FAVES V3: A Three-Tier Regulatory Compliance Detection System for Molecular Databases

Rapid identification of controlled substances and regulatory compliance status is critical for drug discovery pipelines, yet existing tools lack comprehensive coverage of DEA-scheduled compounds. This paper presents FAVES V3 (Fairness, Accountability, Validity, Ethics, Safety-Version 3), a three-tier regulatory compliance detection system that combines whitelist verification, direct molecular matching, and scaffold pattern analysis to identify controlled substances in molecular databases. Validated against a benchmark of 102 compounds comprising 47 DEA-scheduled substances (Schedules I-V) and 55 FDA-approved non-controlled drugs, FAVES V3 achieved 100% sensitivity and 100% specificity with zero false positives and zero false negatives. The system has been deployed to screen a database of 122 million molecules, enabling real-time compliance checking via API. The validation benchmark dataset and methodology are released to enable reproducibility and community validation.