Hi,

You can use the getSeq and alphabetFrequency functions from Bioconductor's Biostrings package to obtain the sequence and compute the frequency of each nucleotide, and then, just divide the sum of G+C by the total number of nucleotides.

To use getSeq the easiest way is to create a GRanges object from your matrix. I use toGRanges from regioneR package, but there are other ways to build it.

library(regioneR)
library(BSgenome.Hsapiens.UCSC.hg19)


ranges <- matrix(c("chr1", "10025",  "10525",  ".", "chr1", "13252",  "13752",  "."),ncol = 4, byrow = TRUE)
colnames(ranges) <- c("chr", "start", "end", "strand")   

#Build a GRanges from your matrix
ranges <- toGRanges(data.frame(ranges))

#Get the sequences and compute the GC content
freqs <- alphabetFrequency(getSeq(BSgenome.Hsapiens.UCSC.hg19, ranges))
gc <- (freqs[,'C'] + freqs[,'G'])/rowSums(freqs)

Here gc is a vector with the GC frequency of each range in the original matrix.

Hope this helps

Bernat

A solution using GenomicRanges and BSgenome from Bioconductor. You will need to transform your data to a GRanges object for this.

library(GenomicRanges)
library(BSgenome.Hsapiens.UCSC.hg19)
library(BSgenome)

gr <- GenomicRanges::GRanges(seqnames = c("chr1", "chr2"),
                             ranges = IRanges(start = c(123456,123490), 
                                              end = c(500020, 600020)))

GetGC <- function(bsgenome, gr){

  seqs <- BSgenome::getSeq(bsgenome, gr)
  return(as.numeric(Biostrings::letterFrequency(x = seqs, letters = "GC", as.prob = TRUE)))


}

> gr
GRanges object with 2 ranges and 0 metadata columns:
      seqnames        ranges strand
         <Rle>     <IRanges>  <Rle>
  [1]     chr1 123456-500020      *
  [2]     chr2 123490-600020      *
  -------
  seqinfo: 2 sequences from an unspecified genome; no seqlengths

> GetGC(bsgenome = BSgenome.Hsapiens.UCSC.hg19, gr = gr)
[1] 0.2823072 0.4355687

thank you all, these solutions are great. I also found a function that calculates GC content using BSgenome.Hsapiens.UCSC.hg19.

require(BSgenome.Hsapiens.UCSC.hg19)
require(BSgenome)
library(Repitools)
Granges<- makeGRangesFromDataFrame(data.frame(ranges))
gc<- gcContentCalc(Granges , organism=Hsapiens, verbose=TRUE)