Mercurial > repos > mvdbeek > r_goseq_1_22_0

annotate get_length_and_gc_content.r @ 11:1802ccf465b8 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 Thu, 25 Feb 2016 06:20:59 -0500
parents 7f8d888e3355
children 02e88556ce1d
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
1
3ab168143b69 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/
6
d4b5942ed347 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents: 5
diff changeset
2 sink(stdout(), type = "message")
1
3ab168143b69 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
3ab168143b69 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)
3ab168143b69 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)
3ab168143b69 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)
3ab168143b69 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)
3ab168143b69 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
3ab168143b69 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 option_list <- list(
3ab168143b69 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 make_option(c("-g","--gtf"), type="character", help="Input GTF file with gene / exon information."),
3ab168143b69 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("-f","--fasta"), type="character", default=FALSE, help="Fasta file that corresponds to the supplied GTF."),
3ab168143b69 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("-o","--output"), type="character", default=FALSE, help="Output file with gene name, length and GC content.")
3ab168143b69 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 )
3ab168143b69 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
3ab168143b69 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 parser <- OptionParser(usage = "%prog [options] file", option_list=option_list)
3ab168143b69 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 args = parse_args(parser)
3ab168143b69 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
5
f9b964d1d386 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents: 1
diff changeset
18 GTFfile = args$gtf
f9b964d1d386 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents: 1
diff changeset
19 FASTAfile = args$fasta
10
7f8d888e3355 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents: 7
diff changeset
20 output_file = args$output
1
3ab168143b69 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
21
3ab168143b69 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
22 #Load the annotation and reduce it
6
d4b5942ed347 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents: 5
diff changeset
23 GTF <- import.gff(GTFfile, format="gtf", genome=NA, feature.type="exon")
7
15ce6435ab83 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents: 6
diff changeset
24 grl <- reduce(split(GTF, elementMetadata(GTF)$gene_id))
1
3ab168143b69 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 reducedGTF <- unlist(grl, use.names=T)
7
15ce6435ab83 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents: 6
diff changeset
26 elementMetadata(reducedGTF)$gene_id <- rep(names(grl), elementLengths(grl))
1
3ab168143b69 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
27
3ab168143b69 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
28 #Open the fasta file
3ab168143b69 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
29 FASTA <- FaFile(FASTAfile)
3ab168143b69 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
30 open(FASTA)
3ab168143b69 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
31
3ab168143b69 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
32 #Add the GC numbers
3ab168143b69 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
33 elementMetadata(reducedGTF)$nGCs <- letterFrequency(getSeq(FASTA, reducedGTF), "GC")[,1]
3ab168143b69 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)$widths <- width(reducedGTF)
3ab168143b69 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
35
3ab168143b69 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
36 #Create a list of the ensembl_id/GC/length
3ab168143b69 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
37 calc_GC_length <- function(x) {
3ab168143b69 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
38 nGCs = sum(elementMetadata(x)$nGCs)
3ab168143b69 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 width = sum(elementMetadata(x)$widths)
3ab168143b69 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
40 c(width, nGCs/width)
3ab168143b69 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents:
diff changeset
41 }
7
15ce6435ab83 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents: 6
diff changeset
42 output <- t(sapply(split(reducedGTF, elementMetadata(reducedGTF)$gene_id), calc_GC_length))
1
3ab168143b69 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 colnames(output) <- c("Length", "GC")
3ab168143b69 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
10
7f8d888e3355 planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit fdd0811efc61c31f88ff17096fbe8ee8cfacd766-dirty
mvdbeek
parents: 7
diff changeset
45 write.table(output, file=output_file, sep="\t")