OdorSig: Automated Molecular Docking Pipeline

OdorSig is a fully automated, open-source Python pipeline for reproducible olfactory receptor-odorant molecular docking analysis. The system integrates AutoDock Vina, Open Babel, PyMOL, and Biopython behind a Streamlit interface to deliver an end-to-end workflow requiring only receptor and ligand names as input. Applied to 40 human olfactory receptors and 12 structurally diverse odorants, OdorSig generated a large-scale 480-pair binding affinity dataset, in which 80.6% of pairs demonstrated high docking reproducibility (sigma <= 0.2 kcal/mol) and 36 high-confidence strong binders (delta-G <= -7.0 kcal/mol) were identified. Pipeline stages: 1. Receptor sequence retrieval (NCBI) and ligand retrieval (PubChem) 2. Homology model generation and structural validation (SWISS-MODEL, Ramachandran-based quality filtering) 3. Ligand preparation and format conversion (Open Babel) 4. Docking configuration (grid box, exhaustiveness, seed mode) 5. Triplicate AutoDock Vina docking execution 6. Reproducibility-aware scoring: mean binding energy and standard deviation computed automatically for every receptor-odorant pair 7. Results parsing, PyMOL pose visualization, and CSV/figure export By reporting mean binding energy and standard deviation as first-class outputs across triplicate runs, OdorSig turns stochastic single-run docking scores into consistent, comparable receptor-odorant interaction profiles. Source code: https://github.com/ODOR-SIG/Automated-Molecular-Docking Code archive DOI (concept, latest release): 10.5281/zenodo.17814494 Dataset archive DOI (480-pair dataset, raw docking logs): 10.5281/zenodo.20373888