High-throughput workflow leveraging spectral flow cytometry and CITE-seq to characterize human bone marrow across race and gender

Characterizing human bone marrow is challenging due to heterogeneity of cell types, and the rarity of some populations. To address this, we developed a workflow that bridges CITE seq and flow cytometry. First, we rigorously titrated 277 TotalSeq-A antibodies (BioLegend) using Laminar Wash technology (Curiox Biosystems) to efficiently wash unbound antibodies in a hands-free manner and improve reproducibility. We resolved the utility of 133 TotalSeq-A antibodies (and 5 isotype controls) for CITE-seq of primary human bone marrow samples. We describe the application of these 133 markers in CITE seq of primary marrow samples (varying sex and race) to create a human bone marrow atlas with both transcriptional and antibody-based markers for over 70 cell states. Notably, we also discovered race-based differences in molecular markers and population frequencies. Next, we exploited a machine learning tool (pyInfinityFlow) to validate these molecular/informatics findings. We used a 22-color spectral flow panel tailored to human bone marrow (developed by Cytek Biosciences) as the backbone, then incorporated 111 PE-conjugated “Infinity markers” guided by the CITE-seq results. The analysis resolved a single final FCS file to co-visualize 133 markers, and provided concordance between CITE-seq values measured by sequencing and flow values measured by fluorescence intensity for marrow populations. In summary, we present a streamlined workflow widely adaptable to bone marrow disease studies and transferrable to other tissues. We deliver an atlas-level high-resolution reference with transcriptome-defined populations and immunophenotyping, enabled by both sequencing and spectral flow cytometry.