[BioC] [limma] Error: nestedF method can't handle NA p-values

[Gordon K Smyth]

> Date: Wed, 24 Jan 2007 14:45:27 -0200
> From: "Marcelo Laia" <marcelolaia at gmail.com>
> Subject: [BioC] [limma] Error: nestedF method can't handle NA p-values
> To: Bioconductor <bioconductor at stat.math.ethz.ch>
>
> Hi,
>
> I suppose that I was done a mistake when I run:
>
> results <- decideTests(fit.3, method = "nestedF")
> Error: nestedF method can't handle NA p-values
>
> NA p-value is common?

Well, if you introduce missing values (NA) into your data by background subtraction or filtering
so that there are no observations left for some of the genes, then the p-values for those genes
must also be NA.

> if I use global instead nestedF the error not appear.
>
> Please, could you point me out what I am wrong doing?

Well, plainly, you need to either stop introducing NAs into your data or else accept that not all
statistical methods will be available to you.  It will always be the case that some of the more
sophisticated or complex statistical methods will rely on complete data.  In another Bioconductor
post today

  https://www.stat.math.ethz.ch/pipermail/bioconductor/2007-January/015792.html

you explain that you are already aware of how to fix the problem, but don't want to do so.  If you
insist on using background subtraction, which I view as very crude and which has often been
advised against on this mailing list, then you have to live with the consequences.  Fair enough?

Best wishes
Gordon

> Thanks very much
>
> --
> Marcelo Luiz de Laia
> Ph.D Candidate
> S?o Paulo State University (http://www.unesp.br/eng/)
> School of Agricultural and Veterinary Sciences
> Department of  Technology
> Via de Acesso Prof. Paulo Donato Castellane s/n
> 14884-900   Jaboticabal - SP - Brazil
> Phone: +55-016-3209-2675
> Cell: +55-016-97098526