BioMed Skill Router — Подбор навыков для биомедицинских задач

--- name: biomed-skill-router description: > Find the most suitable skill for a given biomedical task. Use this skill when: (1) You are unsure which skill to use for a specific biomedical task, (2) You want to discover available skills for a particular domain, (3) You need to compare multiple skills for a given use case. license: MIT category: utilities tags: [skill-discovery, routing, recommendation, navigation] --- # BioMed Skill Router Find the most suitable skill for your biomedical task by analyzing your request and matching it against available skills. ## When to Use - User describes a biomedical task but doesn't know which skill to use - User wants to discover what skills are available for a particular domain - User needs help choosing between multiple skills for a similar task - User asks "What can I do with OpenBioMed?" or "Which skill should I use for X?" ## How It Works The agent analyzes the user's natural language request and determines the most suitable skill based on: 1. Understanding the user's intent and goal 2. Identifying what data/inputs the user has 3. Matching against available skill capabilities ## Available Skills ### Biochemistry & Drug Discovery | Skill | Description | |-------|-------------| | drug-candidate-discovery | Generate diverse druggable molecules for a given target or disease using AI-powered drug discovery tools. | | drug-lead-analysis | Analyze drug candidate molecules for drug-likeness (QED, Lipinski), ADMET properties, BBB penetration, and safety profiles. | | target-based-lead-design | Generate diverse lead compounds for a specific protein target using structure-based drug design with MolCraft. | | admet-prediction | Predict comprehensive ADMET properties (BBB penetration, side effects, Caco-2 permeability, half-life, LD50 toxicity). | | retrosynthesis-planning | Expert-in-the-loop retrosynthetic planning workflow for breaking down target molecules into available starting materials. | | drug-drug-interaction-analysis | Analyze potential drug-drug interactions (DDI) for up to 5 drugs using KEGG DDI database. | | text-based-molecule-editing | Modify molecules based on natural language descriptions using MolT5/BioT5 models. | ### Protein Analysis & Engineering | Skill | Description | |-------|-------------| | protein-mutation-analysis | Analyze protein mutations by retrieving protein data, explaining mutation effects with MutaPLM, and predicting structure. | | mutation-design-aav | Design high-fitness and high-diversity mutants of AAV VP1 capsid protein. | | mutation-design-gfp | Design high-fluorescence and high-diversity GFP mutants. | | functional-protein-design | Generate functional protein sequences using CodeFun with Gene Ontology (GO) tag guidance. | | similar-protein-retrieval | Retrieve proteins with similar structures or sequences from UniProt, PDB, and AFDB databases. | | structure-prediction-boltz-2 | Predict protein complex structures and protein-ligand complexes with binding affinity using Boltz-2. | | protein-structure-design-boltzgen | All-atom protein design using BoltzGen diffusion model for binder design and peptide design. | | antibody-structure-prediction-tfold | Predict antibody/nanobody structures and antigen-antibody complex structures. | | antibody-design-iggm | Epitope-conditioned de novo antibody design and affinity maturation. | | binding-affinity-prediction-prodigy | Predict binding affinity scores for protein complexes using Prodigy. | | protein-ligand-binding-analysis-plip | Analyze protein-ligand interactions in PDB structures using PLIP (Protein-Ligand Interaction Profiler). | | protein-function-prediction | Predict protein function and properties from amino acid sequence using BioT5. | | protein-subcellular-localization-prediction-biot5 | Predict protein subcellular localization from amino acid sequence using BioT5. | ### Single-Cell Omics Data Analysis | Skill | Description | |-------|-------------| | single-cell-foundation-model-scrna-seq-geneformer | Geneformer workflows for tokenization, cell/gene classification, embedding extraction, and perturbation analysis. | | single-cell-foundation-model-scrna-seq-langcell | LangCell for zero-shot and few-shot cell type annotation with multimodal cell-text matching. | | single-cell-foundation-model-scrna-seq-scgpt | scGPT for preprocessing, binning, cell embedding extraction, fine-tuning, and reference mapping. | | spatial-transcriptomics-foundation-model-stofm | SToFM for spatial transcriptomics preprocessing and cell embedding generation. | | single-cell-scrna-seq-analysis-scanpy | Complete scRNA-seq analysis workflow with Scanpy including QC, normalization, clustering, and marker gene identification. | | single-cell-multi-omics-analysis-scvi | Probabilistic deep learning for single-cell multi-omics analysis including scVI, scANVI, totalVI. | | cellxgene-census-query | Query CZ CELLxGENE Census (61M+ cells) for single-cell expression data. | | spatial-transcriptomics-spatial-data-io | Load spatial transcriptomics data from Visium, Xenium, MERFISH, Slide-seq, and other platforms. | | single-cell-atac-seq-qc-processing | Trim adapters, align reads, remove duplicates, and evaluate chromatin accessibility data quality. | | single-cell-atac-seq-peak-calling-annotaion | Call accessible chromatin peaks with MACS2 and identify differentially accessible regions. | | single-cell-proteomics-data-processing | Load, inspect, centroid, and extract features from raw LC-MS/MS data files using pyOpenMS. | | single-cell-proteomics-peptide-identification | Search MS2 spectra against protein databases with MSFragger/Comet. | | single-cell-multi-omics-data-harmonization | Prepare multi-omics datasets for joint integration with normalization and batch correction. | ### Data Retrieval & Knowledge | Skill | Description | |-------|-------------| | pubchem-query | Query PubChem database for chemical structures, similar compounds, and bioactivity data. | | biomedical-literature-search | Search PubMed and bioRxiv for biomedical research papers. | | uniprot-query | Query UniProt database for protein sequences, metadata, and search by criteria. | | kegg-query | Query KEGG database for drug information, pathway analysis, and disease-drug-target discovery. | | ppi-string-query | Query STRING database for protein-protein interactions with confidence scores. | | chembl-query | Query ChEMBL database for bioactivity data on drug-like compounds. | ### Molecule Analysis | Skill | Description | |-------|-------------| | iupac-name-identification-biot5 | Identify the IUPAC name of a molecule using BioT5 question answering model. | | molecule-biochemical-significance-query-biot5 | Query a molecule's biochemical significance and roles in biology and chemistry. | ### Disease & Target Analysis | Skill | Description | |-------|-------------| | disease-drug-intelligence | Comprehensive disease-to-innovative-drug analysis: targets, drugs, clinical progress, and mechanism trends. | | target-drug-report | Generate comprehensive drug development progress reports for disease therapeutic targets. | ### Utilities | Skill | Description | |-------|-------------| | biomed-skill-creator | Create new biomedical skills or improve existing ones. | ## Routing Process When a user asks for help finding a skill: 1. **Understand the request** - Analyze what the user wants to accomplish 2. **Identify available inputs** - Determine what data the user has (molecule, protein, sequence, etc.) 3. **Match to skills** - Select the most appropriate skill(s) based on the task 4. **Provide recommendation** - Suggest the best skill and explain why ## Example Interactions **User:** "I want to design a drug for BCL2 protein" **Agent:** Based on your request to design drugs for the BCL2 protein, I recommend: **`target-based-lead-design`** - This skill generates diverse lead compounds for a specific protein target using structure-based drug design. You'll need a PDB ID for BCL2. Would you like me to help you find one? --- **User:** "What can I use to analyze my scRNA-seq data?" **Agent:** For scRNA-seq analysis, several skills are available depending on your goal: - **`single-cell-scrna-seq-analysis-scanpy`** - Complete analysis workflow (QC, clustering, marker genes) - **`single-cell-foundation-model-scrna-seq-langcell`** - Cell type annotation - **`single-cell-foundation-model-scrna-seq-geneformer`** - Embedding extraction and perturbation analysis - **`single-cell-foundation-model-scrna-seq-scgpt`** - Embedding extraction and reference mapping What specific analysis do you want to perform? --- **User:** "I have a molecule and want to check if it's toxic" **Agent:** For toxicity analysis, use **`admet-prediction`**. This skill predicts comprehensive ADMET properties including toxicity (LD50), side effects, and BBB penetration. Please provide your molecule as a SMILES string. ## See Also - `skills_overview.md` - Full skill catalog with detailed descriptions - `biomed-skill-creator` - Create new skills when none match your needs