Abstract
Single-cell transcriptomics provides a powerful tool to understand cell classes in neural circuits. In cortex, however, the identification of distinct biological cell types based on transcriptomic data is complicated by the existence of many incompletely-understood and closely-related neural classes, which can also show continuous variation in gene expression. The inhibitory cells of hippocampal area CA1 have been extensively characterized, leading to 23 currently defined cell types that provide a “Rosetta Stone” for transcriptomic analysis. Here, we studied the transcriptomes of 3638 CA1 inhibitory cells. Novel clustering methods identified all 23 previously described CA1 inhibitory types, while also suggesting 6 new inhibitory classes. Latent-factor analysis revealed a common continuum of expression of many genes within and between classes, which we hypothesized correlates with a continuum from faster-spiking cells that proximally target pyramidal cells, to slower active cells targeting pyramidal distal dendrites or interneurons. In vitro patch-seq of Pvalb interneurons supported hypothesis.