🧬 Shared and organ-specific gene expression signatures of fibrotic diseases

🔍 Overview

This repository contains the code corresponding to our manuscript:

citation for preprint

We conducted a large-scale meta-analysis of single-cell transcriptomic data from human healthy and fibrotic tissues to identify both shared and organ-specific transcriptomic profiles. Using datasets from the heart, kidney, lung, and liver, we constructed a single-cell fibrosis atlas of over five million cells from 20 studies, covering more than 25 etiologies across four organs.
Through systematic comparison of these datasets, we identified organ-specific as well as cross-organ fibrotic gene expression profiles in major cell types and disease fibroblast subpopulations, characterized by the excessive production of extracellular matrix, revealing a shared fibrotic response across tissues.

💻 Code overview

The code in structured into a snakamake pipeline. Configurations are found in the /profile directory. Snakemake rules, envrionments and analysis code is found in the /workflow directory.

The code was modulatized into the following modules:

• preprocessing
  • processing & harmonization of raw datasets
• analysis
  • core analysis of scRNA/snRNA datasets
• integration
  • GPU-dependent steps (for cluster configurations, see /profile/slurm2 - contains cell type annotation transfer and mesenchymal cell integration
• myofib
  • processing & analysis of disease fibroblast subset of the data
• spatial
  • preprocessing & analysis of spatial datasets
• plotting
  • final processing and visualization of data
• dataformat
  • formatting of data for publication

📂 Where to get the data

Non-processed data is found in the publications of indiviual datasets. Processed pseudobulks and analysis results can be accessed in zenodo:

zenodo link

🌐 User-interface to explore the data

Have a look at our interactive website where you can look up how your gene of interest behaves in fibrotic disease tissues:

website link