Question

Rna-Seq P-Values

0

Entering edit mode

11.9 years ago

siddharth.sethi5 ▴ 290

Hi ,

geneA

geneB

Can anyone tell me why Dbp(geneA) has a high (>0.05) p value and Prokr2(geneB) has a low p -value.
We can see there is a difference in RPKM for Dbp between the wild type and Mutant, then why didn't Dbp gets a good p -value assigned to it like Prokr2.

I find many other genes like this, need to know if there is any explanation for this.
I have done the differential expression testing using DESeq, EdgeR & Cufflinks, and this appears to be in all of them.
There are 4 replicates for Wt and Mut.

Thank you

rna-seq p-value • 3.8k views

ADD COMMENT • link updated 11.9 years ago by Istvan Albert 102k • written 11.9 years ago by siddharth.sethi5 ▴ 290

1

Entering edit mode

Sorry the images are not displayed, fixing it !!

ADD REPLY • link 11.9 years ago by siddharth.sethi5 ▴ 290

1

Entering edit mode

I fixed the images for you.

ADD REPLY • link 11.9 years ago by Damian Kao 16k

1

Entering edit mode

It most likely has to do with the amount of variation among replicates. It looks like your dbp gene has large variations among your mutant and wt replicates while your prokr2 gene has pretty consistent mutant replicates and large variation in wt replicates.

ADD REPLY • link 11.9 years ago by Damian Kao 16k

score 1 · Answer 1 · 2013-01-17

The first thing to keep in mind is that depending on the statistic and methodology that you are using the assumptions behind P values may not work all that well when the distributions are skewed or the data has large non-normal variations.

As Damian Kao points out in your second plot you have a distribution that has a very narrow width (frankly I think it is suspiciously narrow) that makes it unlikely that both samples would have been drawn from the same population, hence the low p-value.