Quantcast

false discovery rate !

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

false discovery rate !

A Ezhil
Dear All,

It is not a typical R question (though I use R for this) but I thought someone will help me. For the list of P values, I have calculated FDR using p.adjust() in R (bioconductor). But my FDR values are same for all the P values. When do we get same FDR values? Does the smallest P values should less than 1/N? (where N is the number of P values)

Thanks in advance.

Kind regards,
Ezhil

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: false discovery rate !

john seers (IFR)
 

Hi

Are all your input p values the same? If so your output FDR values would
be the same.

Or are all your p-values relatively large? Then (nearly) all your FDR
values might be 1.

Why don't you put a small example up of what you did? Then we could see
what method you used etc.

Regards


JS



 
---
-----Original Message-----
From: [hidden email] [mailto:[hidden email]]
On Behalf Of A Ezhil
Sent: 10 July 2008 15:41
To: [hidden email]
Subject: [R] false discovery rate !

Dear All,

It is not a typical R question (though I use R for this) but I thought
someone will help me. For the list of P values, I have calculated FDR
using p.adjust() in R (bioconductor). But my FDR values are same for all
the P values. When do we get same FDR values? Does the smallest P values
should less than 1/N? (where N is the number of P values)

Thanks in advance.

Kind regards,
Ezhil

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: false discovery rate !

john seers (IFR)
 

Hi Ezhil

Quite possibly.

Look at this simple example based on what you describe:

> p.adjust(c(0.0002, 0.41, 0.4, 0.42,0.43))
[1] 0.001 1.000 1.000 1.000 1.000
 
Does this fit what you are seeing?

Regards


JS


---
Web sites:

www.ifr.ac.uk  
www.foodandhealthnetwork.com

-----Original Message-----
From: A Ezhil [mailto:[hidden email]]
Sent: 10 July 2008 16:23
To: [hidden email]; john seers (IFR)
Subject: RE: [R] false discovery rate !

Dear John,

My P values are not same and the smallest P value = 0.0002. My P value
distribution is not that great (see the attached file), most of them are
> 0.4. Do you think this is the reason for getting same FDR vlaues?

Thanks again,
Ezhil


--- On Thu, 7/10/08, john seers (IFR) <[hidden email]> wrote:

> From: john seers (IFR) <[hidden email]>
> Subject: RE: [R] false discovery rate !
> To: [hidden email], [hidden email]
> Date: Thursday, July 10, 2008, 8:27 PM Hi
>
> Are all your input p values the same? If so your output FDR values
> would be the same.
>
> Or are all your p-values relatively large? Then (nearly) all your FDR
> values might be 1.
>
> Why don't you put a small example up of what you did?
> Then we could see
> what method you used etc.
>
> Regards
>
>
> JS
>
>
>
>  
> ---
> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]]
> On Behalf Of A Ezhil
> Sent: 10 July 2008 15:41
> To: [hidden email]
> Subject: [R] false discovery rate !
>
> Dear All,
>
> It is not a typical R question (though I use R for this) but I thought

> someone will help me. For the list of P values, I have calculated FDR
> using p.adjust() in R (bioconductor). But my FDR values are same for
> all the P values. When do we get same FDR values? Does the smallest P
> values should less than 1/N? (where N is the number of P values)
>
> Thanks in advance.
>
> Kind regards,
> Ezhil
>
> ______________________________________________
> [hidden email] mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.


     

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Loading...