Organ Fibrosis Collaboration (Saez-Rodriguez Lab)

In progress

Organ Fibrosis Collaboration (Saez-Rodriguez Lab)

Overview

Integrate the dismech knowledge base with the cross-organ fibrosis atlas from the Saez-Rodriguez lab (Heidelberg / EMBL-EBI), an open single-cell resource mapping shared and organ-specific fibrotic gene programs across heart, liver, kidney, and lung.

The atlas provides quantitative molecular data (gene signatures, TF activities, cell-cell communication, fibroblast subpopulations) while dismech provides structured, evidence-backed disease mechanism context (causal chains, phenotypes, treatments, clinical trials). The two are complementary: the atlas answers "what changes?" and dismech answers "why it matters clinically."

The Organ Fibrosis Atlas

Resource: https://organfibrosis.saezlab.org/ Preprint: Küchenhoff, Kim, Lanzer, Kretzler, Ramirez Flores & Saez-Rodriguez. "Shared and organ-specific gene expression programs of fibrotic diseases." bioRxiv 2026.03.09.709232v1. Code: https://github.com/saezlab/organfibrosis License: CC-BY-NC 4.0 (data), GPL-3.0 (code)

What it provides

Aspect Detail
Scale >5 million cells, 20 studies, 25+ disease etiologies
Organs Heart, liver, kidney, lung
Data type Single-cell / single-nucleus RNA-seq meta-analysis
Key outputs Cross-organ gene expression programs, disease fibroblast subpopulations with ECM signatures, organ-specific vs conserved TF activity profiles, intercellular communication patterns
Pipeline Snakemake, standard scRNA-seq tooling (likely scanpy/decoupler/liana ecosystem)
Interactive Dash-based web app for gene-of-interest queries across fibrotic tissues

Existing dismech coverage

Disorders with explicit fibrotic mechanisms (21 files mention myofibroblasts/stellate cells)

Disorder Organ Fibrosis mechanism
Liver Cirrhosis Liver HSC → myofibroblast transdifferentiation, collagen/ECM deposition
Primary Myelofibrosis Bone marrow TGF-β/PDGF/bFGF → reticulin/collagen deposition
Systemic Sclerosis Skin, lung, heart, kidney Multi-organ fibrosis via myofibroblast activation
Chronic Kidney Disease Kidney Tubulointerstitial fibrosis
Hepatitis B/C Liver Viral-driven hepatic fibrosis
Heart Failure Heart Cardiac fibrosis/remodeling
Dilated Cardiomyopathy Heart Myocardial fibrosis
Hypertensive Heart Disease Heart Pressure-overload cardiac fibrosis
Chronic Pancreatitis Pancreas Pancreatic stellate cell activation
Bird Fancier's Lung Lung Hypersensitivity pneumonitis → pulmonary fibrosis
IgG4-Related Disease Multi-organ Storiform fibrosis
Crohn Disease GI tract Intestinal stricturing/fibrosis
Polycystic Kidney Disease Kidney Cystic renal fibrosis
Cystic Fibrosis Lung, pancreas, liver Epithelial-driven (not fibroblast-centric)

Broader fibrosis mentions

Key gap: no Idiopathic Pulmonary Fibrosis (IPF) entry

IPF is the prototypical fibrotic lung disease and a major focus of the organ fibrosis atlas. Creating this entry would be a high-value first deliverable.

Integration strategy

Value exchange

dismech → Atlas Atlas → dismech
Curated disease mechanisms with causal chains and PMID evidence Quantitative gene signatures per cell type per organ
Ontology-grounded cell types (CL), processes (GO), genes (HGNC) Disease fibroblast subpopulation definitions
Treatment-mechanism linkages (drug → target pathway → phenotype) Cross-organ conserved vs organ-specific programs
Clinical trial connections to fibrotic phenotypes TF activity profiles and intercellular communication data
Comorbidity trajectories between fibrotic diseases Validated gene sets for computational model enrichment

Alignment points

The dismech schema already supports everything needed for integration:

  1. Cell types → CellTypeDescriptor with CL ontology. Atlas fibroblast subpopulations (e.g., ECM-producing disease fibroblasts) can be mapped to CL terms and added to pathophysiology entries.

  2. Gene signatures → GeneDescriptor with HGNC. Cross-organ fibrosis genes identified by the atlas can populate genes slots in pathophysiology entries.

  3. Biological processes → BiologicalProcessDescriptor with GO. Conserved fibrotic programs (ECM organization, TGF-β signaling, wound healing) already have GO terms.

  4. Datasets → Dataset class with SINGLE_CELL_RNA_SEQ type. Atlas study accessions (GEO/Zenodo) can be referenced directly.

  5. Computational models → ComputationalModel class. The Saez lab's multi-omics network models and TF activity inference could be represented.

  6. Cross-organ comparison → The comorbidities field and Disease Trajectories integration can capture cross-organ fibrotic progression patterns.


Tasks

Tier 1: Low Effort (data annotation, new entries)

Tier 2: Medium Effort (new entries, deeper annotation)

Tier 3: High Effort (schema-level, cross-cutting)

Tier 4: Aspirational (collaborative research)


Cross-project synergies

Project Synergy
VIRTUAL_CELL GeneProgram class needed for both; CELLxGENE datasets overlap with atlas source data; scGPT embeddings for fibroblast subpopulations
COMORBIDITIES Cross-organ fibrotic progression (liver → cardiac, kidney → cardiac) informed by both EHR trajectories and atlas shared programs
GWAS_MECHANISMS Fibrosis GWAS loci can be mapped to atlas cell-type-specific expression to assign mechanism context

Contact / Collaboration

Notes

2026-03-12 (Project Creation)

Project created to explore collaboration with Saez-Rodriguez lab organ fibrosis atlas.

Key observations: 1. The preprint was posted March 9, 2026 — only 3 days ago. This is an ideal time to reach out for collaboration before publication. 2. dismech already has substantial fibrotic disease coverage (21 files with myofibroblast/stellate cell mechanisms, 71 files mentioning fibrosis). 3. The atlas covers 4 organs (heart, liver, kidney, lung) — dismech has entries for fibrotic diseases in all 4 but lacks IPF, the archetypal pulmonary fibrosis. 4. Schema is fully ready: CellTypeDescriptor, GeneDescriptor, Dataset (with scRNA-seq type), ComputationalModel all exist. No schema changes needed for Tier 1-2 work. 5. The Saez lab ecosystem (decoupler for TF/pathway activity, liana for cell-cell communication, liver disease atlas, ReHeaT) provides multiple integration points beyond the organ fibrosis atlas itself. 6. CC-BY-NC 4.0 license on the data is compatible with knowledge base annotation (citing/referencing, not redistributing raw data).