OpenClaw Scientific Research

<div align="center"> # 🔬 OpenClaw Scientific Skill **A comprehensive AI-powered scientific research toolkit for OpenClaw** [](https://github.com/openclaw/openclaw) [](LICENSE) [](#features) *Transform your AI assistant into a scientific research powerhouse* </div> --- ## ✨ Features ### 🧬 Bioinformatics & Genomics - Sequence analysis with BioPython - Single-cell RNA-seq with Scanpy - Variant annotation and interpretation - Phylogenetic analysis ### 🧪 Cheminformatics & Drug Discovery - Molecular manipulation with RDKit - Virtual screening and ADMET prediction - Molecular docking workflows - Chemical database access (ChEMBL, PubChem, ZINC) ### 🏥 Clinical Research & Precision Medicine - ClinicalTrials.gov integration - Variant interpretation (ClinVar, COSMIC) - Pharmacogenomics workflows - Cancer genomics (cBioPortal, DepMap) ### 🤖 Machine Learning & AI - Deep learning with PyTorch Lightning - Time series forecasting with TimesFM - Bayesian methods with PyMC - Model interpretability with SHAP ### 📚 250+ Scientific Databases - Literature: PubMed, OpenAlex, bioRxiv - Proteins: UniProt, PDB, AlphaFold DB - Chemicals: ChEMBL, PubChem, DrugBank - Clinical: ClinVar, COSMIC, ClinicalTrials.gov ### 📊 Scientific Visualization & Publishing - Publication-quality figures - LaTeX document generation - Presentation slides - Citation management ### ⚛️ Physics & Quantum Computing - Astronomical data analysis (Astropy) - Quantum computing (Qiskit, PennyLane) - Symbolic mathematics (SymPy) ### ⚙️ Engineering & Optimization - Multi-objective optimization (PyMOO) - Metabolic modeling (COBRApy) - Materials science (Pymatgen) --- ## 🚀 Quick Start ### Installation ```bash # The skill is automatically loaded by OpenClaw # Just copy to your skills directory cp -r scientific ~/.openclaw/skills/ ``` ### Example Usage **Single-cell RNA-seq Analysis:** ```python import scanpy as sc # Load and analyze adata = sc.read_10x_mtx('./data/') sc.pp.filter_cells(adata, min_genes=200) sc.tl.pca(adata) sc.tl.umap(adata) sc.tl.leiden(adata) sc.pl.umap(adata, color='leiden') ``` **Query PubMed:** ```python from Bio import Entrez Entrez.email = "[email protected]" handle = Entrez.esearch(db="pubmed", term="EGFR lung cancer", retmax=20) results = Entrez.read(handle) ``` **Fetch AlphaFold Structure:** ```bash python scripts/fetch_alphafold.py P00533 --output EGFR.pdb ``` --- ## 📁 Structure ``` scientific/ ├── SKILL.md # Main skill definition ├── references/ │ ├── bioinformatics.md # Genomics, single-cell, phylogenetics │ ├── cheminformatics.md # Drug discovery, molecular modeling │ ├── clinical.md # Clinical trials, variant interpretation │ ├── ml.md # Deep learning, time series, Bayesian │ ├── databases.md # 250+ database access patterns │ ├── documents.md # Visualization, LaTeX, publishing │ ├── physics.md # Astronomy, quantum computing │ └── engineering.md # Optimization, metabolic modeling └── scripts/ ├── install_deps.py # Install domain-specific dependencies ├── query_pubmed.py # PubMed search utility └── fetch_alphafold.py # AlphaFold structure downloader ``` --- ## 🔧 Dependencies Install dependencies by domain: ```bash # Core scientific stack pip install numpy scipy pandas matplotlib seaborn # Bioinformatics pip install biopython scanpy pysam # Cheminformatics pip install rdkit deepchem # Machine Learning pip install torch pytorch-lightning scikit-learn # Or use the provided script python scripts/install_deps.py --domain bioinformatics python scripts/install_deps.py --domain ml python scripts/install_deps.py --domain all ``` --- ## 🗃️ Database Coverage | Category | Databases | |----------|-----------| | **Literature** | PubMed, OpenAlex, bioRxiv, CrossRef | | **Proteins** | UniProt, PDB, AlphaFold DB, InterPro | | **Chemicals** | ChEMBL, PubChem, DrugBank, ZINC, BindingDB | | **Genomics** | Ensembl, NCBI Gene, GEO, GTEx, gnomAD | | **Clinical** | ClinVar, COSMIC, ClinicalTrials.gov, ClinPGx | | **Pathways** | KEGG, Reactome, STRING | | **Cancer** | cBioPortal, DepMap | | **Economic** | FRED, Alpha Vantage | --- ## 📖 Documentation - [Bioinformatics Reference](references/bioinformatics.md) - Detailed genomics workflows - [Cheminformatics Reference](references/cheminformatics.md) - Drug discovery pipelines - [Clinical Reference](references/clinical.md) - Precision medicine workflows - [ML Reference](references/ml.md) - Machine learning guides - [Databases Reference](references/databases.md) - API patterns for 250+ databases --- ## 🤝 Comparison with Claude Scientific Skills | Feature | Claude Scientific Skills | OpenClaw Scientific Skill | |---------|-------------------------|---------------------------| | Format | Agent Skills standard | Agent Skills standard ✅ | | Domains | 170+ skills | 8 comprehensive domains | | Code Examples | ✅ | ✅ | | Scripts | Limited | 3+ executable scripts | | Database Access | 250+ | 250+ ✅ | | License | MIT (varies) | MIT | --- ## 📝 Example Workflows ### Find EGFR Inhibitors ``` Query ChEMBL for EGFR inhibitors (IC50 < 50nM), analyze SAR with RDKit, perform virtual screening ``` ### Single-cell Analysis ``` Load 10X data with Scanpy, QC filtering, PCA/UMAP, Leiden clustering, marker identification ``` ### Variant Interpretation ``` Parse VCF, annotate with Ensembl VEP, query ClinVar for pathogenicity, check COSMIC for cancer mutations ``` --- ## 📜 License MIT License - Free for academic and commercial use. --- ## 🙏 Acknowledgments **Inspired by** [Claude Scientific Skills](https://github.com/K-Dense-AI/claude-scientific-skills) by K-Dense. This is an **independent implementation** with original code examples. See [ACKNOWLEDGMENTS.md](ACKNOWLEDGMENTS.md) for details. Built on amazing open-source projects: - [BioPython](https://biopython.org/) - [Scanpy](https://scanpy.readthedocs.io/) - [RDKit](https://www.rdkit.org/) - [PyTorch](https://pytorch.org/) - [scikit-learn](https://scikit-learn.org/) --- <div align="center"> **Made with ❤️ for the scientific community** [Report Bug](https://github.com/Zhaozilongxa/openclaw-scientific-skill/issues) · [Request Feature](https://github.com/Zhaozilongxa/openclaw-scientific-skill/issues) · [ClawHub](https://clawhub.com) </div>