A velocity-informed framework for resolving functional stratification in rare human stem cells.

A velocity-informed framework for resolving functional stratification in rare human stem cells - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Search Log in Primary site navigation Search Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice Leukemia . 2026 Apr 21;40(6):1284–1295. doi: 10.1038/s41375-026-02956-9 A velocity-informed framework for resolving functional stratification in rare human stem cells Justyna Jarczak 1 , Mariusz Z Ratajczak 1, 2 , Magdalena Kucia 1, ✉ Author information Article notes Copyright and License information 1 Laboratory of Regenerative Medicine, Medical University of Warsaw, Warsaw, Poland 2 Stem Cell Institute at Graham Brown Cancer Center, University of Louisville, Kentucky, KY USA ✉ Corresponding author. Received 2025 Oct 28; Accepted 2026 Mar 27; Issue date 2026. © The Author(s) 2026 Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ . PMC Copyright notice PMCID: PMC13233286  PMID: 42014860 Abstract Single-cell transcriptomics has transformed our understanding of stem cell biology; however, the analysis of extremely rare and quiescent populations remains limited by the low cell numbers and functional inaccessibility. Here, we present a velocity-informed framework to resolve transcriptional stratification in ultra-rare stem cells, using very small embryonic-like stem cells (VSELs) from human umbilical cord blood as a proof-of-concept model. By integrating single-cell RNA sequencing with dynamical RNA velocity analysis, we identify two kinetically distinct VSEL subsets: a CD34⁺ fraction characterized by deep quiescence, and a CD133⁺ fraction exhibiting transcriptional priming toward lineage commitment. Pathway-level analyses further reveal state-specific programs of stress response, metabolic regulation, and developmental gene networks. Importantly, our approach highlights how kinetic modeling can uncover functional continua even in populations not amenable to conventional assays. This framework is broadly applicable to rare or controversial stem cell types, providing a generalizable strategy to dissect cellular heterogeneity and state transitions where direct functional validation is constrained. Subject terms: Haematopoietic stem cells, Cell signalling Introduction Very small embryonic-like stem cells (VSELs) are a rare, quiescent population of developmentally primitive cells identified in adult tissues, including bone marrow (BM) and umbilical cord blood (UCB). Initially described by Kucia et al. [ 1 – 3 ], VSELs are defined by their small size (3–5 µm), high nuclear-to-cytoplasmic ratio, and expression of key pluripotency markers such as OCT4, NANOG, SOX2, and SSEA-4. These cells are hypothesized to be remnants of the epiblast deposited in adult tissues during early embryogenesis. Multiple independent studies [ 4 – 16 ] have confirmed the presence of VSEL-like stem cells in human and murine tissues, their low abundance, and their potential to give rise to cells of all three germ layers in vitro and in vivo. However, because of their rarity and phenotypic similarity to hematopoietic progenitors, the biological relevance of VSELs has remained controversial. Classical flow cytometry-based approaches, while useful for phenotypic isolation (e.g., CD34⁺Lin⁻CD45⁻ or CD133⁺Lin⁻CD45⁻), offer limited insight into functional heterogeneity or lineage potential. Thus, there is a compelling need to apply high-resolution molecular techniques that can resolve the transcriptional complexity and developmental trajectories of VSELs at single-cell resolution. Single-cell RNA sequencing of rare stem cells The advent of single-cell RNA sequencing (scRNA-seq) has transformed our ability to profile rare and heterogeneous stem cell populations. Unlike bulk RNA-seq, which yields an averaged expression profile, scRNA-seq captures the gene expression landscape of individual cells, enabling the identification of distinct subpopulations, differentiation hierarchies, and cell state transitions. In the context of HSCs and other early progenitors, scRNA-seq has been instrumental in revealing developmental continua rather than discrete stages [ 17 , 18 ]. In our recent study [ 19 ], we applied scRNA-seq to human VSELs isolated from UCB, revealing transcriptional heterogeneity suggestive of lineage priming and pluripotent-like regulatory programs. However, as is typical of snapshot-based omics, scRNA-seq alone does not capture the temporal dynamics of transcription or directionality of fate decisions. This limitation is especially relevant for quiescent or slowly cycling stem cells like VSELs, whose plasticity may not be readily inferred from static transcriptomic profiles. RNA velocity: inferring dynamics from static snapshots To address this limitation, we turn to RNA velocity analysis—a computational approach introduced by La Manno et al. [ 20 ], which estimates the future transcriptional state of individual cells based on the relative abundance of spliced (mature) and unspliced (nascent) transcripts. By quantifying transcriptional kinetics, RNA velocity allows us to infer vector fields that describe the direction and magnitude of gene expression change within individual cells. This enables the reconstruction of developmental trajectories, lineage bifurcations, and differentiation timing, even from a single timepoint. RNA velocity has since been extended to dynamical modeling approaches [ 20 – 23 ], which estimate transcriptional rates of induction, repression, and decay on a gene-specific basis. These methods have been used to investigate diverse biological processes, including neuronal differentiation [ 24 ], pancreatic islet formation [ 25 ], lung epithelial regeneration [ 26 ], and hematopoietic cell state transitions [ 27 , 28 ]. Importantly, RNA velocity has shown value in studying stem cells and progenitors, where cells are transcriptionally dynamic but not always sharply separated by surface phenotype. Despite the utility of RNA velocity in characterizing stem cell dynamics, no prior study has applied this method to VSELs. This is a critical gap, given that VSELs have been proposed to occupy an early developmental niche with latent pluripotent capacity. Whether VSELs undergo active transcriptional shifts, lineage priming, or cell cycle entry—and how these processes differ between immunophenotypically distinct subsets—remains unknown. Importantly, the single-cell RNA-seq dataset analyzed in the present study has been previously reported and extensively validated [ 19 ]. In that work, transcriptomic quality control and molecular characterization of human umbilical cord blood–derived VSELs demonstrated pronounced heterogeneity, including subpopulations annotated to germline-associated compartments, regulation by parental imprinting, and activation of early developmental fate programs. These features are considered defining molecular hallmarks of bona fide VSELs and provided transcriptomic validation beyond surface-marker–based isolation. The present study builds directly upon this validated dataset and extends it by applying RNA velocity–based kinetic modeling to resolve transcriptional dynamics and state transitions within the previously characterized VSEL compartment. We applied the scVelo pipeline to estimate gene-specific velocity vectors, infer pseudotime trajectories, evaluate cell cycle activity, and identify dynamically regulated transcripts. By integrating splicing kinetics with transcriptomic clustering, we aimed to (i) reconstruct potential differentiation trajectories of VSELs; (ii) compare the dynamic transcriptomic behavior of CD34⁺ and CD133⁺ VSEL subsets; (iii) identify key genes driving early transcriptional transitions; and (iv) explore whether VSELs occupy a transcriptionally stable or transitional state. This study offers new insight into the kinetic architecture of VSELs, demonstrating that RNA velocity can uncover dynamic features of rare stem cell populations previously masked by static transcriptomic approaches. The framework we present here lays the foundation for future functional studies of VSEL biology and their role in tissue homeostasis and regeneration. Materials and methods Ethics statement and sample collection Human umbilical cord blood (hUCB) was collected from a healthy newborn at the Department of Obstetrics and Gynecology, Medical University of Warsaw. Ethical approval was obtained from the Bioethics Committee of the Medical University of Warsaw (Approval No. KB/50/2022). All procedures were conducted under the Declaration of Helsinki. Informed consent was obtained from the donor’s legal guardian. Due to anonymization, donor sex and ethnicity data were not available. Isolation and sorting of VSELs Blood was handled on ice, avoiding cell disruption. It was then diluted with PBS, carefully layered over Ficoll-Paque (GE Healthcare, Chicago, IL, United States), and centrifuged for 30 min at 400 x g at 4 °C. The mononuclear phase was collected, washed, and used for further analysis. In the next step, MNCs were stained with the following antibodies: Lineage (Lin) cocktail of antibodies, each FITC-conjugated: CD235a (clone GA-R2 [HIR2]), anti-CD2 (clone RPA-2.10), anti-CD3 (clone UCHT1), anti-CD14 (clone M5E2), anti-CD16 (clone 3G8), anti-CD19 (clone HIB19), anti-CD24 (clone ML5), anti-CD56 (clone NCAM16.2) and anti-CD66b (clone G10F5) (all BD Biosciences, San Jose, CA, United States); PE-Cy7-conjugated anti-CD45 (clone HI30, BioLegend, San Diego, CA, United States); APC-conjugated anti-CD133 (clone CD133, MiltenyiBiotec, Gladbach, Germany) and PE-conjugated anti-CD34 (clone 581, BioLegend, San Diego, CA, United States). Finally, cells were sorted on the MoFlo Astrios EQ cell sorter (Beckman Coulter, Brea, CA, United States) according to the strategy described in our previous studies [ 8 , 29 , 30 ]. Library preparation and sequencing Directly after sorting, cells were used for gel beads preparation with the use of Chromium X Controller (10X Genomics, USA) and Chromium Next GEM Chip G Single Cell Kit (10X Genomics, USA). Next, Chromium Next GEM Single Cell 3’ GEM, Library & Gel Bead Kit v3.1, and Single Index Kit T Set A (10X Genomics, USA) were used for library preparation according to the manufacturer’s guidelines. Single cell sequencing was run on Illumina NextSeq 1000/2000 (Illumina, San Diego, CA, USA) in P2 flow cell chemistry (200 cycles) with paired-end sequencing mode (read 1–28 bp, read 2–90 bp, index 1–10 cycles, index 2–10 cycles), assuming 25,000 reads per cell. Preprocessing and quality control Raw sequencing data (BCL files) were demultiplexed and converted to fastq files using the bcl2fastq (version v2.20.0.422) within the 10X Genomics Cell Ranger mkfastq pipeline (CellRanger version 7.2.0; https://www.10xgenomics.com/ ). Then, CellRanger count, and aggregate functions were used for the quality control and processing fastq data. Cells with less than 200 and more than 2500 transcripts, and those with more than 5% of mitochondria-related transcripts were excluded from the analysis. In addition to standard filtering criteria, transcriptomic identity of VSELs was evaluated in the context of prior molecular characterization of the same dataset, including expression of pluripotency-associated, early developmental, chromatin-regulatory, and imprinting-related gene programs, as well as low expression of mature hematopoietic lineage markers, as previously reported [ 19 ]. Downstream scRNA-seq analysis Sequencing results were mapped to a human genome GRCh38 (version 2020-A) acquired from the 10× Genomics website ( https://www.10xgenomics.com/support/software/cell-ranger/downloads ). Gene expression measurements for each cell were normalized by the total number of transcripts in the cell, multiplied by a default scale factor, and the normalized values were log transformed (“LogNormalize” method). The log-normalized data were reduced to the first 2000 most highly variable genes. Both VSELs samples (CD133+lin-CD45-, CD34+lin-CD45-) were analyzed separately with the use of CellRanger count pipeline (CellRanger Software, version 7.2.0) to perform alignment, filtering, barcode counting, and UMI counting. Cell trajectories using RNA velocity To obtain the relative abundance of spliced (mature) and unspliced (nascent) transcripts, we initially proceed velocyto pipeline [ 20 ] based on CellRanger outputs. Further analysis was then computed using scVelo [ 20 – 23 ] to recover and visualize RNA velocity. The pipeline includes loading the data, checking of spliced and unspliced counts, filtering and normalization, moment estimation, and RNA velocity computation. Velocities were computed using both steady-state and dynamical models. The main part of the analysis, based on the estimated RNA velocity concerned important genes e.g., pluripotency and stemness-associated, lineage specification and epigenetic regulators, that may help to explain the vector field and inferred lineages. We also identified velocities in cycling progenitors, analyzing the mean expression levels of phase (S and G2M) marker genes. Finally, we estimated both velocity latent time and pseudotime as well as the speed of differentiation and coherence of the vector field. Genes identified as dynamically regulated based on RNA velocity analysis were subjected to pathway enrichment analysis using Reactome. To further explore functional relationships among these genes, interactome networks (interactomes) were constructed and visualized in Cytoscape (version 3.10.4). All data processing and analysis steps were performed using the Python programming language (version 3.10) [ 31 ] within Jupyter Notebook [ 32 ], implemented in Anaconda (Anaconda Software Distribution, Anaconda Inc. 2024). The data sets supporting the results of this article are available in the Sequence Read Archive (SRA) ( https://www.ncbi.nlm.nih.gov/sra ) repository and assigned unique persistent identifiers: PRJNA1126429 ( https://www.ncbi.nlm.nih.gov/sra/PRJNA1126429 ) and PRJNA1128409 ( https://www.ncbi.nlm.nih.gov/sra/PRJNA1128409 ). Results RNA velocity of CD34+lin-CD45- cells and CD133+lin-CD45- cells RNA velocity analysis was performed on a transcriptomically validated VSEL dataset previously shown to display hallmark molecular features of very small embryonic-like stem cells, including developmental, germline-associated, and imprinting-regulated gene expression programs [ 19 ], comprising CD34⁺Lin⁻CD45⁻ and CD133⁺Lin⁻CD45⁻ cells isolated from umbilical cord blood. Single-cell transcriptomes were analyzed using standard scRNA-seq pipelines (Cell Ranger Software, version 7.2.0), followed by velocity modeling with velocyto and scVelo [ 20 – 23 ]. First, *.loom-format files containing annotated matrices of spliced and unspliced transcript counts per single cell were used as input for downstream RNA velocity analysis in scVelo (v0.3.2), implemented in Python (v3.9.13). Initial inspection of spliced and unspliced counts proportions revealed differences in transcriptional states between the two analyzed subsets (Fig. 1A, F ). For CD34⁺Lin⁻CD45⁻ cells, approximately 61% of transcripts were spliced, and 36% were unspliced, with the remainder ambiguous. CD133⁺Lin⁻CD45⁻ cells displayed a slightly higher splicing ratio (67% spliced, 30% unspliced), consistent with a higher proportion of transcriptionally stable cells. These ratios are within expected ranges for rare stem/progenitor cells but suggest divergent transcriptional kinetics between subsets. We also examined the variations on the cluster level to verify consistency in splicing efficiency. In both datasets, we identified clusters with higher proportions of spliced and unspliced transcripts (clusters 1, 3 and 11 in CD34⁺Lin⁻CD45⁻, and 7, 10, and 15 in CD133⁺Lin⁻CD45⁻). The lowest number of spliced transcripts were found at cluster 9 and 8, in CD34⁺Lin⁻CD45⁻ and CD133⁺Lin⁻CD45⁻, respectively (Fig. 1A, F ). High spliced, low unspliced levels of transcripts mean that the cell maintains a set of genes typical of its parental “identity” but does not initiate new transcriptional programs. However, it is essential to identify the specific genes for which these proportions are specified. Fig. 1. Visualization of transcriptional dynamics in very small embrionic-like stem cells (VSELs). Open in a new tab RNA velocity reveals transcriptional dynamics and cell-state transitions in CD34⁺Lin⁻CD45⁻ ( A – E ) and CD133⁺Lin⁻CD45⁻ ( F – J ) very small embryonic-like stem cells (VSELs). Proportion of spliced and unspliced mRNA transcripts within and between identified cell clusters, reflecting transcriptional activity within the CD34 + VSEL population ( A ) and CD133 + Lin − CD45 − ( B ). UMAP embedding of RNA velocity vectors illustrating the directionality of gene expression dynamics across individual cells within the CD34 + VSEL population ( B ) and CD133 + Lin − CD45 − ( G ). Arrows indicate inferred future transcriptional states, suggesting potential differentiation pathways and cell-state transitions. Differentiation speed, quantifying the rate of transcriptional change at the single-cell level within the CD34 + VSEL population ( C ) and CD133 + Lin − CD45 − ( H ). Vector coherence, measuring the consistency of RNA velocity direction across neighboring cells and highlighting coordinated lineage progression within the CD34 + VSEL population ( D ) and CD133 + Lin − CD45 − ( I ). Pseudotime trajectory reconstruction derived from RNA velocity, depicting progression along dynamic transcriptional programs for CD34 + VSEL population ( E ) and CD133 + Lin − CD45 − ( J ). Unspliced transcripts, which retain intronic sequences, typically constitute approximately 10–25% of the total transcript pool in single-cell RNA-seq datasets [ 25 ]. Observed variations in the ratio of unspliced to spliced RNA across distinct cell subpopulations are consistent with heterogeneity in transcriptional kinetics and RNA processing dynamics. Notably, cells at earlier developmental stages tend to exhibit lower proportions of unspliced transcripts, whereas more transcriptionally active or lineage-committed cells often display elevated unspliced-to-spliced ratios, indicative of increased transcriptional flux [ 25 ]. Velocity embeddings reveal putative lineage trajectories RNA velocity vectors were embedded onto a UMAP constructed directly from the spliced and unspliced count matrices (Fig. 1B, G ). This dynamic embedding strategy, rather than using pre-computed UMAPs, allowed us to highlight cell transitions informed by transcriptional activity rather than static expression states. In this analysis, we intentionally did not utilize pre-computed UMAP embeddings or predefined cluster annotations. Instead, we implemented dynamic UMAP embedding derived from RNA velocity estimates, incorporating both unspliced and spliced transcript abundances. This approach enabled the generation of velocity-informed clusters that more accurately reflect the directionality of cell state transitions and underlying lineage relationships. Unlike static embeddings based solely on spliced transcript abundance, this dynamic approach provides enhanced resolution of developmental trajectories and transient cellular states that may otherwise remain obscured. For both populations, the velocity fields showed a coherent directional flow among clusters, indicating putative developmental trajectories. For CD34⁺Lin⁻CD45⁻ cells, the velocity graph revealed a relatively short vector field with low directional acceleration, suggesting these cells occupy a stable, transcriptionally quiescent state with limited progression. In contrast, CD133⁺Lin⁻CD45⁻ cells showed longer and more divergent velocity vectors, particularly across two major clusters, indicating higher transcriptional activity and transition potential. Cell cycle-associated transcript dynamics To investigate whether proliferative activity contributes to the transcriptional heterogeneity of VSELs, we quantified the expression of canonical S-phase and G2/M markers and projected standardized scores onto velocity-informed UMAPs (Fig. 2A, B ). RNA velocity dynamics within proliferative progenitor populations were determined by evaluating phase-specific gene expression signatures. Velocity estimates were inferred based on standardized scores of average expression levels of established cell cycle phase markers. Distinct subclusters in both datasets displayed elevated expression of mitotic regulators such as CCNB1 and TOP2A, consistent with localized entry into cell-cycle progression (Figs. 3 A, B and 4A, B ). Fig. 2. Cell-cycle dynamics in very small embrionic-like stem cells (VSELs). Open in a new tab RNA velocity reveals transcriptional dynamics of cell cycle progression with emphasis on S phase and G2/M transition and inference of latent time along the CD34+lin-CD45- ( A , C ) and CD133+lin-CD45- ( B , D ) VSELs. UMAP projection of CD34 + Lin − CD45 − VSELs (A) and CD133 + Lin − CD45 − ( B ). illustrating cell cycle–associated transcriptional dynamics. Cells in S phase, marked by active DNA replication, display moderate transcriptional momentum. In contrast, cells in the G2/M phase exhibit increased velocity coherence, indicative of synchronized transcriptional programs linked to mitotic entry and progression. These patterns highlight the utility of RNA velocity in capturing dynamic transitions during cell cycle progression. Latent time analysis was used to reconstruct the transcriptional progression within and between identified stem cell clusters of CD34+lin-CD45- ( C ) and CD133+lin-CD45- ( D ). By leveraging RNA velocity–based modeling, we reconstructed a pseudotemporal ordering of single cells that reflects their intrinsic differentiation dynamics. Fig. 3. Expression of cell cycle–related genes, pluripotency and stemness genes, as well as fate decision genes overlaid on RNA velocity embeddings in CD34 + Lin − CD45 − very small embryonic-like stem cells (VSELs). Open in a new tab UMAP projections of RNA velocity embeddings displaying the expression patterns of key cell cycle–associated genes: CCNB1 ( A ) , TOP2A ( B ) , PCNA ( C ) , and TYMS ( D ); cell cycle regulation and mitotic progression: SOX17 ( E ) , KLF4 ( F ) , ZNF10 ( G ) , and POU2F1 ( H ); and fate decision ID1 ( I ) , LMO4 ( J ) , EZH1 ( K ) , and PRDM2 ( L ). Gene expression intensity is mapped onto the velocity landscape, highlighting spatial and dynamic transcriptional activity across the CD34⁺Lin⁻CD45⁻ VSEL population. The plot displays the relationship between gene expression levels and RNA velocity estimates for individual cells. Each point represents a single cell, with color indicating normalized expression of the selected gene denoting the predicted future transcriptional state based on spliced/unspliced transcript ratios. Fig. 4. Expression of mitosis-associated genes overlaid on RNA velocity embeddings in CD133 + Lin − CD45 − very small embryonic-like stem cells (VSELs). Open in a new tab UMAP projections of RNA velocity embeddings illustrating the expression patterns of selected genes involved in key cell cycle-associated genes: CCNB1 ( A ) , TOP2A ( B ) , AURKA ( C ) , and LMO4 ( D ); cell cycle regulation and mitotic progression genes: GATA2 ( E ) , FOXO1 ( F ) , KLF4 ( G ) , and POU2F1 ( H ) ; chromatin organization genes: TET2 ( I ) , DNMT3A ( J ) , HDAC4 ( K ) , and SMARCA5 ( L ). Expression levels are mapped onto the velocity landscape, revealing spatial heterogeneity and dynamic transcriptional activity associated with proliferative states within the CD133⁺Lin⁻CD45⁻ VSEL population. The plot displays the relationship between gene expression levels and RNA velocity estimates for individual cells. Each point represents a single cell, with color indicating normalized expression of the selected gene denoting the predicted future transcriptional state based on spliced/unspliced transcript ratios. CD34⁺Lin⁻CD45⁻ cells expressed, CCNB1 and TOP2A, particularly within low-pseudotime clusters (Fig. 3A, B ), indicating that a small fraction of otherwise quiescent cells engages the cell cycle early in the trajectory. Other proliferative markers, PCNA and TYMS were weakly expressed (Fig. 3C, D ), and their velocity graphs suggested steady-state expression, implying minimal active transcriptional upregulation. These features support the view that CD34⁺ VSELs remain largely transcriptionally stable, with only rare proliferative subsets. By contrast, CD133⁺Lin⁻CD45⁻ cells showed distinct expression of AURKA - a key G2/M checkpoint regulator (Fig. 4C ), accompanied by higher overall G2/M scores relative to CD34⁺ cells. Notably, S-phase markers remained comparatively low, suggesting that CD133⁺ cells are enriched for populations preparing for mitotic entry rather than undergoing active DNA replication. Together, these analyses indicate that CD34⁺ VSELs are largely maintained in a transcriptionally quiescent state with isolated proliferative outliers, whereas CD133⁺ VSELs constitute a more dynamically active, division-ready compartment poised for cell-cycle re-entry. Differentiation speed and vector confidence reveal stable versus transition competent VSEL states To evaluate the extent and reliability of transcriptional dynamics, we quantified two key RNA velocity–derived metrics: velocity length, serving as a proxy for the magnitude of transcriptional activity, and velocity confidence, reflecting the local consistency and smoothness of the vector field. These parameters provide insight into both the intensity and directional reliability of cell state transitions (Fig. 1 C–D, H–I ). Nonetheless, a subset of cells showed modest increases in velocity magnitude, particularly within clusters expressing CCNB1 and TOP2A, markers of G2/M phase progression, indicating localized transcriptional activation consistent with early proliferative engagement. These clusters also corresponded with intermediate pseudotime values (Fig. 1E ), supporting the interpretation that they may represent transitional states between naïve and lineage-committed phenotypes. In contrast, CD133⁺Lin⁻CD45⁻ cells exhibited markedly higher velocity lengths across multiple clusters (Fig. 1J ), indicative of robust transcriptional flux and enhanced differentiation potential. This dynamic activity was accompanied by substantial imbalances between unspliced and spliced transcripts in key genes such as SOX17, associated with endodermal specification, and AURKA, a regulator of mitotic entry, pointing to ongoing state transitions. Importantly, these transcriptionally dynamic clusters were aligned with higher pseudotime values (Fig. 1J ), reinforcing their identity as lineage-primed subpopulations. Elevated velocity confidence in both datasets (Fig. 1 D, I) confirmed that the inferred vector fields were robust and biologically meaningful, underscoring the contrast between the stable transcriptional landscape of CD34⁺ VSELs and the dynamic, transition-competent profiles of CD133⁺ VSELs. These findings prompted us to reconstruct pseudotime trajectories, enabling a more precise delineation of early and late transcriptional states within each VSEL subset. Pseudotime inference highlights early and late VSEL states To further resolve transcriptional progression within VSELs, we reconstructed pseudotime trajectories from the velocity-inferred graph (Fig. 1 E, J ). In CD34⁺Lin⁻CD45⁻ cells, most cells were positioned near the trajectory root, exhibiting uniformly low pseudotime values consistent with a deeply quiescent, primitive state. Only a small cluster shifted toward higher pseudotime, suggestive of rare cells initiating departure from dormancy and engaging lineage-commitment programs. By contrast, CD133⁺Lin⁻CD45⁻ cells demonstrated a bimodal pseudotime distribution. One subpopulation clustered near the trajectory origin, while another was shifted toward advanced pseudotime values, indicating transcriptional divergence within the compartment. This bifurcated distribution points to early lineage branching and a higher degree of developmental plasticity compared to CD34⁺ cells. These pseudotime dynamics, when integrated with the velocity-based measures of transcriptional activity and gene-specific regulation, further support a model in which both subsets retain early progenitor characteristics, but CD133⁺ VSELs exhibit greater transcriptional heterogeneity and are more developmentally advanced than their CD34⁺ counterparts. These findings motivated us to apply latent time analysis, allowing us to quantify the intrinsic pace of transcriptional transitions and compare activation kinetics between the two subsets. Latent time dynamics distinguish activation kinetics of CD34 + and CD133 + VSELs To quantify the intrinsic pace of transcriptional transitions, we applied latent time analysis, which integrates splicing kinetics into a continuous measure of differentiation progression (Fig. 2 C, D ). By leveraging RNA velocity–based modeling, we reconstructed a pseudotemporal ordering of single cells that reflects their intrinsic differentiation dynamics. In both stem cell populations, latent time increased monotonically along the inferred trajectories, consistent with a continuum from quiescent or early activated states toward lineage-committed states. Comparative analysis revealed differences in the distribution of latent times between the two populations, suggesting variation in the pace or extent of transcriptional progression. These differences may indicate distinct activation kinetics, heterogeneity in differentiation potential, or asynchronous responses to microenvironmental cues. This continuum provided a framework to examine gene-specific velocity portraits, enabling us to identify key regulators whose dynamic induction or repression underlie the observed heterogeneity in VSEL state transitions. Dynamic expression of regulatory genes reveals lineage priming in VSELs To dissect gene-level transcriptional dynamics within VSELs, we examined velocity phase portraits, which plot unspliced against spliced transcript abundances to distinguish actively induced, repressed, or steady-state genes. This approach enables kinetic interpretation beyond static expression, capturing whether regulatory genes are being transcriptionally engaged or silenced. These visualizations enable interpretation of gene-specific kinetics, distinguishing actively transcribed genes (high unspliced relative to spliced counts), transcriptionally repressed genes (low unspliced, high spliced), and those in transcriptional equilibrium or steady state (linear relationship between spliced and unspliced counts). Among pluripotency and lineage specification markers, SOX17 (Fig. 3E ) exhibited elevated unspliced-to-spliced transcript ratios in specific subpopulations of CD34 + VSELS, suggesting active transcriptional induction of this gene in cells undergoing early endodermal priming. Similar pattern for RNA velocity and expression was observed for KLF4 (Fig. 3F ), also in the second cell population CD133 + VSELs (Fig. 4G ). The presence of velocity vectors pointing toward these SOX17, and KLF4-high cells supports the notion that they may represent a transitional state along a differentiation trajectory. Among other pluripotency and stemness genes, we found POU2F1 to be expressed in both datasets (Figs. 3 H, 4H ), suggesting the presence of cell clusters, likely representing progenitor cells actively transitioning toward a POU2F1-driven fate (potentially lymphoid-biased progenitors). Subpopulations of CD34 + VSELs with the concurrent high expression and RNA velocity of ID1, EZH1, LMO4, and PRDM2 (Fig. 3I–L ) suggest an actively remodeling transcriptional state poised toward future fate decisions. ID1, a helix–loop–helix inhibitor, is known to restrain premature lineage commitment and sustain progenitor plasticity, while EZH1, a catalytic component of the Polycomb repressive complex, contributes to chromatin-mediated gene silencing and maintenance of stemness. LMO4 functions as a transcriptional adaptor that modulates lineage-associated gene networks, often in cooperation with other hematopoietic regulators, and PRDM2 is a histone methyltransferase implicated in cell cycle control and tumor suppression. The relationship of high expression and positive velocity values may indicate that these genes are not only abundantly transcribed but also dynamically upregulated, highlighting subpopulations of stem cells that are in transition toward lineage priming while maintaining features of self-renewal and epigenetic plasticity. Importantly, we noted the positive velocity values together with up-regulation of ID1 (Fig. 3I ) in CD34 + VSELs, which occur for the same cells that show a similar pattern for SOX17, likely undergoing early endodermal priming. Interestingly, in the case of CD133 + VSELs, we observed a unique pattern for the expression and RNA velocity for TET2, DNMT3A, HDAC4 and SMARCA5 (Fig. 4I–L ) in different subpopulations. It points to an active reprogramming of the epigenetic landscape during dynamic state transitions within these cells, strong RNA velocity and up-regulation of TET2 were found in opposite clusters compared to DNMT3A (Fig. 4J ). TET2 and DNMT3A act in opposing directions to regulate DNA methylation—TET2 promoting demethylation and epigenetic plasticity, while DNMT3A establishes de novo methylation patterns critical for lineage specification. The expression of HDAC4 was confirmed in the same cell cluster as for TET2 (Fig. 4I, K ). HDAC4, a histone deacetylase, contributes to chromatin condensation and transcriptional repression, while the expression of SMARCA5 (Fig. 4L ) was relatively evenly distributed throughout the population. SMARCA5 is an ATP-dependent chromatin remodeler that alters nucleosome positioning to enable or restrict access to regulatory elements. The simultaneous upregulation and high velocity of these genes suggest that CD133 + VSELs are not in a static maintenance phase but are instead undergoing rapid chromatin and DNA methylation remodeling, balancing repression, and accessibility to prime differentiation pathways. Together, these gene-specific velocity analyses confirm that transcriptional heterogeneity within VSELs is functionally structured, with subsets displaying dynamic activation of lineage-specifying and cell cycle–associated programs. Importantly, CD133⁺Lin⁻CD45⁻ cells consistently demonstrated greater transcriptional flux and phase-specific gene activation compared to CD34⁺Lin⁻CD45⁻ cells, reinforcing the conclusion that the former subset may represent a more transcriptionally plastic or lineage-primed population. Comparative reactome pathway enrichment reveals shared cores and subset-specific programs To extend gene-level kinetic insights to systems-level regulation, we performed Reactome pathway enrichment, which was then visualized and analyzed in Cytoscape as interactome networks, on the most variable, velocity-informed genes from both VSEL subsets. This analysis revealed a common core of transcriptional programs, including RUNX-mediated regulation, RHO GTPase signaling, chromatin remodeling (Fig. 5 ), and DNA mismatch repair (Table S1 ), underscoring a conserved stem/progenitor architecture across both populations. Fig. 5. Interactome profi ling of VSEL regulatory networks. Open in a new tab Interactome of key chromatin remodelers based on the pathways found in both CD34⁺Lin⁻CD45⁻ and CD133⁺Lin⁻CD45⁻ very small embryonic-like stem cells (VSELs). Interactomes were constructed using Cytoscape (version 3.10.4), which depicts protein–protein interactions centered on chromatin remodelers (A) identified as critical in the comparative analysis found in both CD34⁺Lin⁻CD45⁻ and CD133⁺Lin⁻CD45⁻. Interaction confidence thresholds were applied as described in “Methods”. Beyond this shared foundation, each subset displayed distinct pathway enrichments (Table S2 and S3 ). CD34⁺ VSELs were uniquely enriched for developmental signaling cascades (WNT, TGF-β, receptor tyrosine kinase), cell-cycle checkpoint control (p53/p21), FOXO-mediated transcription (Supplementary Fig. 1A ), and stress-response programs such as oxidative and oncogene-induced senescence (Table S2 ). These modules suggest a bias toward genome integrity, developmental regulation, and fate restriction. The CD34⁺ population was uniquely enriched for signaling programs linked to lineage specification and developmental signaling, including WNT, TGF-β, and receptor tyrosine kinase pathways, as well as oxidative stress–induced and cellular and oncogene-induced senescence modules. Notably, multiple FOXO-mediated transcriptional programs, cell cycle checkpoint pathways (p53 and p21 activation) and megakaryocyte differentiation signatures were exclusive to this subset, indicating a bias toward cell fate determination and stress-response regulation (Table S2 ). In contrast, the CD133⁺ population demonstrated unique enrichment in neuronal development–related modules, clathrin-mediated endocytosis, chromatin remodelers (Supplementary Fig. 1B ) and RNA surveillance mechanisms (e.g., nonsense-mediated decay), suggesting broader functional plasticity and engagement in immune–neuroregulatory crosstalk (Table S3 ). Quantitatively, approximately 38–42% of enriched pathways were unique to each population, while the remainder formed a conserved core signature. These findings highlight the molecular heterogeneity between CD34⁺ and CD133⁺ subsets and suggest that CD133⁺ cells possess an immune-primed, stress-adaptive transcriptomic profile, whereas CD34⁺ cells maintain a transcriptional program more closely aligned with developmental regulation and cellular homeostasis. Shared chromatin organization programs define a conserved epigenetic core in VSELs To determine whether CD34⁺Lin⁻CD45⁻ and CD133⁺Lin⁻CD45⁻ VSELs share common epigenetic regulatory mechanisms, we performed Reactome pathway enrichment focusing on genes associated with chromatin organization and genome regulation. Network-based visualization of Reactome-curated interactions revealed a highly interconnected core of chromatin-modifying factors common to both populations (Fig. 5 ). This shared network includes key components of histone acetylation and methylation machinery (KAT2A, KAT6A, KMT2C/D, KDM6A), chromatin remodeling complexes (ARID2, CHD3, SMARCA5), and transcriptional co-regulators (EP300, CREBBP), indicating preservation of fundamental epigenetic maintenance programs across VSEL subsets. The presence of this conserved chromatin organization module suggests that both CD34⁺ and CD133⁺ VSELs maintain a common epigenetic framework, upon which subset-specific transcriptional kinetics and lineage-priming programs are superimposed. Collectively, these results reveal transcriptionally encoded stratification within human VSELs, where CD34⁺ cells function as a deeply quiescent reservoir and CD133⁺ cells constitute a transcriptionally plastic, lineage-primed compartment. Discussion Resolving how rare stem cells balance dormancy with activation is a central question in developmental and regenerative biology. Here, by applying RNA velocity to human VSELs we uncover a structured continuum of transcriptional states that stratifies this elusive population into functionally distinct tiers. CD34⁺ VSELs emerge as a deeply quiescent reservoir with minimal transcriptional flux, whereas CD133⁺ VSELs constitute a transcriptionally dynamic, lineage-primed compartment enriched for mitotic and developmental programs. These tiering challenges the view of VSELs as a homogeneous pool and instead support a model of internal functional hierarchy encoded at the level of transcriptional kinetics. Our findings reveal a transcriptionally encoded functional tiering within the VSEL compartment, aligning with hierarchical structures observed in other rare stem cell systems. CD34⁺Lin⁻CD45⁻ cells exhibit minimal transcriptional flux, short velocity vectors, and stable pseudotime positioning, consistent with a deeply quiescent reserve state analogous to long-term hematopoietic stem cells (LT-HSCs) or label-retaining epithelial stem cells. In contrast, CD133⁺Lin⁻CD45⁻ cells display elevated RNA velocity, transcriptional heterogeneity, and activation of mitotic and lineage-specifying genes (AURKA, SOX17), reflecting a transition-competent, lineage-primed tier poised for cell cycle entry or fate specification. This internal tiering refutes the assumption of a uniform VSEL pool and instead supports a model in which transcriptional plasticity is hierarchically encoded even within ultra-rare stem cell populations. Notably, this dynamic stratification is evident without prior lineage bias or exogenous stimulation—suggesting that kinetic heterogeneity is an intrinsic feature of VSEL biology. The discovery of this tiering mechanism extends principles from well-characterized systems (e.g., HSCs, neural stem cells, muscle satellite cells) to human VSELs, offering a unifying framework for dissecting transcriptional readiness and lineage potential across rare stem cell types. Beyond descriptive stratification, the kinetic heterogeneity identified here provides insight into the functional organization of the VSEL compartment. The transcriptional heterogeneity revealed by RNA velocity is not intended as a purely descriptive catalog of VSEL states, but instead reflects a structured, physiologically meaningful organization. The identification of deeply quiescent CD34⁺ VSELs and transcriptionally primed CD133⁺ VSELs mirrors reserve-to-activated hierarchies described in other rare stem cell systems, including long-term versus short-term hematopoietic stem cells, quiescent neural stem cells, and label-retaining epithelial progenitors. Such an organization provides a plausible framework for reconciling prior observations of VSEL dormancy, stress responsiveness, and context-dependent differentiation, suggesting that functional diversity within the VSEL compartment is intrinsically encoded at the level of transcriptional kinetics rather than surface phenotype alone. RNA velocity resolves transcriptional heterogeneity within VSELs Trajectory inference from RNA velocity uncovered distinct transcriptomic states between VSEL subsets. CD34⁺Lin⁻CD45⁻ cells exhibited low RNA velocity magnitudes, uniform pseudotime, and a predominance of steady-state expression across key cell cycle and stemness-associated genes. This is consistent with a deeply quiescent, genome-preserving reservoir that mirrors phenotypes observed in long-term hematopoietic stem cells (LT-HSCs), muscle satellite cells, and adult neural stem cells [ 33 – 35 ]. In contrast, CD133⁺Lin⁻CD45⁻ cells displayed elevated transcriptional flux, broader pseudotime distributions, and dynamic expression of genes associated with cell cycle progression and early lineage specification. This included SOX17 (endodermal priming), AURKA (G2/M transition), TOP2A (chromosome condensation), and PCNA (DNA replication). These features position CD133⁺ VSELs as a transcriptionally poised, activation-competent compartment analogous to short-term HSCs or transitional progenitors. The continuous rather than discrete nature of pseudotime and velocity profiles argues against rigid compartmentalization and instead supports a fluid model in which VSELs undergo gradual state transitions, guided by both intrinsic splicing kinetics and extrinsic signals. Our findings reveal transcriptionally encoded stratification within the VSEL compartment, with CD34⁺ cells occupying a deeply quiescent state and CD133⁺ cells displaying increased transcriptional dynamics and lineage priming. These kinetic distinctions suggest an intrinsic organization analogous to reserve-to-activated hierarchies described in other rare stem cell systems. While RNA velocity–defined states do not necessarily correspond one-to-one with existing surface marker combinations, the identification of membrane-associated candidate markers such as TM4SF18 provides a rational starting point for future efforts to enrich and functionally interrogate transcriptionally defined VSEL subpopulations. A model of functional tiering in rare stem cell populations The contrasting kinetic profiles of CD34⁺ and CD133⁺ VSELs suggest a tiered structure reminiscent of reserve–activated architectures described in muscle [ 36 ], intestinal [ 37 ], and epithelial stem cells [ 38 ]. Within this framework, CD34⁺ VSELs function as long-lived, metabolically inactive cells that preserve genome integrity, while CD133⁺ VSELs constitute an activation-ready pool capable of proliferation and early lineage engagement. These dynamics recapitulate broader features of rare stem cell systems: (i) extended cell cycle dormancy, (ii) readiness for asymmetric division upon stimulation, and (iii) latent developmental multipotency revealed only under regenerative or stress conditions. The splicing-driven velocity profiles observed here—especially those involving SOX17, AURKA, and CCNB1—illustrate how lineage priming and cell cycle engagement are tightly coordinated in transcriptionally rare cell states. Theoretical integration with cell cycle–aware velocity models and DeepCycle Interpretation of velocity vectors in stem cells requires careful disambiguation between transcriptional state transitions and cell cycle–driven expression oscillations. Standard velocity models (e.g., scVelo) assume linear splicing kinetics and may conflate mitotic gene induction with differentiation. To refine these inferences, we propose integration with cell cycle–aware velocity models, such as those implemented in DeepCycle [ 39 ], which apply deep neural networks to predict cell cycle stage and phase transitions from unspliced-spliced transcript ratios. In our data, CD133⁺ VSELs showed elevated levels of G2/M-related gene velocity (AURKA, TOP2A), suggesting synchronous activation of mitotic programs. DeepCycle-derived cell cycle phase predictions could theoretically enhance the resolution of these clusters and separate stochastic transcriptional bursts from genuine lineage priming. Future work integrating latent time estimates and cycle-aware kinetic modeling will enable sharper dissection of how quiescence exit, and fate specification are temporally coupled in rare stem cells. Moreover, velocity-informed pseudotime distributions in CD133⁺ cells exhibited bifurcation patterns consistent with early lineage divergence. Comparative modeling using DeepCycle or latent time algorithms [ 21 ] could validate whether these bifurcations reflect true developmental branch points or stochastic fluctuations in transcriptional entropy. A notable finding of this study is the presence of a conserved chromatin organization network shared between CD34⁺ and CD133⁺ VSELs. Despite pronounced differences in transcriptional kinetics, RNA velocity, and lineage priming, both subsets retained a common epigenetic core comprising histone acetyltransferases and methyltransferases (KAT2A, KAT6A, KMT2C/D), chromatin remodelers (CHD3, ARID2, SMARCA5), and transcriptional co-activators (EP300, CREBBP). The preservation of this chromatin-regulatory architecture suggests that functional stratification within the VSEL compartment arises not from wholesale rewiring of epigenetic programs, but from differential deployment of a shared regulatory framework. This organization is consistent with stem cell systems in which quiescent and activation-competent states coexist on a common epigenetic backbone, enabling rapid transcriptional responsiveness while maintaining genome stability and developmental plasticity. In this context, the transition from CD34⁺ to CD133⁺ VSELs likely reflects kinetic modulation of transcription within a stable chromatin landscape rather than irreversible commitment, providing a mechanistic basis for the observed continuum of transcriptional states. Implications for rare stem cell systems across tissues The transcriptional continuum described here in VSELs parallels patterns observed in other rare stem cell niches. In adult neurogenesis, quiescent neural stem cells exhibit low RNA content, slow splicing kinetics, and rare mitotic gene induction—features replicated in CD34⁺ VSELs. Similarly, label-retaining cells in hair follicles and the hematopoietic system cycle infrequently but can rapidly activate upon injury—a dynamic mimicked by CD133⁺ VSELs. Thus, our findings in VSELs provide a generalizable model for studying other transcriptionally sparse stem cell systems, especially when sample limitations preclude time-series or lineage tracing. RNA velocity and phase portrait analyses offer a non-perturbative, inference-based approach to resolving plasticity and functional heterogeneity in stem cell populations with low transcript counts and high cellular noise. Limitations and future directions This study is constrained by key limitations. First, RNA velocity modeling depends on assumptions of uniform splicing kinetics, which may be violated in rare or low-RNA-content cells like VSELs. Second, the analysis was restricted to a single donor due to the extreme scarcity of VSELs and the requirement for fresh sample processing. Accordingly, the conclusions of this study are intentionally limited to within-sample comparisons of transcriptional states and kinetic stratification rather than inter-individual variability. Third, our analysis was restricted to neonatal UCB; VSEL dynamics in adult tissues, aging, or disease remain unexplored. To build upon these findings, future studies would integrate multiomic modalities (scATAC-seq, scCUT&Tag) to relate transcriptional kinetics to chromatin state and spatial transcriptomics to map VSEL subsets in situ to assess niche influence on their dynamic behavior. Future studies integrating multi-donor single-cell transcriptomics with orthogonal modalities such as single-cell ATAC-seq will be essential to validate the generalizability of the identified kinetic states and to link transcriptional dynamics with chromatin accessibility and gene regulatory architecture. Conclusion By applying RNA velocity modeling to a rare stem cell population, we reveal that human VSELs exist along a structured continuum of transcriptional states—from deeply quiescent CD34⁺ cells to lineage-primed CD133⁺ subsets. Velocity-informed analysis of SOX17, AURKA, CCNB1, and PCNA uncovers coordinated regulation of cell cycle and differentiation programs at single-cell resolution. These findings not only advance our molecular understanding of VSELs but also contribute a broadly applicable framework for studying transcriptional kinetics in rare, heterogeneous stem cell systems. Our work demonstrates that RNA velocity—especially when integrated with cell cycle–aware models—can unlock dynamic dimensions of stemness previously obscured in static single-cell datasets. Our study introduces a generalizable RNA velocity–based framework for resolving functional heterogeneity in ultra-rare and quiescent stem cells. While VSELs remain a debated population, they provide a stringent test case for applying kinetic modeling to uncover transcriptional tiering that is otherwise experimentally intractable. The identification of quiescent versus transcriptionally primed states within VSELs parallels similar continua described in hematopoietic, neural, and muscle stem cells, underscoring the cross-lineage relevance of our approach. By positioning RNA velocity as a tool for rare stem cell analysis, we provide a methodological advance with broad utility: from investigating poorly characterized tissue-resident progenitors to guiding hypothesis generation in systems where classical assays are not feasible. Ultimately, this framework bridges a critical gap in single-cell biology by enabling the study of functional stratification in populations defined by scarcity, quiescence, or controversy. Supplementary information supplementary materials (508.9KB, pdf) Acknowledgements This research was funded by National Science Center NCN OPUS 2021/41/B/NZ6/01590 grant from the National Science Center, Poland to MK. Author contributions JJ—Investigation, Methodology, Formal analysis, Writing—original draft. MZR—Supervision, Writing—original draft, Writing—review and editing. MK—Conceptualization, Project administration, Supervision, Investigation, Writing—original draft, Writing—review and editing. Data availability The data sets supporting the results of this article are available in the Sequence Read Archive (SRA) ( https://www.ncbi.nlm.nih.gov/sra ) repository and assigned unique persistent identifiers: PRJNA1126429 ( https://www.ncbi.nlm.nih.gov/sra/PRJNA1126429 ) and PRJNA1128409 ( https://www.ncbi.nlm.nih.gov/sra/PRJNA1128409 ). Competing interests The authors declare no competing interests. Ethics statement The study was conducted under the Declaration of Helsinki (ethical principles for medical research involving human subjects), and approved by the Medical University of Warsaw Bioethics Committee (permission number KB/50/2022). Informed consent was obtained from the donor’s legal guardian. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Supplementary information The online version contains supplementary material available at 10.1038/s41375-026-02956-9. References 1. Kucia M, Reca R, Campbell FR, Zuba-Surma E, Majka M, Ratajczak J, et al. A population of very small embryonic-like (VSEL) CXCR4(+)SSEA-1(+)Oct-4+ stem cells identified in adult bone marrow. Leukemia. 2006;20:857–69. [ DOI ] [ PubMed ] [ Google Scholar ] 2. Kucia M, Wu W, Ratajczak MZ. Bone marrow-derived very small embryonic-like stem cells: their developmental origin and biological significance. Dev Dyn. 2007;236:3309–20. [ DOI ] [ PubMed ] [ Google Scholar ] 3. Kucia M, Wysoczynski M, Ratajczak J, Ratajczak MZ. Identification of very small embryonic like (VSEL) stem cells in bone marrow. Cell Tissue Res. 2008;331:125–34. [ DOI ] [ PubMed ] [ Google Scholar ] 4. Shin DM, Suszynska M, Mierzejewska K, Ratajczak J, Ratajczak MZ. Very small embryonic-like stem-cell optimization of isolation protocols: an update of molecular signatures and a review of current in vivo applications. Exp Mol Med. 2013;45:e56. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 5. Suszynska M, Adamiak M, Thapa A, Cymer M, Ratajczak J, Kucia M, et al. Purinergic signaling and its role in mobilization of bone marrow stem cells. Methods Mol Biol. 2023;2567:263–80. [ DOI ] [ PubMed ] [ Google Scholar ] 6. Chang YJ, Tien KE, Wen CH, Hsieh TB, Hwang SM. Recovery of CD45/Lin/SSEA-4 very small embryonic-like stem cells by cord blood bank standard operating procedures. Cytotherapy. 2014;16:560–5. [ DOI ] [ PubMed ] [ Google Scholar ] 7. Chen ZH, Lv X, Dai H, Liu C, Lou D, Chen R, et al. Hepatic regenerative potential of mouse bone marrow very small embryonic-like stem cells. J Cell Physiol. 2015;230:1852–61. [ DOI ] [ PubMed ] [ Google Scholar ] 8. Domingues A, Rossi E, Bujko K, Detriche G, Richez U, Blandinieres A, et al. Human CD34(+) very small embryonic-like stem cells can give rise to endothelial colony-forming cells with a multistep differentiation strategy using UM171 and nicotinamide acid. Leukemia. 2022;36:1440–3. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 9. Golipoor Z, Mehraein F, Zafari F, Alizadeh A, Ababzadeh S, Baazm M. Migration of bone marrow-derived very small embryonic-like stem cells toward an injured spinal cord. Cell J. 2016;17:639–47. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 10. Guerin CL, Loyer X, Vilar J, Cras A, Mirault T, Gaussem P, et al. Bone-marrow-derived very small embryonic-like stem cells in patients with critical leg ischaemia: evidence of vasculogenic potential. Thromb Haemost. 2015;113:1084–94. [ DOI ] [ PubMed ] [ Google Scholar ] 11. Havens AM, Sun H, Shiozawa Y, Jung Y, Wang J, Mishra A, et al. Human and murine very small embryonic-like cells represent multipotent tissue progenitors, in vitro and in vivo. Stem Cells Dev. 2014;23:689–701. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 12. Kassmer SH, Jin H, Zhang PX, Bruscia EM, Heydari K, Lee JH, et al. Very small embryonic-like stem cells from the murine bone marrow differentiate into epithelial cells of the lung. Stem Cells. 2013;31:2759–66. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 13. Lahlil R, Scrofani M, Barbet R, Tancredi C, Aries A, Henon P. VSELs Maintain their Pluripotency and Competence to Differentiate after Enhanced Ex Vivo Expansion. Stem Cell Rev Rep. 2018;14:510–24. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 14. Monti M, Imberti B, Bianchi N, Pezzotta A, Morigi M, Del Fante C, et al. A Novel Method for Isolation of Pluripotent Stem Cells from Human Umbilical Cord Blood. Stem Cells Dev. 2017;26:1258–69. [ DOI ] [ PubMed ] [ Google Scholar ] 15. Shaikh A, Nagvenkar P, Pethe P, Hinduja I, Bhartiya D. Molecular and phenotypic characterization of CD133 and SSEA4 enriched very small embryonic-like stem cells in human cord blood. Leukemia. 2015;29:1909–17. [ DOI ] [ PubMed ] [ Google Scholar ] 16. Wu JH, Wang HJ, Tan YZ, Li ZH. Characterization of rat very small embryonic-like stem cells and cardiac repair after cell transplantation for myocardial infarction. Stem Cells Dev. 2012;21:1367–79. [ DOI ] [ PubMed ] [ Google Scholar ] 17. Nestorowa S, Hamey FK, Pijuan Sala B, Diamanti E, Shepherd M, Laurenti E, et al. A single-cell resolution map of mouse hematopoietic stem and progenitor cell differentiation. Blood. 2016;128:e20–31. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 18. Velten L, Haas SF, Raffel S, Blaszkiewicz S, Islam S, Hennig BP, et al. Human haematopoietic stem cell lineage commitment is a continuous process. Nat Cell Biol. 2017;19:271–81. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 19. Jarczak J, Bujko K, Ratajczak MZ, Kucia M. scRNA-seq revealed transcriptional signatures of human umbilical cord primitive stem cells and their germ lineage origin regulated by imprinted genes. Sci Rep. 2024;14:29264. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 20. La Manno G, Soldatov R, Zeisel A, Braun E, Hochgerner H, Petukhov V, et al. RNA velocity of single cells. Nature. 2018;560:494–8. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 21. Bergen V, Lange M, Peidli S, Wolf FA, Theis FJ. Generalizing RNA velocity to transient cell states through dynamical modeling. Nat Biotechnol. 2020;38:1408–14. [ DOI ] [ PubMed ] [ Google Scholar ] 22. Luecken MD, Theis FJ. Current best practices in single-cell RNA-seq analysis: a tutorial. Mol Syst Biol. 2019;15:e8746. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 23. Zheng SC, Stein-O’Brien G, Boukas L, Goff LA, Hansen KD. Pumping the brakes on RNA velocity by understanding and interpreting RNA velocity estimates. Genome Biol. 2023;24:246. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 24. Jerber J, Seaton DD, Cuomo ASE, Kumasaka N, Haldane J, Steer J, et al. Population-scale single-cell RNA-seq profiling across dopaminergic neuron differentiation. Nat Genet. 2021;53:304–12. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 25. Bastidas-Ponce A, Tritschler S, Dony L, Scheibner K, Tarquis-Medina M, Salinno C, et al. Comprehensive single cell mRNA profiling reveals a detailed roadmap for pancreatic endocrinogenesis. Development. 2019;146:dev173849. [ DOI ] [ PubMed ] 26. Strunz M, Simon LM, Ansari M, Kathiriya JJ, Angelidis I, Mayr CH, et al. Alveolar regeneration through a Krt8+ transitional stem cell state that persists in human lung fibrosis. Nat Commun. 2020;11:3559. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 27. Herault L, Poplineau M, Duprez E, Remy E. A novel Boolean network inference strategy to model early hematopoiesis aging. Comput Struct Biotechnol J. 2023;21:21–33. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 28. Jassinskaja M, Gonka M, Kent DG. Resolving the hematopoietic stem cell state by linking functional and molecular assays. Blood. 2023;142:543–52. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 29. Bujko K, Ciechanowicz AK, Kucia M, Ratajczak MZ. Molecular analysis and comparison of CD34(+) and CD133(+) very small embryonic-like stem cells purified from umbilical cord blood. Cytom A. 2023;103:703–11. [ DOI ] [ PubMed ] [ Google Scholar ] 30. Ratajczak MZ, Liu R, Marlicz W, Blogowski W, Starzynska T, Wojakowski W, et al. Identification of very small embryonic/epiblast-like stem cells (VSELs) circulating in peripheral blood during organ/tissue injuries. Methods Cell Biol. 2011;103:31–54. [ DOI ] [ PubMed ] [ Google Scholar ] 31. Van Rossum Guido DFL Python 3 Reference Manual: CreateSpace, Scotts Valley, CA. 2009. 32. Kluyver Thomas R-KB, Pérez F, Granger B, Bussonnier M, Frederic J, Kelley K, et al. Jupyter Notebooks – a publishing format for reproducible computational workflows. IOS Press. 2016:87-90. 33. Blasco-Chamarro L, Farinas I. Fine-tuned rest: unveiling the regulatory landscape of adult quiescent neural stem cells. Neuroscience. 2023;525:26–37. [ DOI ] [ PubMed ] [ Google Scholar ] 34. Cabezas-Wallscheid N, Klimmeck D, Hansson J, Lipka DB, Reyes A, Wang Q, et al. Identification of regulatory networks in HSCs and their immediate progeny via integrated proteome, transcriptome, and DNA methylome analysis. Cell Stem Cell. 2014;15:507–22. [ DOI ] [ PubMed ] [ Google Scholar ] 35. Urbán N, Cheung TH. Stem cell quiescence: the challenging path to activation. Development. 2021;148:dev165084. [ DOI ] [ PMC free article ] [ PubMed ] 36. Chakkalakal JV, Christensen J, Xiang W, Tierney MT, Boscolo FS, Sacco A, et al. Early forming label-retaining muscle stem cells require p27kip1 for maintenance of the primitive state. Development. 2014;141:1649–59. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 37. Buczacki SJ, Zecchini HI, Nicholson AM, Russell R, Vermeulen L, Kemp R, et al. Intestinal label-retaining cells are secretory precursors expressing Lgr5. Nature. 2013;495:65–9. [ DOI ] [ PubMed ] [ Google Scholar ] 38. Mascre G, Dekoninck S, Drogat B, Youssef KK, Brohee S, Sotiropoulou PA, et al. Distinct contribution of stem and progenitor cells to epidermal maintenance. Nature. 2012;489:257–62. [ DOI ] [ PubMed ] [ Google Scholar ] 39. Riba A, Oravecz A, Durik M, Jimenez S, Alunni V, Cerciat M, et al. Cell cycle gene regulation dynamics revealed by RNA velocity and deep-learning. Nat Commun. 2022;13:2865. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials supplementary materials (508.9KB, pdf) Data Availability Statement The data sets supporting the results of this article are available in the Sequence Read Archive (SRA) ( https://www.ncbi.nlm.nih.gov/sra ) repository and assigned unique persistent identifiers: PRJNA1126429 ( https://www.ncbi.nlm.nih.gov/sra/PRJNA1126429 ) and PRJNA1128409 ( https://www.ncbi.nlm.nih.gov/sra/PRJNA1128409 ). Articles from Leukemia are provided here courtesy of Nature Publishing Group ACTIONS View on publisher site PDF (3.3 MB) Cite Collections Permalink RESOURCES Similar articles Cited by other articles Links to NCBI Databases Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top

Related documents

• Low Depth Epigenetic Mapping of Maturation Versus Retrodifferentiation in HepaRG Cells#347757
• iAODE for benchmarking and continuum modeling of single-cell chromatin accessibility.#1014
• Stratified evaluation of blood RNA sequencing in a rare disease cohort#318528
• Single cell trajectories#212307
• Genetic interaction mapping with microfluidic-based single cell sequencing#609452