Hallmarks of Transcriptional Heterogeneity in Human Hematopoietic Stem Cells across 382 Donors
Bibliographic record
Abstract
Our understanding of human hematopoietic stem cells (HSC) has been shaped through decades of experimental research aimed at purifying HSCs to functional homogeneity. However, recent studies have uncovered both functional and molecular heterogeneity within the human HSC pool, particularly in the context of ontogeny, aging, inflammation, and somatic mosaicism. Yet, a comprehensive understanding of human HSC variation and their relevance in health and disease is lacking. Here, we carried out a systematic analysis of human HSC heterogeneity at an unprecedented scale by applying an approach to purify human HSC and multipotent progenitors (MPPs) in-silico (Zeng et al., biorxiv 2023) to publicly available and in-house single cell RNA sequencing datasets of human hematopoiesis, excluding acute leukemia. Hence, we assembled a large database of 526,993 single cell HSC/MPP transcriptomes from over 382 donors spanning ontogeny, age, and disease status. To uncover gene expression programs that vary across HSC/MPPs, we employed consensus non-negative matrix factorization across each dataset identifying 553 programs. As most programs were recurrent across datasets, these were collapsed to form 14 consensus meta-programs constituting the transcriptional hallmarks of human HSCs. We evaluated the biological relevance of this new catalog of recurrent human HSC variation with known functional properties of HSCs. First, we uncovered classical HSC quiescence control meta-programs related to quiescence exit enriched for CTCF targets, as well as lineage associated programs related to myeloid-lymphoid or megakaryocyte-erythroid priming. Second, two distinct meta-programs corresponding to stemness, were enriched in human HSC versus progenitors (NES > 3.00, p < 1e-36) as well as serial repopulating murine HSCs defined by lineage tracing (Rodriguez-Fraticelli et al., 2020; NES > 1.50, p < 0.01). Interestingly, one of these stemness-related meta-programs was strongly enriched for genes defining quiescence including CDKN1A which encodes the cell cycle inhibitor p21. Third, we identified one meta-program underlying HSC trafficking, enriched in circulating HSCs (umbilical cord blood, peripheral blood, and mobilized peripheral blood) compared to bone marrow (BM) HSCs. Notably, pathway analysis of the top genes driving this program revealed enrichment of lymphoid priming signatures, suggesting a path to lymphopoiesis. Finally, we uncovered seven novel HSC programs; five are associated with inflammatory response and upregulated with human HSC aging (NES > 2, p < 1e-07; >40y donor HSCs vs <40y donor HSCs). Three meta-programs are also enriched in HSCs enriched in TET2 and DNMT3A clonal hematopoiesis compared to age-matched controls (NES > 1.5, padj < 1e-03). These include two meta-programs underlying our recent discovery of a human HSC inflammatory memory population (AUC > 0.90), which retains epigenetic and transcriptional memory of prior inflammation (Zeng, Nagree, Jakobsen et al., biorxiv 2023). Pathway analysis revealed convergent and divergent patterns of enrichment, with one program notably significant for TGF-B signaling and SMAD motifs. Importantly, we identified one meta-program upregulated in myelofibrosis and VEXAS syndrome (NES > 2, p < 1e-09), with key genes including CD83, the NR4A family and proinflammatory cytokines, showcasing the relevance for our approach to uncover new understanding of disease. Collectively, we have systematically uncovered transcriptional hallmarks of human HSC heterogeneity, providing a new framework for understanding how HSCs may vary and become dysregulated in hematological disease.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".