[BioC] package R.oo (or R.00?), within package R.classes, and function robustPca

[Sean Davis]

On Wed, Oct 8, 2008 at 6:56 AM, Tillinghast, Guy, M.D.
<Guy.Tillinghast at rivhs.com> wrote:
>
> I am having problem installing the package: R.classes
> Which contains package R.oo
> which is needed to run function robustPca
>
> How can I download package R.classes?
> It is supposed to be "100%R"
>
> I have R version 2.7.2
>
>
>
>
> History:
> I followed the instructions/vignette for using pcaMethods package
> when I attempted the robustPCA function, as in
>
> resRobSvd<-pca(mdcOut,method = "robustPca", nPcs = 5, center = FALSE)
>
> I received an error message requesting loading of aroma.light package.
>
> I loaded this package
>
> Then I received an error message requesting loading of R.oo package.
> R.oo package not on the Bioconductor site.
> I searched Google and found R.oo package part of R.classes package, and
> is "100%R"
> I searched the Bioconductor site and could not find package R.classes
> I Googled R.classes and first entry instructed me to use function
> install.package()
> install.packages("R.classes",contriburl="http://www.maths.lth.se/help/R/
> R.classes/")
> got error message: subscript out of bounds

It is best to use biocLite when installing packages from Bioconductor
AND R since Bioconductor is designed to work with a specific version
of R.  Had you done so, when you installed pcaMethods, all the
dependencies would have been installed simultaneously.  So, go back
and use biocLite to reinstall pcaMethods and everything will come with
it except aroma.light which is only suggested and not required for
pcaMethods.  Then use biocLite to install aroma.light.  That should do
it.

The take-home message is to ALWAYS use biocLite (when possible, which
is basically ALWAYS).

Sean