[BioC] error when using waveTiling package

[James W. MacDonald]

Hi Chris,

In the future please try to follow the posting guide, and supply us with 
something reproducible and your sessionInfo(). Otherwise, you are in 
essence expecting somebody who might want to help to either already know 
what that means, or to have the time and inclination to figure it out.

On 4/28/2014 1:46 PM, chris blazier wrote:
> I am running into an error at the "sigProbes" step of the waveTiling pipeline. If I skip the background correction and quantile normalization step I don't have the error (though it is not a good idea just to skip these).
>
>
> Here is the step where I get the error (P. 3 of the waveTiling "vignette" PDF file):> delta <- log(1.2,2)
>> leafdevInfCompare <- wfm.inference(leafdevFit,
> + contrasts="compare",delta=c("median",delta))
>
>
> The error is:
> Error in sigProbes[[i]][1:(length(sigProbes[[i]]) - 1)] :
>    only 0's may be mixed with negative subscripts

Consider the following:

 > (1:10)[1:(length(character(0))-1)]
Error in (1:10)[1:(length(character(0)) - 1)] :
   only 0's may be mixed with negative subscripts

This is the error you get when you try to programmatically subset 
something and you don't consider the situation where your (in this case, 
sigProbes) subsetting item might have a length of zero.

If this is something you are doing in your own code, then I would 
suggest trapping that error (e.g., in the case that there are no 
sigProbes, you want to return NULL or NA, or 0, or whatever.

In addition, it is a better idea to use e.g.,

seq(along = something)

rather than

1:(length(something))

for that reason.

If this is not your code, and is cropping up from within waveTiling, I 
would contact the maintainer and see if they will agree that this is if 
not a bug, then an infelicity that should be fixed.

Best,

Jim



> Thanks for your help,
>
> Chris Blazier
> 	[[alternative HTML version deleted]]
>
>
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at r-project.org
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor

-- 
James W. MacDonald, M.S.
Biostatistician
University of Washington
Environmental and Occupational Health Sciences
4225 Roosevelt Way NE, # 100
Seattle WA 98105-6099