Chip-seq data analysis in R with bioconductor - change of character from "*" to "+"

Question

[herna.dewit]

I have a chip-seq data file, where the QC and trimming has already been done, as well as the mapping to the reference genome. Now I have data as follow:

 GRanges object with 2288 ranges and 0 metadata columns:
          seqnames                 ranges strand
             <Rle>              <IRanges>  <Rle>
      [1]     chr1   [16174360, 16175609]      *
      [2]     chr1   [20811067, 20812439]      *
      [3]     chr1   [22783676, 22784321]      *
      [4]     chr1   [26735004, 26735693]      *
      [5]     chr1   [27023038, 27023988]      *

How do I go about changing the "*" character to "+" seeing that the fraglen <- estimate.mean.fraglen(reads, method = "correlation") gives the following error:

Error in .local(x, ...) : x must have named elements '+' and '-'

I have tried using gsub, sub and chartr, and specifying the column, but then the whole data frame changes. Any recommendations?

[Jean-Karim Heriche]

What have you tried exactly ? And what do you mean by 'the whole data frame changes' ?

Answer 1

[Devon Ryan]

reads = reads[which(strand(reads) != "*")] or something along those lines. Of course this leads to the question of how you got unstranded alignment information to begin with...

Answer 2

[ejm32]

This should do the trick:

gr0 <- GRanges(Rle(c("chr2", "chr2", "chr1", "chr3"), c(1, 3, 2, 4)), IRanges(1:10, width=10:1))
gr0
strand(gr0)<-Rle("+", length(gr0))
gr0