Transcriptome-based systematic discovery of GABAergic neurons in the neocortex



 DESCRIPTION (provided by applicant): The integrated sensory, motor, and cognitive abilities that guide adaptive behavior in mammals emerge from neural circuit operations in the neocortex. Understanding the organization of cortical circuits requires comprehensive knowledge of the basic cellular components. The neocortex consists of approximately 80% glutamatergic pyramidal neurons and 20% GABAergic neurons. Although a minority, GABA interneurons are exceptionally diverse, and this diversity may be crucial in regulating the balance and functional operations of cortical circuits. However, systematic identification, enumeration and classification of GABAergic neurons have been a challenging goal. We hypothesize that distinct transcription programs underlie GABA prototype identity and diversity as defined by their position, morphology and basic innervation pattern. Thus we suggest that transcription profiling provides a fundamental starting point and efficient strategy for cell type discovery. Here we propose a multi-faceted approach that integrates genetic targeting, single cell transcriptomics, statistical and computational analysis, morpho-physiological studies to systematically identify and classify GABAergic neurons. We focus on GABA neurons derived from the embryonic medial ganglionic eminence (MGE), which constitute two-third of cortical interneurons, and for which we have built effective genetic tools. We have established a robust single cell RNAseq (scRNAseq) method that allows high resolution transcriptome profiling through single mRNA counting using nucleotide barcodes. We will take a two-step "Targeted-Saturation" cell screen approach toward systematic discovery of cortical GABA neurons. First, we will apply scRNAseq to a set of GABA subpopulations, captured by intersectional genetic targeting, and discover their distinct transcription signatures. With these phenotype- characterized populations, we hone our statistical analysis to distinguish biological signal vs experimental noise and artifacts, and shape our computation algorithm based on biological ground truth. Thus in contrast to a unsupervised clustering approach to transcriptome analysis, we incorporate extensive empirical information that enable a biology-motivated supervised approach, where well-delineated phenotypes play the key role of training the algorithm and classifier. Second, we will apply scRNASeq to increasingly broader genetic-defined populations of MGE-derived GABA neurons in the primary motor cortex. We will discover transcriptome-predicted cell types and build 2nd round driver lines that target and validate a subset of novel cell types. Our study will build a comprehensive catalog of a major cohort of cortical GABAergic neurons by integrating transcription profiles and basic cell phenotypes. This will establish a cellular foundation for studying inhibitory circuit organization, function, and dysfunction. 1