Ligand self-energy uncouples apparent binding affinity from site-specific engagement at RAGE and HIF-1α

Dataset for: Ligand self-energy uncouples apparent binding affinity from site-specific engagement at RAGE and HIF-1α Overview This repository contains the computational dataset, raw calculation outputs, parameter files, and analytical scripts supporting the manuscript "Ligand self-energy uncouples apparent binding affinity from site-specific engagement at RAGE and HIF-1α". The study challenges the reliance on aggregate scoring metrics in virtual screening by resolving the interaction energy of Oxalis corniculata phytochemicals at two structurally unrelated targets: the receptor for advanced glycation end-products (RAGE, PDB 6XQ8) and hypoxia-inducible factor 1-alpha (HIF-1α, PDB 4H6J). The dataset provides all the computational evidence demonstrating how extreme electrostatic polarization and ligand self-energy can artificially inflate whole-complex binding scores (MM-GBSA) without producing genuine site-specific engagement. Dataset Structure The repository is organized into 6 main directories corresponding strictly to the multiscale in silico workflow: 01_BIOLOGICAL_ACTIVITY/: Raw prediction spectra from PASS Online and PASS Gero for 163 Oxalis corniculata constituents, including the probability-of-activity (Pa) and probability-of-inactivity (Pi) matrices used for initial library prioritization. 02_ADMET/: Raw output datasets from the five consensus platforms evaluated (SwissADME, ADMETlab 2.0, admetSAR 2.0, ADMET-AI, ProTox-3.0), detailing physicochemical descriptors, blood-brain-barrier permeability consensus, and mutagenicity/toxicity flags. 03_MOLECULAR_DOCKING/: Pre-processed receptor structures (.pdb) for RAGE VC1 and HIF-1α PAS-B, ligand 3D structures, and raw docking output files for both protocols evaluated: CB-Dock2 (blind docking) and SwissDock AC (site-directed). 04_DYNAMIC_MOLECULAR/: Contains GROMACS (v2023.4) simulation parameters (.mdp, .top, .gro) for the 300-ns trajectories, structural stability metrics, and all end-point thermodynamic calculations (gmx_MMPBSA v1.6.0). As detailed in the sub-folder structure, this includes: 01_RMSD_BACKBONE / 02_RMSD_LIGAND / 03_RADIUS_OF_GYRATION / 04_SASA / 05_HYDROGEN_BOND / 06_RMSF (Raw time-series data for structural descriptors). 07_MMPBSA_HIF1_RAGE (Consensus-stage aggregate binding free energy calculations). 08_DECOMPOSITION / 09_PER_RESIDUE (Frame-by-frame per-residue free energy decomposition data showing the uncoupling effect). 10_CODE (Custom analytical scripts used for data extraction and statistical block-averaging). (Note: Raw, uncompressed 300-ns .xtc/.trr trajectory files are excluded due to Zenodo file size limitations but are available from the corresponding authors upon reasonable request). 05_DFT/: ORCA 6.1.1 output files (.out) containing the geometry optimizations and single-point evaluations at the B3LYP-D3BJ/def2-TZVP//SMD(water) level. Also includes Multiwfn post-processing outputs detailing dipole moments (e.g., the 201 D anomaly), condensed Fukui functions, and global reactivity descriptors. 06_NETWORK_PHARMACOLOGY/: Target prediction datasets (SwissTargetPrediction, ProfhEX, COMET), Cytoscape 3.10.3 session files (.cys), and annotated compound-target edge lists demonstrating the exploratory polypharmacology network. Software and Reproducibility Context The data provided herein can be processed or visualized using standard computational biophysics and cheminformatics software: Molecular Dynamics: GROMACS 2023.4 (CHARMM36m / CGenFF) Free Energy Calculations: gmx_MMPBSA 1.6.0 (igb=5 model) Quantum Mechanics: ORCA 6.1.1, Multiwfn Network Analysis: Cytoscape 3.10.3 (AutoAnnotate 1.5.2) Usage Notes Researchers are free to download, analyze, and repurpose this dataset under the terms of the applied Open Access license (e.g., CC-BY 4.0). If you use this data or the structural hypotheses derived from it, please cite the original manuscript linked to this repository via its DOI: 10.5281/zenodo.21983259.