WGCNA -RNAseq data- strong correlation with TIN
1
0
Entering edit mode
5 months ago
Geor • 0

Hello, I am using WGCNA on RNA sequencing data and have correlated modules with clinical traits. I got very significant (e-15,e-18) values for TIN and some modules. I run it on residuals matrix as well but I got even more sign values(e-35!). Can I do sth about it? Or maybe I making mistakes here?Thanks for help. Georg

WGCNA TIN RNAsequencing • 367 views
ADD COMMENT
0
Entering edit mode

I think this is quite common for very small modules that includes genes strongly correlated with one of your traits

ADD REPLY
0
Entering edit mode

Thanks Andres! In my case the the most significant values I see with TIN and the 1st biggest module(8k genes), followed by 2nd biggest module (3k genes) and 4rd biggest (300 genes)...

ADD REPLY
0
Entering edit mode

how many genes do your have in the TIN module?

edit: I see know... TIN is not a module but the trait of interest. If TIN is the Transcript integrity number, you might have serious problem here

ADD REPLY
1
Entering edit mode
5 months ago
cera.fisher ▴ 10

How many samples do you have?

How wide is the range in the TIN among your samples?

If you have enough samples, maybe you can afford to filter out samples with low TIN values. It's well known that RNA species will degrade differentially, AND most transcript abundance estimation methods are taking into account expected transcript length. Your degraded samples' expression profiles will look very different from intact samples. So if some samples have low TIN (which is a measure of sample degradation) then WGCNA is probably seeing a change in gene regulation across your samples that does, in fact, correlate with degradation.

I don't think that there is anything that you can do to "clean up" the data to fix this. My move would be to pick a threshold for TIN, and analyze low TIN and high TIN samples separately. I would compare module preservation between the two, to see if there are any modules detected in the high TIN samples still detected in the low TIN samples. But I would probably end up just going forward with analysis on high TIN (=high transcript intactness) samples.

"Enough samples" for WGCNA purposes is at least 12, mind.

Best of luck! You can't fix biology problems with computers.

ADD COMMENT

Login before adding your answer.

Traffic: 2694 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