Mercurial > repos > prog > lcmsmatching

comparison MassdbConn.R @ 2:20d69a062da3 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 d4048accde6bdfd5b3e14f5394902d38991854f8
author prog
date Thu, 02 Mar 2017 08:55:00 -0500
parents 253d531a0193
children
comparison
equal deleted inserted replaced
1:253d531a0193 2:20d69a062da3
1 if ( ! exists('MassdbConn')) { 1 #####################
2 # CLASS DECLARATION #
3 #####################
2 4
3 source('BiodbConn.R') 5 MassdbConn <- methods::setRefClass("MassdbConn", contains = "BiodbConn")
4 6
5 ##################### 7 ###############################
6 # CLASS DECLARATION # 8 # GET CHROMATOGRAPHIC COLUMNS #
7 ##################### 9 ###############################
10
11 # Get a list of chromatographic columns contained in this database.
12 # compound.ids A list of compound IDs used to filter results.
13 # The returned value is a data.frame with two columns : one for the ID (BIODB.ID) and another one for the title (BIODB.TITLE).
14 MassdbConn$methods( getChromCol = function(compound.ids = NULL) {
15 stop("Method getChromCol() is not implemented in concrete class.")
16 })
17
18 #################
19 # GET MZ VALUES #
20 #################
21
22 # Returns a numeric vector of all masses stored inside the database.
23 MassdbConn$methods( getMzValues = function(mode = NULL, max.results = NA_integer_) {
24 stop("Method getMzValues() not implemented in concrete class.")
25 })
26
27 ################
28 # GET NB PEAKS #
29 ################
30
31 # Returns the number of peaks contained in the database
32 MassdbConn$methods( getNbPeaks = function(mode = NULL, compound.ids = NULL) {
33 stop("Method getNbPeaks() not implemented in concrete class.")
34 })
35
36 #########################
37 # FIND COMPOUND BY NAME #
38 #########################
39
40 # Find a molecule by name
41 # name A vector of molecule names to search for.
42 # Return an integer vector of the same size as the name input vector, containing the found molecule IDs, in the same order.
43 MassdbConn$methods( findCompoundByName = function(name) {
44 stop("Method findCompoundByName() not implemented in concrete class.")
45 })
46
47 ####################################
48 # FIND SPECTRA IN GIVEN MASS RANGE #
49 ####################################
50 # Find spectra in the given mass range.
51 # rtype the type of return, objects, dfspecs data.frame of spectra, dfpeaks data.frame of peaks.
52 MassdbConn$methods( searchMzRange = function(mzmin, mzmax, rtype = c("objects","dfspecs","dfpeaks")){
53 stop("Method searchMzRange() not implemented in concrete class.")
54 })
55
56 ####################################
57 # FIND SPECTRA IN GIVEN MASS RANGE #
58 ####################################
59 MassdbConn$methods( searchMzTol = function(mz, tol, tolunit=BIODB.MZTOLUNIT.PLAIN, rtype = c("objects","dfspecs","dfpeaks")){
60 stop("Method searchMzTol() not implemented in concrete class.")
61 })
62
63 ######################################################
64 # FIND A MOLECULES WITH PRECURSOR WITHIN A TOLERANCE #
65 ######################################################
66 MassdbConn$methods( searchSpecPrecTol = function(mz, tol, tolunit=BIODB.MZTOLUNIT.PLAIN, mode = NULL){
67 stop("Method searchSpecPrecTol not implemented in concrete class.")
68 })
69
70 #################################
71 #perform a database MS-MS search#
72 #################################
73
74 ### spec : the spec to match against the database.
75 ### precursor : the mass/charge of the precursor to be looked for.
76 ### mtol : the size of the windows arounf the precursor to be looked for.
77 ### ppm : the matching ppm tolerance.
78 ### fun :
79 ### dmz : the mass tolerance is taken as the minium between this quantity and the ppm.
80 ### npmin : the minimum number of peak to detect a match (2 recommended)
81
82 MassdbConn$methods( msmsSearch = function(spec, precursor, mztol, tolunit,
83 ppm, fun = BIODB.MSMS.DIST.WCOSINE,
84 params = list(), npmin=2, dmz = 0.001,
85 mode = BIODB.MSMODE.POS, return.ids.only = TRUE){
86
8 87
9 MassdbConn <- setRefClass("MassdbConn", contains = "BiodbConn") 88 # TODO replace by msms precursor search when available.
89 lspec <- .self$searchSpecPrecTol( precursor, mztol, BIODB.MZTOLUNIT.PLAIN, mode = mode)
90 rspec <- lapply(lspec,function(x){
91 peaks <- x$getFieldValue(BIODB.PEAKS)
92
93 ####Getting the correct fields
94 vcomp <- c(BIODB.PEAK.MZ, BIODB.PEAK.RELATIVE.INTENSITY, BIODB.PEAK.INTENSITY)
95
96 foundfields <- vcomp %in% colnames(peaks)
97 if(sum(foundfields ) < 2){
98 stop(paste0("fields can't be coerced to mz and intensity : ",colnames(peaks)))
99 }
100
101 peaks <- peaks[ , vcomp[which( foundfields ) ] ]
102
103 peaks
104 })
105
106 # TODO Import compareSpectra into biodb and put it inside massdb-helper.R or hide it as a private method.
107 res <- compareSpectra(spec, rspec, npmin = npmin, fun = fun, params = params)
108
109 if(is.null(res)) return(NULL) # To decide at MassdbConn level: return empty list (or empty data frame) or NULL.
110 ###Adiing the matched peaks and the smimlarity values to spectra.
111
112 lret <-vector(length(lspec),mode = "list")
113 vsimilarity <- numeric( length( lspec ) )
114 vmatched <- vector( mode = "list", length( lspec ) )
115
116 if( return.ids.only ){
117 lret <- sapply( lspec, function( x ) {
118 x$getFieldValue( BIODB.ACCESSION )
119 })
120 }else{
121 ###TODO implement three types of return.
122 lret <- lspec
123 }
124
125 ###Reordering the list.
126 lret <- lret[ res$ord ]
127
10 128
11 ############################### 129 return( list(measure = res$similarity[ res$ord ], matchedpeaks = res$matched [ res$ord ], id = lret))
12 # GET CHROMATOGRAPHIC COLUMNS # 130 })
13 ###############################
14
15 # Get a list of chromatographic columns contained in this database.
16 # compound.ids A list of compound IDs used to filter results.
17 # The returned value is a data.frame with two columns : one for the ID (BIODB.ID) and another one for the title (BIODB.TITLE).
18 MassdbConn$methods( getChromCol = function(compound.ids = NULL) {
19 stop("Method getChromCol() is not implemented in concrete class.")
20 })
21
22 #################
23 # GET MZ VALUES #
24 #################
25
26 # Returns a numeric vector of all masses stored inside the database.
27 MassdbConn$methods( getMzValues = function(mode = NULL, max.results = NA_integer_) {
28 stop("Method getMzValues() not implemented in concrete class.")
29 })
30
31 }