| Description |
Leukemia is a fatal hematopoietic malignancy with heterogeneous cell populations and complex progressions. Here, we analyzed >80,000 single cell transcriptomes of bone marrow mononuclear cells (BMMCs) from 4 healthy samples and 8 leukemia patients. A comprehensive cell atlas with fine resolution of BMMCs and early hematopoietic stem and progenitor cells (HSPCs) was constructed for revealing hematopoiesis and leukemogenesis. Trajectory analysis along with coordinated genes and TFs of HSPCs showed hematopoietic stem cells continuously differentiated into seven cell lineages with hierarchical structures, based on which we proposed a continuous transitional model to explain hematopoietic process. We developed a new approach, a composite likelihood of multiple tracing statistics (CMTS), to characterize the cell types or progressive status of leukemias, by projecting leukemia subclones to normal cell atlas. We identified various compositions and abundance in leukemia cells of both acute myeloid leukemia (AML) and acute lymphoblastic leukemia (ALL). The putative cell types of leukemia subclones in most patients are proximate in either myeloid lineage or lymphocyte lineage. Thus, we hypothesized the leukemia initial cells with mutated genome partially maintain its developmental trajectory and terminate at different stages due to dysfunctional genes, which lead a serial of leukemia subpopulations, instead of a homozygous population. At last, analysis of an ALL patient with 4-time points vividly revealed the dynamics of leukemia progress during treatment and relapse. Emerging mutations and transcriptional abnormality were identified in relapsed leukemia cells. We further found the relapsed leukemia cells were derived from a tiny subpopulation of early stage. Our study combined with the data source of BMMCs from both normal and leukemia samples, will greatly facilitate precision diagnosis and treatment of leukemia in the future. |
