Single-cell-level spatial gene expression in the embryonic neural differentiation niche.

With the rapidly increasing availability of high-throughput in situ hybridization images, how to effectively analyze these images at high resolution for global patterns and testable hypotheses has become an urgent challenge. Here we developed a semi-automated image analysis pipeline to analyze in situ hybridization images of E14.5 mouse embryos at single-cell resolution for more than 1600 telencephalon-expressed genes from the Eurexpress database. Using this pipeline, we derived the spatial gene expression profiles at single-cell resolution across the cortical layers to gain insight into the key processes occurring during cerebral cortex development. These profiles displayed high spatial modularity in gene expression, precisely recapitulated known differentiation zones, and uncovered additional unknown transition zones or cellular states. In particular, they revealed a distinctive spatial transition phase dedicated to chromatin remodeling events during neural differentiation, which can be validated by genomic clustering patterns, epigenetic modifications switches, and network modules. Our analysis further revealed a role of mitotic checkpoints during spatial gene expression state transition. As a novel approach to analyzing at the single-cell level the spatial modularity, dynamic trajectory, and transient states of gene expression during embryonic neural differentiation and to inferring regulatory events, our approach will be useful and applicable in many different systems for understanding the dynamic differentiation processes in vivo and at high resolution.