Closed kvshams closed 2 years ago
If you lower the logfc.threshold
(eg, logfc.threshold=0.01
), can you identify any peaks that change in the other direction (<0
?)
Yes, I can see that. logfc
is highly skewed towards one direction >0
(Both the data has comparable sequencing depth)
Ok, I'm not sure if you're suggesting there's something wrong in the function or just a general question about why this could happen? It seems like this is not a bug in the FindMarkers()
function, especially if you can identify peaks with negative logFC when lowering the threshold. Are there any clearly differentially accessible regions that are being missed by the test?
Any update here?
Hi @timoast
This is in continuation with #744. After following the merging vignette, DA is not capturing the biology and perhaps the LogFC using
FindMarkers
, all region showed in the same direction (>0
). In the previous integration (#744) the DA regions were consistent with the previously known biology.What would have went wrong?. Below is the complete script that I have used.