Thank you for creating an awesome package. I was optimizing my workflow using this package and noticed in our spatial data that only has 1k targets we were throwing out most of the gene sets from our gmt. Digging into your code, I noticed that these gene sets are removed after calculating the AUC.
I added a preemptive filtering step to our gmt that mimics your default filtering and was able to cut hours off our pathway step. I know this was built for full transcriptome scRNA-seq data and not 1k targets so you wouldn't have expected this level of filtering but is there a reason for filtering after doing the calculations and not before the bulk of the function?
Hi,
Thank you for creating an awesome package. I was optimizing my workflow using this package and noticed in our spatial data that only has 1k targets we were throwing out most of the gene sets from our gmt. Digging into your code, I noticed that these gene sets are removed after calculating the AUC. I added a preemptive filtering step to our gmt that mimics your default filtering and was able to cut hours off our pathway step. I know this was built for full transcriptome scRNA-seq data and not 1k targets so you wouldn't have expected this level of filtering but is there a reason for filtering after doing the calculations and not before the bulk of the function?
Thanks, Maddy