{"id":49640294,"url":"https://github.com/S2026NG/BioAgentFlow","last_synced_at":"2026-05-22T08:00:51.333Z","repository":{"id":346010277,"uuid":"1188166712","full_name":"S2026NG/BioAgentFlow","owner":"S2026NG","description":"Lightweight AI drug virtual screening platform for biologic industry. Nearly 20 years of expertise in biopharma R\u0026D, manufacturing, quality and regulatory affairs.","archived":false,"fork":false,"pushed_at":"2026-03-21T20:18:58.000Z","size":122,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-03-22T09:11:43.471Z","etag":null,"topics":["bioinformatics","diffdock","drug-discovery","esmfold","molecular-docking","protein-structure","virtual-screening"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/S2026NG.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":"CONTRIBUTING.md","funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-03-21T17:51:33.000Z","updated_at":"2026-03-21T20:19:01.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/S2026NG/BioAgentFlow","commit_stats":null,"previous_names":["s2026ng/bioagentflow"],"tags_count":null,"template":false,"template_full_name":null,"purl":"pkg:github/S2026NG/BioAgentFlow","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/S2026NG%2FBioAgentFlow","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/S2026NG%2FBioAgentFlow/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/S2026NG%2FBioAgentFlow/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/S2026NG%2FBioAgentFlow/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/S2026NG","download_url":"https://codeload.github.com/S2026NG/BioAgentFlow/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/S2026NG%2FBioAgentFlow/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":33334777,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-21T12:23:38.849Z","status":"online","status_checked_at":"2026-05-22T02:00:06.671Z","response_time":265,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["bioinformatics","diffdock","drug-discovery","esmfold","molecular-docking","protein-structure","virtual-screening"],"created_at":"2026-05-05T19:00:27.292Z","updated_at":"2026-05-22T08:00:51.326Z","avatar_url":"https://github.com/S2026NG.png","language":"Python","funding_links":[],"categories":["生物医学研究与药物发现"],"sub_categories":[],"readme":"\u003c!-- BioAgentFlow Logo --\u003e\n\u003cp align=\"center\"\u003e\n  \u003cimg src=\"docs/assets/logo_placeholder.png\" alt=\"BioAgentFlow Logo\" width=\"320\"/\u003e\n\u003c/p\u003e\n\n\u003ch1 align=\"center\"\u003eBioAgentFlow\u003c/h1\u003e\n\n\u003cp align=\"center\"\u003e\n  \u003cstrong\u003eLightweight AI-Powered Drug Virtual Screening Platform\u003c/strong\u003e\u003cbr/\u003e\n  \u003cstrong\u003e轻量级 AI 药物虚拟筛选平台\u003c/strong\u003e\n\u003c/p\u003e\n\n\u003cp align=\"center\"\u003e\n  \u003ca href=\"https://github.com/BioAgentFlow/BioAgentFlow/actions\"\u003e\u003cimg src=\"https://img.shields.io/github/actions/workflow/status/BioAgentFlow/BioAgentFlow/ci.yml?branch=main\u0026label=CI\u0026logo=github\" alt=\"CI\"\u003e\u003c/a\u003e\n  \u003ca href=\"https://pypi.org/project/bioagentflow/\"\u003e\u003cimg src=\"https://img.shields.io/pypi/v/bioagentflow?color=blue\u0026logo=pypi\u0026logoColor=white\" alt=\"PyPI\"\u003e\u003c/a\u003e\n  \u003ca href=\"https://github.com/BioAgentFlow/BioAgentFlow/blob/main/LICENSE\"\u003e\u003cimg src=\"https://img.shields.io/badge/License-Apache%202.0-green.svg\" alt=\"License\"\u003e\u003c/a\u003e\n  \u003ca href=\"https://pypi.org/project/bioagentflow/\"\u003e\u003cimg src=\"https://img.shields.io/pypi/pyversions/bioagentflow?logo=python\u0026logoColor=white\" alt=\"Python\"\u003e\u003c/a\u003e\n\u003c/p\u003e\n\n---\n\n## Overview\n\nBioAgentFlow is a lightweight, open-source platform for AI-powered drug virtual screening. It connects the full pipeline from protein target discovery through molecular docking to ADMET filtering, orchestrated by modular agents that automate each stage of the workflow. The platform is designed to run on standard hardware -- no GPU is required for basic screening tasks -- making computational drug discovery accessible to a much wider audience.\n\nBioAgentFlow is built for pharmaceutical researchers, computational biologists, and academic labs who need a practical, reproducible screening workflow without the overhead of large-scale infrastructure. Whether you are exploring a new protein target or triaging a compound library, BioAgentFlow provides a clear, auditable path from hypothesis to ranked hit list.\n\nWhat sets BioAgentFlow apart is its modular, plugin-based architecture. Each pipeline stage (target discovery, structure prediction, docking, scoring) is an independent agent that can be swapped, extended, or run in isolation. Combined with a clean CLI, a Python API, and an interactive Streamlit web UI, BioAgentFlow adapts to workflows ranging from quick single-target screens to large batch campaigns.\n\n\u003e BioAgentFlow 是一个轻量级、开源的 AI 药物虚拟筛选平台，涵盖从靶点发现到分子对接再到 ADMET 筛选的完整流程。无需 GPU 即可在普通硬件上运行，适合药物研发人员、计算生物学家和学术实验室使用。\n\n---\n\n## Key Features\n\n- **Protein Target Discovery** -- search and retrieve targets from UniProt, PDB, and AlphaFold DB (蛋白靶点发现)\n- **Protein Structure Prediction** -- predict 3D structures via ESMFold and IgFold when experimental structures are unavailable (蛋白结构预测)\n- **Molecular Docking** -- run blind docking with DiffDock; supports classical docking backends as plugins (分子对接)\n- **ADMET Property Prediction** -- evaluate drug-likeness with Lipinski and Veber rules, plus customizable filters (ADMET 属性预测)\n- **Compound Sourcing** -- query ChEMBL and ZINC15 for candidate ligands (化合物sourcing)\n- **Interactive Web UI** -- explore results, visualize structures, and adjust filters in a Streamlit dashboard (交互式 Web 界面)\n- **CLI for Batch Processing** -- script and automate large screening campaigns from the command line (命令行批处理)\n- **Docker Support** -- reproducible, one-command deployment with Docker and Docker Compose (Docker 支持)\n- **Extensible Plugin Architecture** -- add new data sources, docking engines, or scoring functions without modifying core code (可扩展插件架构)\n\n---\n\n## Architecture\n\n```\n┌─────────────────────────────────────────────────────────────────────┐\n│                        BioAgentFlow Pipeline                        │\n│                                                                     │\n│  ┌──────────┐   ┌──────────┐   ┌──────────┐   ┌──────────┐        │\n│  │  Target   │──▶│ Structure│──▶│  Ligand  │──▶│ Docking  │        │\n│  │ Discovery │   │Prediction│   │Preparation│   │ Engine   │        │\n│  └──────────┘   └──────────┘   └──────────┘   └──────────┘        │\n│       │              │              │               │               │\n│       ▼              ▼              ▼               ▼               │\n│   UniProt/PDB    ESMFold/       ChEMBL/        DiffDock /          │\n│   AlphaFold DB   IgFold        ZINC15         Plugin Backends      │\n│                                                     │               │\n│                                                     ▼               │\n│                                              ┌──────────┐          │\n│                                              │ Scoring \u0026 │          │\n│                                              │ Filtering │          │\n│                                              └──────────┘          │\n│                                                     │               │\n│                                                     ▼               │\n│                                              ┌──────────┐          │\n│                                              │  Report   │          │\n│                                              │Generation │          │\n│                                              └──────────┘          │\n│                                                                     │\n│  Interfaces:  CLI  |  Python API  |  Streamlit UI                   │\n└─────────────────────────────────────────────────────────────────────┘\n```\n\n**Pipeline summary (管线概览):**\n\n```\nTarget Discovery → Structure Prediction → Ligand Preparation → Docking → Scoring → Filtering → Report\n```\n\n---\n\n## Quick Start\n\n### Install from PyPI\n\n```bash\npip install bioagentflow\n```\n\n### Run with Docker\n\n```bash\ndocker pull bioagentflow/bioagentflow:latest\ndocker run -p 8501:8501 bioagentflow/bioagentflow:latest\n```\n\n### Install from Source\n\n```bash\ngit clone https://github.com/BioAgentFlow/BioAgentFlow.git\ncd BioAgentFlow\npython -m venv .venv \u0026\u0026 source .venv/bin/activate\npip install -e \".[dev]\"\n```\n\n\u003e For a detailed walkthrough, see the [Quickstart Guide](docs/quickstart.md) (快速入门指南).\n\n---\n\n## Usage Examples\n\n### CLI\n\nRun a full screening pipeline against a UniProt target:\n\n```bash\n# Single target screen\nbioagentflow screen --target P12345 --library zinc15 --out results/\n\n# Batch mode with a target list\nbioagentflow screen --targets targets.csv --library chembl --out results/ --workers 4\n\n# Filter existing results by ADMET rules\nbioagentflow filter --input results/docking_scores.csv --rules lipinski veber --out results/filtered.csv\n```\n\n### Python API\n\n```python\nfrom bioagentflow import Pipeline\n\npipeline = Pipeline.from_config(\"config.yaml\")\n\n# Run end-to-end\nreport = pipeline.run(target=\"P12345\", library=\"zinc15\")\n\n# Inspect top hits\nfor hit in report.top_hits(n=10):\n    print(f\"{hit.compound_id}  score={hit.score:.3f}  MW={hit.mw:.1f}\")\n```\n\n### Streamlit UI\n\n```bash\nbioagentflow ui\n# opens http://localhost:8501\n```\n\n\u003c!-- Screenshot placeholder --\u003e\n\u003cp align=\"center\"\u003e\n  \u003cem\u003e[UI screenshot placeholder -- docs/assets/ui_screenshot.png]\u003c/em\u003e\n\u003c/p\u003e\n\n---\n\n## Data Sources\n\n| Source | Description | URL |\n|--------|-------------|-----|\n| **UniProt** | Comprehensive protein sequence and function database (蛋白序列与功能数据库) | https://www.uniprot.org |\n| **PDB** | Worldwide repository of experimentally determined 3D structures (实验测定的三维结构库) | https://www.rcsb.org |\n| **AlphaFold DB** | AI-predicted protein structures from DeepMind (AI 预测蛋白结构) | https://alphafold.ebi.ac.uk |\n| **ChEMBL** | Curated bioactivity database of drug-like molecules (药物活性数据库) | https://www.ebi.ac.uk/chembl |\n| **ZINC15** | Free library of commercially available compounds for virtual screening (可商购化合物库) | https://zinc15.docking.org |\n\n---\n\n## Tech Stack\n\n| Component | Role |\n|-----------|------|\n| **ESMFold** | Single-sequence protein structure prediction |\n| **IgFold** | Fast antibody structure prediction |\n| **DiffDock** | Diffusion-based molecular docking |\n| **RDKit** | Cheminformatics, molecular descriptors, ADMET filters |\n| **BioPython** | Sequence retrieval, PDB parsing |\n| **Streamlit** | Interactive web dashboard |\n| **Pydantic** | Configuration and data validation |\n| **Click** | CLI framework |\n\n---\n\n## Project Structure\n\n```\nBioAgentFlow/\n├── bioagentflow/\n│   ├── __init__.py\n│   ├── cli.py                  # CLI entry points\n│   ├── config.py               # Pydantic configuration models\n│   ├── pipeline.py             # Pipeline orchestrator\n│   ├── agents/\n│   │   ├── target_discovery.py # UniProt / PDB / AlphaFold agents\n│   │   ├── structure.py        # ESMFold / IgFold agents\n│   │   ├── ligand_prep.py      # ChEMBL / ZINC15 sourcing\n│   │   ├── docking.py          # DiffDock and docking plugins\n│   │   ├── scoring.py          # Scoring and ranking\n│   │   └── filtering.py        # ADMET filters (Lipinski, Veber)\n│   ├── plugins/                # Third-party plugin directory\n│   ├── ui/\n│   │   └── app.py              # Streamlit application\n│   └── utils/\n│       ├── io.py\n│       ├── logging.py\n│       └── molecular.py\n├── tests/\n├── docs/\n│   ├── architecture.md\n│   ├── quickstart.md\n│   └── assets/\n├── docker/\n│   ├── Dockerfile\n│   └── docker-compose.yml\n├── pyproject.toml\n├── LICENSE\n├── README.md\n└── CONTRIBUTING.md\n```\n\n---\n\n## Contributing\n\nWe welcome contributions of all kinds -- bug reports, feature requests, documentation improvements, and code. Please read our [Contributing Guide](CONTRIBUTING.md) before opening a pull request.\n\n---\n\n## License\n\nBioAgentFlow is released under the [Apache License 2.0](LICENSE).\n\n---\n\n## Acknowledgments\n\nBioAgentFlow builds on the work of many outstanding open-source projects and public databases:\n\n- **ESMFold** (Meta AI) -- protein structure prediction\n- **IgFold** (Johns Hopkins) -- antibody structure prediction\n- **DiffDock** (MIT CSAIL) -- diffusion-based molecular docking\n- **RDKit** -- open-source cheminformatics\n- **BioPython** -- computational molecular biology toolkit\n- **UniProt, PDB, AlphaFold DB, ChEMBL, ZINC15** -- public biological and chemical databases\n\nWe are grateful to the maintainers and contributors of these projects.\n\n---\n\n## Disclaimer\n\n\u003e **BioAgentFlow is a research tool and is NOT intended for clinical decision-making.** All virtual screening results are computational predictions and must be validated by qualified domain experts through appropriate experimental methods before any downstream use. The authors assume no liability for decisions made based on the output of this software.\n\u003e\n\u003e **免责声明：** BioAgentFlow 是科研工具，不可用于临床决策。所有虚拟筛选结果均为计算预测，使用前须经专业人员通过实验验证。\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FS2026NG%2FBioAgentFlow","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FS2026NG%2FBioAgentFlow","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FS2026NG%2FBioAgentFlow/lists"}