부산대학교 도서관

• A neuronal type is a testable hypothesis: similarity indicates functional equivalency. • Image registration together with morphology algorithms defines types across datasets. • Single-cell transcriptomics produces cell-atlases with consistent clusters. • Specific genetic lines bridge transcriptomic clusters to morphological types. • Genetic driver lines, connectomics and transcriptomics make a full Drosophila neuronal parts list imminent. At around 150 000 neurons, the adult Drosophila melanogaster central nervous system is one of the largest species, for which a complete cellular catalogue is imminent. While numerically much simpler than mammalian brains, its complexity is still difficult to parse without grouping neurons into consistent types, which can number 1–1000 cells per hemisphere. We review how neuroanatomical and gene expression data are being used to discover neuronal types at scale. The correlation among multiple co-varying neuronal properties, including lineage, gene expression, morphology, connectivity, response properties and shared behavioral significance is essential to the definition of neuronal cell type. Initial studies comparing morphological and transcriptomic definitions of neuronal type suggest that these are highly consistent, but there is much to do to match these approaches brain-wide. Matched single-cell transcriptomic and morphological data provide an effective reference point to integrate other data types, including connectomics data. This will significantly enhance our ability to make functional predictions from brain wiring diagrams as well facilitating molecular genetic manipulation of neuronal types. [ABSTRACT FROM AUTHOR]