiling, and Bioassay Activity Mapping.
Abstract
Background
Chalcones are α,β-unsaturated ketone scaffolds with broad documented bioactivity. No prior study has simultaneously integrated structural validation, PAINS assessment, chemical-space clustering, human-target classification, bioassay hit-rate benchmarking, and patent analysis within a single computational framework.
Methods
Seven hundred and twenty-two PubChem chalcone compounds were curated by SMARTS-based E-configured enone validation, ECFP4 k-means clustering, Bemis-Murcko decomposition, QED/Fsp³ profiling, and PAINS filtering. Protein target profiling (476 proteins; 145 human), bioassay benchmarking (12,981 assays; 81,667 data points; ChEMBL 34), and patent analysis (2,148 records) were integrated across PubChem exports.
Results
Among 722 compounds, 686 (95.0 %) were E-configured; 50 (6.9 %) were PAINS-flagged; 195 Bemis-Murcko scaffolds (SDI = 0.270); mean QED = 0.444 ± 0.175; k = 2 clusters were statistically optimal. PAINS-flagged compounds showed significantly lower QED (0.392 vs. 0.449; p < 0.001). Among 145 human protein targets across 25 functional classes, key categories included tubulin isoforms (15), ABC transporters, MAO-A/B, and AChE. Bioassay analysis yielded 18,438 active hits (22.6 % overall); hit rates were: MAO 38.6 %, ABC transporters 32.2 %, AChE 31.8 %, and anticancer 31.5 %. China dominated patent filings (38.4 %); A61P35 (anticancer) was the leading IPC category.
Discussion
Chalcones demonstrate robust polypharmacological engagement across multiple target classes. The low PAINS prevalence (6.9 %) confirms chemical tractability, while the inverse correlation between hit rate and assay dataset size underscores the need to account for reporting bias. The growing patent portfolio highlights the continued translational relevance of this scaffold.
Conclusion
Chalcones represent a tractable, polypharmacological scaffold with target engagement validated across 145 human proteins. Hit-rate patterns provide reliable evidence even after accounting for reporting bias, although translational claims require independent experimental validation.