SEQanswers

SEQanswers (http://seqanswers.com/forums/index.php)
-   Bioinformatics (http://seqanswers.com/forums/forumdisplay.php?f=18)
-   -   change default FDR on edgeR (http://seqanswers.com/forums/showthread.php?t=62687)

tinguzman 09-14-2015 07:47 PM

change default FDR on edgeR
 
Hi,

Could anyone please tell me how to change the FDR from default (0.05) to 0.0001 in edgeR?

et <- exactTest(y, pair=c("1","2"))
out.adjusted <- topTags(et, n=Inf, adjust.method="BH")
out.adjusted.keep <- out$table$FDR <= 0.0001

I'm not sure if i understand this correctly, topTags function with adjust.method will adjust the pvalues from et, using BH/FDR set at 5%, which is the default. then, the FDR column of the output file are the adjusted p.values, on this case retaining only those with <=0.0001 values.

appreciate your help!

dariober 09-15-2015 02:10 AM

The BH/FDR method corrects your pvalues according to their number and distribution, there is no threshold to set there. Maybe the 5% you refer to is the usual threshold people use to filter results. So, if you want to keep genes with FDR < 0.0001 it seems to me that your code is fine (by the way <0.0001 is very stringent, if you have many genes maybe consider shareholding also on logFC and logCPM).


All times are GMT -8. The time now is 12:35 PM.

Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2018, vBulletin Solutions, Inc.