AbstractStem cell populations within developing embryos are diverse, composed of many different sub-populations of cells with varying developmental potential. The structure of stem cell populations in cell culture remains poorly understood and presents a barrier to differentiating stem cells for therapeutic applications. In this paper we develop a framework for controlling the architecture of stem cell populations in cell culture using high-throughput single cell mRNA-seq and computational analysis. We find that the transcriptional diversity of neural stem cell populations collapses in cell culture. Cell populations are depleted of committed neuron progenitor cells and become dominated by a single pre-astrocytic cell population. By analyzing the response of neural stem cell populations to forty distinct signaling conditions, we demonstrate that signaling environments can restructure cell populations by modulating the relative abundance of pre-astrocyte and pre-neuron subpopulations according to a simple linear code. One specific combination of BMP4, EGF, and FGF2 ligands switches the default population balance such that 70% of cells correspond to the committed neurons. Our work demonstrates that single-cell RNA-seq can be applied to modulate the diversity of in vitro stem cell populations providing a new strategy for population-level stem cell control.HighlightsNatural progenitor diversity in the brain collapses during in vitro culture to a single progenitor typeLoss of progenitor diversity alters fate potential of cells during differentiationLarge scale single-cell signaling screen identifies signals that reshape population structure towards neuronal cell typesSignals regulate population structure according to a simple log-linear model
Retinoid Machinery in Distinct Neural Stem Cell Populations with Different Retinoid Responsiveness
