RNASEQ GE analysis, p-values zero whereas fold changes are not??
0
0
Entering edit mode
8.3 years ago

Hi all, I am quite new around here, so if it was asked sorry for creating a duplicated post, but I could not find it. I have TCGA data downloaded and trying to analyze differentially expressed genes by using edgeR.

in BRCA, two of genes have p-values and FDR zero (I don't know if it is an underflow or literally zero) and fold changes around -3,5 and -2,4. I don't know how to interpret these extremely low p-values. Are they because of a biological variation or a statistical error? Maybe I should have made some kind of normalization? Additionally, my p-values of other genes are extremely low, too (around e-305) and for other cancers I found p-values around e-60. So maybe, I am missing something.

Thanks!

edit: I am working with patient tumor/normal tissue pairs.

RNA-Seq R edgeR • 3.2k views
ADD COMMENT
0
Entering edit mode

https://support.bioconductor.org/p/64787/#64792 seems to answer your question.

ADD REPLY
0
Entering edit mode

Thanks for your interest, however my common dispersion value is 0,6 for this type of cancer and it is consistent with other types (they are also around 0,57-0,6). I observed these radical p-values only in BRCA.

ADD REPLY
0
Entering edit mode

Small p-values indicate significant change. Fold change (assuming log2) of -3.5 or 2.4 indicates decrease in expression of genes. So the genes of your interest are significantly downregulated.

ADD REPLY
0
Entering edit mode

Thanks for clarification, I just found strange these extremely low p-values.

ADD REPLY
0
Entering edit mode

Yes but p-values far smaller than reasonable are a reason to be suspicious.

ADD REPLY

Login before adding your answer.

Traffic: 2069 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