Validation systems and reference files for the H5MD biophysics metadata module

Validation dataset of the H5MD biophysics module, a metadata schema that stores the topology, force field, MD parameters, selection groups and observables of a molecular dynamics simulation together with its trajectory in one H5MD (HDF5) file. It accompanies the paper A Hierarchical Metadata Schema for Self-Contained Molecular Dynamics Trajectory Archives. validation_systems.zip contains the GROMACS inputs of the 58 chemical systems used in the paper's technical validation, reduced to what grompp needs: topology, force field, coordinates or run input, index groups and MD parameters. The systems range from 711 to 1,066,628 particles, cover AMBER, CHARMM, GROMOS, OPLS-AA and MARTINI force fields, and come in six groups by origin: self_generated (8): built with GROMACS from PDB entries bena (10): sessions of the Bena dataset (10.5281/zenodo.17632513) jcp_benchmark (6): GROMACS benchmark systems (10.5281/zenodo.3893789) hexabc (8): hexABC nucleic-acid and protein–nucleic-acid deposits from MDposit memprotmd (6): MemProtMD membrane proteins, atomistic and coarse-grained misato (20): MISATO protein–ligand complexes (10.5281/zenodo.7711953) Run under 23 simulation protocols, these inputs give the 379 reference files of the validation. The files can be regenerated with the validation scripts (github.com/miemiemmmm/h5md-biophysics-validation) and the GROMACS implementation published with the paper. Each group's MANIFEST.md credits its source and states the changes made; provenance/ holds the scripts that selected and prepared the systems. Four of the reference files are included as examples, each a short trajectory with the full metadata. They also serve as test data for the validator in gitlab.com/gromacs/h5md-biophysics: hexabc_a001u2_rna_protein__nvt-vrescale.h5md bench_villin__tcoupl-multigroup.h5md bench_villin__sd-langevin.h5md bench_villin__selection-noncontiguous.h5md