What problem does it solve?
This Skill helps you prioritize and triage compound libraries by applying widely used medicinal chemistry filters and structural alert checks so you can quickly focus on higher-quality candidates.
Core Features & Use Cases
- Drug-likeness rule filtering: Apply Lipinski (Rule of Five), Veber, Oprea lead-likeness, CNS, REOS, and related rule sets to screen for appropriate physicochemical profiles.
- Structural alert and demerit assessment: Run PAINS filtering, common structural alerts, and Lilly demerits to deprioritize problematic or assay-interfering chemotypes.
- Medicinal chemistry scoring support: Compute molecular complexity metrics (e.g., Bertz/Whitlock/Barone) and enforce property constraints (e.g., MW, logP, TPSA, HBD/HBA, rotatable bonds).
- Structure pattern detection: Detect specific chemical groups (e.g., hinge binders, phosphate binders, Michael acceptors) and match against named catalogs.
- Workflow orchestration: Use a query language to combine rules, alerts, and property filters into an expressive screening workflow.
Quick Start
Use the medchem skill to filter a candidate molecule library for compounds that pass Rule of Five while excluding PAINS: read the medchem guide, then ask the assistant to apply rule_of_five and pains_filter to your molecules.
