[BioC] multtest error - class label problem

Wolfgang Huber huber at ebi.ac.uk
Thu Jun 22 19:47:22 CEST 2006


Dear Yuhong,

the error message says it all, you specify
	clabel<-c(0,0,0)
which corresponds to one group of size three, while mt.maxT expects two 
groups.

Katie - There seems to be a buglet in the code that produces the error 
message, making it a bit hard to read, could you check this?

If you only have three replicates, I would recommend to a moderated 
t-statistics such as available with the limma package (eBayes function):

  library(limma)
  ? eBayes
  example(eBayes)

see also http://www.bepress.com/bioconductor/paper7


Best wishes
  Wolfgang



yuhong ning wrote:
> Hi, 
> 
> I am using "multtest" to analyze proteomics data. The
> experiment is to compare protein differential
> expression between Wild Type and mutant cell types. MS
> software generates a ratio of intensity of
> heavy-isotope/light-isotope which correspondent to
> wild type vs. mutant. It is a triplicate experiment,
> but for some proteins, ratio data are missing. My R
> script is very simple:
> 
>> dd<-read.delim("log2_data.txt", header=TRUE)
>> clabel<-c(0,0,0)
>> res<-mt.maxT(dd, clabel, B=20)
> Error in mt.checkclasslabel(classlabel, test) : 
>         in t test, every number in class label needs
> to be 0 or 1 and neither of the 0 set or 1 set can be
> empty set
>  The folllowing is your setting of classlabel 0 
> in t test, every number in class label needs to be 0
> or 1 and neither of the 0 set or 1 set can be empty
> set
>  The folllowing is your setting of classlabel 0 
> in t test, every number in class label needs to be 0
> or 1 and neither of the 0 set or 1 set can be empty
> set
>  The folllowing is your setting of classlabel 0 
> 
> 
> Could any one help?
> 
> Thanks a lot.
> 
> yuhong
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor


-- 
------------------------------------------------------------------
Wolfgang Huber  EBI/EMBL  Cambridge UK  http://www.ebi.ac.uk/huber



More information about the Bioconductor mailing list