alg-precision-therapeutics

Files (33)

• plugin.json · .claude-plugin/
• .python-version
• README.md
• SKILL.md
• USAGE.md
• alg_precision_therapeutics.py
• causal-chain-flow.tsx · dashboard/components/
• causal-graph.tsx · dashboard/components/
• evidence-chain.tsx · dashboard/components/
• evidence-search.tsx · dashboard/components/
• gaps-panel.tsx · dashboard/components/
• investigations-list.tsx · dashboard/components/
• mechanism-card.tsx · dashboard/components/
• phenome-table.tsx · dashboard/components/
• lib.ts · dashboard/
• page.tsx · dashboard/pages/apt/disease/[id]/
• layout.tsx · dashboard/pages/apt/
• page.tsx · dashboard/pages/apt/
• route.ts · dashboard/routes/disease/
• route.ts · dashboard/routes/evidence/
• route.ts · dashboard/routes/gaps/
• route.ts · dashboard/routes/genes/
• route.ts · dashboard/routes/investigations/
• route.ts · dashboard/routes/mechanisms/
• route.ts · dashboard/routes/phenome/
• route.ts · dashboard/routes/search/
• route.ts · dashboard/routes/treatments/
• hooks.json · hooks/
• pyproject.toml
• schema.tql
• skill.yaml
• SKILL.md · skills/alg-precision-therapeutics/
• uv.lock

SKILL.md (first lines)

# Algorithm for Precision Therapeutics (APT)

Use this skill when the diagnosis is KNOWN and the goal is to understand the
**mechanism of harm** and identify **therapeutic strategies**. Follows Matt Might's
APM Phase 2 (Therapeutic Phase) applied systematically to a MONDO disease ID.

Central innovation: `apt-mechanism` is a **first-class entity** linking
gene -> pathway -> phenotype -> drug, not just a relation attribute.

**Triggers:** "mechanism of harm", "investigate disease X", "MONDO disease",
"therapeutic landscape", "rare disease mechanism", "what can treat X",
"APM analysis", "how does X cause disease", "disease knowledge graph",
"build disease KG", "what genes cause X", "rational therapy design",
"precision therapeutics", "Algorithm for Precision Medicine"

**NOT for:** differential diagnosis, variant pathogenicity, or finding a diagnosis
from symptoms - use clinical tools for those.

## Prerequisites

- TypeDB running (install `alhazen-core` first and run `/alhazen-core:init`)
- `uv` installed

> **Path note:** Replace `<skill-path>` with your installation directory
> (e.g. `~/.claude/plugins/cache/alg-precision-therapeutics/` when installed as a plugin).

## Quick Start

```bash
SCRIPT="<skill-path>/alg_precision_therapeutics.py"

# 1. Search for MONDO ID
uv run --project <skill-path> python $SCRIPT search-disease --query "NGLY1 deficiency"

# 2. Initialize investigation
uv run --project <skill-path> python $SCRIPT init-investigation MONDO:0800044

# 3. Ingest all external data
uv run --project <skill-path> python $SCRIPT ingest-disease --mondo-id MONDO:0800044

# 4. View mechanism map (after Claude adds mechanisms)
uv run --project <skill-path> python $SCRIPT show-mechanisms --mondo-id MONDO:0800044

# 5. View therapeutic map
uv run --project <skill-path> python $SCRIPT show-therapeutic-map --mondo-id MONDO:0800044
```

Read `USAGE.md` before executing commands.

From sciknow-io/alhazen-skill-examples, MIT.

More from sciknow-io/alhazen-skill-examples