scATAC-seq based references?

[jeremymsimon]

I've been using ProjecTILs and the CD4/CD8 human T cell reference atlases in some scRNA-seq work lately and it has been very useful for us. I wonder whether you or @ncborcherding (via https://github.com/ncborcherding/utility) have any plans to make any scATAC-seq-based references?

I know that I could use promoter accessibility as a proxy for gene activity and then perhaps perform the projections based on that, but I'd imagine this would lose/ignore a lot of distal regulatory elements important for cell-type-specific activity. It would be great if there was a way to project directly in the regulatory element space to enable an analysis like this!

[mass-a]

Hello Jeremy,
thanks for your interest in our tools.
I agree that scATAC-seq references would be useful (possibly multimodal references even more so!). However that would take not only building the actual maps, but some considerable rewriting of the projection algorithm. From our side I don't see it happening in the near future, but would be happy to collaborate if someone else is willing to take it up.