Pork- and pea-derived peptides in an enteral nutrition formula as candidate umami/kokumi and metabolic-functional ligands: a multi-scale in silico prioritization

The complete screening dataset — all 347 identified peptides with their sequences, physicochemical descriptors, ToxinPred toxicity predictions, AllerTOP allergenicity predictions, PeptideRanker bioactivity scores, and AutoDock Vina binding energies against CaSR and TAS1R1/TAS1R3, organised by biological source — is provided as a Supplementary data file.All input files required to reproduce the computational analyses are openly available here. The repository contains: (i) the energy-minimised starting structures and topologies for all nine simulated systems (the five candidate peptides, the FLFK and L-Trp controls, and the peptide-free 7DTV and 7DTW references), together with the GROMACS parameter files used for energy minimisation, NVT and NPT equilibration and production dynamics; (ii) the complete set of GFN2-xTB cluster models, comprising, for each system, replica and sampled time point (70, 80 and 90 ns), the extracted complex together with the isolated peptide, the 6 Å binding pocket and the pocket–peptide fragments for both symmetry-equivalent binding sites; The receptor structures used in this work are publicly available from the Protein Data Bank under accession codes 7DTV, 7DTW and 5X2N. The full molecular dynamics trajectories are available from the corresponding author upon reasonable request owing to their size.