Question

What doest this p.value distribution mean?

0

Entering edit mode

chipolino • 0

@chipolino-15565

Last seen 4.5 years ago

Hi everyone,

I am doing an RNA-seq analysis with DESeq2. I have 3 patients with disease (affected) and 3 healthy patients (unaffected). After following standard DESeq2 pipeline (as described here), I got my table with results and p.value column. I decided to check the distribution of p.values and here it is:

I do know that we expect to see something similar to uniform distribution, apparently we don't see it here... Does it mean that there is a problem with the data or an experiment design (I used simple formula, which was taking into account genders of the patients: ~ Gender + Condition)?

Thank you

deseq2 pvalue • 1.7k views

ADD COMMENT • link updated 5.9 years ago by Michael Love 41k • written 5.9 years ago by chipolino • 0

1

Entering edit mode

Can you show a PCA plot? This sometimes comes from batch effects, or other related correlations or differences in the variance than from the model (e.g. if one group had much larger dispersion, or an outlier).

ADD REPLY • link 5.9 years ago Michael Love 41k

0

Entering edit mode

PCA plot is not great either... Zeros are unaffected patients, Ones are affected. PC2 divides them by gender.

ADD REPLY • link 5.9 years ago chipolino • 0

score 3 · Answer 1 · 2018-06-11

3

Entering edit mode

Michael Love 41k

@mikelove

Last seen 4 hours ago

United States

I guess the non-uniformity could have to do with the partial confounding of sex and condition. This seems like a good case where you could try to alleviate the distribution using fdrtool using Bernd Klaus' tutorial:

http://www-huber.embl.de/users/klaus/Teaching/DESeq2Predoc2014.html#inspection-and-correction-of-pvalues