convert bedgraph to Granges object

Question

[storm1907]

Hello,

I have bedgraph files in following format:

track type=bedGraph description=center_label visibility=full graphType=points color=200,100,0 altColor=0,100,200            
chr1  923391  923526  -0.4838883 
chr1  924813  925002  -0.45778788 
chr1  930090  930401  -0.09189046 
chr1  935707  935961  1.25254348 
chr1  939207  939525  3.51303879

Is there any tool, with whom I can convert bedgraph to Granges object? Desired format is something like this:

> grl.data
GRangesList object of length 10:
$s1
GRanges object with 1318 ranges and 1 metadata column:
         seqnames      ranges strand |     value
            <Rle>   <IRanges>  <Rle> | <numeric>
     [1]        1       1-576      * |   2.03030
     [2]        1    577-1112      * |   1.94532
     [3]        1   1113-1511      * |   1.92982
     [4]        1   1512-2113      * |   1.86865
     [5]        1   2114-2573      * |   1.93882
     ...      ...         ...    ... .       ...
  [1314]       22 19081-19153      * |   2.00131
  [1315]       22 19154-20059      * |   1.95003
  [1316]       22 20060-21926      * |   2.00952
  [1317]       22 21927-23554      * |   1.95225
  [1318]       22 23555-24465      * |   2.01210

Currently I was not able to find any related solution, so will be happy for any help

Thank you!

Answer 1

[rpolicastro]

the import function from rtracklayer

[storm1907]

I work on remote Linux server, and have difficulties with installing and running rtracklayer. Will try to do it on my local machine, thanks

[rpolicastro]

If you don't want to install rtracklayer you could also build the GRanges semi-manually.

library("GenomicRanges")

bg <- read.table("file.bedgraph", sep="\t", skip=1)
colnames(bg) <- c("seqnames", "start", "end", "value")

gr <- makeGRangesFromDataFrame(bg, ignore.strand=TRUE, keep.extra.columns=TRUE)

[storm1907]

Thank you, this works for one sample! What could be appropriate approach, if I need to make one Grange list object from multiple bedgraphs?

[rpolicastro]

library("GenomicRanges")

# A vector of bedGraph files.
# Can be constructed with list.files for convenience.
bg_files <- c("file1.bedgraph", "file2.bedgrah")

# Load the bedgraphs files into a list of GRanges.
bgs <- lapply(bg_files, function(x) {
  bg <- read.table(x, sep="\t", skip=1)
  colnames(bg) <- c("seqnames", "start", "end", "value")
  gr <- makeGRangesFromDataFrame(bg, ignore.strand=TRUE, keep.extra.columns=TRUE)
  return(gr)
})

# Convert the list of GRanges into a GRangesList.
gr_list <- GRangesList(bgs)

[storm1907]

Thank you!

At first it is working, however when trying to run R package CINdex, I still get error

> gr_list
GRangesList object of length 4:
[[1]]
GRanges object with 200255 ranges and 1 metadata column:
           seqnames            ranges strand |      value
              <Rle>         <IRanges>  <Rle> |  <numeric>
       [1]     chr1     923391-923526      * |   0.544787
       [2]     chr1     923551-924817      * |   0.565173
       [3]     chr1     924813-925002      * |   1.072250
       [4]     chr1     925877-926078      * |  -0.511750
       [5]     chr1     930090-930401      * |   0.270485
       ...      ...               ...    ... .        ...
  [200251]    chr22 50768711-50768939      * |  1.2515725
  [200252]    chr22 50769840-50770081      * |  0.2780226
  [200253]    chr22 50775707-50775916      * |  0.0372048
  [200254]    chr22 50776605-50776814      * | -0.2125530
  [200255]    chr22 50777887-50778046      * |  0.3990656
  -------
  seqinfo: 22 sequences from an unspecified genome; no seqlengths

...
<3 more elements>
> run.cin.chr(gr_list)
Error in segments.to.profile(segments) :
  Wrong representation of segments.
In addition: Warning message:
In dir.create(out.folder.name) : 'output_chr_cin' already exists

[storm1907]

Hello again,

This is the result of my commands (my_gr). However, is it possible to get result as in the lower sample (grl.data)?

in grl.data file start there are two additional lines:

GRangesList object of length 10:
$s1

When I try to run downstream analysis, everything is ok with file grl.data, but for my_gr I get output:

> run.cin.chr(grl.seg = ivf_gr)
Error in run.cin.chr(grl.seg = ivf_gr) :
  Input 'grl.seg' must be a GRangesList object

grl.data is a sample file, provided by CINdex package

[storm1907]

OK, I found this function for making Grange lists: https://rdrr.io/bioc/GenomicRanges/man/makeGRangesListFromDataFrame.html

However, still the question, how to make one Grange object list from multiple bedgraphs

I also tried to make Granges list object from 1 bedgraph file

Trying to run analysis with this new file I got following error:

> run.cin.chr(grl.seg = ivf_gr1)
Error in segments.to.profile(segments) :
  Wrong representation of segments.
In addition: Warning message:
In dir.create(out.folder.name) : 'output_chr_cin' already exists

[Ram]

Stop adding answers unless you're answering the top level post. Use Add Comment/Add Reply instead. I've cleaned up the post now but please be more careul in the future.

[Ram]

Please do not paste screenshots of plain text content, it is counterproductive. You can copy paste the content directly here (using the code formatting option shown below), or use a GitHub Gist if the content volume exceeds allowed length here.