Cross-species single cell analysis of convergent cell differentiation

Abstract

Since the widespread use of single-cell RNA-sequencing (scRNA-seq) technology, researchers have conducted sequencing with various species and conditions of tissues, leading to an accumulation of diverse scRNA-seq datasets. However, integrating and comparing these data poses challenges due to differences in experimental conditions of each data. Combining multiple datasets for comparison is crucial in single-cell analysis to make them comparable, including normalization and batch effect removal using consistent criteria. Particularly for multi-species data, aligning transcriptomic annotations across species is crucial because of the genomic and transcriptomic differences of each species. This dissertation focuses on addressing these challenges to compare scRNA-seq data of similar cell types across various conditions. Initially, I analyzed the temporal changes in cell composition during the development of mucociliary epithelium (MCE) in Xenopus laevis, comparing them over time and compared the gene expression with MCEs from human and mouse airways. These findings demonstrate the similarities in cell composition between mammalian airways and Xenopus MCE tissue, as well as the gene expression in individual cell types. Next, I compared the transcriptomes of multiciliated cells (MCCs), a common cell type found in various organs, including the airway, oviduct, and ependyma. To ensure a fair comparison, scRNA- seq data of each tissue from human and mouse were subjected to unsupervised comparison, revealing shared core MCC marker genes across different organs while tissue-specific gene expression differences still exist, particularly in precursor cell markers. Further comparison using developmental time course data from human airway and ependyma confirmed convergent differentiation of MCCs despite differences in precursor cell populations during development. Lastly, I examined the changes occurring during spinal cord regeneration in Xenopus tropicalis tadpoles and compared them with data from the closely related species Xenopus laevis. This comparison revealed the emergence of two microglia subtypes during spinal cord regeneration, with a significant increase in the population of one subtype. By gene ontology analysis using marker genes for each subtype, I observed that the subtype involved in regeneration exhibited pronounced expression of genes related to morphogenesis. Validation through HCR staining of subtype marker genes confirmed distinct population changes of subtypes in Xenopus tropicalis tadpoles. In summary, this study highlights the importance of normalization, batch effect removal, and cross-species transcriptomic annotation alignment in comparing scRNA-seq data, providing insights into the similarities and differences in cell composition and differentiation processes across various conditions and species.