[BioC] globaltest - multinomial model
J.J.Goeman at lumc.nl
J.J.Goeman at lumc.nl
Fri Aug 17 13:18:31 CEST 2007
Dear Hans-Ulrich,
This is a minor bug, which has, I believe, been fixed in the latest
devel version. The worst thing that can happen is that you miss a
warning that the null model did not converge before the maximum number
of iterations.
If you want you can check this by asking for
fit(gt)@iter
after gt <- globaltest(Golub, "Source")
If this is smaller than 20 the null model has converged.
Regards,
Jelle
> -----Original Message-----
> From: Hans-Ulrich Klein [mailto:h.klein at uni-muenster.de]
> Sent: 16 August 2007 19:31
> To: bioconductor at stat.math.ethz.ch
> Subject: [BioC] globaltest - multinomial model
>
> Hi List,
>
> I get a warning message whenever I use the function globaltest with a
> multinomial model. Below is a code example taken from the vignette
> leading to the warning message:
> Warning message:
> $ operator not defined for this S4 class, returning NULL in: fit$iter
>
> Should I be concerned about the warning message?
>
> Regards,
> Hans-Ulrich
>
>
>
> > library(globaltest)
> > library(golubEsets)
> > library(vsn)
> > data(Golub_Merge)
> > Golub <- Golub_Merge[, 1:34]
> > exprs(Golub) <- exprs(vsn2(Golub_Merge[, 1:34]))
>
> > globaltest(Golub, "Source")
> Global Test result:
> Data: 34 samples with 7129 genes; 1 gene set
> Model: multinomial
> Method: Asymptotic distribution
>
> Genes Tested Statistic Q Expected Q sd of Q P-value
> all 7129 7129 23.126 10 1.9632 2.5091e-05
> Warning message:
> $ operator not defined for this S4 class, returning NULL in: fit$iter
>
>
> > sessionInfo()
> R version 2.5.0 (2007-04-23)
> x86_64-unknown-linux-gnu
>
> locale:
> C
>
> attached base packages:
> [1] "splines" "tools" "stats" "graphics"
> "grDevices" "utils"
> [7] "datasets" "methods" "base"
>
> other attached packages:
> vsn limma affy affyio golubEsets globaltest
> multtest
> "2.2.0" "2.10.0" "1.14.2" "1.4.1" "1.4.3" "4.6.0"
> "1.16.1"
> survival Biobase
> "2.32" "1.14.1"
>
>
>
More information about the Bioconductor
mailing list