Re: [Bioc-devel] IRanges should support long vectors

2019-05-28 Thread Pages, Herve
On 5/28/19 07:57, Pariksheet Nanda wrote: > Hi Hervé, > > Indeed, an IRanges with 2^31 elements is 17.1 GB. > The reason I was interested in IRanges, was GRanges are needed to create > the BSgenome::BSgenomeViews. > More broadly, my use case is chopping up a large genome into a fixed > kmer size

Re: [Bioc-devel] IRanges should support long vectors

2019-05-28 Thread Pariksheet Nanda
Hi Hervé, Indeed, an IRanges with 2^31 elements is 17.1 GB. The reason I was interested in IRanges, was GRanges are needed to create the BSgenome::BSgenomeViews. More broadly, my use case is chopping up a large genome into a fixed kmer size so that repetitive "unmappable" regions can be removed. h

Re: [Bioc-devel] IRanges should support long vectors

2019-05-28 Thread Pages, Herve
Hi Pariksheet, On 5/25/19 12:49, Pariksheet Nanda wrote: Hello, R 3.0 added support for long vectors, but it's not yet possible to use them with IRanges. Without long vector support it's not possible to construct an IRanges object with more than 2^31 elements: ir <- IRanges(start = 1:(2^31

[Bioc-devel] IRanges should support long vectors

2019-05-25 Thread Pariksheet Nanda
Hello, R 3.0 added support for long vectors, but it's not yet possible to use them with IRanges. Without long vector support it's not possible to construct an IRanges object with more than 2^31 elements: > ir <- IRanges(start = 1:(2^31 - 1), width = 1) > ir <- IRanges(start = 1:2^31, width = 1)