gRNAde MCP Server

{ "project_name": "grnade-mcp", "full_name": "gRNAde MCP (Geometric RNA Design)", "description": "Model Context Protocol server for generative inverse design of RNA structure and function with gRNAde", "version": "1.0.0-dev", "status": "in_development", "tools": [ { "name": "design_rna_sequences", "type": "sync", "description": "Generate RNA sequences from 3D structure using gRNAde", "example_prompt": "Design 100 RNA sequences for @examples/data/riboswitch.pdb with target secondary structure", "max_limit": "1000 designs", "runtime": "< 10 minutes" }, { "name": "filter_rna_designs", "type": "sync", "description": "Filter and rank RNA designs based on quality metrics", "example_prompt": "Filter designs in @results/designs.fasta using openknot score > 80", "parameters": ["designs_file", "filter_metric", "threshold"] }, { "name": "evaluate_rna_designs", "type": "sync", "description": "Evaluate RNA designs using computational tools like RibonanzaNet", "example_prompt": "Evaluate RNA designs in @results/designs.fasta using RibonanzaNet", "methods": ["ribonanzanet", "rhofold", "viennarna"] }, { "name": "submit_rna_design_campaign", "type": "submit", "description": "Large-scale design generation for background processing", "example_prompt": "Submit design campaign for @examples/data/structure.pdb generating 10,000 designs", "runtime": "> 10 minutes" }, { "name": "submit_batch_rna_design", "type": "submit", "description": "Process multiple structures in parallel", "example_prompt": "Submit batch design for all PDB files in @examples/data/structures/", "runtime": "> 10 minutes" }, { "name": "validate_rna_structure", "type": "sync", "description": "Validate PDB structure file format and content", "example_prompt": "Validate structure file @examples/data/sample.pdb", "runtime": "< 1 minute" }, { "name": "get_job_status", "type": "management", "description": "Check job progress", "parameters": ["job_id"] }, { "name": "get_job_result", "type": "management", "description": "Get results when completed", "parameters": ["job_id"] }, { "name": "get_job_log", "type": "management", "description": "View execution logs", "parameters": ["job_id", "tail"] }, { "name": "cancel_job", "type": "management", "description": "Cancel running job", "parameters": ["job_id"] }, { "name": "list_jobs", "type": "management", "description": "List all jobs", "parameters": ["status"] }, { "name": "get_example_data", "type": "utility", "description": "Get information about available example datasets", "example_prompt": "What example data is available for testing?" } ], "scripts": [ { "name": "design_rna.py", "description": "Generate RNA sequences from 3D structures", "example_command": "python scripts/design_rna.py --pdb examples/data/riboswitch.pdb --output results/ --n-designs 1000", "parameters": ["--pdb", "--target-structure", "--n-designs", "--mode", "--output"], "status": "placeholder_implemented" }, { "name": "filter_designs.py", "description": "Filter and rank generated designs", "example_command": "python scripts/filter_designs.py --input results/designs.fasta --filter-metric openknot_score --threshold 80", "parameters": ["--input", "--output", "--filter-metric", "--threshold"], "status": "placeholder_implemented" }, { "name": "evaluate_designs.py", "description": "Evaluate designs with computational tools", "example_command": "python scripts/evaluate_designs.py --input results/designs.fasta --method ribonanzanet", "parameters": ["--input", "--output", "--method", "--reference-structure"], "status": "placeholder_implemented" } ], "demo_data": [ { "file": "riboswitch.pdb", "description": "ZMP riboswitch structure for testing RNA design", "size": "~70 nucleotides", "used_by": ["design_rna_sequences", "design_rna.py"], "status": "planned" }, { "file": "aptamer.pdb", "description": "Small aptamer structure for quick testing", "size": "~30 nucleotides", "used_by": ["design_rna_sequences", "validate_rna_structure"], "status": "planned" }, { "file": "ribozyme.pdb", "description": "RNA polymerase ribozyme for functional design", "size": "~100+ nucleotides", "used_by": ["submit_rna_design_campaign"], "status": "planned" } ], "configs": [ { "file": "default_design.yaml", "description": "Default RNA design parameters", "parameters": { "mode": "3d", "total_samples": 1000, "filter_metric": "openknot_score", "pass_threshold": 80, "temperature_range": [0.1, 1.0] }, "status": "implemented" }, { "file": "high_throughput.yaml", "description": "High-throughput design configuration", "parameters": { "mode": "3d", "total_samples": 100000, "parallel_jobs": 4, "batch_size": 1000, "pass_threshold": 70 }, "status": "implemented" } ], "features": [ "3D RNA Design from PDB structures", "Expert-Level Accuracy on Pseudoknots", "Functional Design Campaigns", "High-Throughput Pipeline with Screening", "Batch Processing Support", "Multiple Evaluation Metrics", "Job Management for Long-Running Tasks" ], "dependencies": { "python_version": "3.10+", "pytorch": "2.1.0", "pytorch_geometric": "latest", "mcp": "fastmcp>=0.2.0", "bioinformatics": ["biopython", "biotite", "mdanalysis"], "scientific": ["numpy", "pandas", "scipy", "matplotlib"], "utilities": ["pyyaml", "tqdm", "wandb", "python-dotenv"] }, "installation": { "environment": "conda/mamba environment with Python 3.10", "gpu_support": "CUDA 11.8/12.1 recommended", "memory_requirements": "8GB RAM minimum, 16GB+ recommended", "storage": "Model checkpoints ~2GB from HuggingFace" }, "original_project": { "name": "geometric-rna-design", "url": "https://github.com/chaitjo/geometric-rna-design", "paper": "Generative inverse design of RNA structure and function with gRNAde", "license": "MIT", "authors": "Chaitanya K. Joshi et al." }, "development_status": { "phase": "Phase 1 - Foundation Setup", "completed": [ "Repository structure created", "README documentation", "Configuration files", "MCP server skeleton", "Script placeholders" ], "in_progress": [ "MCP tool implementations", "Script functionality", "Example data preparation" ], "planned": [ "RibonanzaNet integration", "RhoFold integration", "Batch processing system", "Job management system" ] } }