ScienceSystems BiologyPeer-reviewed journal article
Whole-body single-cell atlas of an adult vertebrate in homeostasis and regeneration
Kutay Deniz AtabayiD, Patrick AoudeiD, Chanyoung ParkiD, Mykola KadobianskyiiD, Olivier H. Paugois, Benjamin JudkewitziD, Peter W. ReddieniD
In-depth summary
Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 2 (17 Jul 2026) · CC BY 4.0
Region-stratified scRNA-seq of 607,370 cells plus 500-gene MERFISH maps adult Danionella cerebrum cell types body-wide, revealing fibroblast Hox and limb-identity programs and telencephalon regeneration dynamics.
No adult vertebrate had its full anatomy mapped at single-cell transcriptomic resolution, partly because common vertebrate models are too large and complex for whole-body molecular coverage. Using the roughly 12 mm, transparent, regenerative teleost Danionella cerebrum, the work supplies a spatially anchored whole-body atlas plus a telencephalon injury time course as a reference for neuroscience, regeneration, and evolution.
Key results
ShownA whole-body scRNA-seq atlas of adult Danionella cerebrum resolves cells from all major tissue classes and recovers very rare populations, supporting near-saturation coverage of the transcriptomic states that survive dissociation.
607,370 cells passed QC across 12 samples, forming 40 broad classes; subclustering gave neural 390,222 cells/95 clusters, epithelial 92,214/47, connective 110,902/42, muscle 10,038/27. pth, rln3a, avp, galn+ neurons (30 to 150 cells CNS-wide) were detected. DESeq2 confirmed 95.7% (internal organs) and 85.4% (CNS) of top-20 Seurat markers. · Single-cell RNA sequencing and profiling cell…; Neural tissues; Non-neural tissues; 10X single-cell mRNA sequencing
ShownAdult connective tissue, especially fibroblasts, keeps regional Hox expression along the body axis, and fin fibroblasts persistently express fin-type-appropriate limb-identity factors, consistent with constitutive adult positional information.
Among 31,756 trunk fibroblasts (19 clusters), Hox genes were top anterior-vs-posterior DE genes; epcam+ epithelial and myod1+ muscle cells lacked this. Pectoral (vs anal) fin enriched tbx5a, lmx1b, hoxc8a; pelvic (vs pectoral) tbx4, pitx1; caudal (vs pectoral) hoxc12a, hoxb13a, hoxc13a; dorsal (vs anal) zic1. HCR-FISH showed tbx5a, hoxc9a, lmx1b in pectoral fin. · Constitutive adult positional information and…; Figure legends
SuggestedAcross the adult CNS, neuronal transcriptomic similarity tracks brain region more than neurotransmitter class, and SAMap links Danionella neuron and glia clusters to mouse and zebrafish counterparts largely by regional or developmental identity.
About 269,000 CNS neurons were collapsed into 206 metacells; hierarchical clustering on PCs of 3000 HVGs split first by region, with neurotransmitter classes nested within region blocks; batch and ARI checks also favored region, especially for glutamatergic cells. GABAergic clusters matched mouse 09-CNU-LGE and 08-CNU-MGE classes; telencephalic glutamatergic clusters matched zebrafish Dp and Dm. · Regional identity drives neuronal transcriptomic…; CNS Region and neurotransmitter class comparison…; Variance partitioning (PERMANOVA) test; Cross-species analysis
ShownUnilateral telencephalon injury triggers a staged program on the injured side (wound and myeloid response, radial glia proliferation, remodeling, circuit maturation) and a distinct metabolic/stress response contralaterally, with new glutamatergic neurons integrating.
127,168 cells in 53 clusters across 5 timepoints (12 hours to 30 days); acute changes concentrated between 12 hours and 7 days, and the contralateral side lacked myeloid recruitment. HCR-FISH confirmed upregulated fabp11a, fabp7a, abcd4a and epd (n=4 for the day 3 panel); EdU labeling marked mature slc17a6a+ glutamatergic neurons in the injured hemisphere. · Regenerative dynamics in the Danionella cerebrum…; Brain injury experiments and regeneration…; Figure legends
SuggestedAdult Danionella cell types occupy intermediate, mosaic developmental positions between larval and adult zebrafish references, with larval-like blood and cartilage programs alongside adult-like cardiomyocytes and germ cells.
Spearman-based developmental-position scoring against the ZMAP larval atlas (0 to 5 days post-fertilization) and 11 pooled adult zebrafish datasets placed Danionella globally between them; erythroid and Schwann cells scored larval-like, cardiomyocytes and germ cells adult-like; embryonic hemoglobins (hbbe2, hbae1.3) and cartilage genes col2a1b and matn4 persisted. · Molecular analysis of paedomorphic features…; Comparative paedomorphic comparison analysis; Developmental-position scoring (per-cell-type…; Figure legends
Methods, data and reproducibility
Cells from 12 microdissected body regions were profiled separately by 10x Genomics scRNA-seq, binned anatomically, and merged into neural, epithelial, connective, and muscle classes for Seurat/Leiden subclustering; MERFISH and HCR-FISH localized markers, alongside SAMap cross-species mapping, Monocle3 pseudotime, and a telencephalon injury time course.
- Whole-body adult scRNA-seq atlas from 12 microdissected regions (brain divisions, eyes, internal organs, cranium, trunk, fins) · Danionella cerebrum · n = 607,370 cells after QC · SRA PRJNA1402358; Zenodo record 18236172
- Telencephalon unilateral-injury regeneration time course, injured and contralateral sides (12 hours to 30 days) · Danionella cerebrum · n = 127,168 cells; 5 timepoints · SRA PRJNA1402358
- Additional fin scRNA-seq: pelvic fins and physically separated pectoral fin regions · Danionella cerebrum · SRA PRJNA1402358
- MERFISH whole-body and brain cryosections (16 µm) with a 500-gene panel, plus whole-mount HCR-FISH · Danionella cerebrum · Interactive portal: digifish.wi.mit.edu/atlas
- Allen Brain Cell Atlas mouse whole-brain 10Xv3 data (20250531 release), subsampled to 25% of cells per cluster for SAMap · mouse
- Adult zebrafish telencephalon scRNA-seq (Anneser et al., 2024) used for SAMap comparison · zebrafish
- ZMAP larval zebrafish atlas (0 to 5 days post-fertilization) used as the larval reference · zebrafish
- Adult zebrafish reference assembled from independently generated, published single-cell datasets · zebrafish · n = 11 datasets
- Methods & tools
- 10x Genomics Chromium v3.1 (HT and standard), MERFISH (Vizgen MERSCOPE), HCR-FISH, F-ara-EdU labeling, Cell Ranger v7.2.0, CellBender 0.3.0, scDblFinder v1.12.0, Seurat v5.0.3, DESeq2, SAMap v2.0.2 with eggNOG-mapper v2.1.12, Monocle3 v1.4.26, vegan adonis2 PERMANOVA, variancePartition, Harmony
- Data & accessions
- https://zenodo.org/records/18236172
- Reagents & environment
- Cell Ranger v7.2.0, CellBender 0.3.0, Seurat v5.0.3, scDblFinder v1.12.0, Monocle3 v1.4.26, SAMap v2.0.2, eggNOG-mapper v2.1.12, ComplexHeatmap v2.14.0, gffread v0.12.7; no container or environment file described.
- Reproducibility gaps
- Read mapping used a Danionella cerebrum genome scaffold (FASTA/GFF3) generated by the Judkewitz Lab, for which no accession or download location is given.
Limitations (6)
- Near-saturation is claimed only for states that withstood enzymatic dissociation and droplet capture; cell types lost during tissue processing would be missing from the atlas despite recovery of rare neuropeptidergic and ciliated populations. · stated by the authors
- In the CNS data, brain region is partially correlated with 10x chemistry/experiment, so region-versus-neurotransmitter conclusions rest on cross-checks with uncorrected and Harmony-integrated embeddings. · stated by the authors
- The adult zebrafish reference pools heterogeneous datasets, including telencephalon-only brain data and skin sci-RNA-seq from the squamation transition rather than mature adults, which the authors flag as relevant when reading developmental-position scores. · stated by the authors
- Permutation p-values for regeneration cluster-composition (Jensen-Shannon divergence) tests are floor-limited by narrow null distributions, so they only confirm non-random structure and were not used to rank clusters. · stated by the authors
- Fin- and trunk-specific expression of tbx5a, tbx4, pitx1, and Hox genes is correlative; no perturbation tests whether these factors encode adult positional memory or direct fin-type-appropriate regeneration. · noted in this summary
- CNS libraries A1–A12 are aliquots split from regional pools of 20 animals, so DESeq2 marker confirmation that treats library preparations as replicates rests partly on technical rather than biological replication. · noted in this summary
Read this if you study vertebrate cell-type evolution, adult positional memory in fins or limbs, or neural regeneration and want a whole-body single-cell and MERFISH reference for a transparent miniature fish.
- Citations
- 6
- FWCI
- 12.26
- Field percentile
- top 1%
Abstract
A complete cell-type transcriptome atlas of a vertebrate could promote understanding of animal cell-type composition, organization, and evolution. The miniaturized, transparent, and regenerative teleost Danionella cerebrum brings whole-vertebrate single-cell profiling experiments within reach. We performed region-stratified single-cell RNA sequencing across the adult Danionella body and mapped cell types and gene expression at single-cell resolution using spatial transcriptomics. We delineated spatially-distinct neural cell types based on their regional gene expression signatures. The body-wide atlas elucidated cell types harboring adult positional information, uncovered paedomorphic features, and revealed conserved body-region and appendage-specification programs in adult connective tissue. Comparative analyses revealed conserved neural cell types, and regeneration datasets uncovered expression dynamics during telencephalon regeneration. This whole-vertebrate transcriptome atlas yields a comprehensive resource for myriad questions in biology and neuroscience.
- organism: multiple species
- organism: zebrafish
- assay: scRNA-seq
- assay: spatial transcriptomics
- assay: imaging
- research focus: cell type annotation
- research focus: database / resource
- research focus: trajectory inference
- research focus: phylogenetics / evolution
- approach: experimental assay
- approach: statistical model
Reproducibility ledger
Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.
Sign in to file a reproducibility report.
Related papers
- A whole-body cell type atlas mapped into an electron microscopy volume of an annelid worm
Oel Adam Phillip, Cyril Cros, … Detlev Arendt·bioRxiv·Oct 30, 2025
- Cnidarian cell type diversity revealed by whole-organism single-cell RNA-seq analysis
Arnau Sebé-Pedrós, Elad Chomsky, … Heather Q. Marlow·bioRxiv·Oct 11, 2017
- Multimodal reference brain atlas of adult Danionella cerebrum
Mykola Kadobianskyi, Jörg Henninger, … Benjamin Judkewitz·bioRxiv·Mar 11, 2026
- Atlas of adult Danionella fish telencephalon reveals retention of larval characters
Jonathan T. Perelmuter, Andrew H. Bass·bioRxiv·Sep 29, 2026
- Spatiotemporal transcriptome at single-cell resolution reveals key radial glial cell population in axolotl telencephalon development and regeneration
Xiaoyu Wei, Sulei Fu, … Ying Gu·bioRxiv·Oct 23, 2021
- Spatially resolved cell atlas of the teleost telencephalon and deep homology of the vertebrate forebrain
Brianna E. Hegarty, George W. Gruenhagen, … Jeffrey Todd Streelman·Communications BiologyPeer-reviewed·May 21, 2024