CoolFace
Apppublic

Tsimech2000/virulent_molecular_structures

sourceHugging Facemitupdated 2mo agoView on Hugging Face
0likes
App README

Molecular Interaction Research Workbench

A transparent research workbench for:

  • retrieving and inspecting experimental structures from RCSB PDB;
  • comparing homologous protein structures with sequence-aware Cα alignment; and
  • running reproducible protein–ligand docking with AutoDock Vina.

Scientific scope

This application is a hypothesis-generation tool. A predicted pose or favorable Vina score does not prove binding, establish biological relevance, or equal an experimental binding free energy. Appropriate positive and negative controls, orthogonal computational analysis, and experimental validation are required.

The protein comparison workflow is not protein–protein docking. It aligns homologous residues and reports a Cα RMSD for structural comparison.

Reproducibility

Every completed docking run provides a ZIP bundle containing:

  • unmodified uploaded inputs and their SHA-256 hashes;
  • prepared PDBQT structures;
  • all generated poses;
  • a score table;
  • search-box and Vina parameters;
  • software-version metadata; and
  • an interpretation warning.

Runs are processed in unique temporary directories and are not intentionally retained by the application.

Required preparation decisions

Before using results for research, inspect and document:

  1. 1.biological assembly and chain selection;
  2. 2.missing residues and atoms;
  3. 3.protonation states and experimental pH;
  4. 4.cofactors, metals, buffer molecules, and structural waters;
  5. 5.ligand bond orders, stereochemistry, tautomer and charge state; and
  6. 6.the biological justification for the docking search box.

Open Babel preparation is provided for convenience and does not replace a project-specific preparation protocol.

Local use

bash
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
streamlit run app.py

Methods

  • RCSB PDB Search/Data APIs for structure discovery and retrieval.
  • Biopython for PDB parsing and sequence-aware structural alignment.
  • Open Babel for PDBQT conversion and hydrogen/charge assignment.
  • AutoDock Vina 1.2.x for ligand docking.
  • py3Dmol for interactive visualization.

Citation

If results contribute to research, cite RCSB PDB, AutoDock Vina, Open Babel, Biopython, and the original experimental structure publications as appropriate.