Mercurial > repos > prog > lcmsmatching

annotate KeggCompound.R @ 0:3afe41d3e9e7 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
author prog
date Mon, 11 Jul 2016 09:12:03 -0400
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
1 if ( ! exists('KeggCompound')) { # Do not load again if already loaded
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
2
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
3 source('BiodbEntry.R')
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
4
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
5 #####################
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
6 # CLASS DECLARATION #
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
7 #####################
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
8
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
9 KeggCompound <- setRefClass("KeggCompound", contains = 'BiodbEntry')
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
10
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
11 ###########
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
12 # FACTORY #
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
13 ###########
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
14
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
15 createKeggCompoundFromTxt <- function(contents, drop = TRUE) {
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
16
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
17 library(stringr)
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
18
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
19 compounds <- list()
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
20
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
21 # Define fields regex
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
22 regex <- character()
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
23 regex[[RBIODB.NAME]] <- "^NAME\\s+([^,;]+)"
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
24 regex[[RBIODB.CHEBI.ID]] <- "^\\s+ChEBI:\\s+(\\S+)"
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
25 regex[[RBIODB.LIPIDMAPS.ID]] <- "^\\s+LIPIDMAPS:\\s+(\\S+)"
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
26
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
27 for (text in contents) {
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
28
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
29 # Create instance
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
30 compound <- KeggCompound$new()
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
31
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
32 lines <- strsplit(text, "\n")
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
33 for (s in lines[[1]]) {
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
34
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
35 # Test generic regex
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
36 parsed <- FALSE
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
37 for (field in names(regex)) {
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
38 g <- str_match(s, regex[[field]])
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
39 if ( ! is.na(g[1,1])) {
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
40 compound$setField(field, g[1,2])
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
41 parsed <- TRUE
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
42 break
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
43 }
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
44 }
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
45 if (parsed)
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
46 next
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
47
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
48 # ACCESSION
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
49 {
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
50 # ENZYME ID
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
51 g <- str_match(s, "^ENTRY\\s+EC\\s+(\\S+)")
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
52 if ( ! is.na(g[1,1]))
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
53 compound$setField(RBIODB.ACCESSION, paste('ec', g[1,2], sep = ':'))
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
54
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
55 # ENTRY ID
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
56 else {
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
57 g <- str_match(s, "^ENTRY\\s+(\\S+)\\s+Compound")
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
58 if ( ! is.na(g[1,1]))
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
59 compound$setField(RBIODB.ACCESSION, paste('cpd', g[1,2], sep = ':'))
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
60
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
61 # OTHER ID
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
62 else {
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
63 g <- str_match(s, "^ENTRY\\s+(\\S+)")
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
64 if ( ! is.na(g[1,1]))
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
65 compound$setField(RBIODB.ACCESSION, g[1,2])
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
66 }
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
67 }
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
68
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
69 # ORGANISM
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
70 g <- str_match(s, "^ORGANISM\\s+(\\S+)")
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
71 if ( ! is.na(g[1,1]))
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
72 compound$setField(RBIODB.ACCESSION, paste(g[1,2], compound$getField(RBIODB.ACCESSION), sep = ':'))
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
73 }
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
74 }
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
75
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
76 compounds <- c(compounds, compound)
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
77 }
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
78
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
79 # Replace elements with no accession id by NULL
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
80 compounds <- lapply(compounds, function(x) if (is.na(x$getField(RBIODB.ACCESSION))) NULL else x)
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
81
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
82 # If the input was a single element, then output a single object
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
83 if (drop && length(contents) == 1)
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
84 compounds <- compounds[[1]]
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
85
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
86 return(compounds)
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
87 }
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
88 }