Answers are grounded in this paper and cite their sources.
Open-source pipeline integrates multiple tools for drug discovery, enhancing Korean traditional medicine research.
Correction (new version, 2026-07-21) — citation correction. Correction (2026-07-21) - withdrawn calibration removed from the contribution list. The 'unique system contribution' list credited a 'ChEMBL rho=0.72 calibration' as part of the single-pipeline integration. That calibration (n=15) was computed against a ligand panel whose potency annotations were established as fabricated (2026-07-16 audit) and has been withdrawn from the source methodology preprint; it is removed from this list. The remaining pipeline components are unaffected. No result, ranking or figure value changes; this is a provenance and scope-of-claim correction. Full audit: preprints/_metadata/DANGLING_CITATION_AUDIT_2026_07_21.md; underlying finding: preprints/_metadata/FABRICATED_PANEL_SCOPE_2026_07_16.md. Modern AI-driven drug discovery is built on a heterogeneous stack of open-source tools. Specialized application domains require tool selection, integration, and adaptation. We describe Genesis_Medicine, an open-source pipeline for AI-driven Korean traditional medicine (한약·생약) drug discovery. We honestly distinguish: (i) a 7-tool active core that produces all real pipeline outputs in this work (Boltz-2 cofold, REINVENT 4 generative, ADMET-AI v2 property prediction, OpenMM 8 + openmmtools alchemical sampling, RDKit chemistry, ChEMBL read-only, Open Targets v4 GraphQL); and (ii) a 40+ adapter scaffold catalog organized into 12 functional tiers, written as Python modules and exposed through a natural-language agent interface but not yet actively used in primary pipeline runs — including AlphaFold-3-class cofold ensemble (Chai-1, Protenix-v2, OpenFold-3), generative SOTA (PocketXMol, FlowMol3, DiffSBDD, TurboHopp), ML potentials (MACE-OFF24, AIMNet2), pose validation (PoseBusters), and specialized adapters (PROTAC designer, chronotherapy, TxGNN repurposing). The active core supports the complete workflow from natural-product curation through corrected ABFE; the adapter scaffold provides ready integration points for prioritized future expansion. We discuss the design philosophy — 3-pillar institutional integration (HAN PREDICT, Inc. AI healthcare platform; Recover Korean Medicine Clinic), commitment to honest in silico reporting, and explicit method limitations including MMP-1 zinc handling and Boltz-2 binary-classifier vs IC₅₀ distinction. The pipeline is open-source under Apache-2.0 at <https://github.com/crazat/genesis_medicine>. Keywords: open-source pipeline, Korean traditional medicine, AI drug discovery, REINVENT4, Boltz-2, ADMET-AI, ABFE, OpenMM, natural products, integrated stack. ---
No takes yet. Share an insight, caveat, or question.
Cheongwoo Han (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: