Tooluniverse Drug Target Validation

Drug Target Validation Pipeline

Validate drug target hypotheses using multi-dimensional computational evidence before committing to wet-lab work. Produces a quantitative Target Validation Score (0-100) with priority tier classification and GO/NO-GO recommendation.

Reasoning Before Searching

A valid drug target must pass 4 gates in order. Failing an early gate makes later gates irrelevant:

1. Genetic evidence linking it to disease: Does human genetic data (GWAS, rare variant studies, Mendelian genetics) support this target's role? Genetic evidence is the strongest predictor of clinical success. Use OpenTargets and GWAS catalog before anything else. If no genetic link exists, the hypothesis is speculative — document this clearly. 2. Druggability: Can a molecule reach and modulate the target? Check structure availability (PDB, AlphaFold), binding pocket prediction (ProteinsPlus), target class (kinase, GPCR, nuclear receptor = favorable; transcription factor, scaffold protein = difficult), and existing chemical probes. 3. Safety — essentiality in normal tissue: Is the target expressed in critical tissues (heart, liver, bone marrow)? Is knockout lethal in mice? High expression in essential tissue or lethality in mouse models is a strong safety red flag even before any clinical data. 4. Competitive landscape: Are other drugs already approved or in late-stage trials for this target? If so, the bar is differentiation, not first-in-class. Check ChEMBL, DrugBank, and ClinicalTrials.gov early.

Do not proceed to Phase 3 (Chemical Matter) before completing Phase 1 (Disease Association). Gate 1 failures should prompt a NO-GO or pivot recommendation.

LOOK UP DON'T GUESS: Never assume a target is druggable based on its protein family alone, never assume expression is low in a tissue without checking GTEx or HPA, never assume no competitors without searching ClinicalTrials.gov.

RUN THE ML MODELS, DON'T SKIP THEM: When deep-learning predictors are available (ADMET-AI, ESMFold, AlphaFold, DoGSite, DynaMut2, DeepGO), run them even when database lookups or experimental data already cover the same property. The ML predictions provide an orthogonal, mechanistically-grounded estimate that's a first-class output of this skill — not a fallback. A target-validation report missing ML predictions is incomplete regardless of how much database evidence is present.

COMPUTE, DON'T DESCRIBE

When analysis requires computation (statistics, data processing, scoring, enrichment), write and run Python code via Bash. Don't describe what you would do — execute it and report actual results. Use ToolUniverse tools to retrieve data, then Python (pandas, scipy, statsmodels, matplotlib) to analyze it.

Key Principles

1. Report-first - Create report file FIRST, then populate progressively 2. Target disambiguation FIRST - Resolve all identifiers before analysis 3. Evidence grading - Grade all evidence as T1 (experimental) to T4 (computational) 4. Disease-specific - Tailor analysis to disease context when provided 5. Modality-aware - Consider small molecule vs biologics tractability 6. Safety-first - Prominently flag safety concerns early 7. Quantitative scoring - Every dimension scored numerically (0-100 composite) 8. Negative results documented - "No data" is data; empty sections are failures 9. Source references - Every statement must cite tool/database 10. English-first queries - Always use English terms in tool calls; respond in user's language

When to Use

Apply when users ask about:

• "Is [target] a good drug target for [disease]?"
• Target validation, druggability assessment, or target prioritization
• Safety risks of modulating a target
• Chemical starting points for target validation
• GO/NO-GO recommendation for a target

Not for (use other skills): general target biology (tooluniverse-target-research), drug compound profiling (tooluniverse-drug-research), variant interpretation (tooluniverse-variant-interpretation), disease research (tooluniverse-disease-research).

Input Parameters

Reference Files

• SCORING_CRITERIA.md - Detailed scoring matrices, evidence grading, priority tiers, score calculation
• REPORT_TEMPLATE.md - Full report template, completeness checklist, section format examples
• TOOL_REFERENCE.md - Verified tool parameters, known corrections, fallback chains, modality-specific guidance, phase-by-phase tool lists
• QUICK_START.md - Quick start guide

---

Scoring Overview

Total: 0-100 points across 5 dimensions (details in SCORING_CRITERIA.md):

Priority Tiers: 80-100 = Tier 1 (GO) | 60-79 = Tier 2 (CONDITIONAL GO) | 40-59 = Tier 3 (CAUTION) | 0-39 = Tier 4 (NO-GO)

Evidence Grades: T1 (clinical proof) > T2 (functional studies) > T3 (associations) > T4 (predictions)

---

Pipeline Phases

Phase 0: Target Disambiguation (ALWAYS FIRST)

Resolve target to ALL identifiers before any analysis.

Steps: 1. MyGene_query_genes - Get initial IDs (Ensembl, UniProt, Entrez) 2. ensembl_lookup_gene - Get versioned Ensembl ID (species="homo_sapiens" REQUIRED) 3. ensembl_get_xrefs - Cross-references (HGNC, etc.) 4. OpenTargets_get_target_id_description_by_name - Verify OT target 5. ChEMBL_search_targets - Get ChEMBL target ID 6. UniProt_get_function_by_accession - Function summary (returns list of strings) 7. UniProt_get_alternative_names_by_accession - Collision detection

Output: Table of verified identifiers (Gene Symbol, Ensembl, UniProt, Entrez, ChEMBL, HGNC) plus protein function and target class.

Phase 1: Disease Association (0-30 pts)

Quantify target-disease association from genetic, literature, and pathway evidence.

Key tools:

• OpenTargets_get_diseases_phenotypes_by_target_ensembl - Disease associations
• OpenTargets_target_disease_evidence - Detailed evidence (needs efoId + ensemblId)
• OpenTargets_get_evidence_by_datasource - Evidence by data source
• gwas_get_snps_for_gene / gwas_search_studies - GWAS evidence
• gnomad_get_gene_constraints - Genetic constraint (pLI, LOEUF)
• PubMed_search_articles - Literature (returns plain list of dicts)
• OpenTargets_get_publications_by_target_ensemblID - OT publications (uses entityId)

Phase 2: Druggability (0-25 pts)

Assess whether the target is amenable to therapeutic intervention.

Key tools:

• OpenTargets_get_target_tractability_by_ensemblID - Tractability (SM, AB, PR, OC)
• OpenTargets_get_target_classes_by_ensemblID - Target classification
• Pharos_get_target - TDL: Tclin > Tchem > Tbio > Tdark
• DGIdb_get_gene_druggability - Druggability categories
• alphafold_get_prediction (param: qualifier) / alphafold_get_summary
• ProteinsPlus_predict_binding_sites - Pocket detection
• OpenTargets_get_chemical_probes_by_target_ensemblID - Chemical probes
• OpenTargets_get_target_enabling_packages_by_ensemblID - TEPs
• TCDB_get_transporter - For SLC/ABC transporter targets: TC classification, family, PDB structures (param: uniprot_accession)
• TCDB_search_by_substrate - Find transporters by substrate (param: substrate_name)

Phase 3: Chemical Matter (feeds Phase 2 scoring)

Identify existing chemical starting points for target validation.

Key tools:

• ChEMBL_search_targets + ChEMBL_get_target_activities - Bioactivity data (note: target_chembl_id__exact with double underscore)
• BindingDB_get_ligands_by_uniprot - Binding data (affinity in nM)
• PubChem_search_assays_by_target_gene + PubChem_get_assay_active_compounds - HTS data
• OpenTargets_get_associated_drugs_by_target_ensemblID - Known drugs (size REQUIRED)
• ChEMBL_search_mechanisms - Drug mechanisms
• DGIdb_get_gene_info - Drug-gene interactions

Phase 3b: ADMET-AI Deep-Learning Profile (REQUIRED)

For each lead / approved compound identified above, run all ten ADMET-AI Chemprop-GNN endpoints. This is a required deliverable of the skill, not optional:

Required output — ADMET head-to-head table: when two or more candidate drugs exist (approved or late-stage), produce a side-by-side comparison table with every endpoint in the same row and a "Winner" column flagging which drug is safer. This table is the primary visual of the report and must not be abbreviated or summarized into prose.

ADMET-AI fallback (IMPORTANT): If MCP calls to ADMETAI_predict_* fail, return empty, or timeout, run them via Bash + Python SDK instead:

from tooluniverse import ToolUniverse
tu = ToolUniverse()
tu.load_tools()
for endpoint in ['physicochemical_properties','toxicity','BBB_penetrance','CYP_interactions',
                 'bioavailability','clearance_distribution','nuclear_receptor_activity',
                 'stress_response','solubility_lipophilicity_hydration']:
    r = tu.run_one_function({'name': f'ADMETAI_predict_{endpoint}',
                              'arguments': {'smiles_list': [SMILES_DRUG_A, SMILES_DRUG_B]}})
    print(f'{endpoint}: {r}')

This SDK path bypasses the CLI subprocess and avoids segfault issues with torch. Always try MCP first; use this fallback if MCP returns no data.

Phase 4: Clinical Precedent (0-15 pts)

Assess clinical validation from approved drugs and clinical trials.

Key tools:

• FDA_get_mechanism_of_action_by_drug_name / FDA_get_indications_by_drug_name
• drugbank_get_targets_by_drug_name_or_drugbank_id (ALL params required: query, case_sensitive, exact_match, limit)
• search_clinical_trials (query_term REQUIRED)
• OpenTargets_get_drug_warnings_by_chemblId / OpenTargets_get_drug_adverse_events_by_chemblId

Phase 5: Safety (0-20 pts)

Identify safety risks from expression, genetics, and known adverse events.

Key tools:

• OpenTargets_get_target_safety_profile_by_ensemblID - Safety liabilities
• GTEx_get_median_gene_expression - Tissue expression (operation="median" REQUIRED)
• HPA_search_genes_by_query / HPA_get_comprehensive_gene_details_by_ensembl_id
• OpenTargets_get_biological_mouse_models_by_ensemblID - KO phenotypes
• FDA_get_adverse_reactions_by_drug_name / FDA_get_boxed_warning_info_by_drug_name
• OpenTargets_get_target_homologues_by_ensemblID - Paralog risks

Critical tissues to check: heart, liver, kidney, brain, bone marrow.

Phase 6: Pathway Context

Understand the target's role in biological networks and disease pathways.

Key tools:

• Reactome_map_uniprot_to_pathways (param: id, NOT uniprot_id)
• STRING_get_protein_interactions (param: protein_ids as array, species=9606)
• intact_get_interactions - Experimental PPI
• OpenTargets_get_target_gene_ontology_by_ensemblID - GO terms
• STRING_functional_enrichment - Enrichment analysis

Assess: pathway redundancy, compensation risk, feedback loops.

Phase 7: Validation Evidence (0-10 pts)

Assess existing functional validation data.

Key tools:

• DepMap_get_gene_dependencies - Essentiality (score < -0.5 = essential)
• PubMed_search_articles - Search for CRISPR/siRNA/knockout studies
• CTD_get_gene_diseases - Gene-disease associations

Phase 8: Structural Insights

Leverage structural biology for druggability and mechanism understanding. ALWAYS run both the deep-learning predictors (ESMFold, DoGSite) AND retrieve experimental structures, even when high-resolution PDB entries already exist. The ML models give an independent pLDDT/druggability score that is a required output of this phase.

Required tool calls (every run):

• ESMFold_predict_structure — Meta ESM-2 language-model structure prediction from the UniProt sequence. Report: model pLDDT, worst-residue confidence, RMSD vs. reference PDB if available.
• alphafold_get_prediction / alphafold_get_summary — DeepMind AlphaFold model + per-residue pLDDT.
• ProteinsPlus_predict_binding_sites — DoGSite deep-learning pocket scoring. Report: top 3 pockets with volume, druggability score, residue composition.

Supporting tools:

• UniProt_get_entry_by_accession - Extract PDB cross-references
• get_protein_metadata_by_pdb_id / pdbe_get_entry_summary / pdbe_get_entry_quality
• InterPro_get_protein_domains / InterPro_get_domain_details - Domain architecture

Phase 9: Literature Deep Dive

Comprehensive collision-aware literature analysis.

Steps: 1. Collision detection: Search "{gene_symbol}"[Title] in PubMed; if >20% off-topic, add filters (AND protein OR gene OR receptor) 2. Publication metrics: Total count, 5-year trend, drug-focused subset 3. Key reviews: review[pt] filter in PubMed 4. Citation metrics: openalex_search_works for impact data 5. Broader coverage: EuropePMC_search_articles

Phase 10: Validation Roadmap (Synthesis)

Synthesize all phases into actionable output: 1. Target Validation Score (0-100) with component breakdown 2. Priority Tier (1-4) assignment 3. GO/NO-GO Recommendation with justification 4. Recommended Validation Experiments 5. Tool Compounds for Testing 6. Biomarker Strategy 7. Key Risks and Mitigations 8. Deep-Learning Models Contributing — explicit attribution table listing every ML predictor invoked during the run and what each produced. Example format:

Only list models actually called during the run. This section makes the ML content first-class for a scientific or investor audience.

---

Report Output

Create file: [TARGET]_[DISEASE]_validation_report.md

Use the full template from REPORT_TEMPLATE.md. Key sections:

• Executive Summary (score, tier, recommendation, key findings, critical risks)
• Validation Scorecard (all 12 sub-scores with evidence)
• Sections 1-14 covering each phase
• Completeness Checklist (mandatory before finalizing)

Complete the Completeness Checklist (in REPORT_TEMPLATE.md) before finalizing to verify all phases were covered, all scores justified, and negative results documented.

Drug Target Validation - Quick Start Guide

Overview

This skill validates drug target hypotheses using computational evidence from 60+ ToolUniverse tools. It produces a Target Validation Score (0-100) with a GO/NO-GO recommendation.

Basic Usage

Use Case 1: Full Target Validation (Disease-Specific)

Is KRAS a druggable target for pancreatic cancer?

What happens: Full 10-phase analysis with disease-specific evidence, produces validation score and GO/NO-GO recommendation.

Use Case 2: General Target Druggability Assessment

Assess the druggability of BTK as a therapeutic target

What happens: Focuses on druggability, chemical matter, structural tractability, and clinical precedent without disease-specific filtering.

Use Case 3: Target Safety Assessment

What are the safety risks of inhibiting IDH1?

What happens: Emphasizes safety phase - expression in critical tissues, knockout phenotypes, known adverse events, paralogs.

Use Case 4: Chemical Starting Points

What compounds are available for PCSK9 validation?

What happens: Deep dive into chemical matter - ChEMBL activities, BindingDB ligands, PubChem bioassays, chemical probes.

Use Case 5: Target Comparison

Compare PD-1 vs PD-L1 as immunotherapy targets

What happens: Runs validation pipeline for both targets, produces side-by-side comparison of scores.

Use Case 6: Novel Target Evaluation

Is TEAD4 a viable target for solid tumors? Consider small molecule approach.

What happens: Full validation with modality-specific analysis, extra focus on structural tractability for small molecules.

Output

Every analysis produces a markdown report file with:

1. Executive Summary with GO/NO-GO recommendation 2. Validation Scorecard (0-100 composite score) 3. 10 Detailed Sections covering all evidence dimensions 4. Validation Roadmap with recommended experiments 5. Risk Assessment with mitigation strategies 6. Completeness Checklist showing analysis coverage

Priority Tiers

Evidence Grades

• [T1]: Direct mechanistic evidence (crystal structure, patient mutations, FDA-approved drug)
• [T2]: Functional studies (CRISPR KO, mouse model, biochemical assay)
• [T3]: Association (GWAS hit, DepMap essentiality, expression correlation)
• [T4]: Annotation (database entry, computational prediction, review article)

Example Report Structure

EGFR_NSCLC_validation_report.md

Executive Summary
  Score: 90/100 | Tier 1 | GO
  Strong genetic + clinical validation for NSCLC

Validation Scorecard
  Disease Association:   28/30
  Druggability:          24/25
  Safety Profile:        14/20
  Clinical Precedent:    15/15
  Validation Evidence:    9/10

[10 detailed sections with tool outputs and evidence]

Validation Roadmap
  - Recommended experiments
  - Tool compounds
  - Biomarker strategy
  - Risk mitigations

Key Databases Queried

Drug Target Validation - Report Template

Use this template to create the validation report file: [TARGET]_[DISEASE]_validation_report.md

---

Report Structure

# Drug Target Validation Report: [TARGET]

**Target**: [Gene Symbol] ([Full Name])
**Disease Context**: [Disease Name] (if provided)
**Modality**: [Small molecule / Antibody / etc.] (if specified)
**Generated**: [Date]
**Status**: In Progress

---

## Executive Summary

**Target Validation Score**: [XX/100]
**Priority Tier**: [Tier X] - [Description]
**Recommendation**: [GO / CONDITIONAL GO / CAUTION / NO-GO]

**Key Findings**:
- [1-sentence disease association strength with evidence grade]
- [1-sentence druggability assessment]
- [1-sentence safety profile]
- [1-sentence clinical precedent]

**Critical Risks**:
- [Top risk 1]
- [Top risk 2]

---

## Validation Scorecard

| Dimension | Score | Max | Assessment | Key Evidence |
|-----------|-------|-----|------------|--------------|
| **Disease Association** | | 30 | | |
| - Genetic evidence | | 10 | | |
| - Literature evidence | | 10 | | |
| - Pathway evidence | | 10 | | |
| **Druggability** | | 25 | | |
| - Structural tractability | | 10 | | |
| - Chemical matter | | 10 | | |
| - Target class | | 5 | | |
| **Safety Profile** | | 20 | | |
| - Expression selectivity | | 5 | | |
| - Genetic validation | | 10 | | |
| - Known ADRs | | 5 | | |
| **Clinical Precedent** | | 15 | | |
| **Validation Evidence** | | 10 | | |
| - Functional studies | | 5 | | |
| - Disease models | | 5 | | |
| **TOTAL** | **XX** | **100** | **[Tier]** | |

---

## 1. Target Identity
[Researching...]

## 2. Disease Association Evidence
### 2.1 OpenTargets Disease Associations
[Researching...]
### 2.2 GWAS Genetic Evidence
[Researching...]
### 2.3 Constraint Scores (gnomAD)
[Researching...]
### 2.4 Literature Evidence
[Researching...]

## 3. Druggability Assessment
### 3.1 Tractability (OpenTargets)
[Researching...]
### 3.2 Target Classification
[Researching...]
### 3.3 Structural Tractability
[Researching...]
### 3.4 Chemical Probes & Enabling Packages
[Researching...]

## 4. Known Modulators & Chemical Matter
### 4.1 Approved/Clinical Drugs
[Researching...]
### 4.2 ChEMBL Bioactivity
[Researching...]
### 4.3 BindingDB Ligands
[Researching...]
### 4.4 PubChem Bioassays
[Researching...]
### 4.5 Chemical Probes
[Researching...]

## 5. Clinical Precedent
### 5.1 FDA-Approved Drugs
[Researching...]
### 5.2 Clinical Trial Landscape
[Researching...]
### 5.3 Failed Programs & Lessons
[Researching...]

## 6. Safety & Toxicity Profile
### 6.1 OpenTargets Safety Liabilities
[Researching...]
### 6.2 Expression in Critical Tissues
[Researching...]
### 6.3 Knockout Phenotypes
[Researching...]
### 6.4 Known Adverse Events
[Researching...]
### 6.5 Paralog & Off-Target Risks
[Researching...]

## 7. Pathway Context & Network Analysis
### 7.1 Biological Pathways
[Researching...]
### 7.2 Protein-Protein Interactions
[Researching...]
### 7.3 Functional Enrichment
[Researching...]
### 7.4 Pathway Redundancy Assessment
[Researching...]

## 8. Validation Evidence
### 8.1 Target Essentiality (DepMap)
[Researching...]
### 8.2 Functional Studies
[Researching...]
### 8.3 Animal Models
[Researching...]
### 8.4 Biomarker Potential
[Researching...]

## 9. Structural Insights
### 9.1 Experimental Structures (PDB)
[Researching...]
### 9.2 AlphaFold Prediction
[Researching...]
### 9.3 Binding Pocket Analysis
[Researching...]
### 9.4 Domain Architecture
[Researching...]

## 10. Literature Landscape
### 10.1 Publication Metrics
[Researching...]
### 10.2 Key Publications
[Researching...]
### 10.3 Research Trend
[Researching...]

## 11. Validation Roadmap
### 11.1 Recommended Validation Experiments
[Researching...]
### 11.2 Tool Compounds for Testing
[Researching...]
### 11.3 Biomarker Strategy
[Researching...]
### 11.4 Clinical Biomarker Candidates
[Researching...]
### 11.5 Disease Models to Test
[Researching...]

## 12. Risk Assessment
### 12.1 Key Risks
[Researching...]
### 12.2 Mitigation Strategies
[Researching...]
### 12.3 Competitive Landscape
[Researching...]

## 13. Completeness Checklist
[To be populated post-audit...]

## 14. Data Sources & Methodology
[Will be populated as research progresses...]

---

Completeness Checklist (MANDATORY)

Before finalizing the report, verify all items:

Phase Coverage

• [ ] Phase 0: Target disambiguation (all IDs resolved)
• [ ] Phase 1: Disease association (OT + GWAS + gnomAD + literature)
• [ ] Phase 2: Druggability (tractability + class + structure + probes)
• [ ] Phase 3: Chemical matter (ChEMBL + BindingDB + PubChem + drugs)
• [ ] Phase 4: Clinical precedent (FDA + trials + failures)
• [ ] Phase 5: Safety (OT safety + expression + KO + ADRs + paralogs)
• [ ] Phase 6: Pathway context (Reactome + STRING + GO)
• [ ] Phase 7: Validation evidence (DepMap + literature + models)
• [ ] Phase 8: Structural insights (PDB + AlphaFold + pockets + domains)
• [ ] Phase 9: Literature (collision-aware + metrics + key papers)
• [ ] Phase 10: Validation roadmap (score + recommendations)

Data Quality

• [ ] All scores justified with specific data
• [ ] Evidence grades (T1-T4) assigned to key claims
• [ ] Negative results documented (not left blank)
• [ ] Failed tools with fallbacks documented
• [ ] Source citations for all data points

Scoring

• [ ] All 12 score components calculated
• [ ] Total score summed correctly
• [ ] Priority tier assigned
• [ ] GO/NO-GO recommendation justified

---

Section-Specific Report Formats

Chemical Matter Section Example

### 4. Known Modulators & Chemical Matter

#### 4.1 Approved Drugs
| Drug | ChEMBL ID | Mechanism | Phase | Indication | Source |
|------|-----------|-----------|-------|------------|--------|
| Erlotinib | CHEMBL553 | Inhibitor | 4 | NSCLC | [T1] OpenTargets |

#### 4.2 ChEMBL Bioactivity Summary
**Total Activities**: 12,456 datapoints across 2,341 assays
**Most Potent Compound**: CHEMBL413456 (IC50 = 0.3 nM) [T1]
**Chemical Series**: 8 distinct scaffolds with pChEMBL >= 7.0

#### 4.3 BindingDB Ligands
**Total Ligands**: 856 with measured affinity
**Affinity Distribution**: <1nM: 23, 1-10nM: 89, 10-100nM: 234, 100nM-1uM: 510

#### 4.4 Chemical Probes
| Probe | Source | Potency | Selectivity | Use |
|-------|--------|---------|-------------|-----|
| SGC-1234 | SGC | IC50=5nM | >100x | In vitro |

Pathway Context Section Example

### 7. Pathway Context & Network Analysis

#### 7.1 Key Pathways
| Pathway | Reactome ID | Relevance to Disease | Evidence |
|---------|-------------|---------------------|----------|
| EGFR signaling | R-HSA-177929 | Driver pathway in NSCLC | [T1] |

#### 7.2 Protein-Protein Interactions
**Total Interactors**: 45 (STRING confidence > 0.7)
**Key Interactors**: GRB2, SHC1, PLCG1, PIK3CA, STAT3

#### 7.3 Pathway Redundancy Assessment
**Compensation Risk**: MODERATE
- Parallel pathways: HER2, HER3 can compensate
- Feedback loops: RAS activation bypasses EGFR

Target Identity Section Example

## 1. Target Identity

| Database | Identifier | Verified |
|----------|-----------|----------|
| Gene Symbol | EGFR | Yes |
| Full Name | Epidermal growth factor receptor | Yes |
| Ensembl | ENSG00000146648 | Yes |
| Ensembl (versioned) | ENSG00000146648.18 | Yes |
| UniProt | P00533 | Yes |
| Entrez Gene | 1956 | Yes |
| ChEMBL | CHEMBL203 | Yes |
| HGNC | HGNC:3236 | Yes |

**Protein Function**: [from UniProt_get_function_by_accession]
**Subcellular Location**: [from UniProt_get_subcellular_location_by_accession]
**Target Class**: [from OpenTargets_get_target_classes_by_ensemblID]

Drug Target Validation - Scoring Criteria

Detailed scoring matrices, evidence grading, and priority tier definitions for the Target Validation Score (0-100).

---

Score Components (Total: 0-100)

Disease Association (0-30 points)

Genetic Evidence (0-10):

• GWAS hits for specific disease: +3 per significant locus (max 6)
• Rare variant evidence (ClinVar pathogenic): +2
• Somatic mutations in disease: +2
• pLI > 0.9 (essential gene): +2

Literature Evidence (0-10):

• >100 publications on target+disease: 10
• 50-100 publications: 7
• 10-50 publications: 5
• 1-10 publications: 3
• 0 publications: 0

Pathway Evidence (0-10):

• OpenTargets overall score > 0.8: 10
• Score 0.5-0.8: 7
• Score 0.2-0.5: 4
• Score < 0.2: 1

---

Druggability (0-25 points)

Structural Tractability (0-10):

• High-res co-crystal structure with ligand: 10
• PDB structure available, pockets detected: 7
• AlphaFold only, confident pocket prediction: 5
• AlphaFold low confidence / no structure: 2
• No structural data: 0

Chemical Matter (0-10):

• Known drug-like compounds (IC50 < 100nM): 10
• Tool compounds (IC50 < 1uM): 7
• HTS hits only (IC50 > 1uM): 4
• No known ligands: 0

Target Class Bonus (0-5):

• Validated druggable family (kinase, GPCR, nuclear receptor): 5
• Enzyme, ion channel: 4
• Protein-protein interaction, transporter: 2
• Novel/unknown class: 0

---

Safety Profile (0-20 points)

Tissue Expression Selectivity (0-5):

• Target restricted to disease tissue: 5
• Low expression in heart/liver/kidney/brain: 4
• Moderate expression in 1-2 critical tissues: 2
• High expression in multiple critical tissues: 0

Genetic Validation (0-10):

• Mouse KO viable, no severe phenotype: 10
• Mouse KO viable with mild phenotype: 7
• Mouse KO has concerning phenotype: 3
• Mouse KO lethal: 0
• No KO data, low pLI (<0.5): 5
• No KO data, high pLI (>0.9): 2

Known Adverse Events (0-5):

• No known safety signals: 5
• Mild, manageable ADRs: 3
• Serious ADRs reported: 1
• Black box warning or drug withdrawal: 0

---

Clinical Precedent (0-15 points)

• FDA-approved drug for SAME disease: 15
• FDA-approved drug for DIFFERENT disease: 12
• Phase 3 clinical trial: 10
• Phase 2 clinical trial: 7
• Phase 1 clinical trial: 5
• Preclinical compounds only: 3
• No clinical development: 0

Adjustment factors:

• Failed clinical program for safety: -3
• Drug withdrawal: -5
• Multiple approved drugs (validated class): +2

---

Validation Evidence (0-10 points)

Functional Studies (0-5):

• CRISPR KO shows disease-relevant phenotype: 5
• siRNA knockdown shows phenotype: 4
• Biochemical assay validates mechanism: 3
• Overexpression study only: 2
• No functional data: 0

Disease Models (0-5):

• Patient-derived xenograft (PDX) response: 5
• Genetically engineered mouse model: 4
• Cell line model: 3
• In silico model only: 1
• No model data: 0

---

Priority Tiers

---

Evidence Grading System

---

Score Calculation (Pseudocode)

def calculate_validation_score(phase_results):
    score = {
        'disease_genetic': 0,      # 0-10
        'disease_literature': 0,   # 0-10
        'disease_pathway': 0,      # 0-10
        'drug_structural': 0,      # 0-10
        'drug_chemical': 0,        # 0-10
        'drug_class': 0,           # 0-5
        'safety_expression': 0,    # 0-5
        'safety_genetic': 0,       # 0-10
        'safety_adverse': 0,       # 0-5
        'clinical': 0,             # 0-15
        'validation_functional': 0, # 0-5
        'validation_models': 0,    # 0-5
    }
    total = sum(score.values())

    if total >= 80:
        tier, rec = "Tier 1", "GO - Highly validated target"
    elif total >= 60:
        tier, rec = "Tier 2", "CONDITIONAL GO - Needs focused validation"
    elif total >= 40:
        tier, rec = "Tier 3", "CAUTION - Significant validation needed"
    else:
        tier, rec = "Tier 4", "NO-GO - Consider alternatives"

    return total, tier, rec, score

---

Example Scores

EGFR for NSCLC (well-validated target)

• Disease Association: ~28/30 (strong genetic + pathway + literature)
• Druggability: ~24/25 (kinase, many structures, abundant compounds)
• Safety: ~14/20 (widely expressed but manageable toxicity)
• Clinical Precedent: 15/15 (multiple approved drugs)
• Validation Evidence: ~9/10 (extensive functional data)
• Total: ~90/100 = Tier 1

Novel understudied kinase

• Disease Association: ~8/30 (limited GWAS, few publications)
• Druggability: ~15/25 (kinase family bonus, AlphaFold structure)
• Safety: ~12/20 (limited data, unknown KO phenotype)
• Clinical Precedent: 0/15 (no clinical development)
• Validation Evidence: ~2/10 (minimal functional data)
• Total: ~37/100 = Tier 4

Drug Target Validation - Tool Reference

Verified tool parameters, known corrections, fallback chains, and modality-specific tool guidance.

---

Known Parameter Corrections

---

Verified Tool Parameters (Quick Reference)

---

Fallback Chains

When a primary tool fails, use these fallbacks in order:

---

Modality-Specific Tool Focus

Small Molecule

• Emphasize: binding pockets, ChEMBL compounds, Lipinski compliance
• Key tractability: OpenTargets SM tractability bucket
• Structure: co-crystal structures with small molecule ligands
• Chemical matter: IC50/Ki/Kd data from ChEMBL/BindingDB

Antibody

• Emphasize: extracellular domains, cell surface expression, glycosylation
• Key tractability: OpenTargets AB tractability bucket
• Structure: ectodomain structures, epitope mapping
• Expression: surface expression in disease vs normal tissue

PROTAC

• Emphasize: intracellular targets, surface lysines, E3 ligase proximity
• Key tractability: OpenTargets PROTAC tractability
• Structure: full-length structures for linker design
• Chemical matter: known binders + E3 ligase binders

---

Phase-by-Phase Tool Lists

Phase 0: Target Disambiguation

• MyGene_query_genes - initial ID resolution
• UniProt_get_entry_by_accession - protein details
• ensembl_lookup_gene - Ensembl ID + version
• ensembl_get_xrefs - cross-references
• OpenTargets_get_target_id_description_by_name - OT target info
• ChEMBL_search_targets - ChEMBL target ID
• UniProt_get_function_by_accession - function summary
• UniProt_get_alternative_names_by_accession - collision detection

Phase 1: Disease Association

• OpenTargets_get_diseases_phenotypes_by_target_ensembl - disease associations
• OpenTargets_get_disease_id_description_by_name - disease ID lookup
• OpenTargets_target_disease_evidence - detailed evidence
• OpenTargets_get_evidence_by_datasource - evidence by source
• gwas_get_snps_for_gene - GWAS associations
• gwas_search_studies - GWAS studies
• gnomad_get_gene_constraints - genetic constraint (pLI, LOEUF)
• PubMed_search_articles - literature evidence
• OpenTargets_get_publications_by_target_ensemblID - OT publications

Phase 2: Druggability

• OpenTargets_get_target_tractability_by_ensemblID - tractability
• OpenTargets_get_target_classes_by_ensemblID - target classification
• Pharos_get_target - target development level (TDL)
• DGIdb_get_gene_druggability - druggability categories
• alphafold_get_prediction / alphafold_get_summary - structure prediction
• ProteinsPlus_predict_binding_sites - pocket detection
• OpenTargets_get_chemical_probes_by_target_ensemblID - chemical probes
• OpenTargets_get_target_enabling_packages_by_ensemblID - TEPs

Phase 3: Chemical Matter

• ChEMBL_search_targets / ChEMBL_get_target_activities - bioactivity
• BindingDB_get_ligands_by_uniprot - binding data
• PubChem_search_assays_by_target_gene - HTS assays
• PubChem_get_assay_summary / PubChem_get_assay_active_compounds - assay details
• OpenTargets_get_associated_drugs_by_target_ensemblID - known drugs
• ChEMBL_search_mechanisms - drug mechanisms
• DGIdb_get_gene_info - drug-gene interactions

Phase 4: Clinical Precedent

• FDA_get_mechanism_of_action_by_drug_name - FDA MoA
• FDA_get_indications_by_drug_name - FDA indications
• drugbank_get_targets_by_drug_name_or_drugbank_id - DrugBank targets
• drugbank_get_safety_by_drug_name_or_drugbank_id - DrugBank safety
• search_clinical_trials - active trials
• OpenTargets_get_drug_warnings_by_chemblId - drug warnings
• OpenTargets_get_drug_adverse_events_by_chemblId - adverse events

Phase 5: Safety

• OpenTargets_get_target_safety_profile_by_ensemblID - safety profile
• GTEx_get_median_gene_expression - tissue expression
• HPA_search_genes_by_query / HPA_get_comprehensive_gene_details_by_ensembl_id - HPA expression
• OpenTargets_get_biological_mouse_models_by_ensemblID - KO phenotypes
• gnomad_get_gene_constraints - essentiality proxy
• FDA_get_adverse_reactions_by_drug_name - ADRs
• FDA_get_warnings_and_cautions_by_drug_name / FDA_get_boxed_warning_info_by_drug_name - warnings
• OpenTargets_get_target_homologues_by_ensemblID - paralogs

Phase 6: Pathway Context

• Reactome_map_uniprot_to_pathways / Reactome_get_pathway - pathways
• STRING_get_protein_interactions - PPI network
• intact_get_interactions - experimental PPI
• OpenTargets_get_target_interactions_by_ensemblID - OT interactions
• OpenTargets_get_target_gene_ontology_by_ensemblID - GO terms
• GO_get_annotations_for_gene - GO annotations
• STRING_functional_enrichment - enrichment analysis

Phase 7: Validation Evidence

• DepMap_get_gene_dependencies - essentiality screen
• PubMed_search_articles - functional studies literature
• CTD_get_gene_diseases - gene-disease associations

Phase 8: Structural Insights

• UniProt_get_entry_by_accession - PDB cross-references
• get_protein_metadata_by_pdb_id / pdbe_get_entry_summary - PDB metadata
• pdbe_get_entry_quality / pdbe_get_entry_experiment - quality metrics
• alphafold_get_prediction / alphafold_get_summary - AlphaFold
• ProteinsPlus_predict_binding_sites - binding pockets
• ProteinsPlus_generate_interaction_diagram - interaction diagrams
• InterPro_get_protein_domains / InterPro_get_domain_details - domains

Phase 9: Literature Deep Dive

• PubMed_search_articles - collision-aware searches
• EuropePMC_search_articles - broader coverage
• openalex_search_works - citation metrics