comparison goseq.r @ 0:fe71b97cc1a5 draft

planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/goseq_1_22_0 commit 85e0f34aec3045f48ca9b1a00db02de199a99c95-dirty
author mvdbeek
date Wed, 24 Feb 2016 15:54:37 -0500
parents
children 8ce951313688
comparison
equal deleted inserted replaced
-1:000000000000 0:fe71b97cc1a5
1 sink(stdout(), type = "message")
2 library(goseq)
3 library(optparse)
4
5 option_list <- list(
6 make_option(c("-d", "--dge_file"), type="character", help="Path to file with differential gene expression result"),
7 make_option(c("-w","--wallenius_tab"), type="character", help="Path to output file with P-values estimated using wallenius distribution."),
8 make_option(c("-s","--sampling_tab"), type="character", default=FALSE, help="Path to output file with P-values estimated using wallenius distribution."),
9 make_option(c("-n","--nobias_tab"), type="character", default=FALSE, help="Path to output file with P-values estimated using wallenius distribution and no correction for gene length bias."),
10 make_option(c("-l","--length_bias_plot"), type="character", default=FALSE, help="Path to length-bias plot."),
11 make_option(c("-sw","--sample_vs_wallenius_plot"), type="character", default=FALSE, help="Path to plot comparing sampling with wallenius p-values."),
12 make_option(c("-padj", "--p_adj_column"), type="integer",help="Column that contains p. adjust values"),
13 make_option(c("-c", "--cutoff"), type="double",dest="p_adj_cutoff",
14 help="Genes with p.adjust below cutoff are considered not differentially expressed and serve as control genes"),
15 make_option(c("-r", "--repcnt"), type="integer", default=100, help="Number of repeats for sampling"),
16 make_option(c("-g", "--genome"), type="character", help = "Genome [used for looking up correct gene length]"),
17 make_option(c("-i", "--gene_id"), type="character", help="Gene ID of gene column in DGE file")
18 )
19 parser <- OptionParser(usage = "%prog [options] file", option_list=option_list)
20 args = parse_args(parser)
21
22 # Vars:
23 dge_file = args$dge_file
24 p_adj_column = args$p_adj_colum
25 p_adj_cutoff = args$p_adj_cutoff
26 genome = args$genome
27 gene_id = args$gene_id
28 wallenius_tab = args$wallenius_tab
29 sampling_tab = args$sampling_tab
30 nobias_tab = args$nobias_tab
31 length_bias_plot = args$length_bias_plot
32 sample_vs_wallenius_plot = args$sample_vs_wallenius_plot
33 repcnt = args$repcnt
34
35
36 # format DE genes into vector suitable for use with goseq
37 dge_table = read.delim(dge_file, header = TRUE, sep="\t", check.names = FALSE)
38 genes = as.integer(dge_table[,p_adj_column]<p_adj_cutoff)
39 names(genes) = dge_table[,1] # Assuming first row contains gene names
40
41 # Estimate PWF
42
43 pdf(length_bias_plot)
44 pwf=nullp(genes, genome , gene_id)
45 dev.off()
46 # Null dstribution wallenius
47 GO.wall=goseq(pwf, genome, gene_id)
48
49 GO.nobias=goseq(pwf, genome, gene_id, method="Hypergeometric")
50
51 # Sampling dsitribution
52 GO.samp=goseq(pwf,genome, gene_id, method="Sampling",repcnt=repcnt)
53
54 # Compare sampling with wallenius
55 pdf(sample_vs_wallenius_plot)
56 plot(log10(GO.wall[,2]), log10(GO.samp[match(GO.samp[,1],GO.wall[,1]),2]),
57 xlab="log10(Wallenius p-values)",ylab="log10(Sampling p-values)",
58 xlim=c(-3,0))
59 abline(0,1,col=3,lty=2)
60 dev.off()
61
62
63 write.table(GO.wall, wallenius_tab, sep="\t", row.names = FALSE, quote = FALSE)
64 write.table(GO.samp, sampling_tab, sep="\t", row.names = FALSE, quote = FALSE)
65 write.table(GO.nobias, nobias_tab, sep="\t", row.names = FALSE, quote = FALSE)
66
67 # Use the following to get a list of supported genomes / gene ids
68
69 # write.table(supportedGenomes(), "available_genomes.tab", row.names = FALSE, quote=FALSE)
70 # write.table(supportedGeneIDs(), "supported_gene_ids.tab", row.name = FALSE, quote = FALSE)
71 # write.table(table.summary, "input_gene_count_matrix.tab", row.names = FALSE, quote = FALSE)