Mercurial > repos > mvdbeek > r_goseq_1_22_0

annotate get_length_and_gc_content.r @ 0:a6427f7893b0 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
author mvdbeek
date Fri, 26 Feb 2016 06:57:47 -0500
parents
children 0e9424413ab0
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
1 # originally by Devon Ryan, https://www.biostars.org/p/84467/
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
2 sink(stdout(), type = "message")
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
3
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
4 library(GenomicRanges)
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
5 library(rtracklayer)
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
6 library(Rsamtools)
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
7 library(optparse)
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
8 library(data.table)
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
9
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
10 option_list <- list(
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
11 make_option(c("-g","--gtf"), type="character", help="Input GTF file with gene / exon information."),
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
12 make_option(c("-f","--fasta"), type="character", default=FALSE, help="Fasta file that corresponds to the supplied GTF."),
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
13 make_option(c("-o","--output"), type="character", default=FALSE, help="Output file with gene name, length and GC content.")
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
14 )
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
15
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
16 parser <- OptionParser(usage = "%prog [options] file", option_list=option_list)
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
17 args = parse_args(parser)
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
18
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
19 GTFfile = args$gtf
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
20 FASTAfile = args$fasta
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
21 output_file = args$output
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
22
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
23 #Load the annotation and reduce it
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
24 GTF <- import.gff(GTFfile, format="gtf", genome=NA, feature.type="exon")
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
25 grl <- reduce(split(GTF, elementMetadata(GTF)$gene_id))
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
26 reducedGTF <- unlist(grl, use.names=T)
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
27 elementMetadata(reducedGTF)$gene_id <- rep(names(grl), elementLengths(grl))
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
28
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
29 #Open the fasta file
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
30 FASTA <- FaFile(FASTAfile)
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
31 open(FASTA)
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
32
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
33 #Add the GC numbers
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
34 elementMetadata(reducedGTF)$nGCs <- letterFrequency(getSeq(FASTA, reducedGTF), "GC")[,1]
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
35 elementMetadata(reducedGTF)$widths <- width(reducedGTF)
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
36
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
37 #Create a list of the ensembl_id/GC/length
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
38 calc_GC_length <- function(x) {
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
39 nGCs = sum(elementMetadata(x)$nGCs)
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
40 width = sum(elementMetadata(x)$widths)
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
41 c(width, nGCs/width)
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
42 }
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
43 output <- t(sapply(split(reducedGTF, elementMetadata(reducedGTF)$gene_id), calc_GC_length))
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
44 output <- setDT(data.frame(output), keep.rownames = TRUE)[]
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
45 colnames(output) <- c("#gene_id", "length", "GC")
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
46
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
47 write.table(output, file=output_file, row.names=FALSE, quote=FALSE, sep="\t")
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
48
a6427f7893b0 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
49 sessionInfo()