Single spectral flow cytometry panel for simultaneous detection of hematopoietic stem, progenitor, and mature lineages in mouse bone marrow

Hematopoiesis occurs in the bone marrow of adult mammals and is supported by hematopoietic stem cells that sustain lifelong blood cell production. Pathological conditions can disrupt HSC differentiation, causing anemia, immunodeficiency, and other cytopenias. Therefore, precise and simultaneous identification of hematopoietic populations from stem to mature cells is essential for understanding disease mechanisms and developing targeted therapies. In order to study these alterations, flow cytometry is generally the reference technique. However, most cytometry panels are designed to study a specific population, leaving out potential discoveries on other populations from the same microenvironment. Here we present the design of a 19-marker spectral flow cytometry panel capable of simultaneously identifying HSPCs, erythroid, myeloid and lymphoid cells within a single murine bone marrow sample. This integrated approach replaces multiple conventional panels and enables comprehensive mapping of hematopoietic differentiation from a single assay. Validation confirmed accurate detection of long-term and short-term HSCs, multipotent progenitors, common myeloid and lymphoid progenitors, and erythroid populations from proerythroblasts to reticulocytes. UMAP visualization captured the continuous trajectory of differentiation. We validated our method on aged and β-thalassemic mouse models showing a clear detection of hematopoiesis and erythropoiesis alterations respectively. This panel provides a robust and flexible tool suitable for both basic research and translational studies.

The bone marrow is the primary site of hematopoiesis in adult mammals, sustained by hematopoietic stem cells (HSCs) that generate all mature blood lineages throughout life. Pathological conditions can alter HSC differentiation, leading to cytopenias such as anemia or immunodeficiency. Therefore, the ability to precisely and simultaneously identify hematopoietic populations, from HSCs to fully differentiated cells, is essential to understand the mechanisms underlying these disorders and to develop targeted interventions.

Flow cytometry remains the reference standard technique to study hematopoiesis, providing high-resolution quantitative analysis of cellular diversity within the bone marrow. However, most conventional panels are designed to study a specific lineage, typically hematopoietic stem and progenitor cells (HSPCs) (1–3), erythroid populations (4, 5), or myeloid and lymphoid compartments (6, 7). This strategy requires multiple samples, increases biological variability, and limits the integrated assessment of hematopoietic differentiation within a single biological sample.

To overcome these limitations, we developed and optimized a high-dimensional, single-sample spectral flow cytometry panel composed of 19 markers, enabling the comprehensive characterization of hematopoietic populations in a single murine bone marrow sample. This integrated panel simultaneously identifies HSPCs, myeloid, lymphoid and erythroid populations across all stages of maturation, from proerythroblasts to reticulocytes, within one assay. Although spectral flow cytometry panels have been developed to study specific hematopoietic compartments (3, 5, 6, 8), there is currently no published spectral panel that provides a unified, single-sample overview integrating stem cells, progenitors, erythroid maturation stages, and mature immune populations within the same and unique murine bone marrow assay. The novelty of our work therefore relies not in the use of spectral cytometry itself, but in the design and validation of a multi-lineage panel enabling global overview of hematopoiesis in a single sample.

Validation experiments confirmed accurate detection and discrimination of long-term and short-term HSCs (LT-HSCs, ST-HSCs), multipotent progenitors (MPP), common myeloid and lymphoid progenitors, as well as all erythroid maturation stages. UMAP-based visualization faithfully captured the continuous hierarchy of differentiation from stem to mature cells. Importantly, this single-tube, high-parameter approach consolidates three conventional panels into one, minimizes sample consumption, and allows reproducible mapping of hematopoietic trajectories within the same biological sample.

We further validated our approach under pathological conditions using aged mice known to display hematopoietic alterations, especially HSC dysfunction and a myeloid lineage bias (9). We also used a mouse model of β-thalassemia (beta-thal), a congenital anemia caused by mutations in globin genes leading to defective erythropoiesis and chronic hemolytic anemia (10–12). The panel successfully detected hematopoietic abnormalities previously described in these models, including progenitor expansion and erythroid maturation deficiency, confirming its sensitivity and applicability to pathological hematopoiesis.

Beyond a broad phenotypic characterization, our panel was explicitly designed to support the evaluation of emerging treatments by intentionally leaving specific fluorescence channels available for the detection of reporter proteins such as mCherry or GFP. As an example, therapeutic options for hematopoietic alterations, such as anemia remain limited, and many patients continue to experience suboptimal long-term outcomes. Recent advances in molecular and genetic therapies, including gene therapy and RNA-based approaches, rely on an expanding range of viral and non-viral delivery vectors to correct disease-causing defects at the level of hematopoietic stem and progenitor cells (13, 14). As these approaches are primarily developed and optimized in preclinical models prior to clinical translation, there is a critical need for robust and comprehensive tools to accurately evaluate hematopoietic progenitor targeting, as well as the global impact of these interventions on hematopoiesis and erythropoiesis. In this preclinical context, our single-tube, high-parameter cytometry panel provides a robust, flexible, and scalable platform ideally suited for early-stage therapeutic evaluation. By enabling the simultaneous characterization of all major hematopoietic compartments within a single assay, it allows (i) precise monitoring of hematopoietic progenitor targeting by therapeutic delivery vectors using reporter-based readouts, and (ii) an integrated assessment of treatment-induced alterations across the entire hematopoietic hierarchy.

This comprehensive and system-level view makes our panel particularly valuable for preclinical studies of next-generation therapies for hematopoietic disorders such as congenital anemias, and supports informed decision-making prior to clinical translation.

Reagents and equipment

All reagents and equipment used in this study are listed in Table 1.

Antibodies

All antibodies used are listed in Table 2 including fluorochrome, supplier, clone, catalogue number and titration.

Cytometer

The Cytek ® Aurora cytometer was used in this study. Cytek ® Aurora cytometer is configured with 5 lasers (355 nm (UV), 405 nm (Violet), 488 nm (Blue), 561 nm (Yellow-Green) and 640 nm (Red)) and a full-spectrum detection system comprising 64 avalanche photodiode detectors (UV1-UV16; V1-V16; B1-B14; YG1-YG10; R1-R8). Spectral reference controls and daily quality control (QC) were performed according to the manufacturer’s guidelines.

Software

Data analysis was performed in the SpectroFlow software (Cytek) and OMIQ (license software) and Cytek Cloud web tool. GraphPad prism was used to generate the graphs.

Animals

Male mice Wild-Type (WT) (young (3 months) and aged (18 months), C57BL/6J background, Charles River) and beta-thalassemic (beta-thal, strain B6.D2-Hbbd3th/BrkJ, C57BL/6J background, Jackson Laboratory) were housed under controlled conditions of temperature (21 ± 1 °C), hygrometry (60 ± 10%) and lighting (12 h per day). Animals were acclimatized in the laboratory for one week before the start of the experiments. All animals received care according to institutional guidelines, and all experiments were approved by the Institutional Ethics committee number 14, France (beta-thal mice: license number #49620-2024052208518465). Mice were euthanatized by cervical dislocation and femur and tibia, and blood were collected and processed for further evaluations.

Statistical analysis

Data are expressed as mean values ± standard error of the mean (SEM). Statistical significance was tested using either one or two-way analysis of variance (ANOVA) with Fisher multiple comparison test. The results were considered significant if the p-value was <0.05.

Methods and results

Objective and panel design

This panel was primarily designed in order to precisely identify hematopoietic stem cells (long-term LT-HSC, short-term ST-HSC, multipotent progenitor MPP2, MPP3, MPP4), progenitor subsets (common myeloid progenitor CMP, megakaryocyte erythroid progenitor MEP, granulocyte-monocyte progenitor GMP and common lymphoid progenitor CLP) and erythroid populations (from pro-erythroblasts to reticulocytes) using a single high-dimensional spectral flow cytometry panel (Figure 1). In addition, the panel allowed a broader identification, of mature myeloid cells, granulocytes, dendritic cells, macrophages, and lymphoid cells, (B and T cells). The stepwise workflow for spectral flow cytometry experiment is illustrated below and in Figure 1.

Defining bone marrow cells to analyze

The initial step in panel development consists of defining the relevant cell populations within the biological sample, in this case, murine bone marrow, and identifying the most informative surface markers for their discrimination. Based on this analysis, we identified and selected 19 markers to distinguish 19 cell types of the mouse bone marrow (Figures 2A, B). The scientific references and validation for the selected cell-surface markers are provided in Section 3.3 (Data Analysis).

Panel design

We then designed the panel (Figure 3A) by selecting the most appropriate fluorochrome for each antibody based on several criteria. First, fluorochromes were evaluated according to their full emission spectra (Figure 3B) and their similarity index profiles (Figure 3C), using the Cytek Cloud online tool to minimize spectral overlap. The similarity index was used to assess spectral proximity between fluorochromes, and combinations with a similarity index above >0.85 were avoided to prevent unmixing errors and loss of resolution. We applied a similarity index threshold of 0.85 based on internal validation and expertise to further optimize panel design. In our panel, the highest Similarity Index values were intentionally assigned to non-co-expressed markers (e.g., ViaDyeRed, Figure 3C). The similarity index, calculated using the previously described similarity method, ranges from 0 (distinct spectra) to 1 (highly overlapping spectra). Although Cytek commonly recommends a threshold of 0.98, this value is primarily reported in the literature as manufacturer guidance rather than as a formally validated cutoff (15–17). Additional considerations included fluorochrome brightness and antigen expression pattern, low-fluorescence for high-expressed markers, and commercial availability of the required antibody (Table 2).

During panel design, we also characterized the intrinsic autofluorescence of bone marrow cells to guide fluorochrome selection (Figure 3D). By identifying the channels with the highest autofluorescence (V7 and B3), we strategically assigned well expressed antigens to these channels, ensuring optimal signal detection while minimizing interference from the background autofluorescence.

Antibody titration and optimization

Once the panel has been designed, all antibodies were titrated to determine their optimal concentration in order to achieve an optimal balance between signal level for positive populations and negative populations (Table 2). Titration aimed to achieve maximal resolution between positive and negative populations while minimizing background signal and spillover. If a selected fluorochrome was not optimal, either due to overlap with another fluorochrome leading to excessive spectral spillover, insufficient signal intensity or limited availability, the panel was adjusted accordingly, replacing the fluorochrome or redistributing markers to achieve an optimal panel design. The overall similarity index for our 19-fluorochrome panel was 7.9 (Figure 3C).

Step-by-step experimental procedure

Step 1. Isolation of bone marrow cells.

Tibia and femur were collected in PBS 1x and divided into fragments, each bone was cut in half and placed with the cut surface down into a pierced tube nested within an unpierced tube (pierced tubes were prepared using a heated 21G needle) (8) (Figure 1). 500 µL of PBS 1x was added per tube. Samples were centrifuged at 3500g for 5 min to flush bone marrow into the lower tube, and the supernatant was removed. Pellets were resuspended in 1 mL PBS 1x, filtered through a 100-µm strainer, centrifuged at 400g for 5 min, and resuspended again in PBS. An aliquot of cell suspension was taken and incubated with RBC lysis buffer before counting. The equivalent of 5 million cells from the non-lysed fraction was transferred into 35-µm filter-cap FACS tubes to obtain cell suspensions for staining (Figure 1). Of note, if not interested in erythroid cell measurement, one may choose to remove red blood cells by using RBC lysis buffer on the whole bone marrow cell suspension.

Step 2. Preparation and staining of cells.

Viability staining. Cells were first washed once in PBS in order to removed excess of soluble proteins in the supernatant, then incubated for 20 minutes with 100µL of the viability dye ViaDyeRed mix, prepared at a 1:50,000 dilution in PBS. Following incubation, cells were washed with PBS, and subsequent staining steps were performed as described below. Antibody dilutions are listed in Table 2.

Pre-incubation with CD34 and CD127 antibodies. Cells were initially incubated with CD34 in a total volume of 20 µL for 5 minutes at room temperature. Then, without any intermediate wash, an additional 20 µL containing CD127 was added, and the cells were incubated for 15 additional minutes at room temperature. This two-step pre-incubation with CD34 and CD127 insured a better signal during acquisition.

Surface staining – Mix 1 (Mix Blue–YG–Red). Without any prior washing, cells were sequentially stained with two antibody mixes. First, 30 µL of the Blue–YG–Red Mix 1 (2× concentrated) was added to each sample. This mix contained antibodies against CD44, CD48, Ter119, CD135/Flt3, CD4,

c-Kit/CD117, and Gr-1, as well as the TrueStain Monocyte Blocker to reduce nonspecific fluorochrome-cells interaction such as Cy7 with myeloid cells.

Surface staining – Mix 2 (Mix BUV–V). Immediately after, 30 µL of the BUV–V mix 2 (2× concentrated) was added. This second mix included antibodies for CD16/32/FcyR, CD71, F4/80, CD3, CD45, Sca-1, CD19, CD11b, and CD150. Brilliant Stain Buffer was incorporated into this staining step to minimize fluorochrome interactions, following the manufacturer’s recommendations.

Final incubation and fixation. Samples were incubated for 20 minutes at room temperature in the dark, washed to remove excess antibodies, and finally resuspended in PBS containing 2% paraformaldehyde to fix the cells prior to flow cytometry acquisition.

Step 3. Data acquisition.

Data were acquired on a Cytek Aurora Spectral Flow Cytometer (Cytek Biosciences) equipped with 5 lasers ultraviolet, violet, blue, yellow-green, and red, allowing for high-dimensional spectral analysis.The instrument’s Cytek assay settings were configured to ensure optimal resolution of cell populations. Forward Scatter Area (FSC-A), which correlates with cell size, allowing clear discrimination between small debris and intact cells. Side Scatter Area (SSC-A), reflecting cellular granularity ensure proper cell doublets discrimination in combination with FSC signals. A signal threshold of 50,000 was applied to exclude sub-threshold particles and debris from acquisition.

Data analysis

Spectral unmixing and compensation

Spectral unmixing and fluorescence compensation were performed using the SpectroFlo software (Cytek Biosciences). For the unmixing process, the cellular signature of the unstained control mouse bone marrow sample was used as the reference to accurately capture and extract the sample’s intrinsic autofluorescence (Figure 3D). This autofluorescence was then computationally subtracted by the software (18). Compensation beads (Invitrogen 01-3333-42) were included as reference controls for each antibody used in the panel.

Following unmixing, fluorescence compensation was applied, requiring only minimal adjustments (Supplementary Figure 1). Once unmixing was completed, the unmixed gated “all events” were exported as FCS files and uploaded into the OMIQ software for further analysis.

Manual gating strategy using OMIQ software

Downstream analysis and identification of cell populations were carried out using OMIQ software. Standard gating strategies (FSC-A vs SSC-A and viability dye vs SSC-A) were applied to exclude doublets and dead cells, followed by manual (supervised) population identification of hematopoietic stem and progenitor cells (HSPCs), erythroid, myeloid and lymphoid mature cells (Figures 4–6 respectively).

Quality controls

Prior to further analysis, we recommend to perform different quality controls. Signals may fluctuate during the time of sample acquisition, generating erroneous events. These events, which include clogs, flow speed, etc…, may influence data analyses at various range such as generating “new populations” (worst and rarer cases), altering fluorescence intensity, increase sample variability. Here, we have used the PeacoQC algorithm available in OmiQ (19). This algorithm will remove events presenting abnormal signals in any of the parameters during time of acquisition, for each sample, generating a more reliable dataset (Supplementary Figure 2).

Hematopoietic stem and progenitor cells

Hematopoietic stem cells (HSCs) and progenitors were isolated using a lineage-negative (called here HSC-Prog-Lin -) pre-gating strategy to exclude mature hematopoietic populations (Figure 4A). Cells expressing CD71/Tfrc, Ter119, CD11b, CD19, CD3, CD4, Gr1, or F4/80 were excluded, thereby removing erythroid, myeloid, lymphoid differentiated lineages from the analysis.

Within the HSC-Prog-Lin - fraction, the HSPC subsets were further defined based on expression of Sca1, c-Kit (CD117), CD150, CD48, and CD135, as previously described (1, 2, 8). Using these markers, long-term HSCs (LT-HSCs), short-term HSCs (ST-HSCs), and multipotent progenitors (MPP2, MPP3, MPP4) were distinguished, enabling precise separation of self-renewing stem cells from more differentiated progenitor populations (Figure 4B).

Progenitors engaged in myeloid and lymphoid lineages were identified based on expression of CD34, CD127, CD16/32 and c-Kit (CD117) within the HSC-Prog-Lin - fraction, as previously described (3). This allowed the identification of the common myeloid progenitors (CMP), myelo-erythroid progenitors (MEP), granulocyte-monocyte progenitors (GMP) and common lymphoid progenitor (CLP) (Figure 4C).

For high-dimensional, unsupervised analysis, a random subsampling of 2,700 LSK (Lin -, Sca1 +, c-Kit +) and a random subsampling of 14,000 c-Kit +, Sca1 + cells were performed and were used to generate a Uniform Manifold Approximation and Projection (UMAP) representation of HSPCs (Figures 4D, E). Antibody expression levels were visualized by using color-continuous visualization, which allowed assessment of the relative fluorescence intensity for each marker across cell populations. The fluorochrome associated with each antibody was mapped onto the Z-axis, providing a three-dimensional view of marker expression and ensuring accurate identification of each HSPC subsets (Supplementary Figures 3A, B).

Erythroid cells

Erythroid cells were labeled and gated based on surface expression of Ter119 and CD71, as previously described (5, 20). Prior to erythroid-specific gating, a lineage-negative for erythroid cells (Erythro-Lin -) pre-gating strategy was applied to exclude non-erythroid cells. Cells positive for CD19, CD11b, CD4, CD3, F4/80 or Gr1 were excluded, thereby removing B cells, T cells, monocytes/macrophages, and granulocytes from the analysis (Figure 5A). Within the Erythro-Lin - population, Ter119-positive cells were then identified as erythroid lineage (Figure 5B). Within this population, sequential analysis of CD71 expression allowed discrimination of distinct stages of terminal erythroid differentiation: the morphologic characteristics broadly corresponded to pro-erythroblasts in the Ter119 int CD71 high cell population (Figure 5B, region I), basophilic erythroblasts in the Ter119 high CD71 high cell population (Figure 5B, region II), late basophilic and polychromatophilic erythroblasts in the Ter119 high CD71 int cell population (Figure 5B, region III), and orthochromatophilic erythroblasts/mature reticulocytes in the Ter119 high CD71 low cell population (Figure 5B, region IV) (5, 20). Cell surface marker expression levels were validated using a color-continuous visualization for each marker across the cell populations (Supplementary Figure 4). We then performed a subsampling of 50,000 Ter119 + cells to generate a UMAP representation of the erythroid populations, allowing visualization of the continuous progression of erythroid maturation (Figure 5C).

Myeloid and lymphoid mature cells

Finally, myeloid and lymphoid cells were labeled and gated based on specific surface markers, as previously described (6, 7). Prior to subset-specific gating, a myelo-lineage-negative pre-gating strategy (Myelo-Lin - = CD19 -, Sca1 -, Ter119 -) was applied to exclude erythroid and stem/progenitor cells (Figure 6A). Within the Myelo-Lin - population, a CD11b positive gated was drawn to focus on the myeloid lineage. The granulocytes-enriched fraction was then defined based on Gr-1 expression, and monocytes and macrophages were distinguished using F4/80 and CD16/32 (Figure 6B). Notably, the use of Gr-1 antibody which targets both Ly-6C and Ly-6G antigen may preclude a clear distinction between the prominent population of neutrophils (Ly-6C- Ly-6G+) and inflammatory monocytes (Ly-6C+ Ly-6G-) which are less abundant. A non-supervised analysis (Figure 6C), using CD11b, Gr-1, CD16/32, F4/80 markers may refine the cell-type definition, although one should also consider adding markers such as Ly-6C, Ly-6G, SIGLEC-7, CD115 to better define the myeloid cell content. Cell surface marker expression levels were validated using a color-continuous visualization for each marker across the myeloid populations, which could reinforce the tentative cell type annotation that we performed with the UMAP (Supplementary Figure 5). Lymphocytes were identified using CD19, CD3 and CD4 cell surface marker (Figure 6D). The use of markers such as CD1d, NK1.1, or anti-TCRγδ, would allow a better distinction of lymphoid subsets, with the risk of panel’s similarity index increase and loss in the quality of the panel.

In multiparameter flow cytometry, separate panels, particularly for mature myeloid populations, are often designed to minimize antibody non-specific binding to Fc-receptors. This is commonly addressed using Fc-block, which in mice consists of an anti-CD16/32 antibody (clone 2.4G2). Because the antibody against CD16/32 is included in our panel, Fc-block would directly interfere with the detection of CD16/32-expressing cells and was therefore not incorporated into our main staining protocol. To evaluate its impact, we compared staining with and without Fc-block using the full panel. UMAP analysis revealed that Fc-block treatment is detrimental to the analysis, as it results in the loss of monocyte and macrophage clusters (Supplementary Figure 6A). Furthermore, Fc-block also compromised the resolution of MEP, CMP, and GMP populations (Supplementary Figures 6B–D) compared to the untreated condition (Supplementary Figures 6A–C). Noteworthy, the resolution or identification of other cell populations is not affected by the use or absence of use of Fc-block (Supplementary Figures 6A, B). Altogether, these results demonstrate that Fc-block is incompatible with our full panel configuration and should not be included in the antibody mixes.

Observation of all hematopoietic cells

In order to fully picture the content of the bone marrow hematopoietic system, we generated an UMAP integrating all identified hematopoietic cells (Figure 7). Cell surface marker expression levels were validated using a color-continuous visualization for each marker across the all-cell populations (Supplementary Figure 7). This visualization clearly showed the most immature cells at the center of the UMAP and more mature cells (Lymphoid, myeloid and erythroid lineages) radiating in separate directions.

Panel validation

We validated our spectral cytometry panel on well-characterized mouse models known to exhibit hematopoietic alterations, in order to assess its ability to resolve changes in hematopoietic heterogeneity under pathological conditions.

First, we analyzed bone marrow cells of young (3 months) and aged (18 months) wild-type mice (C57BL/6J) to confirm the panel capacity to detect age-associated hematopoietic alterations, as aging is known to affect HSC function and to drive a myeloid lineage bias (9). Consistently, our analysis revealed clear age-related changes: aged mice exhibited a significant expansion of LT-HSCs alongside a reduction in MPP4 populations (Figure 8A). These observations are in line with the myeloid bias that emerges during hematopoietic aging, reflecting both the shift in lineage commitment in mice and comparable trends reported in healthy elderly humans (21).

Next, we applied our panel to a beta-thalassemic mouse model (B6.D2-Hbbd3th/BrkJ), which is characterized by severe erythroid defects, ineffective erythropoiesis, and chronic anemia (22, 23). In this model, we observed a pronounced erythroid dysregulation, marked by the accumulation of immature erythroblasts and decreased RIV population (Figure 8B). These results highlight the robustness of our spectral cytometry panel in capturing hematopoietic alterations across both age-related and disease-specific contexts, demonstrating its broad applicability for studying hematopoietic heterogeneity.

In this study, we developed and validated a high-dimensional spectral flow cytometry panel capable of simultaneously identifying hematopoietic stem cells (HSC), progenitor subsets, erythroid populations and mature myeloid or lymphoid cells from a single murine bone marrow sample. This method overcomes a major limitation of classical hematopoietic analysis, which typically requires multiple separate panels to profile these different cell populations. By integrating all these analyses into one single assay, we were able to map hematopoietic cells from progenitors to mature cells (Figure 7).

Using this panel, we successfully identified key HSC and progenitor populations, including long-term (LT-HSC) and short-term HSC (ST-HSC), multipotent progenitors (MPP2, MPP3, MPP4), and common myeloid and lymphoid progenitors (CMP, MEP, GMP, CLP) with clear resolution. Within the HSC-Prog-Lin − fraction, we defined HSPC subsets with a set of 5 well-established markers (Sca1, c-Kit/CD117, CD150, CD48, and CD135), following previously described strategies (1, 2, 8). We applied the exact gating approach from 2, and obtained HSPC distributions consistent with those reported by Wilson and colleagues: in our panel, 6% of LSK cells corresponded to roughly 12% LT-HSC, 15% MPP2, and 60% MPP3, compared to 3% of LSK cells with approximately 6% LT-HSC, 17% MPP2, and 75% MPP3 in Wilson et al. Other groups have developed more extensive panels to better resolve HSPC heterogeneity. For instance, Solomon and colleagues employed a 15-marker spectral flow cytometry panel (CD34, Flk2/CD135, CD105, CD48, CD150, EPCR, CD49b, CD41, Sca1, Gfi1-GFP, among others) to dissect the murine HSC compartment (24). This approach allows a finer distinction between the most primitive long-term HSCs (CD34 − EPCR hi CD49b − Gfi1 hi) and intermediate short-term HSCs (CD34 − EPCR low CD49b +). Although such expanded panels are valuable for detailed subset-specific studies, the 5 markers used in our panel are sufficient to discriminate the main HSPC populations of interest and ar well-suited for a global overview.

Concerning the erythroid compartment, our panel includes the classical markers CD71, CD44 and Ter119. We have chosen to discriminate erythroid sequential maturation stages with the combination of CD71/Ter119 which provided a good discrimination from pro-erythroblasts to reticulocytes (20). Socolovsky et al. introduced a flow cytometry-based assay using Ter119 and CD71 as surface markers to distinguish murine erythroblasts at different differentiation stages. The method was subsequently improved by adding forward scatter, which reflects cell size (25). In 2009, Chen K et al. introduced CD44 as an additional marker and developed a new strategy based on Ter119 and CD44 expression levels changes combined with cell size (26). We are also aware that others have proposed other strategies such as CD44 vs FSC (4). However, in our hands and with the cytometer used, the discrimination was not sufficiently clear, perhaps due to weak CD44 signal and size discrimination (data not shown). Furthermore, because OMIQ does not support UMAP analysis that includes a linear parameter such as size discrimination (FSC), inclusion of CD44 would not have allowed to perform unsupervised analysis as proposed here. In the future, improvement of the staining with CD44, or addition of erythroid-specific cell markers may improve erythroid cell annotation.

Last, myeloid and lymphoid cells were sufficiently resolved to identify fractions enriched in Polymorphonuclear cells (PMN, called here granulocyte-enriched), monocytes, macrophages and lymphocytes (B and T cells), enabling general lineage identification and facilitating the gating of lineage-negative populations for HSC and progenitor analysis. The use of additional markers will definitely improve the cell discrimination for both lineages, although one may consider using a lineage-targeted panel to further focus, pending of the results provided first using our approach. Indeed, numerous distinct myeloid and lymphoid subsets exist, and in our study these populations were characterized only in a broad, general manner. While our current panel provides an overview of mature immune cells, it may be limited in terms of resolution. Importantly, the panel could be expanded in future studies to enable a more refined characterization: for instance, the inclusion of markers such as Ly-6C and Ly-6G for the myeloid lineage, CD8 and TCRαβ for T cells, or B220, CD43, and IgM for B cells would allow a more precise distinction of mature immune subsets (27).

As a proof of validation of the panel, we performed flow cytometry analyses on bone marrow from different mouse models, known to display hematopoietic cell alterations. The panel was applied to anemic disease model, beta-thalassemia mice, demonstrating its ability to capture altered erythropoiesis associated with congenital anemias (22, 23). We also used young and aged animals to capture the related changes in HSC subsets were successfully detected in WT mice, with the expected expansion of long-term HSCs and reduction of multipotent progenitors MPP4 in aged animals compared to young ones (28, 29). These results illustrate that the panel can be reliably used to assess both physiological and pathological hematopoiesis or erythropoiesis, in one shot. Here, the number of animals was intentionally minimized to respect ethical principles of reduction with the primary goal of validating our panel and providing a reliable tool for the research community. Researchers applying this panel in specific studies can include larger cohorts as required by their experimental objectives.

To conclude, the design philosophy of our panel resides in the fact that our panel allows visualization of the entire hematopoietic population in a single panel, which, to our knowledge, has not been previously reported. Notably, the full UMAP presented in Figure 7 offers researchers a global overview of their populations, while also allowing them to modify and/or target specific populations, through genetic manipulations or therapeutic interventions, and simultaneously assess the effects on other populations within the same sample. Notably, our panel reserves the mCherry and FITC/GFP channels for reporter protein detection, an approach successfully demonstrated by Eich et al. (30) for GFP. However, our design integrates this flexibility into a more complex framework, enabling broad lineage profiling (HSPC, myeloid and lymphoid) while maintaining the ability to detect common fluorescent reporters.

Overall, our spectral flow cytometry panel represents a high-resolution tool for preclinical research and for future clinical applications.