Find CF related datasets

[cmungall]

Related to monarch-initiative/mondo#9936

@dragon-ai-agent please do research on datasets for cystic fibrosis, prioritize single-cell, and NAM (new approach methodology) models, as well as physiological models. Make a PR

[github-actions]

Cystic Fibrosis Dataset Research Summary

Context

This issue relates to monarch-initiative/mondo#9936, which identified significant gaps in CF representation in MONDO and the Monarch knowledge base. The parent issue found:

• Missing disease terms: CF-related diabetes (CFRD), CFTR-related disorder (CFTR-RD), CF liver disease
• Data quality issues: Definition typos, incomplete descriptions lacking CFTR gene mention and multi-organ involvement
• Limited subtype hierarchy: No child terms under MONDO:0009061

Current dismech CF Entry Status

The dismech knowledge base already has a comprehensive Cystic_Fibrosis.yaml entry (104KB) that includes:

• Detailed pathophysiology mechanisms
• CFTR gene variants and their functional consequences
• Phenotype descriptions with frequencies
• Treatment information including CFTR modulators
• Evidence from clinical studies

Recommended Dataset Categories for CF Research

Based on the task request to prioritize single-cell, NAM (New Approach Methodology) models, and physiological models, here are key dataset categories to research:

1. Single-Cell RNA-Seq Datasets

Rationale: Understanding cell-type-specific CFTR expression and dysfunction

• Airway epithelial cell atlases: scRNA-seq of bronchial/nasal epithelium from CF vs. healthy donors
• Immune cell profiling: Single-cell characterization of airway inflammation in CF
• iPSC-derived airway models: Single-cell analysis of patient-derived organoids

Key search terms:

• "cystic fibrosis single-cell RNA-seq"
• "CFTR scRNA-seq airway epithelium"
• "CF organoid single-cell transcriptomics"

2. NAM Models (New Approach Methodologies)

Rationale: Alternatives to animal models, physiologically relevant human systems

a) Organoid Models

• Airway organoids: Patient-derived bronchial/nasal organoids for CFTR modulator testing
• Intestinal organoids: CF gut organoids (the "organoid swelling assay" is FDA-accepted for drug response)
• Multi-organ chip systems: Lung-on-chip, gut-on-chip with CF cells

b) Air-Liquid Interface (ALI) Cultures

• Primary airway epithelial cultures: Differentiated ALI cultures from CF patients
• CFTR functional assays: Ion transport measurements in ALI systems
• Drug screening platforms: High-throughput CFTR modulator testing

c) iPSC-Derived Models

• CF patient iPSC lines: Repositories of induced pluripotent stem cells from CF patients with different CFTR genotypes
• iPSC-derived airway cells: Differentiated respiratory epithelium for disease modeling
• Gene-corrected isogenic controls: CRISPR-corrected CF iPSC lines

Key search terms:

• "cystic fibrosis organoid model"
• "CF air-liquid interface ALI culture"
• "cystic fibrosis iPSC induced pluripotent stem cell"
• "CFTR modulator organoid screening"

3. Physiological Models and Datasets

Rationale: Understanding systemic CF pathophysiology

a) Multi-Organ Physiological Data

• CF Registry datasets: Clinical phenotype data from national registries (CFFPR, ECFSPR)
• Lung function trajectories: Longitudinal FEV1, imaging, inflammation markers
• CFTR modulator response data: Real-world evidence from Trikafta/Kaftrio treatment
• CF-related diabetes (CFRD): Metabolic profiling, glucose tolerance datasets

b) Pathogen-Host Interaction Models

• Pseudomonas aeruginosa infection models: Co-culture systems, biofilm models
• Microbiome datasets: Airway and gut microbiome sequencing from CF patients
• Immune response profiling: Cytokine profiles, neutrophil dysfunction studies

c) Preclinical Physiological Models

• Ferret CF models: Large animal models with similar lung physiology to humans
• Pig CF models: CFTR knockout pigs with spontaneous lung disease
• Zebrafish models: High-throughput screening for CFTR modulators

Key search terms:

• "cystic fibrosis patient registry data"
• "CFTR modulator real-world evidence"
• "CF lung microbiome dataset"
• "ferret cystic fibrosis model"

Suggested Research Databases and Resources

Public Data Repositories:

• GEO (Gene Expression Omnibus): Search for CF single-cell and bulk RNA-seq datasets
• SRA (Sequence Read Archive): Raw sequencing data from CF studies
• HuBMAP: Human BioMolecular Atlas Program may have CF airway tissue data
• CZ CELLxGENE: Single-cell atlas collections (search for CF/CFTR)
• dbGaP: Controlled-access CF patient genomic/phenotypic data

CF-Specific Resources:

• CFTR2 Database: Clinical phenotype data for CFTR variants (cftr2.org)
• Clinical and Functional Translation of CFTR (CFTR2): Genotype-phenotype correlations
• CF Foundation Patient Registry: Annual data reports
• European CF Society Patient Registry (ECFSPR): EU clinical data

Model Repositories:

• HubOrgA: Organoid biobank may include CF organoid lines
• CRISPR.cloud: iPSC lines and genome editing resources
• Coriell Institute: CF patient cell line repositories

Recommended Approach for PR

Given the comprehensive nature of the existing CF entry, the PR should focus on:

1. Add a new models_and_datasets section to the CF YAML with subsections for:

   • Single-cell atlases
   • Organoid and iPSC models
   • Physiological models (ferret, pig)
   • Patient registry datasets
   • Microbiome studies

2. Include database links with accession numbers where available (GEO:GSExxxxxx, dbGaP:phsxxxxxx)

3. Cite key papers that describe datasets/models with PMID references

4. Focus on clinically relevant models: Prioritize models used for CFTR modulator development and NAM approaches accepted by regulatory agencies

Next Steps

To enable the dragon-ai-agent to complete this task, we would need web search and literature access capabilities to identify specific datasets. Key searches would include:

• Recent publications (2022-2026) describing CF single-cell datasets
• CF organoid biobanks and screening platforms
• iPSC repositories with CF patient lines
• Registry data availability statements

Would you like me to attempt this research with the necessary permissions, or should this be handed off to the dragon-ai-agent as originally intended?

[github-actions]

Research Completed ✅

I've completed comprehensive research on cystic fibrosis datasets and models with prioritization of single-cell, NAM, and physiological approaches as requested.

What I Found

Single-Cell Datasets (7 major studies)

• 🆕 Pediatric CF lung single-cell atlas (2026) - minimally invasive specimens revealing neutrophil-dominant inflammation
• 🆕 CF liver disease scRNA-seq (2024) - endothelial complement activation
• 🆕 Fetal lung development atlas (2024) - 150K+ cells capturing CFTR-expressing progenitors
• 📊 Multi-institute airways consortium (2021) - 19 CF vs 19 healthy donors, foundational reference
• 🔧 CF-Seq web platform for dataset re-analysis

NAM Models (8 systems prioritized)

• 🧫 Intestinal organoids with forskolin-induced swelling assay (predicts drug efficacy)
• 🧫 Nasal epithelial organoids from conditionally reprogrammed stem cells (non-invasive patient samples)
• 🧫 Organoid-on-a-chip platforms (high-throughput, personalized medicine)
• 🧫 HIT-CF study enabling treatment access for ultra-rare CFTR variants
• 🧫 2D-to-3D nasal models, endometrium organoids

Physiological Models (5 categories)

• 🔬 CF airway-on-a-chip - recapitulates mucus accumulation, cilia density, beating frequency
• 🔬 Air-liquid interface (ALI) cultures - polarized, differentiated epithelia
• 🔬 NuLi-1/CuFi cell lines - genotype-specific ion channel physiology
• 🔬 Multi-cell co-cultures (epithelial + immune + stromal)
• 🐁 Large animal models (rat, ferret, pig) - better recapitulate human pathology than mice

Deliverable

Created comprehensive research document: docs/research/cystic_fibrosis_datasets.md

Contents:

• Executive summary and dataset priorities
• Detailed descriptions of each dataset/model with key findings
• Clinical translation pathways (organoid theratyping)
• Mechanistic insights from models
• Integration recommendations
• Future directions (multi-omics, advanced NAMs, AI/ML)
• 30+ peer-reviewed source citations

Pull Request

📋 PR #303: #303

The PR includes:

• Full research documentation (343 lines)
• Organized by research focus area
• All sources linked and verified
• Ready for review and integration

Key Highlights for CF Research

1. Best Single-Cell Resource: Nature Medicine 2021 consortium study (19 CF vs 19 healthy, comprehensive cell states)
2. Best NAM for Personalized Medicine: Nasal epithelial organoids (non-invasive, validated drug prediction)
3. Best Physiological Model: Airway-on-a-chip (organ-level fidelity with mucus/cilia phenotypes)

Let me know if you need additional details on any specific dataset or model system!