Strategy 1: Target-Based（ターゲット型）
  Disease → Target genes → Known drugs for targets → Validation

Strategy 2: Compound-Based（化合物型）
  Drug → All known targets → Other diseases of targets → Score

Strategy 3: Disease-Driven（疾患駆動型）
  Disease → DEGs/GWAS targets → Pathway → Drugs in pathway

Strategy 4: Mechanism-Based（メカニズム型）
  Known MOA → Drugs with similar MOA → New indications

Strategy 5: Network-Based（ネットワーク型）
  Disease module → Proximity analysis → Proximal drugs

Strategy 6: Phenotype-Based（表現型型）
  Adverse events → Therapeutic potential → Repurpose

Strategy 7: Structure-Based（構造型）
  Active compound → Similar structures → Approved analogs

def target_based_repurposing(disease_name):
    """
    疾患 → ターゲット → 既存薬 のリポジショニングパイプライン。
    """
    pipeline = {
        "step1_disease": "EFO/MONDO IDで疾患を同定",
        "step2_targets": "Open Targets で上位 50 ターゲットを取得",
        "step3_drugs": "ターゲット毎に ChEMBL/DrugBank で既存薬検索",
        "step4_filter": "現在の適応症と異なる薬剤を候補に",
        "step5_validate": "文献エビデンス + 構造的妥当性で検証",
        "step6_score": "多基準スコアリングで最終ランキング",
    }
    return pipeline

# スコアリング関数
def repurposing_score(candidate):
    """
    リポジショニング候補のスコアリング。
    """
    score = 0.0
    weights = {
        "target_association": 0.25,  # Open Targets score
        "drug_approval_status": 0.20,  # FDA approved > Phase3 > Phase2
        "mechanism_relevance": 0.20,  # MOA と疾患の関連性
        "safety_profile": 0.15,       # 既知安全性プロファイル
        "literature_evidence": 0.10,  # PubMed エビデンス
        "structural_fit": 0.10,       # 分子ドッキングスコア
    }
    for key, weight in weights.items():
        score += candidate.get(key, 0) * weight
    return score

import networkx as nx
import numpy as np

def network_proximity(drug_targets, disease_genes, ppi_network):
    """
    ネットワーク近接度法によるリポジショニング評価。
    Guney et al., Nature Communications 2016 に準拠。

    d(S,T) = 1/|T| * Σ min d(s,t) for s in S
    """
    G = ppi_network  # PPI ネットワーク (NetworkX Graph)
    distances = []

    for dt in drug_targets:
        if dt not in G:
            continue
        min_dist = float('inf')
        for dg in disease_genes:
            if dg not in G:
                continue
            try:
                d = nx.shortest_path_length(G, dt, dg)
                min_dist = min(min_dist, d)
            except nx.NetworkXNoPath:
                continue
        if min_dist < float('inf'):
            distances.append(min_dist)

    if not distances:
        return {"proximity": None, "significance": "N/A"}

    proximity = np.mean(distances)

    return {
        "proximity": proximity,
        "significance": "Close" if proximity < 2.0 else "Moderate" if proximity < 4.0 else "Distant",
        "n_drug_targets_in_network": len([t for t in drug_targets if t in G]),
        "n_disease_genes_in_network": len([g for g in disease_genes if g in G]),
    }

## Repurposing Candidate Evaluation

### Candidate: [Drug Name]
| Criterion | Score (0-1) | Weight | Evidence |
|-----------|-------------|--------|----------|
| Target-disease association | | 0.25 | |
| FDA approval status | | 0.20 | |
| Mechanism relevance | | 0.20 | |
| Safety profile | | 0.15 | |
| Literature support | | 0.10 | |
| Structural compatibility | | 0.10 | |
| **Weighted Total** | | | |

### Evidence Summary
- Clinical trials: [NCT IDs]
- Case reports: [PMID]
- Mechanistic studies: [PMID]

### Risk Assessment
- Patent status: ___
- Formulation feasibility: ___
- Regulatory pathway: 505(b)(2) / Orphan / Standard

# Drug Repurposing Report: [DISEASE / DRUG]

**Strategy**: [Target-Based / Network-Based / etc.]
**Date**: [date]

## 1. Background & Rationale

## 2. Methodology
### 2.1 Strategy Selection Justification
### 2.2 Data Sources
### 2.3 Scoring Criteria

## 3. Top Candidates
| Rank | Drug | Original Indication | Score | Key Evidence |
|------|------|---------------------|-------|-------------|
| 1 | | | | |
| 2 | | | | |
| 3 | | | | |

## 4. Detailed Candidate Profiles
### 4.1 Candidate 1
### 4.2 Candidate 2
### 4.3 Candidate 3

## 5. Validation Evidence

## 6. Recommendations
### 6.1 Immediate Next Steps
### 6.2 Clinical Trial Design Considerations
### 6.3 Regulatory Path

## 7. Limitations & Caveats

## 8. Data Sources