---name: organoid-drug-response-agent description: AI-powered analysis of patient-derived organoid (PDO) drug screening for personalized oncology treatment selection and biomarker discovery. license: MIT metadata: author: AI Group version: "1.0.0" created: "2026-01-19" compatibility:

• system: Python 3.10+ allowed-tools:
• run_shell_command
• read_file
• write_file

keywords:

• organoid-drug-response-agent
• automation
• biomedical measurable_outcome: execute task with >95% success rate. ---"

Organoid Drug Response Agent

The Organoid Drug Response Agent provides AI-driven analysis of patient-derived organoid (PDO) drug screening data for personalized treatment selection. It correlates organoid drug responses with patient outcomes and molecular profiles to guide precision oncology decisions.

When to Use This Skill

• When interpreting organoid drug screening results for treatment selection.
• To correlate PDO drug sensitivity with molecular features.
• For identifying combination therapies using organoid co-culture systems.
• When predicting patient response from organoid-derived data.
• To discover biomarkers from large-scale organoid screens.

Core Capabilities

1. Drug Response Analysis: Process organoid viability data to calculate IC50, AUC, and response metrics.

2. Patient-Organoid Concordance: Assess molecular fidelity between PDO and donor tumor.

3. Biomarker Discovery: Identify molecular features predicting drug sensitivity.

4. Combination Screening: Analyze drug synergy from combination matrices.

5. Clinical Translation: Project organoid findings to patient treatment recommendations.

6. Microenvironment Modeling: Analyze immune co-culture and CAF interactions.

Organoid Advantages

Workflow

1. Input: Organoid drug screening data, organoid molecular profiles, patient tumor data.

2. QC: Assess organoid viability and growth metrics.

3. Response Calculation: Compute drug sensitivity metrics.

4. Concordance: Compare organoid to donor tumor molecular profiles.

5. Biomarker Analysis: Correlate sensitivity with molecular features.

6. Translation: Generate patient treatment recommendations.

7. Output: Drug rankings, biomarkers, recommended treatments.

Example Usage

User: "Analyze organoid drug screening results for this colorectal cancer patient and recommend treatments."

Agent Action:

python3 Skills/Oncology/Organoid_Drug_Response_Agent/organoid_analyzer.py \
    --screening_data drug_screen_384well.csv \
    --organoid_rnaseq organoid_expression.tsv \
    --organoid_mutations organoid_variants.maf \
    --patient_tumor patient_expression.tsv \
    --tumor_type colorectal \
    --combination_matrix combo_screen.csv \
    --output organoid_report/

Drug Response Metrics

Organoid-Patient Concordance Studies

Combination Synergy Analysis

Methods:

• Bliss independence
• Loewe additivity
• ZIP (Zero Interaction Potency)
• HSA (Highest Single Agent)

Output:

• Synergy scores
• Combination indices
• Dose-effect surfaces
• Optimal ratio identification

AI/ML Models

Response Prediction:

• Multi-omic features (expression, mutation, CNV)
• Drug structural features
• Graph neural networks for drug-response

Biomarker Discovery:

• LASSO regression for feature selection
• Random forest for interaction detection
• SHAP values for interpretability

Translation Modeling:

• Transfer learning (organoid → patient)
• Concordance-weighted predictions
• Uncertainty quantification

Organoid Co-Culture Systems

Immune Co-Culture:

• T-cell killing assays
• Checkpoint inhibitor testing
• CAR-T efficacy evaluation

Stromal Co-Culture:

• CAF interactions
• Drug resistance mechanisms
• ECM-mediated effects

Prerequisites

• Python 3.10+
• Drug response analysis packages
• Machine learning frameworks
• Organoid molecular databases

Related Skills

• PDX_Model_Analysis_Agent - For complementary models
• Drug_Repurposing - For additional drug candidates
• Multi_Omics_Integration - For molecular characterization

Quality Control Metrics

Clinical Implementation

1. Turnaround Time: 4-8 weeks from biopsy
2. Panel Size: 50-100+ drugs typically tested
3. Decision Support: Ranked drug recommendations
4. Monitoring: Re-screen on progression

Author

AI Group - Biomedical AI Platform