Current capabilities and limitations

The teaching notebooks are complete workflow guides. Their reusable helpers import in the installed ipha_clean environment; the old import-order blocker list no longer describes this checkout. Completion of a tutorial does not validate a force field, a production trajectory or a binding prediction.

Area	Current implementation	Prerequisites / limits
Construction and export	01–04: curated monomers, side-chain/custom design, R checks, SDF/PDB export	RDKit and the installed package; inspect geometry before parameterisation
GAFF2	05A: MOL2/FRCMOD/PRMTOP/INPCRD and command/timing logs	AmberTools; default ABCG2 requires version ≥23; benchmark used AM1-BCC
Quick dry test	05A_quick_check_openmm: load, minimise and short dynamics	05A files + OpenMM; no solution-physics interpretation
Polymer in water	06A: ParmEd conversion to GROMACS, CHARMM-style TIP3P/SOD/CLA	05A inputs, ParmEd, GROMACS; explicit run/write flags
Enzyme–PHA reference	05B → 06B: CGenFF checks and CHARMM-GUI Solution Builder preparation	Website steps manual; packaged ANC55/P3HB4 inputs run by default; reviewed docking and protonation for new systems
Optional enzyme–PHA	05A → 06C: ff19SB/GAFF2/OPC, tleap + staged OpenMM	Docked complex and same-oligomer 05A files; matched conditions and differences; GAFF2/OPC is not validated for PHA
HPC	07: reviewed SLURM execution, restart and benchmarking; YAML helper interface	Prepared run folder; cluster environment/directives must be edited
Polymer benchmark	10: six-system GAFF2/GROMACS batch workflow	External engines and storage; its execution cell runs preparation and may submit with sbatch; separate from a single enzyme simulation
Docking inputs	11: whole-polymer GRO-to-PDB export and manual HADDOCK records	Benchmark frame/topology or other reviewed docking source; no automated submission
Teaching analysis	08/09: GROMACS preprocessing and Rg/end-to-end/SASA; 12: enzyme stability	Existing matching trajectories/topologies; machine-specific paths and selections must be replaced
Research contacts	Separate contact notebook/CLI under src/md_simulation_scripts/enzyme_contacts/	Matching periodic trajectory/topology; sampled proximity does not establish affinity or catalysis
Database construction	PHAPolymerBuilder and prepin/trimer route	Open Babel/AmberTools and manually supplied head/main/tail definitions; distinct from RDKit tutorials
Packing, GUI, script builders	Additional advanced interfaces	Not a unified validated production pipeline

Validation and remaining limits

The maintained tests/ suite is the test entry point. Bare repository-wide pytest also discovers a legacy src/dan_example_scripts/test_openmm_script_builder.py whose constructor call uses an obsolete root_dir argument; it is outside these teaching routes. Research datasets are not distributed as tutorial trajectories. A notebook’s saved outputs do not validate new inputs.

Schema/syntax checks and short execution can establish runnable preparation and restart behaviour. They do not establish full equilibration, convergence or PHA-specific accuracy. 06C retains deliberate force-field and engine differences from 06B. GROMACS and OpenMM output formats also differ; GROMACS analysis cells cannot consume DCD/CSV by changing filenames.

Teaching and research material

Use the notebook catalogue for names, prerequisites and run order. The contact validation record records a particular research preview and its historical environment; it is separate from the current teaching validation. Notebook docking tables and catalytic-residue notes are project context, not validated package predictions.