FEP+ and metadynamics input files, custom force field parameters and analysis scripts for "NMR structure and free-energy simulations for rational design of myotonic dystrophy RNA inhibitors"

Input, parameter and analysis files needed to reproduce the binding pose metadynamics (BPMD) and free energy perturbation (FEP+) calculations reported in "NMR structure and free-energy simulations for rational design of myotonic dystrophy RNA inhibitors" (submitted to the Journal of Chemical Information and Modeling). This record replaces 10.5281/zenodo.20259706, whose custom force field parameter file was corrupted. Contents BPMD_input_files.tar: Desmond binding pose metadynamics inputs. The 20 prepared RNA-ligand complexes from the NMR ensemble of compound 4 (metadynamics_binding_2.maegz) and the job script (metadynamics_binding_2.sh; 10 independent trials per complex). fep_mapper_1_full_set_for_publication.fmp: FEP+ map containing the prepared RNA-ligand complex, all ligand structures, the perturbation graph and simulation settings (Schrödinger 2024-4, OPLS5). ffb_1_publication.tar: complete Force Field Builder job used to generate the custom ligand torsion parameters, including input structures, run script, Jaguar fragment inputs, the job database with QM and fitted torsion energies (ffb_db_v2.sql, SQLite), and the resulting custom parameter archive (custom_2024_4.opls). FFB_parameterisation_export.tar: plain-text export of the parameterisation. Schrödinger stores fitted OPLS parameter values in encrypted form, so this folder provides the complete fitting data instead: parameterised fragments, the ligand torsions each covers, QM torsion scans, force field profiles before and after fitting, fit RMSDs, and plots of all fitted profiles (see README.txt). fig4_fep_vs_spr.py and fig4_benchmark_18_compounds.csv: Python script and input data that reproduce the statistics and correlation plots in all four panels of Figure 4. fig4_benchmark_18_compounds.csv contains the FEP+ predictions for the 18-compound benchmark at 0.15 M NaCl and with the worst BPMD pose, keyed to the compound IDs in SI5_Assay_data.csv, which is provided with the manuscript as Supporting Information and contains the experimental SPR data and final FEP+ predictions. Place SI5_Assay_data.csv in the same folder and run: python fig4_fep_vs_spr.py (requires numpy, pandas and matplotlib). The .opls, .fmp, .maegz and .mae files require a Schrödinger installation to read. All CSV, text and SQLite files can be read with standard tools.