immunogenomics/presto

find genes with expression in cluster A higher by a given FC than any other cluster

kaizen89 opened this issue · 2 comments

kaizen89 commented

Hi,
Thanks for this nice package.
I was wandering if it was possible to add a feature to find genes with an expression in cluster A , let's say 2 times higher than any other cluster among B,C and D, instead of grouping all the cells outside of cluster A which could dilute the expression of the genes.
Thanks!

slowkow commented

@kaizen89

We can re-reun presto::wilcoxauc(mat, group) as many times as we wish with any variables we define. We can decide for ourselves how to define mat (e.g. what cells to include or not). We can also define group however we wish (e.g. 1 when cluster A and 2 when cluster B, C, or D).

kaizen89 commented

Thanks for your answer @slowkow.
So if I have 10 clusters, that would make 55 different comparisons, I was hoping to use another approach.^^
I will try to figure out a solution that would not require to run all the combinations.