Deciphering Hematopoiesis at single cell level through the lens of reduced dimensions

Abstract Hematopoiesis plays a critical role in maintaining a diverse pool of blood cells throughout human life. Despite recent efforts with single-cell data analyses, the nature of the early cell fate decisions and compartmentalization of progenitors remains contentious due to the sparsity and noise of the data. Using publically available single-cell RNA-Seq hematopoietic data from bone marrow, with three different matrix factorization approaches to recover associated gene modules from cell clusters reveals a tri-directional and hierarchically-structured transcriptional landscape of hematopoietic differentiation. We also devised a bootstrap method, which in combination with the above can better characterize the progenitor compartments and retrieve cellular hierarchies.