Skip to content

Latest commit

 

History

History
 
 

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 
 
 

README.md

CRISPR Screen Analysis Skill

Overview

Comprehensive analysis workflow for CRISPR knockout/activation/interference screens. Integrates gene essentiality data from DepMap, pathway enrichment analysis, protein interaction networks, druggability assessment, and clinical relevance to prioritize hits for experimental validation and therapeutic targeting.

Quick Start

Input Format

Provide gene list from your screen:

Analyze these CRISPR hits: KRAS, EGFR, WEE1, PLK1, AURKA, CDK2, CHEK1, MCM2, E2F1, RB1

Or ask about a specific cancer type:

What are the top essential genes for pancreatic cancer?

Or validate a single gene:

Is WEE1 a good therapeutic target for TP53-mutant cancers?

Output

Complete markdown report with:

  • Essentiality Analysis: DepMap scores, pan-cancer vs selective classification
  • Pathway Enrichment: GO, Reactome, KEGG, MSigDB hallmarks
  • PPI Networks: Protein complexes, hub genes, synthetic lethal candidates
  • Druggability: Pharos TDL, approved drugs, clinical compounds, chemical probes
  • Clinical Relevance: COSMIC mutations, expression, biomarkers, trials
  • Hit Prioritization: Top 10 targets with multi-dimensional scores (0-100)
  • Validation Strategy: Recommended experiments, tool compounds, timelines

Key Features

Multi-dimensional prioritization - Integrates essentiality + selectivity + druggability + clinical relevance ✅ Validation recommendations - Specific experiments, tool compounds, expected outcomes ✅ Tier-based ranking - Tier 1 (immediate) > Tier 2 (medium-term) > Tier 3 (long-term) ✅ Pathway-level interpretation - Identifies convergent pathways and protein complexes ✅ Synthetic lethal discovery - Finds combination therapy opportunities ✅ Evidence grading - ★★★ (high) to ★☆☆ (low) based on data quality

Analysis Modes

1. Gene List Analysis (Most Common)

  • Input: 5-100 gene symbols from your CRISPR screen
  • Use case: Prioritize hits from pooled/arrayed screen
  • Output: Comprehensive report with prioritization

2. Cancer Type Query

  • Input: Cancer type name (e.g., "lung cancer", "breast cancer")
  • Use case: Find essential genes for specific cancer
  • Output: Top 20-50 essential genes for that cancer

3. Single Gene Validation

  • Input: One gene symbol
  • Use case: Deep dive on specific target, validate essentiality
  • Output: Target validation report with selectivity analysis

Supported Screens

  • CRISPR knockout (Cas9, dropout/depletion screens)
  • CRISPR activation (CRISPRa, enrichment screens)
  • CRISPR interference (CRISPRi, knockdown screens)
  • shRNA screens (similar workflow)

Examples

See EXAMPLES.md for detailed walkthroughs:

  • Example 1: Lung cancer A549 cell line screen (20 genes)
  • Example 2: Pancreatic cancer essentiality query
  • Example 3: WEE1 target validation (TP53 synthetic lethal)
  • Example 4: RB1 synthetic lethal analysis
  • Example 5: DNA repair pathway analysis
  • Example 6: Lung vs breast cancer comparison
  • Example 7: Prioritizing 100 hits down to 10 for validation

Tools Used

Core Tools

  • DepMap: Gene essentiality data from 1000+ cancer cell lines
  • Enrichr: Pathway and GO enrichment analysis
  • STRING: Protein-protein interaction networks
  • Pharos: Target development level (Tclin/Tchem/Tbio/Tdark)
  • DGIdb: Drug-gene interaction database
  • Open Targets: Chemical probes, tractability, safety
  • ClinicalTrials.gov: Clinical trial status
  • COSMIC: Somatic mutations in cancer

Supporting Tools

  • GTEx: Gene expression in normal tissues
  • TCGA: Gene expression in tumors
  • IntAct: Curated protein interactions
  • ChEMBL: Bioactivity data for tool compounds
  • PubMed: Literature validation

Limitations

Not designed for:

  • RNA-seq differential expression analysis (different workflow)
  • ChIP-seq or epigenomics screens (different context)
  • Metabolic screens (requires different tools)
  • Morphology screens (requires image analysis)

Data limitations:

  • DepMap represents cell lines, not patient tumors (may not reflect in vivo)
  • Tool compounds may not perfectly phenocopy genetic knockout
  • Selectivity in cell lines may differ from patient-level selectivity

Biological limitations:

  • CRISPR artifacts (off-target effects, toxicity of Cas9 cutting)
  • Pathway redundancy not captured in single-gene knockout
  • Context-dependency (nutrient conditions, 2D vs 3D culture)

Validation Recommendations

Tier 1 Targets (High Priority)

  • Existing approved/late-stage drugs available
  • Strong selective essentiality (tissue-specific)
  • Validation timeline: 2-3 weeks
  • Success rate: 80-100%

Tier 2 Targets (Medium Priority)

  • Chemical probes or early-stage compounds available
  • Moderate essentiality, good druggability
  • Validation timeline: 4-8 weeks
  • Success rate: 60-80%

Tier 3 Targets (Lower Priority)

  • Novel mechanisms, no tool compounds
  • Requires alternative strategies (PROTACs, genetic validation only)
  • Validation timeline: 8-16 weeks
  • Success rate: 20-50%

Resources Required

Computational

  • ToolUniverse installation with API keys:
    • DepMap (no key required, public API)
    • Enrichr (no key required)
    • STRING (no key required)
    • Pharos (no key required)
    • ClinicalTrials.gov (no key required)

Experimental Validation

  • Budget: $15-30K for full validation (10 targets)
    • Compounds: $5-10K
    • Cell culture: $5-10K
    • Assays: $5-10K
  • Personnel: 1 postdoc + 1 technician for 3-4 months
  • Equipment: Standard cell culture, plate readers, flow cytometry

Quality Control

Before considering analysis complete, verify:

  • All input genes validated against DepMap registry
  • Essentiality scores retrieved for 100% of valid genes
  • Pathway enrichment completed (GO + Reactome + Hallmarks minimum)
  • Druggability assessed for all hits (Pharos TDL)
  • Top 10 priority list generated with scores
  • Validation recommendations provided for Tier 1 targets
  • Evidence grades assigned (★★★/★★☆/★☆☆)
  • All findings cited to source tools/databases

Citation

When using this skill, cite the underlying databases:

  • DepMap: Broad Institute DepMap Portal (https://depmap.org/)
  • Enrichr: Chen et al., BMC Bioinformatics 2013
  • STRING: Szklarczyk et al., Nucleic Acids Res 2021
  • Pharos: Nguyen et al., Nucleic Acids Res 2021

Support

For issues or questions:

  • Check EXAMPLES.md for common use cases
  • See SKILL.md for detailed methodology
  • Report bugs via ToolUniverse GitHub issues