Heuristic bias in stem cell biology

Abstract

When studying purified hematopoietic stem cells, the urge for mechanisms and reductionist approaches appears to be overwhelming. The prime focus of the field has recently been on the study of highly purified hematopoietic stem cells using various lineage and stem cell-specific markers, all of which adequately and conveniently fit the established hierarchical stem cell model. This methodology is tainted with bias and has led to incomplete conclusions. Much of our own work has shown that the purified hematopoietic stem cell, which has been so heavily studied, is not representative of the total population of hematopoietic stem cells and that rather than functioning within a hierarchical model of expansion the true hematopoietic stem cell is one that is actively cycling through various differentiation potentials within a dynamic continuum. Additional work with increased emphasis on studying whole populations and direct mechanistic studies to these populations is needed. Furthermore, the most productive studies may well be mechanistic at the cellular or tissue levels. Lastly, the application of robust machine learning algorithms may provide insight into the dynamic variability and flux of stem cell fate and differentiation potential.