Abstract
Kidney development requires the coordinated growth and differentiation of multiple cells. Despite recent single cell profiles in nephrogenesis research, tools for data analysis are rapidly developing, and offer an opportunity to gain additional insight into kidney development. In this study, single-cell RNA sequencing data obtained from embryonic mouse kidney were re-analyzed. Manifold learning based on partition-based graph-abstraction coordinated cells, reflecting their expected lineage relationships. Consequently, the coordination in combination with ForceAtlas2 enabled the inference of parietal epithelial cells of Bowman's capsule and the inference of cells involved in the developmental process from the S-shaped body to each nephron segment. RNA velocity suggested developmental sequences of proximal tubules and podocytes. In combination with a Markov chain algorithm, RNA velocity suggested the self-renewal processes of nephron progenitors. NicheNet analyses suggested that not only cells belonging to ureteric bud and stroma, but also endothelial cells, macrophages, and pericytes may contribute to the differentiation of cells from nephron progenitors. Organ culture of embryonic mouse kidney demonstrated that nerve growth factor, one of the nephrogenesis-related factors inferred by NicheNet, contributed to mitochondrial biogenesis in developing distal tubules. These approaches suggested previously unrecognized aspects of the underlying mechanisms for kidney development.