[Bioc-devel] could bsseq::data.frame2GRanges be added to GenomicRanges

[Tim Triche, Jr.]

For any tabular data structure with "chr[om]" and one or more of starts, ends, widths, and strands, there _is_ an obvious mapping, though!  And I personally always have an optional argument to keep the rest as mcols(). It just seems so straightforward. 

The generic granges() method also suggests that people often want a way for  something that is not itself a GRanges to at least emit one. 

granges(foo) is also a lot easier than 

with(foo, GRanges(fairly,
                               Involved(),
                               Constructor,
                               ...)

It would be nice if everything were a bed, wig, BAM or similar. Then rtracklayer would be all anyone needs. But, sometimes in the course of events, an unenlightened soul will have their library emit a table of coordinates bearing information that would be more useful as a GRanges. Given the frequency with which this occurs, it's nice to have that generic. 

Anyways, this is all your fault!  If you hadn't built such a marvelously useful data structure, people wouldn't want to use it in ways you never intended. :-)

No good deed goes unpunished,

--t

> On Oct 6, 2013, at 1:26 PM, Michael Lawrence <lawrence.michael at gene.com> wrote:
> 
> I'm still unconvinced that there is an obvious, general path from
> data.frame -> GRanges. It's usually easy enough to just call GRanges(),
> often of the pattern with(df, GRanges(...)). Moreover, it's unusual for me
> to encounter genomic data in data.frames.
> 
> 
> 
> 
> On Sun, Oct 6, 2013 at 8:37 AM, Kasper Daniel Hansen <
> kasperdanielhansen at gmail.com> wrote:
> 
>> Also, it goes without saying that I am happy to provide a patch for
>> GenomicRanges, and check that it passes R CMD check, to minimize the work
>> of the maintainer.
>> 
>> Kasper
>> 
>> 
>> On Sun, Oct 6, 2013 at 9:28 AM, Kasper Daniel Hansen <
>> kasperdanielhansen at gmail.com> wrote:
>> 
>>> bsseq::data.frame2GRanges does the obvious step of converting a
>> data.frame
>>> to GRanges.  It has a couple of bells and whistles where strand can be
>>> ignored and additional columns (apart from genomic location) may be
>> ignore
>>> in the output object.
>>> 
>>> I (and now quite a few other people) use this function almost every day.
>>> I have seen other implementations in other packages, suggesting this is
>>> not just something I (we) use.
>>> 
>>> I suggests adding this function to GenomicRanges.  I am happy to support
>>> it going forward.
>>> 
>>> Using this function we could also add an as(x, "GRanges") method for
>>> x=data.frame, but I still suggest keeping the basic function for the
>>> extended functionality it provides.
>>> 
>>> Best,
>>> Kasper
>> 
>>        [[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