The code that's there now is already unit tested and should be merged.
The test I mentioned was a small check that the following works (this is
what started this thread):
> m <- matrix(1, 5, 3, dimnames=list(NULL, NULL))
> gr <- GRanges("chr1", IRanges(1:5, 10))
> colData <- DataFrame(x=letters[1:3])
>
> sexp <- SummarizedExperiment(assays=SimpleList(m=m),
+ rowData=gr,
+ colData=colData)
> sexp2 <- SummarizedExperiment(assays=SimpleList(m=m),
+ rowData=GIntervalTree(gr),
+ colData=colData)
>
> olaps1 <- findOverlaps(gr, sexp)
> olaps2 <- findOverlaps(gr, sexp2)
> identical(olaps1, olaps2)
[1] TRUE
>
> sexp <- sexp[1:2,]
> sexp2 <- sexp2[1:2,]
>
> olaps1 <- findOverlaps(gr, sexp)
> olaps2 <- findOverlaps(gr, sexp2)
> identical(olaps1,olaps2)
[1] TRUE
On Fri, Aug 2, 2013 at 1:36 PM, Michael Lawrence
<lawrence.mich...@gene.com>wrote:
> If you don't mind, I think I'll wait till after you're finished testing
> before merging.
>
> Please keep me up to date,
> Michael
>
>
> On Fri, Aug 2, 2013 at 7:43 AM, Hector Corrada Bravo
> <hcorr...@gmail.com>wrote:
>
>> Thanks again Kasper,
>> All of these are now fixed:
>> * empty subject seqlevel,
>> * out-of-order seqinfo's,
>> * show,GIntervalTree-method
>>
>> Michael, please merge from
>> https://github.com/hcorrada/IRanges [version 1.19.20]
>> https://github.com/hcorrada/GenomicRanges [version 1.13.36]
>>
>> I'll test the SummarizedExperiment stuff later today.
>>
>> Cheers,
>> Hector
>>
>>
>>
>> On Thu, Aug 1, 2013 at 5:21 PM, Kasper Daniel Hansen <
>> kasperdanielhan...@gmail.com> wrote:
>>
>>> If you're fixing tonight here is a possible related. I get an error
>>> when I do the findOverlaps and I have two GRanges with the same seqlevels
>>> but in different order (so, I guess, technically not the same seqlevels).
>>> If this is not supported, I suggest an upfront check.
>>>
>>> gr1 = GRanges(seqnames = c("chr1", "chr2"), ranges = IRanges(1:2, width
>>> = 1))
>>> gr2 = gr1
>>> seqlevels(gr2, force = TRUE) = c("chr2", "chr1")
>>> findOverlaps(GIntervalTree(gr1), GIntervalTree(gr2))
>>>
>>> I also got a seqfault on a similar example (but with real data). I am
>>> trying to see if I can make a small reproducible example (which is how I
>>> got here), hypothesizing that it has something to do with the seqinfo slot
>>> in the two GRanges.
>>>
>>> Kasper
>>>
>>>
>>>
>>>
>>>
>>>
>>> On Thu, Aug 1, 2013 at 4:48 PM, Hector Corrada Bravo <hcorr...@gmail.com
>>> > wrote:
>>>
>>>> Thanks. I'll fix tonight.
>>>>
>>>>
>>>> On Thu, Aug 1, 2013 at 4:28 PM, Kasper Daniel Hansen <
>>>> kasperdanielhan...@gmail.com> wrote:
>>>>
>>>>> Bug:
>>>>>
>>>>> > library(GenomicRanges) ## 1.13.35
>>>>>
>>>>> > gr = GRanges(seqnames = c("chr1", "chr2"), ranges = IRanges(1:2,
>>>>> width = 1))
>>>>>
>>>>> > gr.tree <- GIntervalTree(gr)
>>>>>
>>>>>
>>>>> > findOverlaps(gr.tree, gr.tree)
>>>>>
>>>>>
>>>>> Hits of length 2
>>>>>
>>>>>
>>>>> queryLength: 2
>>>>>
>>>>>
>>>>> subjectLength: 2
>>>>>
>>>>>
>>>>> queryHits subjectHits
>>>>>
>>>>>
>>>>> <integer> <integer>
>>>>>
>>>>>
>>>>> 1 1 1
>>>>>
>>>>>
>>>>> 2 2 2
>>>>>
>>>>>
>>>>> !> findOverlaps(gr.tree, gr.tree[1])
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> *** caught segfault ***
>>>>>
>>>>>
>>>>> address (nil), cause 'memory not mapped'
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> Traceback:
>>>>>
>>>>>
>>>>> 1: .Call(.NAME, ..., PACKAGE = PACKAGE)
>>>>>
>>>>>
>>>>> 2: .Call2(fun, object@ptr, ..., PACKAGE = "IRanges")
>>>>>
>>>>>
>>>>> 3: .IntervalForestCall(subject, fun, query, partitionIndices,
>>>>> elementLengths(partitioning), query_ord)
>>>>>
>>>>> 4: .local(query, subject, maxgap, minoverlap, type, select, ...)
>>>>>
>>>>>
>>>>> 5: findOverlaps(qlist, subject@ranges, maxgap = maxgap, minoverlap
>>>>> = minoverlap, type = type, select = "all")
>>>>>
>>>>> 6: findOverlaps(qlist, subject@ranges, maxgap = maxgap, minoverlap
>>>>> = minoverlap, type = type, select = "all")
>>>>>
>>>>> 7: .local(query, subject, maxgap, minoverlap, type, select, ...)
>>>>>
>>>>>
>>>>> 8: findOverlaps(gr.tree, gr.tree[1])
>>>>>
>>>>>
>>>>> 9: findOverlaps(gr.tree, gr.tree[1])
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> Possible actions:
>>>>>
>>>>>
>>>>> 1: abort (with core dump, if enabled)
>>>>>
>>>>>
>>>>> 2: normal R exit
>>>>>
>>>>>
>>>>> 3: exit R without saving workspace
>>>>>
>>>>>
>>>>> 4: exit R saving workspace
>>>>>
>>>>>
>>>>> Selection:
>>>>>
>>>>>
>>>>> On Thu, Aug 1, 2013 at 3:45 PM, Kasper Daniel Hansen <
>>>>> kasperdanielhan...@gmail.com> wrote:
>>>>>
>>>>>> Ah! Super-nice! I'll test as well, when I have time.
>>>>>>
>>>>>> Kasper
>>>>>>
>>>>>>
>>>>>> On Thu, Aug 1, 2013 at 3:34 PM, Hector Corrada Bravo <
>>>>>> hcorr...@gmail.com> wrote:
>>>>>>
>>>>>>> This is not tested, but...
>>>>>>>
>>>>>>> GIntervalTree extends GenomicRanges so, in principle, can be used
>>>>>>> in the rowData slot of SummarizedExperiment. I tried to make
>>>>>>> GIntervalTree
>>>>>>> support as much of the GenomicRanges interface as I could so this may
>>>>>>> work
>>>>>>> already for overlap queries where the SummarizedExperiment object is the
>>>>>>> subject. I'll test this soon and let you know.
>>>>>>>
>>>>>>>
>>>>>>> On Thu, Aug 1, 2013 at 3:12 PM, Kasper Daniel Hansen <
>>>>>>> kasperdanielhan...@gmail.com> wrote:
>>>>>>>
>>>>>>>> Now that we have GIntervalTree, I am really interested in a
>>>>>>>> SummarizedExperiment that stores its index in a slot, so it can be
>>>>>>>> indexed
>>>>>>>> once. I am sure other people are greedily eyeing this as well.
>>>>>>>> Either
>>>>>>>> changing the class or extending it.
>>>>>>>>
>>>>>>>> Unfortunately, I will not have time to work on this for some weeks,
>>>>>>>> but I
>>>>>>>> wanted to put the idea out there.
>>>>>>>>
>>>>>>>> Best,
>>>>>>>> Kasper
>>>>>>>>
>>>>>>>> [[alternative HTML version deleted]]
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> Bioc-devel@r-project.org mailing list
>>>>>>>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>
>
[[alternative HTML version deleted]]
_______________________________________________
Bioc-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel
