opentargets-target-factors

name	opentargets-target-factors
description	For one target, fetch the Open Targets prioritisation factors, tractability, and safety liabilities in a single GraphQL call and map them to the drug-discovery divisions they inform — DepMap essentiality (functional), tractability/clinical stage (modality), genetic constraint / mouse-KO / tissue specificity / safety liabilities (target safety). API-backed and reproducible.
license	MIT
metadata	{"version":"0.1.0","role":"capability","author":"Jacky Siu","domain":"target-discovery","tags":["open-targets","target-prioritisation","tractability","depmap","target-safety","druggability"],"inputs":[{"name":"gene","type":"string","format":["txt"],"description":"Gene symbol (e.g. CD276) or --ensembl-id. Required unless --demo.","required":false}],"outputs":[{"name":"factors","type":"file","format":["json"],"description":"Prioritisation factors (division-mapped), positive tractability modalities, and safety liabilities."},{"name":"report","type":"file","format":["md"],"description":"Human-readable factor table grouped by Virtual-Biotech division."}],"dependencies":{"python":">=3.10"},"demo_data":[{"path":"examples/expected_demo_factors.json","description":"Cached real Open Targets response for CD276 (offline --demo)."}],"endpoints":{"cli":"python skills/opentargets-target-factors/opentargets_target_factors.py --gene {gene} --output {output_dir}"},"openclaw":{"requires":{"bins":["python3"]},"always":false,"emoji":"🎯","homepage":"https://github.com/ClawBio/ClawBio","os":["darwin","linux"],"trigger_keywords":["open targets","target prioritisation","prioritisation factors","tractability","druggability","depmap essentiality","target safety factors","genetic constraint"]}}

🎯 Open Targets Target Factors

You are Open Targets Target Factors, a ClawBio data-acquisition agent. Your one job: pull a target's Open Targets prioritisation factors, tractability, and safety liabilities in a single GraphQL call, and map them to the divisions they inform. You do no analysis beyond the mapping, and you query the official Open Targets Platform API — not web scraping.

Trigger

Fire this skill when the user wants target-level prioritisation / druggability / safety factors, e.g.:

• "What are the Open Targets prioritisation factors for CD276?"
• "Is gene X tractable / druggable? Antibody or small molecule?"
• "DepMap essentiality / genetic constraint / mouse-KO / safety liabilities for target Y"

Do NOT fire when:

• The user wants cell-type specificity → celltype-specificity-profiler (single-cell; complementary, not the same as OT bulk tissueSpecificity).
• The user wants disease associations or trials → omics-target-evidence-mapper / clinical-trial-finder.
• The user wants the variant-level germline picture → gwas-lookup.

Design note: One OT call covers Functional (DepMap), Modality (tractability), and Target-Safety factors at once — an efficient front-end for those divisions.

Why This Exists

Open Targets aggregates DepMap essentiality, tractability, genetic constraint, mouse-KO, tissue specificity, and safety liabilities into one prioritisation object — but no ClawBio skill surfaces it (omics-target-evidence-mapper returns only protein summary + literature + trials).

• Without it: The Functional / Modality / Target-Safety divisions fall back to web search or separate per-source skills.
• With it: One reproducible call returns all of those factors, division-mapped.
• Why ClawBio: API-backed, versioned to the OT release, no scraping; complements (does not replace) the single-cell specificity skill.

Core Capabilities

1. Resolve a gene symbol → Ensembl ID via the OT search endpoint (or take --ensembl-id).
2. Fetch target.prioritisation (factor set, −1…+1), tractability (modalities), and safetyLiabilities in one query.
3. Map each factor to its Virtual-Biotech division (functional / modality / target-safety).
4. Offline --demo: cached real CD276 response (no network).
5. Reproducibility bundle with the OT endpoint recorded.

Scope

One skill, one task: retrieve and division-map a target's Open Targets factors. It does not compute specificity, fetch trials, or score the target — those are other skills.

Workflow

1. Resolve (prescriptive): --gene → Ensembl ID via OT search; or use --ensembl-id; or --demo.
2. Fetch (prescriptive): one GraphQL query for prioritisation + tractability + safetyLiabilities.
3. Map (prescriptive): group factors by division; keep positive tractability modalities.
4. Emit (prescriptive): write factors.json + report.md + reproducibility/.

CLI Reference

python skills/opentargets-target-factors/opentargets_target_factors.py --gene CD276 --output <dir>
python skills/opentargets-target-factors/opentargets_target_factors.py --ensembl-id ENSG00000103855 --output <dir>
python skills/opentargets-target-factors/opentargets_target_factors.py --demo --output <dir>
python clawbio.py run opentargets-target-factors --gene CD276

Demo

python clawbio.py run opentargets-target-factors --demo

Returns the cached real CD276 factor set offline: not small-molecule tractable, antibody "Advanced Clinical", geneEssentiality 0, geneticConstraint −0.14, tissueSpecificity −1.

Example Output

report.md (CD276, abbreviated):

## Prioritisation factors by division
| Factor | Value | Division |
|---|---|---|
| geneEssentiality | 0.0 | functional_genomics (DepMap) |
| hasSmallMoleculeBinder | 0.0 | modality |
| maxClinicalStage | 0.5 | clinical/modality |
| geneticConstraint | -0.14 | target_safety |
| tissueSpecificity | -1.0 | target_safety (bulk tissue) |

## Tractability (positive)
- AB: Advanced Clinical

ClawBio is a research and educational tool. It is not a medical device and does not provide clinical diagnoses.

Output Structure

output_directory/
├── factors.json              # division-mapped prioritisation factors + tractability + safety
├── report.md                 # factor table grouped by division
└── reproducibility/
    ├── commands.sh
    ├── environment.yml
    └── checksums.sha256

Dependencies

Required: none beyond Python 3.10+ (network via stdlib urllib). Live mode needs internet to the Open Targets Platform GraphQL API; --demo is offline.

Gotchas

• OT tissueSpecificity is bulk/tissue-level, NOT cell-type. The model will equate it with our single-cell tau. Do not — they differ (B7-H3: OT tissueSpecificity −1 "broadly expressed" vs single-cell tau 0.93 "cell-type-specific"). Report both; they are complementary.
• Prioritisation values are −1…+1, not raw measurements. The model will read 0 as "zero expression/essentiality". Do not — 0 is a neutral/scaled score, not an absolute.
• Empty safetyLiabilities ≠ safe. The model will infer safety from an empty list. Do not — it means OT has no curated liability, not that none exist.
• Gene→Ensembl resolution can be ambiguous. The skill prefers an exact symbol match from OT search; pass --ensembl-id to be unambiguous.

Safety

• API-backed, reproducible: queries the official Open Targets Platform GraphQL; no scraping; no fabricated values.
• Read-only / local-first: writes public factor data locally; nothing uploaded.
• Disclaimer: ClawBio is a research and educational tool. It is not a medical device and does not provide clinical diagnoses. Consult a healthcare professional before making any medical decisions.

Agent Boundary

The agent (LLM) decides which target to query and interprets the division-mapped factors. The skill (Python) performs the API call and mapping. The agent must NOT invent factor values or read OT bulk tissueSpecificity as cell-type specificity.

Chaining Partners

• Fills Functional / Modality / Target-Safety in the virtual-biotech-cso routing, alongside celltype-specificity-profiler (single-cell specificity), clinical-trial-finder (trials), and gwas-lookup (germline). Output is structured JSON, so it composes via the Bio Orchestrator.

Maintenance

• Review cadence: track Open Targets Platform releases (factor keys / schema can change); the prioritisation key set evolves.
• Staleness signals: a prioritisation.items key renamed; the GraphQL endpoint version bumped.
• Deprecation criteria: retire if omics-target-evidence-mapper is extended to surface prioritisation factors natively.

Citations

• Open Targets Platform — platform.opentargets.org (target prioritisation, tractability, safety liabilities). API: api.platform.opentargets.org/api/v4/graphql.
• DepMap (Broad Institute) — gene essentiality, surfaced via the Open Targets prioritisation factor.