Diffbind fold values and parameter Fold in DBA.report function
1
0
Entering edit mode
5 months ago
Manko47 ▴ 10

Hello. Sorry if this question seems completely trivial but google search/biocondutor help site, didn't help me and I also spent more hours than I'd want actually trying to understand the difference between those two things.

I'm using Diffbind and DBA.report function to generate the table with full report on differentially binding sites. I set Fold parameter to 1, indicating that I want to only include results with absFold > 1. Yet when I set it to 1 it sets the parameter much higher and throws away every single results with absFold > 1.71 or around that (not sure about the exact number it's somewhere in between 1.71 and 1.725).

The default explanation of Fold argument in DBA.report function is "only sites with an absolute log Fold value greater than equal to this will be included in the report. This should be supplied as a log2() value." I even set the parameter to fold = log2(2) - obviously the result is the same. The explanation of Fold column in the object generated by DBA.report makes it clear that it's using log values therefore it should be the same. Can please anybody help with this as I am nearly sure that my mind can't process something that should be obviously simple.

DESEQ2 ATAC-seq DBA.report Diffbind • 328 views
ADD COMMENT
1
Entering edit mode
5 months ago
Manko47 ▴ 10

Thread solved

Fold function sets X as lfcthreshold = X therefore changing expected logFC in null hypothesis when doing the differential analysis from 0 to X, instead of just doing post-hoc filtering of results one would got under the hypothesis that lfcthreshold = 0 to all genes with abs(lfc) > X. In short it works as if you set the lfcthreshold parameter in results function in DESeq2.

ADD COMMENT

Login before adding your answer.

Traffic: 2434 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6