FluxSprout: Single-cell metabolic modeling of vascular sprouting in zebrafish
- 1Iowa State University of Science and Technology: Iowa State University
Abstract
Genome-scale metabolic modeling offers a powerful framework for identifying cell-state-specific metabolic vulnerabilities, yet its application to heterogeneous biological systems remains limited by uncertainty in network annotations, thermodynamic feasibility, transcriptome-to-flux inference, and target specificity. Here, we introduce FluxSprout, a single-cell transcriptomics-informed genome-scale metabolic modeling framework to resolve metabolic heterogeneity during zebrafish sprouting angiogenesis. By integrating single-cell RNA sequencing of 10,090 quality-controlled cells with Zebrafish-GEM, we reconstructed network ensembles for 14 endothelial and hemogenic states. We established four safeguards: stoichiometric auditing, loop auditing, cell-cycle-controlled testing, and ensemble specificity screening. First, we identified four inorganic pyrophosphatase reactions misannotated as oxidative phosphorylation, showing that apparent pathway essentiality arose only with unbounded exchange fluxes. Second, 19.5% of reactions could carry thermodynamically infeasible loop flux, including every candidate of a conventional screen. Third, tip-like and stalk-like cells differed mainly in cell-cycle phase, and elevated glycolysis separated sprouting from established vessels only in gene expression. In contrast, sprouting vessels had higher predicted growth capacity at both constraint strengths, with serine and nucleotide synthesis following in fixed proportion. Fourth, screening 825 genes across four network-extraction settings and 13 non-endothelial lineages left no candidate called by more than one setting. Notably, SLC25A20, that screen's lead candidate, had no predicted effect in any FASTCORE network. Although the data derive from a single pooled clutch and require experimental validation, FluxSprout provides a transferable, uncertainty-aware framework for distinguishing genuine metabolic dependencies from artifacts of network reconstruction, model assumptions, and cellular heterogeneity, and identifies biosynthetic capacity as a testable feature of sprouting vessels.
Related
- Emerging Light-Sheet Fluorescence Microscopy for Cardiac Vascular Imaging in Zebrafish.
- 628 Vascular-metabolic checkpoint derives immunotherapy resistance in solid tumor
- Neuro-Metabolic and Vascular Dysregulation are Functional Biomarkers of Acquired Brain Injury
- An in vitro System for Studying Osteochondrogenic Differentiation of Smooth Muscle Cells and Modeling Intimal Vascular Calcification
- Breeding Utilization of the White-Grained Pre-Harvest Sprouting-Resistant Wheat Landrace Tuotuomai Through Marker-Assisted Selection
BibTeX
@misc{zargar2026fluxsprout,
title = {{FluxSprout: Single-cell metabolic modeling of vascular sprouting in zebrafish}},
author = {M. r. Zargar and R. Chowdhury},
howpublished = {bioRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.24.761567},
url = {https://www.biorxiv.org/content/10.64898/2026.10.24.761567v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - FluxSprout: Single-cell metabolic modeling of vascular sprouting in zebrafish
AU - M. r. Zargar
AU - R. Chowdhury
PY - 2026
DA - 2026/10/24
PB - bioRxiv
DO - 10.64898/2026.10.24.761567
UR - https://www.biorxiv.org/content/10.64898/2026.10.24.761567v1
AB - Genome-scale metabolic modeling offers a powerful framework for identifying cell-state-specific metabolic vulnerabilities, yet its application to heterogeneous biological systems remains limited by uncertainty in network annotations, thermodynamic feasibility, transcriptome-to-flux inference, and target specificity. Here, we introduce FluxSprout, a single-cell transcriptomics-informed genome-scale metabolic modeling framework to resolve metabolic heterogeneity during zebrafish sprouting angiogenesis. By integrating single-cell RNA sequencing of 10,090 quality-controlled cells with Zebrafish-GEM, we reconstructed network ensembles for 14 endothelial and hemogenic states. We established four safeguards: stoichiometric auditing, loop auditing, cell-cycle-controlled testing, and ensemble specificity screening. First, we identified four inorganic pyrophosphatase reactions misannotated as oxidative phosphorylation, showing that apparent pathway essentiality arose only with unbounded exchange fluxes. Second, 19.5% of reactions could carry thermodynamically infeasible loop flux, including every candidate of a conventional screen. Third, tip-like and stalk-like cells differed mainly in cell-cycle phase, and elevated glycolysis separated sprouting from established vessels only in gene expression. In contrast, sprouting vessels had higher predicted growth capacity at both constraint strengths, with serine and nucleotide synthesis following in fixed proportion. Fourth, screening 825 genes across four network-extraction settings and 13 non-endothelial lineages left no candidate called by more than one setting. Notably, SLC25A20, that screen's lead candidate, had no predicted effect in any FASTCORE network. Although the data derive from a single pooled clutch and require experimental validation, FluxSprout provides a transferable, uncertainty-aware framework for distinguishing genuine metabolic dependencies from artifacts of network reconstruction, model assumptions, and cellular heterogeneity, and identifies biosynthetic capacity as a testable feature of sprouting vessels.
ER -