,
8 tweets,
5 min read
Read on Twitter
Very excited to announce two papers from our group are out today. Amazing collaborations with @FertigLab & @sethblackshaw and led by @genesofeve & @Brian_S_Clark. The landscape of mouse retinal development using 10x scRNA-Seq. doi.org/10.1016/j.neur…
doi.org/10.1016/j.cels… (1/n)
doi.org/10.1016/j.cels… (1/n)
We use these data to explore mechanisms of retinal cell fate specification (doi.org/10.1016/j.neur…) and how learned latent spaces can be used to represent knowledge transfer across model systems and experimental paradigms (doi.org/10.1016/j.cels…) (2/n)
In the first manuscript, we define developmental trajectories for each of the major cell lineages in retinal development and identify key genes associated with these fate decisions. We also identify the NFI family of TFs as key players in progenitor competence... (3/n) 
Overexpression and KO of the NFI factors show that these genes contribute to establishment of 'late born' retinal cell type identities, but also help to direct progenitor cell quiescence. Triple Nfia/b/x KO showed proliferating progenitors well into early postnatal period. (4/n)
In the second study <link> we learn a set of latent spaces from these data and show that these patterns are biologically meaningful gene expression representations of cellular features including cell type/state, cell cycle phase, sex, etc. (5/n)
@GenesOfEve led the development of a transfer learning framework to A) validate these latent spaces by projection to other related single cell and bulk RNA-Seq datasets and B) determine how these reduced dimensional feature representations are used in other contexts...(6/n)
...including cross-species comparisons, comparisons between developing retina and brain, across data types (ATAC), and ultimately using ProjectR to explore the use of cellular features across the Tabula Muris dataset, highlighting specific shared features of cells across (7/n)
Very proud of this project, and excited to be a part of the wonderfully collaborative team responsible for these companion papers. Data are available on github (github.com/gofflab/develo…) and scCoGAPS and ProjectR are now available on Bioconductor (bioconductor.org/packages/relea…). (8/8)
