Question: Change the FDR in DEXSeq
gravatar for arom2
13 months ago by
United States
arom20 wrote:

To Whom It May Concern,

Using DEXSeq, I am attempting to run an analysis of differentially used exons between two treatments groups at different False Discovery Rates (10%, 5%, and 1%). Yet the commands for running the test (testForDEU ) and retrieving results (DEXSeqResults) seem to have no argument for alternative FDR's. In the manual, It seems that only running HTML reports supplies an argument for this. Am I missing something where I can run three separate tests that also correspond to matching adjusted p value thresholds?


Results_from_test_10: FDR of 10% with Differentially used exons < 0.10

Results_from_test_05: FDR of 5% with Diff used exons < 0.05

Results_from_test_01: FDR of 1% with Diff used econs < 0.01


Thank you for your time!

A. Romney

ADD COMMENTlink modified 13 months ago by Jakub30 • written 13 months ago by arom20
gravatar for Jakub
13 months ago by
United Kingdom
Jakub30 wrote:

You can subset the results class with any cuttoff you like, as the results object has a "padj" column, which contains the the adjusted p value (also known as the FDR adjusted p-value), by some multiple testing adjustment, here using the Benjamini-Hochberg method.

At the end of chapter 4 of the vignette you will see how to do this:

table( dxr2$padj < 0.1 )

This will give you the number of DEU with a 10% FDR cutoff. Change that to 0.05 for 5% and 0.01 for 1%. If you want to get the subset of results object that meet that FDR cutoff I think something like this should work:

dxr2sig <- dxr2[ which(dxr2$padj < 0.01 ), ]

I am not sure what you mean by "FDR of 10% with Differentially used exons < 0.10" and by "three separate tests", as the differential exon usage test here is only run once, you are just exploring different arbitrary FDR cutoffs?

ADD COMMENTlink modified 13 months ago • written 13 months ago by Jakub30
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 160 users visited in the last hour