Mercurial > repos > proteore > proteore_expression_rnaseq_abbased

comparison add_expression_HPA.R @ 0:234d114cbe3a draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload commit 2ee714c3e3d1cce461125dbc041904e03ea8bac1-dirty
author proteore
date Mon, 12 Nov 2018 11:05:24 -0500
parents
children 03487ed0f458
comparison
equal deleted inserted replaced
-1:000000000000 0:234d114cbe3a
1 # Read file and return file content as data.frame
2 read_file <- function(path,header){
3 file <- try(read.csv(path,header=header, sep="\t",stringsAsFactors = FALSE, quote="\"", check.names = F),silent=TRUE)
4 if (inherits(file,"try-error")){
5 stop("File not found !")
6 }else{
7 return(file)
8 }
9 }
10
11 #convert a string to boolean
12 str2bool <- function(x){
13 if (any(is.element(c("t","true"),tolower(x)))){
14 return (TRUE)
15 }else if (any(is.element(c("f","false"),tolower(x)))){
16 return (FALSE)
17 }else{
18 return(NULL)
19 }
20 }
21
22 add_expression = function(input, atlas, options) {
23 if (all(!input %in% atlas$Ensembl)) {
24 return(NULL)
25 } else {
26 res = atlas[match(input,atlas$Ensembl),c("Ensembl",options)]
27 res <- as.data.frame(apply(res, c(1,2), function(x) gsub("^$|^ $", NA, x))) #convert "" et " " to NA
28 return(res)
29 }
30 }
31
32 main = function() {
33 args <- commandArgs(TRUE)
34 if(length(args)<1) {
35 args <- c("--help")
36 }
37
38 # Help section
39 if("--help" %in% args) {
40 cat("Selection and Annotation HPA
41 Arguments:
42 --inputtype: type of input (list of id or filename)
43 --input: either a file name (e.g : input.txt) or a list of blank-separated
44 ENSG identifiers (e.g : ENSG00000283071 ENSG00000283072)
45 --atlas: path to protein atlas file
46 --column: the column number which you would like to apply...
47 --header: true/false if your file contains a header
48 --select: information from HPA to select, maybe:
49 RNA.tissue.category,Reliability..IH.,Reliability..IF. (comma-separated)
50 --output: text output filename \n")
51 q(save="no")
52 }
53
54 # Parse arguments
55 parseArgs <- function(x) strsplit(sub("^--", "", x), "=")
56 argsDF <- as.data.frame(do.call("rbind", parseArgs(args)))
57 args <- as.list(as.character(argsDF$V2))
58 names(args) <- argsDF$V1
59
60 save(args,file="/home/dchristiany/proteore_project/ProteoRE/tools/add_expression_data_HPA/args.rda")
61 #load("/home/dchristiany/proteore_project/ProteoRE/tools/add_expression_data_HPA/args.rda")
62
63 inputtype = args$inputtype
64 if (inputtype == "copypaste") {
65 input = strsplit(args$input, "[ \t\n]+")[[1]]
66 } else if (inputtype == "tabfile") {
67 filename = args$input
68 ncol = args$column
69 # Check ncol
70 if (! as.numeric(gsub("c", "", ncol)) %% 1 == 0) {
71 stop("Please enter an integer for level")
72 } else {
73 ncol = as.numeric(gsub("c", "", ncol))
74 }
75 header = str2bool(args$header)
76 file = read_file(filename, header)
77 input = unlist(sapply(as.character(file[,ncol]),function(x) rapply(strsplit(x,";"),c),USE.NAMES = FALSE))
78 }
79
80 # Read protein atlas
81 protein_atlas = args$atlas
82 protein_atlas = read_file(protein_atlas, T)
83
84 # Add expression
85 output = args$output
86 options = strsplit(args$select, ",")[[1]]
87 res = add_expression(input, protein_atlas, options)
88
89
90 # Write output
91 if (is.null(res)) {
92 write.table("None of the input ENSG ids are can be found in HPA data file",file=output,sep="\t",quote=FALSE,col.names=TRUE,row.names=FALSE)
93 } else {
94 if (inputtype == "copypaste") {
95 output_content = cbind(as.matrix(input), res)
96 colnames(output_content)[1] = "Ensembl"
97 } else if (inputtype == "tabfile") {
98 output_content = merge(file, res, by.x=ncol, by.y=1, incomparables = NA,all.x=T)
99 output_content = output_content[order(output_content[,ncol],decreasing = T),]
100 }
101 output_content <- as.data.frame(apply(output_content, c(1,2), function(x) gsub("^$|^ $", NA, x)))
102 write.table(output_content, output, row.names = FALSE, sep = "\t", quote = FALSE)
103 }
104 }
105
106 main()