Biaxial plot of genes' expression per cell?

[innesbre]

Sonya's immunology buddies asked if genes could be viewed on a biaxial plot, a la FACS analysis. Given that each cell library is an incomplete (~10%) sample of that cell's transcriptome, this probably isn't the greatest idea, since most cells will pile up on one/both axes since the genes of interest will be missing. However, it could be done, and colour each cell by cluster, as well as including a meta-dot (with whiskers for IQR/SE/SD?) for cluster-wise average gene expression. This might be a way to move toward the "cell-type signature" idea of having multiple genes uniquely representing a cluster, rather than just single marker genes.