[BioC] c()ombining G/IRanges with elementMetadata: feature inquiry/request

[Steve Lianoglou]

Hi,

I just wanted to let it be known that I committed a change to the
GenomicRanges and IRanges packages to implement what we discussed in
this thread: namely, adding an .ignoreElementMetadata argument to the
c()-ombine functions for GRanges and IRanges objects.

Along with the (simple) code to make that happen, I:

(i) updated the corresponding documentation files (IRanges-class.Rd
and GRanges-class.Rd) with appropriate descriptions and examples.
(ii) Added an extra unit test in the test_IRanges_combine function; and
(iii) Added a test_GRanges_combine function to the GenomicRanges
package to test for the new functionality, and other (trivial)
c()-ombining scenarios.

I ran the `R CMD check` for both packages and everything is kosher.

Lastly:

I just pulled in the latest from trunk in both packages and now notice
that there has been a change in the IRanges-setops stuff that fires a
warning message about "multiple methods tables found for ... " (union,
intersect, and setdiff) when the GenomicRanges package is subsequently
loaded.

It looks like this is a result of the generics for these methods now
takes `...`, so I added the dots to the
GenomicRanges:::union,intersect,setdiff functions to sidestep these
warnings.

The version numbers have been bumped for both.

Hope that's OK,

-steve

On Wed, Feb 9, 2011 at 5:07 PM, Steve Lianoglou
<mailinglist.honeypot at gmail.com> wrote:
> On Wed, Feb 9, 2011 at 4:45 PM, Martin Morgan <mtmorgan at fhcrc.org> wrote:
>> On 02/09/2011 01:20 PM, Steve Lianoglou wrote:
>>> Sorry that this thread went cold, but I'm hoping we can come to some consensus:
>>>
>>> On Tue, Feb 8, 2011 at 12:56 PM, Martin Morgan <mtmorgan at fhcrc.org> wrote:
>>> [snip]
>>>> just trying not to get stuck on values / elementMetadata. what about
>>>> append / rbind, thinking either that append works w/out element
>>>> metadata, or that c does and rbind (in the GRanges-are-row-indicies on
>>>> values() interpretation; not suitable for IRanges?) / append tries to do
>>>> something with element metadata.
>>> [/snip]
>>>
>>> I think my preference would be to have all of these treated the same
>>> .. which would mean to error on rbind/c'ing *Ranges with different
>>> cols in their elementMetadata, and include an option to drop their
>>> elementMetadata when you combine them.
>>
>> ok; I'm also fine (in the rather-see-progress-than-stand-in-the-way)
>> with an arg that mentions elementMetadata rather than values.
>
> Ok. I have a patches for IRanges and GRanges on my machine at home for
> the `c` methods using the .ignoreElementMetadata argument.
>
> It's admittedly quite trivial, but I'll double check it and post back
> to list when I get back to home base.
>
> -steve
>
> --
> Steve Lianoglou
> Graduate Student: Computational Systems Biology
>  | Memorial Sloan-Kettering Cancer Center
>  | Weill Medical College of Cornell University
> Contact Info: http://cbio.mskcc.org/~lianos/contact
>



-- 
Steve Lianoglou
Graduate Student: Computational Systems Biology
 | Memorial Sloan-Kettering Cancer Center
 | Weill Medical College of Cornell University
Contact Info: http://cbio.mskcc.org/~lianos/contact