pyrosetta-scoring
/pyrosetta-scoring
Ocenia struktury białek funkcją punktową PyRosetta ref2015 i liczy energię interfejsu, dG, dSASA oraz komplementarność.
- Autor
- ekko8600 ↗
- Repozytorium
- enzyme_design_harness ↗
- Licencja
- Apache-2.0
- Ostatnia zmiana repozytorium
- 08.11.2026
Ocena i bezpieczeństwo
- Pliki w paczce
- 6
- Skrypty
- 2
- Komentarze HTML
- 0
- Adresy URL w SKILL.md
- 0
Brak sygnałów. Repozytorium zmienione 08.11.2026, bez odwołań do wycofanych modeli i API.
Apache-2.0, otwarta: można używać i zmieniać, z zachowaniem licencji i informacji o zmianach.
- 2 skrypty w paczce. Przejrzyj je przed instalacją, bo Claude może je uruchomić.
Skrypty: scripts/decompose_structures.py, scripts/pyrosetta_tools.py
Co robi i kiedy
Ocenia struktury białek funkcją punktową PyRosetta ref2015 i liczy energię interfejsu, dG, dSASA oraz komplementarność.
Oryginalny opis (EN)
Score protein structures with PyRosetta score functions and compute energy or interface metrics. Use this skill for ref2015 scoring, residue/complex assessment, score tables, and PyRosetta interface dG/dSASA/shape complementarity.
Jak działa
- Commands
- Decision Rules
- PyRosetta Performance Policy
Podgląd SKILL.md
---
name: pyrosetta-scoring
description: 'Score protein structures with PyRosetta score functions and compute
energy or interface metrics. Use this skill for ref2015 scoring, residue/complex
assessment, score tables, and PyRosetta interface dG/dSASA/shape complementarity.
'
metadata:
category: stability
tags: pyrosetta, scoring, ref2015, interface, protein-qc
legacy_category: evaluation
taxonomy_version: enzyme-agent/taxonomy/v1
capability_ids:
- science.general_operation
---
# PyRosetta Scoring
Use this skill for Rosetta score functions and interface metrics.
PyRosetta/Rosetta scoring and packing are CPU-bound in normal PyRosetta builds;
do not expect GPU acceleration for saturation scanning. For large scans, speed
comes from process-level parallelism, local repacking, resumable batches, and
limiting full relax to the best or most ambiguous candidates.
## Commands
Score a structure:
```bash
enzyme-core run pyrosetta-scoring --work-dir ./runs/pyrosetta-scoring --operation pyrosetta_tools -- \
score \
--pdb input.pdb \
--scorefxn ref2015 \
--out out/pyrosetta-scoring/scores.tsv
```
Compute interface metrics:
```bash
enzyme-core run pyrosetta-scoring --work-dir ./runs/pyrosetta-scoring --operation pyrosetta_tools -- \
interface \
--pdb complex.pdb \
--partners A_B \
--out out/pyrosetta-scoring/interface.tsv
```
Fast mutation ddG scan with local repacking:
```bash
enzyme-core run pyrosetta-scoring --work-dir ./runs/pyrosetta-scoring --operation pyrosetta_tools -- \
mutation-ddg-scan \
--pdb input.pdb \
--positions 17,61,246 \
--workers 8 \
--pack-radius 8 \
--skip-wt \
--resume \
--out out/pyrosetta-design/mutation_ddg.tsv
```
When the PDB uses a different residue numbering scheme from the target
sequence, pass a Host-provided JSON map such as `{"1": 24, "2": 25}`
with `--numbering-map`. The output then includes `full_sequence_position`
and `full_mutation`; do not infer this mapping from the output pose index.
For a full single-site saturation scan, start with local repack only. Add
`--relax --relax-rounds 25` only for a smaller follow-up set because FastRelax
dominates runtime.
For matched structural interpretation of a small known-variant panel, use the
shared structure-producing companion command:
```bash
enzyme-core run pyrosetta-scoring --work-dir ./runs/pyrosetta-scoring --operation pyrosetta_tools -- \
mutation-relax-panel \
--pdb input.pdb \
--mutation-csv mutation_panel.csv \
--scorefxn ref2015 \
--pack-radius 8 \
--relax-rounds 5 \
--options "-constant_seed -jran 111" \
--out-dir out/pyrosetta-relaxed-panel
```
This emits position-matched WT controls, variant PDBs, a `relaxed_panel.tsv`
and a manifest. The output is a comparative structural proxy, not a measured
stability or activity result.
To inspect which weighted score terms drive the relaxed comparison:
```bash
enzyme-core run pyrosetta-scoring --work-dir ./runs/pyrosetta-scoring --operation pyrosetta_tools -- \
score-terms-panel \
--panel-dir out/pyrosetta-relaxed-panel
```
Term decomposition remains score-function interpretation; it is not a
thermodynamic free-energy decomposition.
For detailed local structural interpretation, pass the matched panel through
the Host postprocessor after the run completes. The postprocessor emits the
existing geometry/context figures plus local coordinate displacement, contact
network turnover, heavy-atom donor/acceptor geometry, and catalytic/Ca spatial
relation figures. The structured JSON keeps per-seed and per-residue details.
These outputs remain computational proxies; a nonconverged panel is reported
as such and cannot be treated as mechanism validation.
```bash
python -m research_plane.cli physchem-relaxed-geometry \
--result-root result \
--project-id <project-id> \
--source-run-id <run-id> \
--panel-dir seed_111=<panel-dir-1> \
--panel-dir seed_222=<panel-dir-2> \
--panel-dir seed_333=<panel-dir-3> \
--site-spec-json '<target-site-specification>'
```
## Decision Rules
- Use `ref2015` unless the workflow requires a specific score function.
- Interface metrics require a correct partners string such as `A_B`.
- Treat Rosetta scores as comparative model scores, not physical free energies.
- Use `mutation-ddg-scan --workers N` rather than one long serial scan.
- Keep `--pack-radius` near 6-10 Angstrom for screening; broader repacking or
FastRelax should be reserved for shortlisted mutations.
## PyRosetta Performance Policy
Standard PyRosetta/Rosetta score functions, rotamer packing, and FastRelax are
CPU-bound and do not use GPU acceleration. For saturation mutagenesis, use the
shared `mutation-ddg-scan` command with multiple CPU workers, local repacking,
`--skip-wt`, and `--resume`; reserve FastRelax for shortlisted or conflicting
mutations.
Pliki w paczce (6)
- SKILL.md 5 KB
- execution_profile.json 1 KB
- scientific_contract.yaml 1 KB
- agents/openai.yaml 1 KB
- scripts/decompose_structures.py 4 KB
- scripts/pyrosetta_tools.py 1 KB
Wersje
W bazie nie ma innych skilli o nazwie pyrosetta-scoring.
Migawka bazy z 13.11.2026 · wersja paczki 4e3c6d0e0a6f (skrót SHA-256 całej paczki)
Pobrania w ostatnich 30 dniach: 0 · ranking
Zgłoś problem z tym skillem: szkodliwa treść, błędny opis albo prośba autora o usunięcie.
Podobne skille
- 100pyrosetta-relax-refine
Udoskonala struktury białek w PyRosetta przez FastRelax, minimalizację i generowanie decoys przed oceną. - 100pyrosetta-pose
Przygotowuje struktury białek w PyRosetta przez wczytywanie PDB, sprawdzanie reszt, współrzędnych i RMSD. - 100protein-qc
Podaje metryki kontroli jakości i progi filtrowania dla projektowania białek: pLDDT, ipTM, PAE, ryzyka sekwencji, metryki interfejsu PyRosetta i właściwości biofizyczne. - 100protein-qc
Dostarcza metryki kontroli jakości i progi filtrowania dla projektów białek, takie jak pLDDT, ipTM i PAE, oraz pułapki sekwencji. - 87bindcraft
Projektuje białka wiążące od zera, łącząc AF2, ProteinMPNN i filtry PyRosetta dla zadanego celu w PDB. - 87invrotzyme
Składa zestawy miejsc aktywnych enzymu wokół ligandu z pliku ograniczeń Rosetta w PyRosetta, jako zarodki do projektowania enzymów.