[Bioc-devel] GenomicRanges: Seqinfo indexing with non-existing key

[Julian Gehring]

Hi Dan,

Subestting your vector returns an element with both value and name set
to 'NA'.  In contrast, subsetting the 'Seqinfo' object returns an
element with 'seqname' set to the key - which can be easily
misinterpreted.  Since the other columns ('seqlengths', 'genome',
'circular') are in many cases 'NA' anyway, it becomes hard to catch
this.  Also, subsetting a GRanges (e.g. 'as(si, "GRanges")') with a
non-existing key throws an error.

Best wishes
Julian



On 27/02/14 14:00, Dan Du wrote:
> Hi,
> 
> To me it sounds like what I am expecting, from a R base point of view.
> 
> x<-1:5
> names(x)<-letters[1:5]
> i<-'f'
> x[i]
> 
> Best,
> Dan
> 
> On Thu, 2014-02-27 at 13:49 +0100, Julian Gehring wrote:
>> Hi,
>>
>> In the current bioc-stable and bioc-devel, a 'Seqinfo' object can be
>> indexed successfully with any character key, even it is non-existing:
>>
>> #+BEGIN_SRC R
>>   library(GenomicRanges)
>>   library(BSgenome.Hsapiens.UCSC.hg19)
>>
>>   si = seqinfo(BSgenome.Hsapiens.UCSC.hg19)
>>
>>   ind2 = "999"
>>   ind2 %in% seqnames(si) ## FALSE, no seqlevel with this name
>>
>>   si[ind2] ## works, but should not
>> #+END_SRC
>>
>> Shouldn't this throw an error?
>>
>> Best wishes
>> Julian
>>
>> _______________________________________________
>> Bioc-devel at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
> 
>