$npx -y skills add ma-compbio-lab/SkillFoundry --skill rdkit-scaffold-analysis-starterUse this skill to compute Murcko scaffold summaries for a small local molecule set with RDKit. Prefer it for deterministic scaffold grouping and smoke-scale cheminformatics checks.
| 1 | ## Purpose |
| 2 | Analyze a small local TSV of SMILES strings with RDKit Murcko scaffolds and emit a compact JSON summary that can feed later smoke integration or scaffold triage workflows. |
| 3 | |
| 4 | ## When to use |
| 5 | - You need a local scaffold grouping summary for a small curated molecule set. |
| 6 | - You want canonical SMILES, Murcko scaffolds, generic scaffolds, and group counts from one deterministic run. |
| 7 | |
| 8 | ## When not to use |
| 9 | - You need large library clustering, matched molecular pair analysis, or SAR interpretation. |
| 10 | - You need remote compound lookup or medicinal-chemistry recommendations. |
| 11 | |
| 12 | ## Inputs |
| 13 | - A TSV file with columns `name` and `smiles` |
| 14 | - Optional JSON output path |
| 15 | |
| 16 | ## Outputs |
| 17 | - JSON with per-molecule canonical SMILES, Murcko scaffold, generic scaffold, scaffold groups, generic scaffold groups, and summary counts |
| 18 | |
| 19 | ## Requirements |
| 20 | - `slurm/envs/chem-tools/bin/python` |
| 21 | - RDKit available in that environment |
| 22 | |
| 23 | ## Procedure |
| 24 | 1. Inspect `examples/molecules.tsv`. |
| 25 | 2. Run `slurm/envs/chem-tools/bin/python skills/drug-discovery-and-cheminformatics/rdkit-scaffold-analysis-starter/scripts/run_rdkit_scaffold_analysis.py --input skills/drug-discovery-and-cheminformatics/rdkit-scaffold-analysis-starter/examples/molecules.tsv`. |
| 26 | 3. Review `molecules`, `scaffold_groups`, and `summary`. |
| 27 | |
| 28 | ## Validation |
| 29 | - The bundled example returns at least one scaffold group with count `>= 2`. |
| 30 | - Invalid SMILES input returns a non-zero exit code with a clear error message. |
| 31 | |
| 32 | ## Failure modes and fixes |
| 33 | - Invalid SMILES: fix the offending row in the input TSV. |
| 34 | - Missing RDKit environment: rerun with `slurm/envs/chem-tools/bin/python`. |
| 35 | - Missing `name` or `smiles` columns: use a header row with exactly those field names. |
| 36 | |
| 37 | ## Safety and limits |
| 38 | - Local scaffold computation only. |
| 39 | - No medicinal-chemistry conclusions are implied by the grouping. |
| 40 | |
| 41 | ## Provenance |
| 42 | - RDKit docs: https://www.rdkit.org/docs/index.html |
| 43 | - RDKit Murcko scaffold API: https://www.rdkit.org/docs/source/rdkit.Chem.Scaffolds.MurckoScaffold.html |
| 44 | - RDKit repository: https://github.com/rdkit/rdkit |
| 45 | |
| 46 | ## Related skills |
| 47 | - `rdkit-molecular-descriptors` |
| 48 | - `rdkit-molecule-standardization` |