Drug Discovery

This skill provides comprehensive documentation and guidance for using datamol, a Python library that simplifies molecular cheminformatics tasks. It covers SMILES parsing, molecular descriptors, fingerprints, clustering, 3D conformer generation, visualization, and chemical reactions with sensible defaults built on top of RDKit.

Provides comprehensive API access to the ZINC database, a UCSF-maintained repository of 230M+ purchasable chemical compounds. Supports searching by ZINC ID or SMILES notation, similarity searches, random sampling, and downloading 3D structures for molecular docking studies in drug discovery workflows.

This skill enables querying the PubChem database, the world's largest freely available chemical database with 110M+ compounds. It supports searching compounds by name, structure (SMILES/InChI), or formula, retrieving molecular properties, performing similarity and substructure searches, and accessing bioactivity data from screening assays.

This skill enables access to the RCSB Protein Data Bank, the worldwide repository for 3D structural data of biological macromolecules. It supports searching for protein and nucleic acid structures by text, sequence, or structural similarity, downloading coordinate files in various formats, and retrieving metadata for structural biology and drug discovery workflows.

This skill enables querying the Open Targets Platform for drug target discovery, allowing users to search for gene targets, disease associations, known drugs, and evidence data. It provides Python helper functions to access tractability assessments, safety liabilities, clinical precedence, and genetic evidence for therapeutic target prioritization.

This skill enables querying the ChEMBL database, a manually curated collection of over 2 million bioactive molecules maintained by the European Bioinformatics Institute. It supports compound searches, target information retrieval, bioactivity data queries (IC50, Ki, EC50), similarity/substructure searches, and drug discovery workflows for medicinal chemistry research.

This skill enables access to the AlphaFold Protein Structure Database, providing code examples and documentation for retrieving AI-predicted 3D protein structures. It supports querying by UniProt ID, downloading coordinate files (PDB/mmCIF), analyzing confidence metrics (pLDDT, PAE), and bulk data access via Google Cloud for structural biology and drug discovery workflows.

TorchDrug is a documentation skill that provides comprehensive guidance for using the TorchDrug PyTorch library in drug discovery and molecular science. It covers graph neural networks for molecules and proteins, including molecular property prediction, protein modeling, knowledge graph reasoning, molecular generation, and retrosynthesis planning with 40+ curated datasets and 20+ model architectures.

RDKit is a comprehensive cheminformatics skill that guides users through molecular analysis tasks including reading/writing chemical structures (SMILES, SDF), calculating molecular descriptors (LogP, TPSA, molecular weight), generating fingerprints for similarity searching, and performing substructure matching. It includes ready-to-use Python scripts for drug-likeness screening, similarity searches, and functional group filtering.

PyTDC (Therapeutics Data Commons) provides AI-ready datasets and benchmarks for drug discovery and development. It offers curated datasets spanning ADME, toxicity, drug-target interactions, and molecular generation with standardized evaluation metrics and meaningful data splits for therapeutic machine learning applications.

Molfeat is a comprehensive guide for molecular featurization in machine learning. It provides documentation for converting chemical structures (SMILES strings) into numerical representations using 100+ featurizers including fingerprints (ECFP, MACCS), descriptors (RDKit, Mordred), and pretrained models (ChemBERTa, GIN). Ideal for QSAR modeling, virtual screening, and cheminformatics tasks.

This skill helps chemists and drug discovery researchers filter and prioritize compound libraries using established medicinal chemistry rules like Lipinski's Rule of Five, PAINS filters, and structural alerts. It can process molecules in batch with parallel processing and generate detailed filtering reports.

DiffDock is a molecular docking skill for computational drug discovery. It helps predict how small molecule ligands bind to protein targets using diffusion-based deep learning, providing binding pose predictions and confidence scores for structure-based drug design workflows.

A comprehensive molecular machine learning skill using DeepChem for predicting chemical properties like solubility and toxicity. It supports graph neural networks, transfer learning with pretrained models (ChemBERTa, GROVER), and MoleculeNet benchmarks for drug discovery and materials science applications.

Autonomous biomedical AI agent framework for executing complex research tasks across genomics, drug discovery, molecular biology, and clinical analysis. Use this skill when conducting multi-step biomedical research including CRISPR screening design, single-cell RNA-seq analysis, ADMET prediction, GWAS interpretation, rare disease diagnosis, or lab protocol optimization. Leverages LLM reasoning with code execution and integrated biomedical databases.