I want to thank Aboozar Monavarfeshani, Seth Blackshaw, Teresa Puthussery, Jared Tangeman, Olivier Marre and others — especially Joshua Sanes and my fantastic PI Karthik Shekhar @karthikshekhar.bsky.social for guiding me through this huge endeavor. It was an honor to lead this work! 13/13

Read the full details in our preprint out today on bioRxiv: www.biorxiv.org/content/10.6... 12/13

In lamprey, cells co-expressed markers of both GABAergic ACs and RGCs, suggesting they share a common evolutionary precursor. Interestingly, glycinergic ACs appear to have diverged first. 11/13

Most of these types trace back to non-mammalian vertebrates like birds, reptiles, amphibians, and fish — meaning these cell types are over 500 million years old! 10/13

Based on conserved transcription factor expression, we derived a regulatory code that may specify this diversity. Below is the code for glycinergic amacrine cells. 9/13

Conservation was remarkably high — most if not all AC types appear universal across mammals. Strikingly, ACs are more conserved than retinal ganglion cells, despite being more diverse. 8/13

To annotate orthotypes, we used known markers and identified many new ones. Because these markers are conserved, researchers can use them to genetically target these novel cell types in their favorite animal model! 7/13

How do we know it’s working? We manually annotated three known types (SAC, A2, VG3) across species — orthotypes recovered them with ~98% precision and ~96% recall. 6/13

Our primary goal was to identify clusters of molecularly conserved cell types. After projecting cells from different species into a common space, we performed unsupervised clustering — calling these clusters “orthotypes” (orthologous types). 5/13

We integrated and analyzed single-cell RNA-sequencing atlases across 24 species. Due to the immense size of this data, we split our analyses into two parts: 1) tetrapods and 2) non-mammalian vertebrates. 4/13

We focused on the retina — an accessible, well-studied brain region. Its primary inhibitory neurons are amacrine cells (ACs). A handful are well-studied across species, but the vast majority (~80%) remain uncharacterized. 3/13

Inhibitory neurons do the computational heavy lifting of neural circuits. As most neuroscience relies on animal models, knowing whether our neural building blocks are conserved will help us study their origin, function, and role in disease. 2/13

Inhibitory neurons are among the most transcriptomically diverse class of neurons in the CNS, with some brain regions having 60+ distinct cell types. Do humans share the same repertoire as rodents? Birds? Fish? 1/13