Molecular Docking AutoDock

实现分子对接全流程自动化，支持输入蛋白结构文件（PDB格式）、小分子SMILES表达式、口袋位置描述（文本描述或坐标范围），输出对接后打分最高的复合物结构及对接打分结果，默认使用AutoDock VINA算法。使用场景：用户需要进行蛋白-小分子对接预测结合模式、筛选小分子配体、获取对接复合物结构时触发。

Security Scan

The skill's code and instructions match the stated molecular‑docking purpose, but the implementation runs many external commands with shell=True and unsanitized user inputs (file paths / pocket descriptions), creating a realistic command‑injection / misuse risk — review or sandbox before use.

Purpose & Capability

Name/description (AutoDock VINA docking pipeline) aligns with included files and commands: preprocessing (RDKit, Meeko, OpenBabel), pocket prediction (p2rank), and vina execution. Dependencies and references are coherent with the stated goal.

Instruction Scope

SKILL.md and the script instruct running external binaries (obabel, vina, p2rank) and a wget download. The runtime instructions and script access local files (protein PDB, output_dir) and spawn subprocesses; they do not request unrelated system secrets, but they allow user-supplied file paths and pocket descriptions to be passed into shell commands without sanitization, which can be abused.

Install Mechanism

No formal install spec in registry (instruction-only), but SKILL.md tells users to install conda packages and to wget a p2rank tarball from a GitHub release. Downloading from GitHub releases is reasonable, but it still performs a network download and adds third‑party tooling to the environment.

Credentials

No environment variables, credentials, or config paths are requested. The required tools and packages (rdkit, openbabel, meeko, vina, p2rank) are proportionate to molecular docking.

Persistence & Privilege

Skill is not forced-always, does not request elevated privileges or modify other skills. It runs as invoked and writes outputs to a user-specified output_dir only.

What to consider before installing

This skill appears to implement the advertised docking workflow, but the Python script builds and runs shell commands (obabel, prank, vina) by interpolating user-supplied paths and descriptions into f-strings with subprocess.run(..., shell=True). That pattern allows command injection if a malicious or malformed file path/pocket string is provided. Before installing or running: (1) review the full scripts yourself or with a developer; (2) run the skill in an isolated environment or container; (3) avoid supplying untrusted file paths or freeform text as the pocket argument; (4) consider patching the code to call external programs with argument lists (subprocess.run([...], shell=False)) or to sanitize/quote inputs (shlex.quote) and validate paths; (5) ensure you trust the binaries you install (conda packages and the p2rank tarball). If you cannot review or sandbox it, treat it as risky for use with sensitive systems or data.

分子对接自动化技能（AutoDock VINA实现）

Overview

本技能提供端到端的蛋白-小分子分子对接自动化流程，无需手动处理中间文件，支持文本描述的口袋位置自动识别，内置AutoDock VINA作为默认对接引擎，输出对接后TopN复合物结构、结合亲和力打分及结果可视化。

依赖安装

运行前确保已安装以下依赖：

conda install -c conda-forge autodock-vina pymol-open-source rdkit openbabel
pip install meeko biopython

若需要从文本描述预测口袋位置，需额外安装p2rank：

# 安装p2rank口袋预测工具
wget https://github.com/rdk/p2rank/releases/download/2.4/p2rank_2.4.tar.gz
tar -xzf p2rank_2.4.tar.gz && export PATH=$PATH:$(pwd)/p2rank_2.4

核心工作流程

步骤1：输入校验

• 校验输入蛋白文件是否为有效PDB格式
• 校验小分子SMILES是否合法可生成3D构象
• 解析口袋位置：支持两种输入形式
  1. 坐标范围：center_x=xxx center_y=xxx center_z=xxx size_x=xxx size_y=xxx size_z=xxx
  2. 文本描述：如"结合位点在蛋白的ATP结合口袋"、"靠近残基LYS123和ASP145的区域"，将自动调用p2rank预测口袋或根据残基编号计算坐标

步骤2：蛋白预处理

• 自动去除蛋白中的水分子、非标准残基
• 加氢、计算电荷
• 输出预处理后的蛋白PDBQT格式文件

步骤3：小分子预处理

• 从SMILES生成最优3D构象
• 加氢、计算Gasteiger电荷
• 输出小分子PDBQT格式文件

步骤4：运行AutoDock VINA对接

• 默认exhaustiveness=8，num_modes=9
• 支持自定义对接参数调整

步骤5：结果输出

• 输出打分最高的1个（可配置）复合物结构（PDB格式，蛋白+配体）
• 输出所有对接结果的结合亲和力打分表
• 可选生成对接结果可视化图片

脚本使用说明

scripts/molecular_docking.py

主执行脚本，调用方式：

python scripts/molecular_docking.py \
  --protein path/to/protein.pdb \
  --smiles "C1=CC=C(C=C1)C(=O)O" \
  --pocket "center_x=10.5 center_y=20.3 center_z=-5.2 size_x=20 size_y=20 size_z=20" \
  --output_dir ./docking_results

参数说明：

• --protein: 输入蛋白PDB文件路径
• --smiles: 小分子SMILES表达式
• --pocket: 口袋位置描述，支持坐标格式或文本描述
• --output_dir: 结果输出目录，默认./docking_results
• --num_modes: 输出对接构象数目，默认1
• --exhaustiveness: 对接搜索穷尽度，默认8

references/

• references/autodock_vina_manual.md: AutoDock VINA官方使用手册
• references/pocket_prediction_guide.md: 口袋位置描述规范及预测工具使用指南