Here, we have explored the transcriptional architecture underlying the core identity of GABAergic neurons in the cerebral cortex. Unlike recent studies that classify neurons using unsupervised statistical clustering of transcriptomes from unbiased (Zeisel et al., 2015) or relatively broad populations (Tasic et al., 2016), we analyzed single cell transcriptomes of 6 genetically labeled and phenotypically well characterized GABAergic types or subpopulations to discover their core molecular features. Using these anatomically and physiologically defined subpopulations as an assay, we designed a computational genomics strategy to screen through the ~620 HGNC (Human Genome Nomenclature Committee) gene families for those whose differential expression reliably distinguish these subpopulations. Remarkably, approximately 40 gene families implicated in regulating synaptic communication best distinguish these subpopulations. These gene families constitute 6 functional categories that include cell adhesion molecules, neurotransmitter and modulator receptors, ion channels, regulatory components of membrane-proximal signaling pathways, neuropeptides and vesicular release components, and transcription factors. Combinatorial and coordinated expression of select family members across these categories shapes a molecular scaffold along the cell membrane that appears to customize the pattern and property of synaptic communication
