Molecular Arms Race Classifier Training, Testing, Clustering, Docking, and Molecular Dynamics Simulation Dataset

This is an update to the MARC dataset including new files from additional analysis in the marc_mdsim_docking and marc_analysis_classification folders .Old:This dataset contains all files, scripts, directories, and computational outputs generated for the Molecular Arms Race Classifier (MARC) project, including training, testing, predicted structures, clustering, docking, and molecular dynamics (MD) simulation data. The dataset supports reproducibility of the MARC classifier workflow and enables visualization and re-analysis of docking and MD-based orthogonal validation outputs. The archive includes: • Training, testing, and prediction datasets used to develop and evaluate the MARC ion-channel specificity classifier.• qTMclust structural clustering outputs (CSV files, processed cluster directories, and raw US-align qTMclust text output).• Docking results generated using Schrödinger PIPER protein–protein docking workflows.• Molecular dynamics simulation directories generated using Schrödinger Maestro’s Desmond engine, including full trajectory folders and job files.• System preparation and membrane-building files, including POPC membrane embedding and SPC solvation stages.• Protein preparation outputs (Schrödinger Protein Preparation Wizard) for all proteins used in docking and MD simulations.• Supporting scripts used for organizing qTMclust outputs, generating cluster directories, and processing MARC classification data. This dataset accompanies the MARC manuscript and is provided to enable full transparency, re-analysis, and independent validation of the structural, biophysical, and machine learning components of the study. All files are organized to reflect the exact directory structure produced by the computational workflows.