---

name: prolif description: "Use ProLIF to prepare molecular inputs, define protein-ligand interactions, compute interaction fingerprints, and create visualization outputs for chemistry and molecular-simulation workflows." disable-model-invocation: true metadata: disco-role: operating license: Apache 2.0

ProLIF

Use this repo skill when a user asks for ProLIF, protein-ligand interaction fingerprints, molecular interaction fingerprints from docking poses or MDAnalysis trajectories, residue-level interaction metadata, or ProLIF plots. ProLIF is a Python API package, not a command-line application.

Start Here

1. Confirm the package imports and optional dependencies with scripts/prolif_environment_check.py when the runtime environment is uncertain.
2. Route the task to the focused sub-skill below before writing workflow code.
3. Keep ProLIF input preparation, interaction selection, fingerprint execution, and plotting as separate decisions; most failures come from mixing those stages.
4. Use bundled references and scripts from this skill tree only; do not rely on the original ProLIF repository checkout.

Route By Task

User goal	Read
Convert RDKit, MDAnalysis, SDF, MOL2, PDBQT, PDB, or CIF-backed structures into ProLIF objects	sub-skills/molecules-and-io/SKILL.md
Choose interaction classes, tune parameters, use implicit hydrogens, water bridges, counts, or direct residue checks	sub-skills/interactions/SKILL.md
Run Fingerprint.run, run_from_iterable, or generate; export DataFrames, bitvectors, countvectors, and pickles	sub-skills/fingerprints/SKILL.md
Create 2D ligand networks, barcode plots, 3D complex views, or residue drawings from completed fingerprints	sub-skills/visualization/SKILL.md

Minimal Environment Check

python scripts/prolif_environment_check.py
python scripts/prolif_environment_check.py --include-plotting

Expected output is JSON with prolif_version, available interactions, and optional plotting dependency statuses. Install the base package for fingerprinting and prolif[tutorials] or prolif[plots] when workflows require RDKit-based tutorials or plotting extras.

Common Workflow Order

1. Prepare molecules: build prolif.Molecule objects or MDAnalysis selections with residue labels that will remain stable.
2. Choose interactions: decide defaults versus named interaction classes, parameters, count=True, implicit_hydrogens=True, WaterBridge, and custom ignore predicates.
3. Run fingerprints: choose trajectory, ligand-iterable, or single-pair execution and validate fp.ifp before exporting.
4. Export or plot: convert results to pandas/RDKit vectors or pass completed fingerprints plus matching molecules to plotting helpers.

Root References

• references/package-overview.md summarizes ProLIF concepts, dependencies, workflow stages, and no-CLI guidance.
• references/troubleshooting.md covers cross-cutting install/import, optional dependency, data, and API misuse failures.
• references/repo-provenance.md records the source snapshot and evidence used to build this skill.
• references/repo-routing-metadata.json is structured router metadata for managed imports.

Cross-Cutting Warnings

• Do not call export or plotting methods before a fingerprint run has populated fp.ifp.
• Do not assume residue strings from one conversion path match another; check chain versus segid and route label questions to molecules-and-io.
• Do not include WaterBridge without providing a water selection or water molecule through interaction parameters.
• Do not debug plotting by recomputing interactions first; validate the completed fingerprint and molecule objects, then route to visualization.