05_Cluster_Unique_Peaks_v2.R - changes needed for identifying unique gene activities

[owenwilkins]

Was just wondering if someone might be able to comment on what changes would be needed to adapt the 05_Cluster_Unique_Peaks_v2.R script to make it appropriate for detecting unique gene activities, rather than using peaks.

From a review of the script and the methods section of the manuscript, it seems like at changing the 1st argument of the uniqueFeatures() function from:
edgeR::cpm(assay(sePB),log=TRUE,prior.count=3)
to
log2(edgeR::cpm(2^assay(sePB)-1) +1) would be required.

Similarly, it seemed like it may be necessary to change the lines 144, 175, 185, 254, and 263 in the createPseudoBulk() function, to first convert back to gene activity scores and then log transform, however I wasn't sure about that.

Would be great to hear anyones thoughts on those changes and any others that might be required to work with gene activities rather than peaks.

Thanks in advance!