[Bioc-devel] [BioC] granges() method for GenomicRanges objects akin to ranges()...

Hervé Pagès hpages at fhcrc.org
Tue Jul 8 17:45:09 CEST 2014


Hi,

On 05/13/2014 01:15 AM, Julian Gehring wrote:
> Hi,
>
> In summary, would it be feasible to add to 'GenomicRanges'?
>
> 1) A 'granges(x, use.mcols=FALSE, ...)' method with signature 'GRanges'
> that converts to a 'GRanges' object and optionally drops the mcols (if
> 'use.mcols' is TRUE)

Will do.

>
> 2) A 'dropMcols' or 'dropmcols' method with signature 'GRanges' that is
> a wrapper for
>    mcols(x) <- NULL

How about setMcols(), which is more general than dropmcols()?

Thanks,
H.

>
> If I can be of help in providing a patch for this, please let me know.
>
> Best wishes
> Julian
>
>
>
> On 05.05.2014 23:29, Hervé Pagès wrote:
>> On 05/05/2014 02:12 PM, Cook, Malcolm wrote:
>>>> On 05/05/2014 01:00 PM, Cook, Malcolm wrote:
>>>   >> Wondering,
>>>   >>
>>>   >> Is it too off the beaten track to expect
>>>   >>
>>>   >> `mcols<-`(x,NULL)
>>>   >
>>>   >   > args(`mcols<-`)
>>>   >   function (x, ..., value)
>>>   >
>>>   >Arguments after the ellipsis must be named:
>>>   >
>>>   >   `mcols<-`(x, value=NULL)
>>>
>>> Herve - Great - of course - so - does this not provide the means
>>> requested by the original poster?
>>
>> I think Tim also wanted 'x' to be downgraded to a GRanges instance,
>> like Julian's grangesPlain() does. We could use granges() for that.
>>
>> Deciding of an idiom that can be used inline for just dropping the
>> mcols would be good too. `mcols<-`(x, value=NULL) is a little bit
>> tricky, ugly, and error prone as you noticed. These are probably
>> enough reasons for not choosing it as *the* idiom. Its only advantage
>> is that it doesn't introduce a new symbol.
>>
>> H.
>>
>>>
>>>   >
>>>   >Nothing we can do about this.
>>>   >
>>>   >Cheers,
>>>   >H.
>>>   >
>>>   >>
>>>   >> to work?
>>>   >>
>>>   >> hint: it does not
>>>   >>
>>>   >>   >-----Original Message-----
>>>   >>   >From: bioc-devel-bounces at r-project.org
>>> [mailto:bioc-devel-bounces at r-project.org] On Behalf Of Hervé Pagès
>>>   >>   >Sent: Monday, May 05, 2014 1:28 PM
>>>   >>   >To: Kasper Daniel Hansen; Michael Lawrence
>>>   >>   >Cc: Johnston, Jeffrey; ttriche at usc.edu;
>>> bioc-devel at r-project.org; bioconductor at r-project.org
>>>   >>   >Subject: Re: [Bioc-devel] [BioC] granges() method for
>>> GenomicRanges objects akin to ranges()...
>>>   >>   >
>>>   >>   >Hi,
>>>   >>   >
>>>   >>   >I have no problem using granges() for that. Just to clarify:
>>>   >>   >   (a) it would propagate the names()
>>>   >>   >   (b) it would drop the metadata()
>>>   >>   >   (c) the mcols() would propagate only if 'use.mcols=TRUE' was
>>>   >>   >       specified ('use.mcols' is FALSE by default)
>>>   >>   >   (d) it would return a GRanges *instance* i.e. input object
>>> 'x'
>>>   >>   >       would be downgraded to GRanges if it extends GRanges
>>>   >>   >
>>>   >>   >@Kasper: granges() on SummarizedExperiment ignores the
>>> 'use.mcols'
>>>   >>   >arg and always propagates the mcols. Alternatively you can use
>>> rowData()
>>>   >>   >which also propagates the mcols. granges() is actually just an
>>> alias
>>>   >>   >for rowData() on SummarizedExperiment objects.
>>>   >>   >
>>>   >>   >H.
>>>   >>   >
>>>   >>   >
>>>   >>   >On 05/05/2014 10:31 AM, Kasper Daniel Hansen wrote:
>>>   >>   >> I agree with Michael on this.
>>>   >>   >>
>>>   >>   >> I can see why, in some usage cases, granges() is convenient
>>> to have with
>>>   >>   >> use.mcols=FALSE (which seems to have been added in the
>>> latest release).
>>>   >>   >>   But in my usage of granges(), where I call granges() on
>>> objects like
>>>   >>   >> SummarizedExperiments and friends, I have been expecting
>>> granges() to give
>>>   >>   >> me the GRange component of the object.  Not a crippled
>>> version of the
>>>   >>   >> GRange component.
>>>   >>   >>
>>>   >>   >> This is - to me - very counter intuitive and I wish I had
>>> seen this
>>>   >>   >> earlier.  It is particular frustrating that this default is
>>> part of the
>>>   >>   >> generic.
>>>   >>   >>
>>>   >>   >> Best,
>>>   >>   >> Kasper
>>>   >>   >>
>>>   >>   >>
>>>   >>   >> On Mon, May 5, 2014 at 12:11 PM, Michael Lawrence
>>> <lawrence.michael at gene.com
>>>   >>   >>> wrote:
>>>   >>   >>
>>>   >>   >>> In my opinion, granges() is not very clear as to the
>>> intent. The mcols are
>>>   >>   >>> part of the GRanges, so why would calling granges() drop
>>> them? I think we
>>>   >>   >>> want something similar to unclass(), unname(), etc. This
>>> why I suggested
>>>   >>   >>> dropmcols().
>>>   >>   >>>
>>>   >>   >>>
>>>   >>   >>>
>>>   >>   >>>
>>>   >>   >>> On Mon, May 5, 2014 at 8:17 AM, Tim Triche, Jr.
>>> <tim.triche at gmail.com
>>>   >>   >>>> wrote:
>>>   >>   >>>
>>>   >>   >>>> That's exactly what I was after -- the generic is already
>>> defined, so why
>>>   >>   >>>> not use it?
>>>   >>   >>>>
>>>   >>   >>>> --t
>>>   >>   >>>>
>>>   >>   >>>>> On May 5, 2014, at 7:42 AM, Julian Gehring
>>> <julian.gehring at embl.de>
>>>   >>   >>>> wrote:
>>>   >>   >>>>>
>>>   >>   >>>>> Hi,
>>>   >>   >>>>>
>>>   >>   >>>>>> On 05.05.2014 16:22, Martin Morgan wrote:
>>>   >>   >>>>>> generalize as setMcols, like setNames? setMcols(x, NULL)
>>>   >>   >>>>>
>>>   >>   >>>>> How about Tim's original suggestion, to add a 'granges'
>>> method that
>>>   >>   >>>> works on a 'GRanges' input?  The current definition
>>>   >>   >>>>>
>>>   >>   >>>>> granges(x, use.mcols=FALSE, ...)
>>>   >>   >>>>>
>>>   >>   >>>>> seem suited for this.
>>>   >>   >>>>>
>>>   >>   >>>>> Best wishes
>>>   >>   >>>>> Julian
>>>   >>   >>>>
>>>   >>   >>>
>>>   >>   >>>          [[alternative HTML version deleted]]
>>>   >>   >>>
>>>   >>   >>> _______________________________________________
>>>   >>   >>> Bioc-devel at r-project.org mailing list
>>>   >>   >>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>>   >>   >>>
>>>   >>   >>
>>>   >>   >>     [[alternative HTML version deleted]]
>>>   >>   >>
>>>   >>   >> _______________________________________________
>>>   >>   >> 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
>>>   >>   >
>>>   >>   >_______________________________________________
>>>   >>   >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
>>>
>>

-- 
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



More information about the Bioc-devel mailing list