[Bioc-devel] "nearest" & related methods for GRangesList & friends?

[Hervé Pagès]

On 05/23/2014 11:20 AM, Ryan C. Thompson wrote:
> Here is my use case (relevant lines highlighted):
>
> https://gist.github.com/anonymous/0cd0c926fb21a8d62fcc#file-annotatepeaks-r-L49-L54
>
>
> The idea is to annotate each peak with the Entrez ID of the nearest
> transcription start site. Each Entrez ID can have multiple transcripts,
> hence multiple TSS, so I basically want to call nearest with a GRanges
> of peaks against a GRangesList of TSS (each TSS is a range with a width
> of 1 containing the first base pair of a transcript). I don't personally
> need a method that works for a GRangesList query.

Good. nearest() between a GRanges and a GRangesList is the easy one:

   togroup(subject, nearest(query, unlist(subject, use.names=FALSE)))

Is this doing what you need?

Cheers,
H.

>
> -Ryan
>
>
> On Fri 23 May 2014 11:13:24 AM PDT, Hervé Pagès wrote:
>>
>> Hi Ryan,
>>
>> On 05/22/2014 03:38 PM, Ryan C. Thompson wrote:
>>>
>>> Hello,
>>>
>>> I recently found myself in want of a nearest method that handles
>>> GRangesList objects. Is there any plan to add one?
>>
>>
>> Not that I know of. I guess most of the times it's probably good enough
>> to call range() on both GRangesList objects before passing them to
>> nearest(). What's the use case for caring about "gaps" within each
>> GRangesList element?
>>
>>>
>>> I just want to define
>>> "nearest" for elements of a GRangesList by the shortest distance between
>>> any query range and any subject range. Obviously I can do this by
>>> unlisting the GRangesList, calling nearest, and then post-processing to
>>> figure out which element of the original GRangesList includes the
>>> nearest range, but it would be nice to have a function that already does
>>> this for me.
>>
>>
>> The easy one is a "nearest" method between a GRanges and a GRangesList.
>> Is it what you need? That would cover finding the nearest gene or
>> transcript for ChIPseq peaks when the exon structure actually matters
>> (seems like it would matter only for peaks that fall inside more than
>> 1 transcript).
>>
>> Or do you need a method that operates on 2 GRangesList objects? Can't
>> think of a use case for that one. Finding the nearest gene or transcript
>> for junction reads?
>>
>> Thanks,
>> H.
>>
>>
>>>
>>>
>>> -Ryan
>>>
>>> _______________________________________________
>>> Bioc-devel at r-project.org mailing list
>>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>

-- 
Hervé Pagès

Program in Computational Biology
Division of Public Health Sciences
Fred Hutchinson Cancer Research Center
1100 Fairview Ave. N, M1-B514
P.O. Box 19024
Seattle, WA 98109-1024

E-mail: hpages at fhcrc.org
Phone:  (206) 667-5791
Fax:    (206) 667-1319