Based on the comment below from #116, we should add in a quick linear regression analysis to the current gene comparison between alevin-fry cr-like and cellranger. This will help us identify if there are any genes that may be found in both sets but are consistently under or over represented in one of these tools.
I think this analysis is good as is: the one other thing I might add if I were doing it would be to just look at the correlation between cellranger and cr-like-em with a linear model and identify any genes with large residuals that might fall somewhere other than at the bottom of the distribution.
That is, I would do something like lm(cr_like_em$logmean_expr ~ cellranger$logmean_expr) for each sample, then look at the residuals to see if there are other genes that are consistently over or undercounted, even if they are not showing up in the "unique to a method" set.
Based on the comment below from #116, we should add in a quick linear regression analysis to the current gene comparison between alevin-fry
cr-likeand cellranger. This will help us identify if there are any genes that may be found in both sets but are consistently under or over represented in one of these tools.