Question

How to simply change chromosome names of a GRanges object?

0

Entering edit mode

changxu.fan ▴ 20

@changxufan-24128

Last seen 12 months ago

United States

I have one simple goal:

gr <- GRanges(seqnames = c("chr1", "chr2"),
              ranges = IRanges(start = 1:2, end = 4:5))

I want to change the chromosome names all to "chrX"

> seqnames(gr) <- "chrX"
Error in .normalize_seqnames_replacement_value(value, x) : 
  levels of supplied 'seqnames' must be identical to 'seqlevels(x)'

Please let me know if there is a simple way to do this.

Best regards

genomicrang GenomicRanges • 2.8k views

ADD COMMENT • link updated 2.3 years ago by Hervé Pagès 16k • written 2.5 years ago by changxu.fan ▴ 20

score 1 · Answer 1 · 2021-10-28

1

Entering edit mode

James W. MacDonald 65k

@james-w-macdonald-5106

Last seen 53 minutes ago

United States

I can't imagine the use case for doing that, but it's easy enough to do. You just can't do it to the original GRanges object, so far as I know. But you can make a new one easily enough.

> gr <- GRanges(seqnames = c("chr1", "chr2"),
              ranges = IRanges(start = 1:2, end = 4:5))

> gr
GRanges object with 2 ranges and 0 metadata columns:
      seqnames    ranges strand
         <Rle> <IRanges>  <Rle>
  [1]     chr1       1-4      *
  [2]     chr2       2-5      *
  -------
  seqinfo: 2 sequences from an unspecified genome; no seqlengths
> gr2 <- GRanges(rep("chrx", length(gr)), ranges(gr))
> gr2
GRanges object with 2 ranges and 0 metadata columns:
      seqnames    ranges strand
         <Rle> <IRanges>  <Rle>
  [1]     chrx       1-4      *
  [2]     chrx       2-5      *
  -------
  seqinfo: 1 sequence from an unspecified genome; no seqlengths

ADD COMMENT • link 2.5 years ago James W. MacDonald 65k

1

Entering edit mode

Exactly.

Note that a more idiomatic way of creating this new GRanges is with:

GRanges(Rle("chrX", length(gr)), ranges(gr, use.mcols=TRUE), strand=strand(gr))

This will ensure that the strand and metadata cols of the original object are preserved.

H.

ADD REPLY • link 2.5 years ago Hervé Pagès 16k

0

Entering edit mode

What's the rationale for using an Rle? Is it more efficient?

ADD REPLY • link 2.5 years ago James W. MacDonald 65k

0

Entering edit mode

This is to make sure that the supplied seqnames has the correct length.

Note that the GRanges() constructor will turn the supplied seqnames into an Rle, and, if its length is not the same as the length of the supplied IRanges, it will recycle the shortest to the length of the longest. This normally works well, except when gr has length 0:

gr0 <- gr[0]
GRanges("chrX", ranges(gr0, use.mcols=TRUE), strand=strand(gr0))
# Error in stop_if_wrong_length(what, ans_len) : 
#   'ranges' must have the length of the object to construct (1) or length
#   1

Supplying the seqnames as rep("chrX", length(gr)) or Rle("chrX", length(gr)) is guaranteed to always work. However using Rle() is more efficient than using rep() and is not more typing.

H.

ADD REPLY • link 2.3 years ago Hervé Pagès 16k