SEQanswers

SEQanswers (http://seqanswers.com/forums/index.php)
-   Bioinformatics (http://seqanswers.com/forums/forumdisplay.php?f=18)
-   -   p-value and fold change (http://seqanswers.com/forums/showthread.php?t=75172)

renuka22 04-02-2017 05:32 PM

p-value and fold change
 
Hello,

I got my rna-seq analysis from the company and they have assigned p-value to each gene. Is it possible to have a p-value from only one set of data since I had sent only one batch and there were no replicates? Also, some of my genes have p>0.05 (not differentially expressed) but high fold-change increase. So which of the two should be considered important - p-value or fold change?

Thanks

dpryan 04-03-2017 01:10 AM

If you only sent them single samples from your groups then the p-values are largely meaningless. There's not a direct correspondence between fold-change and p-value, since lowly expressed genes will randomly show higher fold-changes (i.e., the p-value is a function of expression level and fold change).

renuka22 04-03-2017 07:35 AM

Ok. But can the p-value be between my control and treatment samples as this is what they mentioned in their analysis and p<0.05 means those genes are differentially expressed?

Thanks

dpryan 04-03-2017 11:50 PM

The presumption is that the p-value is from a poisson test, so yes it's between the two samples.

Yes, p<0.05 would normally mean differentially expressed, aside from the fact that the p-values in your case are largely meaningless.

renuka22 04-04-2017 11:22 AM

So I guess the p-value is between technical replicates hence meaningless. Thanks a lot!


All times are GMT -8. The time now is 12:00 PM.

Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2020, vBulletin Solutions, Inc.