Linking transcriptomes with morphological and functional phenotypes in mammalian retinal ganglion cells

Retinal ganglion cells (RGCs) are the brain's gateway to the visual world. They can be classified into different types on the basis of their electrophysiological, transcriptomic, or morphological characteristics. Here, we characterize the transcriptomic, morphological, and functional features of 472 high-quality RGCs using Patch sequencing (Patch-seq), providing functional and morphological annotation of many transcriptomic-defined cell types of a previously established RGC atlas. We show a convergence of different modalities in defining the RGC identity and reveal the degree of correspondence for well-characterized cell types across multimodal data. Moreover, we complement some RGC types with detailed morphological and functional properties. We also identify differentially expressed genes among ON, OFF, and ON-OFF RGCs such as Vat1l, Slitrk6, and Lmo7, providing candidate marker genes for functional studies. Our research suggests that the molecularly distinct clusters may also differ in their roles of encoding visual information.