A molecular atlas of the developing ectoderm defines neural, neural crest, placode, and nonneural progenitor identity in vertebrates

During vertebrate neurulation, the embryonic ectoderm is patterned into lineage progenitors for neural plate, neural crest, placodes and epidermis. Here, we use Xenopus laevis embryos to analyze the spatial and temporal transcriptome of distinct ectodermal domains in the course of neurulation, during the establishment of cell lineages. In order to define the transcriptome of small groups of cells from a single germ layer and to retain spatial information, dorsal and ventral ectoderm was subdivided along the anterior-posterior and medial-lateral axes by microdissections. Principal component analysis on the transcriptomes of these ectoderm fragments primarily identifies embryonic axes and temporal dynamics. This provides a genetic code to define positional information of any ectoderm sample along the anterior-posterior and dorsal-ventral axes directly from its transcriptome. In parallel, we use nonnegative matrix factorization to predict enhanced gene expression maps onto early and mid-neurula embryos, and specific signatures for each ectoderm area. The clustering of spatial and temporal datasets allowed detection of multiple biologically relevant groups (e.g., Wnt signaling, neural crest development, sensory placode specification, ciliogenesis, germ layer specification). We provide an interactive network interface, EctoMap, for exploring synexpression relationships among genes expressed in the neurula, and suggest several strategies to use this comprehensive dataset to address questions in developmental biology as well as stem cell or cancer research.Author summary: Vertebrate embryo germ layers become progressively regionalized by evolutionarily conserved molecular processes. Catching the early steps of this dynamic spatial cell diversification at the scale of the transcriptome was challenging, even with the advent of efficient RNA sequencing. We have microdissected complementary and defined areas of a single germ layer, the developing ectoderm, and explored how the transcriptome changes over time and space in the ectoderm during the differentiation of frog epidermis, neural plate, and neural crest. We have created EctoMap, a searchable interface using these regional transcriptomes, to predict the expression of the 31 thousand genes expressed in neurulae and their networks of co-expression, predictive of functional relationships. Through several examples, we illustrate how these data provide insights in development, cancer, evolution and stem cell biology.