P.adjusted values

Question

[sarahawan92]

I have two questions related to P.adjusted values,

1. In differential analysis, after performing Treat functions, is it necessary to find out the p.adjusted values as this functions itself provide FDR (p.adjusted).
2. If it is necessary to still adjust the P values after TREAT function, how can I find the p.adjusted values. As I am trying by using the P.adjusted formula in R but still not getting my head around it.

Kindly help me.

Thanks.

[Ram]

Hi,

Hello everyone is not a valid tag. Tags are subject matter keywords that help describe the core of your problem so experts can help you better. I've fixed the tags in your post now, please be more careful in the future.

Answer 1

[Istvan Albert]

Not quite clear what you mean - FDR is a type of p-value adjustment so no need to adjust that again.

Answer 2

[ATpoint]

edgeR (I guess that is what you refer to) takes full care of multiple testing. You can use the FDR column directly.

[sarahawan92]

Yep you are right. Can you tell me how can i get the FDR column. After applying Treat function when I export ti excel the FDR column is missing.Even if i checked what information I get it is just showing logfc, unshrunkn.logfc, pvalue but not FDR. Can you tell me how can i get FDR column info. Thanks

Answer 3

[Gordon Smyth]

Assuming that you are using limma, then you just use

topTreat(fit)

where fit is the output from treat(). That will give you a table of DE results with FDR values automatically included. The treat() function itself does not produce adjusted p-values.