Mercurial > repos > proteore > proteore_kegg_pathways_coverage

comparison compute_kegg_pathways.R @ 1:7004924a3686 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload commit 2e8375a0a037aaf4042cdeb5244228d3f9ca583b-dirty
author proteore
date Wed, 14 Nov 2018 05:46:35 -0500
parents 8883a7173cba
children
comparison
equal deleted inserted replaced
0:8883a7173cba 1:7004924a3686
80 geneID_to_kegg <- function(vector,species){ 80 geneID_to_kegg <- function(vector,species){
81 vector <- sapply(vector, function(x) paste(species,x,sep=":"),USE.NAMES = F) 81 vector <- sapply(vector, function(x) paste(species,x,sep=":"),USE.NAMES = F)
82 return (vector) 82 return (vector)
83 } 83 }
84 84
85 to_keggID <- function(id_list,id_type){
86 if (id_type == "ncbi-geneid") {
87 id_list <- unique(geneID_to_kegg(id_list,args$species))
88 } else if (id_type=="uniprot"){
89 id_list <- unique(sapply(id_list, function(x) paste(id_type,":",x,sep=""),USE.NAMES = F))
90 if (length(id_list)>250){
91 id_list <- split(id_list, ceiling(seq_along(id_list)/250))
92 id_list <- sapply(id_list, function(x) keggConv("genes",x))
93 id_list <- unique(unlist(id_list))
94 } else {
95 id_list <- unique(keggConv("genes", id_list))
96 }
97 } else if (id_type=="kegg-id") {
98 id_list <- unique(id_list)
99 }
100 return (id_list)
101 }
102
85 kegg_mapping<- function(kegg_id_list,id_type,ref_ids) { 103 kegg_mapping<- function(kegg_id_list,id_type,ref_ids) {
86
87 #convert to KEGG ID
88 #if (id_type!="kegg-id"){
89 # id_list <- unique(sapply(id_list, function(x) paste(id_type,":",x,sep=""),USE.NAMES = F))
90 # if (length(id_list)>250){
91 # id_list <- split(id_list, ceiling(seq_along(id_list)/250))
92 # id_list <- sapply(id_list, function(x) keggConv("genes",x))
93 # kegg_id_list <- unique(unlist(id_list))
94 # } else {
95 # kegg_id_list <- unique(keggConv("genes", id_list))
96 # }
97 #} else {
98 # kegg_id_list <- unique(id_list)
99 #}
100 104
101 #mapping 105 #mapping
102 map<-lapply(ref_ids, is.element, unique(kegg_id_list)) 106 map<-lapply(ref_ids, is.element, unique(kegg_id_list))
103 names(map) <- sapply(names(map), function(x) gsub("path:","",x),USE.NAMES = FALSE) #remove the prefix "path:" 107 names(map) <- sapply(names(map), function(x) gsub("path:","",x),USE.NAMES = FALSE) #remove the prefix "path:"
104 108
127 #save(args,file="/home/dchristiany/proteore_project/ProteoRE/tools/kegg_pathways_identification/args.Rda") 131 #save(args,file="/home/dchristiany/proteore_project/ProteoRE/tools/kegg_pathways_identification/args.Rda")
128 #load("/home/dchristiany/proteore_project/ProteoRE/tools/kegg_pathways_identification/args.Rda") 132 #load("/home/dchristiany/proteore_project/ProteoRE/tools/kegg_pathways_identification/args.Rda")
129 133
130 ###setting variables 134 ###setting variables
131 header = str2bool(args$header) 135 header = str2bool(args$header)
132 if (!is.null(args$id_list)) {id_list <- get_list_from_cp(args$id_list)} 136 if (!is.null(args$id_list)) {id_list <- get_list_from_cp(args$id_list)} #get ids from copy/paste input
133 if (!is.null(args$input)) { 137 if (!is.null(args$input)) { #get ids from input file
134 csv <- read_file(args$input,header) 138 csv <- read_file(args$input,header)
135 ncol <- as.numeric(gsub("c", "" ,args$id_column)) 139 ncol <- as.numeric(gsub("c", "" ,args$id_column))
136 id_list <- as.vector(csv[,ncol]) 140 id_list <- as.vector(csv[,ncol])
137 id_list <- id_list[which(!is.na(id_list))] 141 id_list <- id_list[which(!is.na(id_list))]
138 } 142 }
139 if (args$id_type == "ncbi-geneid") {
140 id_list <- geneID_to_kegg(id_list,args$species)
141 }
142 143
144 #convert to keggID if needed
145 id_list <- to_keggID(id_list,args$id_type)
143 146
144 #get pathways of species with associated KEGG ID genes 147 #get pathways of species with associated KEGG ID genes
145 pathways_list <- get_pathways_list(args$species) 148 pathways_list <- get_pathways_list(args$species)
146 149
147 #mapping on pathways 150 #mapping on pathways