Mercurial > repos > prog > lcmsmatching

annotate MassbankCompound.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('MassbankCompound')) { # 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 MassbankCompound <- setRefClass("MassbankCompound", 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 createMassbankCompoundFromTxt <- function(contents) {
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 for (text in contents) {
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
22
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
23 # Create instance
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
24 compound <- MassbankCompound$new()
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
25
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
26 # Read text
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
27 lines <- strsplit(text, "\n")
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
28 for (s in lines[[1]]) {
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
29
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
30 # NAME
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
31 if (is.na(compound$getField(RBIODB.NAME))) {
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
32 g <- str_match(s, "^CH\\$NAME:\\s+(.+)$")
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
33 if ( ! is.na(g[1,1]))
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
34 compound$setField(RBIODB.NAME, g[1,2])
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
35 }
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
36
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
37 # CHEBI ID
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
38 g <- str_match(s, "^CH\\$LINK: CHEBI\\s+(.+)$")
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
39 if ( ! is.na(g[1,1]))
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
40 compound$setField(RBIODB.CHEBI.ID, g[1,2])
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
41
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
42 # KEGG ID
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
43 g <- str_match(s, "^CH\\$LINK: KEGG\\s+(.+)$")
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
44 if ( ! is.na(g[1,1]))
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
45 compound$setField(RBIODB.KEGG.ID, g[1,2])
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
46
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
47 # PUBCHEM ID
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
48 g <- str_match(s, "^CH\\$LINK: PUBCHEM\\s+(.+)$")
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
49 if ( ! is.na(g[1,1]))
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
50 compound$setField(RBIODB.PUBCHEM.ID, g[1,2])
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
51
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
52 # INCHI
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
53 g <- str_match(s, "^CH\\$IUPAC:\\s+(.+)$")
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
54 if ( ! is.na(g[1,1]))
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
55 compound$setField(RBIODB.INCHI, g[1,2])
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
56 }
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
57
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
58 compounds <- c(compounds, compound)
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
59 }
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 # 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
62 compounds <- lapply(compounds, function(x) if (is.na(x$getField(RBIODB.NAME))) NULL else x)
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
63
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
64 return(compounds)
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
65 }
3afe41d3e9e7 planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit bb4d3e23d99828bfee16d31d794c49a17313ec2f
prog
parents:
diff changeset
66 }