A differential network analysis approach for lineage specifier prediction in stem cell subpopulations.

backgroundStem cell differentiation is a complex biological process. Cellular heterogeneity, such as the co-existence of different cell subpopulations within a population, partly hampers our understanding of this process. The modern single-cell gene expression technologies, such as single-cell RT-PCR and RNA-seq, have enabled us to elucidate such heterogeneous cell subpopulations. However, the identification of a transcriptional regulatory network (TRN) for each cell subpopulation within a population and genes determining specific cell fates (lineage specifiers) remains a challenge due to the slower development of appropriate computational and experimental workflows. Here, we propose a computational differential network analysis approach for predicting lineage specifiers in binary-fate differentiation events.

methodsThe proposed method is based on a model that considers each stem cell subpopulation being in a stable state maintained by its specific TRN stability core, and cell differentiation involves changes in these stability cores between parental and daughter cell subpopulations. The method first reconstructs topologically different cell-subpopulation specific TRNs from single-cell gene expression data, literature knowledge and transcription factor (TF)-DNA binding-site prediction. Then, it systematically predicts lineage specifiers by identifying genes in the TRN stability cores in both parental and daughter cell subpopulations.

resultsApplication of this method to different stem cell differentiation systems was able to predict known and putative novel lineage specifiers. These examples include the differentiation of inner cell mass into either primitive endoderm or epiblast, different progenitor cells in the hematopoietic system, and the lung alveolar bipotential progenitor into either alveolar type 1 or alveolar type 2.

conclusionsThe method is generally applicable to any binary-fate differentiation system, for which single-cell gene expression data are available. Therefore, it should aid in understanding stem cell lineage specification, and in the development of experimental strategies for regenerative medicine.