Question: fdrtool corrected p-value histograms look weird
0
gravatar for hele7
16 months ago by
hele70
hele70 wrote:

Hi,

I'm performing DE analysis on the miRNA secretome of two cell lines. I'm comparing cells, fraction1,  fraction 2 of the secretome for both of them, resulting in 3 comparisons: fraction1/cells; fraction2/cells; fraction1/fraction2. 

For cell line 1 the p-value histogram before correction look fine (grey barplots), however after p-value correction with fdrtool the distribution is far from perfect (purple barplots) https://ibb.co/ecvUd7 Shouldn't the corrected p-value histograms have the same uniform distribution as uncorrected p-value histograms? Do I need to be worried when using the corrected p-values in defining the DE miRNAs?

Cell line 2 p-value histograms look more problematic https://ibb.co/dby4BS and unfortunately I don't have any idea what could be the reason?? The input gene list has been prefiltered with

keep <- rowMeans (counts(dds)) > 1

resulting in having 469 miRNAs included in the analysis. The dispersion plot can be found here: https://ibb.co/gKwman 

There is definitely smth wrong with my analysis, and considering the histograms, I can't use these corrected p-values to call my DE miRNAs. Yet, I'm unable to track down what the problem could be...Can somebody please look into this? Being not 100% at home with DE analysis/DESeq2, I would appreciate all the help I can get :) Let me know if any piece of code is needed to track down the problem...

ADD COMMENTlink modified 16 months ago by Michael Love24k • written 16 months ago by hele70
1

hi,

This is an fdrtool question, more than a DESeq2 question. I'm changing the tag.

For cell line 1, there is no reason to use p-value correction. For cell line 2, it does seem that the test is being conservative but after fdrtool the histograms don't look much better. I'd stick with the original adjusted p-values produced by DESeq2.

ADD REPLYlink modified 16 months ago • written 16 months ago by Michael Love24k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 16.09
Traffic: 338 users visited in the last hour