Why the maintenance factors are not essential

Why the maintenance factors are not essential for normal dhpg development is unclear. One possibility is that under normal conditions, the inner cell mass is transient, so that even the complete absence of any of these factors does not produce the full detrimental effect. However, if the inner cell mass cells were subjected to a slowed or prolonged growth period, a situation similar to mESC culture in vitro, then the deficiency of these factors would result in the developmental arrest of blastocysts. This hypothesis is supported by the finding that gp130−/− blastocysts were arrested only when they were subjected to diapause, a phenomenon of delayed blastocyst development that has evolved in some mammals, including mice, to get around adverse conditions during pregnancy (Nichols et al., 2001). The following are the supplementary data related to this article.
Acknowledgments This work was partly supported by grants from the National Institutes of Health (NS081684, CA97124, and the Core grant CA016672). P.S. is the recipient of the Dodie P. Hawn Fellowship in Genetics. We thank Dawn Chalaire for editing the manuscript.
Introduction Stem cell differentiation is a complex process that involves a multitude of regulatory mechanisms at different organizational levels. Despite accumulating experimental evidence, identification of lineage specifiers and understanding of the regulatory mechanisms of cell-fate commitments are partially hampered by the heterogeneity in stem cell populations. Indeed, stem cells in tissues and culture exist as a heterogeneous population consisting of different subpopulations, which are characterized by different gene expression states driven by different underlying TRNs. Different TRNs in turn determine different propensities for cell fate decision. Hence, conventional bulk gene expression profiling and ChIP-seq approaches generated from a heterogeneous population of cells appear to be suboptimal for studying stem cell differentiation (Moignard et al., 2013). Recent development of modern technologies for single-cell gene expression studies, such as single-cell RT-PCR and RNA-seq, have made possible gene expression profiling of hundreds of cells. They have been successfully used for elucidating heterogeneity in different stem cell systems, including the early embryonic development (Guo et al., 2010; Tang et al., 2010), hematopoiesis (Moignard et al., 2013; Guo et al., 2013), induced pluripotent stem cells (Buganim et al., 2012) and lung alveolar development (Treutlein et al., 2014). Nevertheless, a remaining challenge is the development of computational methods for elucidating complex molecular interaction networks and predicting lineage specifiers within a heterogeneous cell population. A couple of studies has proposed computational workflows for predicting cell lineage specifiers by reconstructing a single TRN that represents multiple cell types (Xu et al., 2014; Moignard et al., 2015). However, it has been revealed that cell subpopulation-specific TRNs showed significant rewiring during differentiation (Moignard et al., 2013). Hence, TRNs that are differentially reconstructed for different cell subpopulations provide a more realistic picture of underlying transcriptional regulatory mechanisms. Here, we introduce a general method for predicting lineage specifiers in binary-fate differentiation events based on the reconstruction of cell subpopulation-specific TRNs using single-cell gene expression data. Our method is based on a model, in which each stem cell subpopulation is considered to be in a stable state maintained by a TRN stability motif. We particularly focused on a set of circuits known as strongly connected components (SCCs) that we previously used for the prediction of reprogramming determinants (Crespo and Del Sol, 2013). The model further assumes that the stability of a parental stem cell subpopulation, which differentiates into two mutually exclusive daughter cell subpopulations, is maintained by a balance between the two opposing differentiation forces exerted by lineage specifiers for each of the two daughter cell subpopulations. Indeed, this “seesaw model” of stem cell differentiation has been observed during mesendodermal and ectodermal specification of embryonic stem cells (ESCs) (Montserrat et al., 2013; Shu et al., 2013). In this case, the balanced expression of a mesendodermal specifier, Pou5f1, and an ectodermal specifier, Sox2, which mutually activate each other, maintains the pluripotent state. Hence, the method searches for opposing lineage specifier pairs that reside in the TRN stability core of the parental cell subpopulation, and exhibit a significantly unbalanced expression ratio in the daughter cell sub-populations with respect to the parental cell subpopulation.