[BioC] Category package problems with Vignette documentation

Robert Gentleman rgentlem at fhcrc.org
Tue Jun 7 16:58:10 CEST 2005

Hi Jean-Marc,
   You have not told us how you are getting your packages but I suspec 
that it is not using any of the tools provided by R. If you use 
install.packages (or for BioC reposTools and then install.packages2) all 
dependencies should be downloaded. If you choose to do this manually 
without using these tools, then you are going to have to sort the 
dependencies out yourself (and that seems to be most of what is 
happening here). Much of this is documented on the BioC website - I 
refer you there.
   A few more comments below.

Neefs, Jean-Marc [PRDBE] wrote:
> Dear Bioconductor,
> I have installed the Category package 1.0.1 on my Windows 2000 computer,
> after installation of R 2.1.0 and other packages required by Category
> (Biobase 1.5.12; tools -- not listed in R package list;
  tools is a standard R package and comes with every copy of R, if you 
do not have it then you did something wrong at installation time.

  genefilter 1.6.2
> actually loads 1.6.1;

R can only load packages you downloaded. There is no magic here, if you 
download 1.6.1 then that is what you have. If you want 1.6.2 you must 
download it (or as I said above please use the tools that have been 
written and documented).

  KEGG 1.8.1; hgu95av2 1.8.4; ALL 1.0.2; survival 2.17;
> splines -- no longer listed in Bioconductor package list)

  splines is a package that is distributed with R. Every copy of R has a 
copy of splines.
 > packageDescription("splines")
Package: splines
Version: 2.1.0
Priority: base
Title: Regression Spline Functions and Classes
Author: Douglas M. Bates <bates at stat.wisc.edu> and William N. Venables
         <Bill.Venables at cmis.csiro.au>
Maintainer: R Core Team <R-core at r-project.org>
Description: Regression spline functions and classes
License: GPL Version 2 or later.
Built: R 2.1.0; x86_64-unknown-linux-gnu; 2005-05-10 11:52:33; unix

> I would like to use the Category package for microarray data analysis. I am
> not an experienced R user, so I get easily confused.
> The Category vignette code works well until line:
>>Amat = t(PWAmat("hgu95av2"))
> where PWAmat is a function from the 'annotate' package -- not uploaded by
> Category and no error reported when loading the Category package.
> I could find a fix by loading and installing 'annotate' from local zip file.
> Could you make 'Category' report an error when 'annotate' is not loaded?

  Yes, but it would not help. Packages list only direct dependencies, 
other dependencies are determined by the direct dependencies. If you 
want to do manual installation, you are responsible for following the tree.

> Also could you point me to the latest 'splines' package or any other that
> replaces it? -- I found 'splines' yesterday morning but no longer yesterday
> afternoon.

I have no idea what you are splines comes with R (it is a base package), 
you seem to have a very broken installation and can only suggest that 
you try to find out what has gone wrong and fix it.

Best wishes

> Then the code breaks again at line:
>>tA = AmER2 %*% tobs$statistic
> generates Error: Object "tobs" not found
> I suppose 'tobs' is a matrix calculated from 'ttests' described higher in
> the vignette, because of the matrix multiplication operator. I tried:
>>tA = AmER2 %*% ttests$statistic
> which is probably silly because it generates: Error in AmER2 %*%
> ttests$statistic : non-conformable arguments
> Could you include in the vignette the code to obtain 'tobs' and send me a
> copy?
> Thank you very much in advance for your kind help,
> Jean-Marc
> Jean-Marc Neefs, PhD
> Senior Bioinformatics Specialist
> Johnson&Johnson PRD, Janssen Pharmaceutica
> 30 Turnhoutseweg,
> B-2340 Beerse
> tel +32 14 60 6084
> fax +32 14 60 3226
> eml jneef1 at prdbe.jnj.com
> 	[[alternative HTML version deleted]]
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor

More information about the Bioconductor mailing list