Cell-Type Deconvolution of Equine BALF RNA-Seq: A Critical Comparison with Matched Single-Cell Data
Genes · 30.06.2026 · 17 (7) · 773
DOI: 10.3390/genes17070773 ↗Kurz gefasst
Vergleichsstudie in der Zeitschrift Genes, online erschienen am 30.06.2026. Verfasst von Vidhya Jagannathan und 3 weiteren Autor:innen. Beteiligte Einrichtung: Institute of Genetics, Vetsuisse Faculty, University of Bern, 3012 Bern, Switzerland. Hauptthemen laut MEDLINE: Bronchoalveolar Lavage Fluid, RNA-Seq, Asthma, Single-Cell Analysis. Die Arbeit nennt 55 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Swiss National Science Foundation, Internal Research Fund of the Swiss Institute of Equine Medicine.
Abstract (Originalsprache)
Background/Objectives: Bulk RNA sequencing (RNA-seq) averages signals across heterogeneous cell populations. Computational deconvolution methods aim to infer cell type composition and cell type-specific gene expression from bulk data, but their performance in equine samples has not been evaluated. In this study, we assessed the ability of computational deconvolution to recover cellular composition and differential expression signals in bronchoalveolar lavage fluid (BALF) from horses with severe equine asthma (SEA) and controls (CTL).
Methods
Cryopreserved BALF samples from six SEA and five CTL horses previously analyzed by scRNA-seq were used to generate bulk RNA-seq data. The matched scRNA-seq dataset served as the reference for deconvolution. Performance was evaluated by comparing deconvolution raw and mRNA-corrected estimates with scRNA-seq cell proportions. Differential expression between SEA and CTL was analyzed on bulk RNA-seq, deconvoluted expression profiles, and scRNA-seq pseudobulk data.
Results
Deconvolution primarily captured mRNA-derived cell type proportions rather than true cell counts: agreement with scRNA-seq cell counts was moderate (r = 0.62; 95% CI 0.45–0.75) but improved after mRNA content correction (r = 0.83; 95% CI 0.74–0.89). Comparison with mRNA-weighted scRNA-seq proportions showed near-perfect concordance (r = 0.98; 95% CI 0.97–0.99). Cell type-specific performance varied, with stronger correlations for B cells and dendritic cells and weaker performance for neutrophils, T cells and monocytes/macrophages. Recovery of cell type-specific differential expression was inconsistent, frequently showing cross-lineage signal spillover. Although both approaches detected a Th17 signature in SEA, most deconvolution-derived differentially expressed genes overlapped with conventional bulk RNA-seq results.
Conclusions
Deconvolution of bulk RNA-seq did not reliably estimate cell counts or provide substantial biological insight beyond conventional bulk analysis, highlighting the value of scRNA-seq for resolving cell type-specific disease mechanisms in equine asthma.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Substanzen
Autor:innen und Institutionen
- Institute of Genetics, Vetsuisse Faculty, University of Bern, 3012 Bern, Switzerland.
- Next Generation Sequencing Platform, University of Bern, 3012 Bern, Switzerland.
- Swiss Institute of Equine Medicine, Department of Clinical Veterinary Medicine, Vetsuisse Faculty, University of Bern, 3012 Bern, Switzerland.
ORCID-Profile: Vidhya Jagannathan, Tosso Leeb, Sophie E Sage
Förderung
- Swiss National Science Foundation
- Internal Research Fund of the Swiss Institute of Equine Medicine
Zitieren
APA
Jagannathan, V., Leeb, T., Gerber, V., & Sage, S. E. (2026). Cell-Type Deconvolution of Equine BALF RNA-Seq: A Critical Comparison with Matched Single-Cell Data. Genes, 17(7), 773. https://doi.org/10.3390/genes17070773
Vancouver
Jagannathan V, Leeb T, Gerber V, Sage SE. Cell-Type Deconvolution of Equine BALF RNA-Seq: A Critical Comparison with Matched Single-Cell Data. Genes. 2026;17(7):773. doi:10.3390/genes17070773
BibTeX
@article{Jagannathan2026,
author = {Jagannathan, Vidhya and Leeb, Tosso and Gerber, Vinzenz and Sage, Sophie E},
title = {Cell-Type Deconvolution of Equine BALF RNA-Seq: A Critical Comparison with Matched Single-Cell Data},
journal = {Genes},
year = {2026},
volume = {17},
number = {7},
pages = {773},
doi = {10.3390/genes17070773}
}
EUVIMED