Scientists have unveiled the most detailed map of human blood stem cells to date, utilizing game theory and machine learning techniques. With this map, the research team has identified over 80 distinct subgroups of hematopoietic stem and progenitor cells (HSPCs), which are early cells that give rise to mature red and white blood cells and other cells in the human blood system. The latest study holds promise for new therapies for leukemia. The findings were published in the March 21st edition of the journal Nature Immunology.
The team employed various advanced genomic research tools and computational methods to map detailed genomic and proteomic profiles of over 300,000 bone marrow cells from donors of different ethnicities and genders. In addition to single-cell RNA sequencing and flow cytometry, they utilized cutting-edge techniques from the field of machine learning and game theory for crucial statistical analysis.
Among the team's key findings are the identification of 89 clusters that aid in revealing new populations of stem and progenitor cells, the discovery of two new surface markers, C5L2 and TSPAN33, which can be used to identify cells transitioning between progenitor states and mature neutrophils, and the identification of unique stages in human red blood cell formation, namely CD326 and CD235a erythroblasts. They also developed a new online computational tool for predicting the cellular origins and potential triggers of cancers like acute myeloid leukemia.
The team notes that their analysis, which combines information from previous disparate maps, surface markers, genes, and advances in machine learning, reveals previously hidden cellular states, such as the most primitive stem cells. Their precise capture of early HSPCs, intermediate cell states with stromal populations, and the most abundant terminal cell states offers the most in-depth view yet of bone marrow stem and progenitor cells.
The researchers believe that one of the most immediate potential applications of the new map is in treating leukemia patients. They identified 69 antibodies among 23 cell states that could be used for clinical monitoring of leukemia patients. The new map is also expected to accelerate scientists' discovery of more blood diseases' cellular and molecular regulatory factors.
