Apply drug-likeness, structural alerts, and PAINS diagnostics to compound libraries.
npx skills add https://github.com/ovachiever/droid-tings --skill medchem-ovachiever
Please help me install this Agent Skill. Skill: medchem Source: https://github.com/ovachiever/droid-tings/tree/main/skills/medchem Command: npx skills add https://github.com/ovachiever/droid-tings --skill medchem-ovachiever
Medchem offers a suite of medicinal chemistry rules, structural alerts, and filters to triage and prioritize compound libraries for drug design, enabling faster, chemistry-aware decision making.
Use cases include high-throughput triage, lead optimization, and CNS-targeted drug discovery.
Filter a library with Ro5 and common alerts, then review Lilly demerits for top candidates.
High-intent search queries and answers about installing and using this skill.
Drug-likeness filtering applies Ro5, Veber, and lead-like rules to screen compounds by molecular weight, logP, H-bond donors/acceptors, and rotatable bonds. The Skill batch-applies these rules to SMILES strings or RDKit Mol objects, returning per-molecule pass/fail results for rapid library triage in drug discovery workflows.
Structural alerts flag substructures linked to toxicity, instability, or poor druggability; PAINS (Pan-Assay Interference Compounds) identifies compounds likely to generate false positives in screening assays. This Skill detects both via modular alert filters (CommonAlertsFilters, NIBRFilters, LillyDemeritsFilters), enabling chemistry-aware compound prioritization.
Yes. The Skill accepts SMILES, CSV, TSV, and SDF inputs and supports high-performance batch processing via n_jobs parallelization and progress tracking. This enables scalable screening of large molecular datasets for library curation and lead optimization.
Stack rule filters sequentially using RuleFilters and modular alert components to apply drug-likeness, structural alerts, and vendor demerits in a single workflow. The Skill outputs detailed per-molecule results showing which filters each compound passes or fails.
Common alerts flag frequent toxicophores; NIBR and Lilly demerits are vendor-specific structural alert sets reflecting internal drug-discovery criteria. The Skill applies each as independent modular filters, letting you prioritize compounds based on your project's risk tolerance and chemistry constraints.
Yes. The Skill depends on RDKit for molecular parsing and manipulation, pandas for tabular data handling, datamol for cheminformatics utilities, and tqdm for progress display. These handle SMILES parsing, descriptor calculation, and batch processing of compound libraries.
jasrajtulsiParse and standardize molecular structures for cheminformatics workflows.
silversteinApply drug-likeness rules and PAINS filters to prioritize compound libraries.
ovachieverConvert SMILES or RDKit molecules into ML-ready numerical representations.
biomaps-infraTrains and applies DeepChem machine learning models for molecular property prediction.
