R/as.ftmsData.R
In EMSL-Computing/fticRanalysis: Analysis and visualization tools for FT-MS data
Defines functions
as.moduleData
as.reactionData
as.compoundData
.as.peakData
as.peakData
Documented in
as.compoundData
as.moduleData
as.peakData
as.reactionData
#' Convert Data to peakData Class
#'
#' Converts a list object or several data.frames of FT-MS data to an object of the class 'peakData'. Objects of the class 'peakData' are lists with three obligatory components \code{e_data}, \code{f_data}, and \code{e_meta}.
#'
#' @param e_data a \eqn{p \times n + 1} data.frame of expression data, where \eqn{p} is the number of observed peaks and \eqn{n} is the number of samples. Each row corresponds to data for each peak. One column specifying a unique identifier for each peak/mass (row) must be present.
#' @param f_data a data.frame with \eqn{n} rows. Each row corresponds to a sample with one column giving the unique sample identifiers found in e_data column names and other columns providing qualitative and/or quantitative traits of each sample.
#' @param e_meta a data.frame with \eqn{p} rows. Each row corresponds to a peak/mass with one column giving a unique peak/identifier (must be named the same as the column in \code{e_data}) and other columns giving meta information. At a minimum a column giving the mass of each peak and a column giving molecular formulae or columns giving elemental counts must be present.
#' @param edata_cname character string specifying the name of the column containing a unique identifier for each peak/mass in \code{e_data} and \code{e_meta}.
#' @param fdata_cname character string specifying the name of the column containing the sample identifiers in \code{f_data}.
#' @param mass_cname character string specifying the name of the column containing the peak/mass identifiers in \code{e_meta}. Note: this is often the same as \code{edata_cname} for cases where mass is used as a unique identifier.
#' @param ... further arguments
#'
#' @details Objects of class 'peakData' contain some attributes that are referenced by downstream functions. These attributes must be specified (or added using available functions) to reference downstream functions for: Kendrick plots, Van Krevelen plots, and functions involving databases.
#'
#' If your data contains information about isotopic peaks (e.g. C13), you should specify the attribute \code{isotopic_cname} which gives the column in \code{e_meta} that contains an indicator of yes/no for each peak. Additionally, you must specify the attribute \code{isotopic_notation} which is a character string indicating the value in column \code{isotopic_cname} which indicates that a peak is isotopic.
#' Currently, any peaks that are isotopic are removed from the dataset, as available methods (e.g. Van Krevelen plot) are not applicable to these peaks.
#'
#' Attributes giving general information about the data object:
#' \tabular{ll}{
#' data_scale \tab character string giving the scale that the data is on. Valid options include 'log2', 'log10', 'log' (for natural log), 'pres' (for 0/1 presence/absence data), and 'abundance'. Default value is 'abundance'. \cr
#' \tab \cr
#' instrument_type \tab character string giving the type of FT-MS instrument data was generated by. Valid options are: "12T" and "21T". Defaults to "12T". This information is used to determine appropriate plotting functions for Van Krevelen, Kendrick, etc. plots. \cr
#' }
#' Attributes giving extra information in \code{f_data}:
#' \tabular{ll}{
#' extraction_cname \tab character string specifying the name of the column, in \code{f_data}, containing information as to what extraction method was used for a sample. Only necessary if \code{e_data} contains samples from multiple extraction methods. \cr
#' }
#' Attributes giving extra information in \code{e_meta}:
#' \tabular{ll}{
#' mass_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the mass information for each peak. \cr
#' \tab \cr
#' mf_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the mass (empirical) formula for a peak/mass. \cr
#' \tab \cr
#' element_col_names \tab named list of character strings specifying element/isotope column names, in \code{e_meta}, containing the respective count for each peak/mass. \cr
#' \tab \cr
#' isotopic_cname \tab character string specifying the name of the column, in \code{e_meta}, containing information about whether each peak is isotopic or not. \cr
#' \tab \cr
#' isotopic_notation \tab character string specifying the value used in column \code{isotopic_cname} which indicates that a peak is isotopic. \cr
#' \tab \cr
#' ratio_cnames \tab named list of character strings specifying the name of the column, in \code{e_meta}, containing the respective ratio of two elements or isotopes for each peak/mass. \cr
#' \tab \cr
#' kmass_cname \tab a possibly named character vector specifying the name of the columns, in \code{e_meta}, containing the Kendrick Mass for each peak/mass. Names should be any of 'CH2', 'CO2', 'H2', 'H2O', 'CHO' and correspond to the base compounds used to calculate each of the Kendrick Masses \cr
#' \tab \cr
#' kdefect_cname \tab a possibly named character vector specifying the name of the column, in \code{e_meta}, containing the Kendrick Defect for each peak/mass. Names should be any of 'CH2', 'CO2', 'H2', 'H2O', 'CHO' and correspond to the base compounds used to calculate each of the Kendrick Masses\cr
#' \tab \cr
#' nosc_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the NOSC value for each peak/mass \cr
#' \tab \cr
#' gfe_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the Gibb's Free Energy value for each peak/mass \cr
#' \tab \cr
#' mfname_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the name/description for each peak/mass \cr
#' \tab \cr
#' aroma_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the aromaticity value for each peak/mass \cr
#' \tab \cr
#' modaroma_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the modified aromaticity value for each peak/mass \cr
#' \tab \cr
#' dbe_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the double-bond equivalent values for each peak/mass \cr
#' \tab \cr
#' dbeo_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the double-bond equivalent minus oxygen value for each peak/mass \cr
#' \tab \cr
#' dbeai_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the double-bond equivalent aromaticity index value for each peak/mass \cr
#' \tab \cr
#' elcomp_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the general elemental composition of each peak/mass \cr
#' \tab \cr
#' check_rows \tab logical indicating whether to remove peaks with no nonzero entries. Defaults to FALSE \cr
#' }
#'
#'@author Lisa Bramer
#'
#'@export
as.peakData <- function(e_data, f_data, e_meta, edata_cname, fdata_cname, mass_cname, ...){
.as.peakData(e_data, f_data, e_meta, edata_cname, fdata_cname, mass_cname, ...)
}
# Need to change the col names param here
.as.peakData <- function(e_data, f_data, e_meta, edata_cname, fdata_cname, mass_cname,
extraction_cname = NULL, mf_cname = NULL,
element_col_names = list("C"="C", "H"="H", "O"=NULL, "N"=NULL, "S"=NULL, "P"=NULL),
isotopic_cname = NULL, isotopic_notation = NULL,
ratio_cnames = list("O:C" = NULL, "H:C" = NULL), kmass_cname = NULL,
kdefect_cname = NULL, nosc_cname = NULL, gfe_cname = NULL, mfname_cname = NULL,
aroma_cname = NULL, modaroma_cname = NULL, dbe_cname = NULL, dbeo_cname = NULL, dbeai_cname = NULL,
elcomp_cname = NULL, instrument_type = "12T", data_scale = "abundance", check_rows = FALSE){
# check that e_data, f_data, and e_meta are data.frames #
# if(!inherits(e_data, "data.frame")) stop("e_data must be of the class 'data.frame'")
# if(!inherits(f_data, "data.frame")) stop("f_data must be of the class 'data.frame'")
# if(!inherits(e_meta, "data.frame")) stop("e_meta must be of the class 'data.frame'")
# make sure e_data, f_data, and e_meta are data.frames #
e_data <- as.data.frame(e_data)
f_data <- as.data.frame(f_data)
e_meta <- as.data.frame(e_meta)
# check that the peak column exists in e_data and e_meta (if applicable) #
if(!(edata_cname %in% names(e_data))) stop(paste("Peak/Mass column ", edata_cname," not found in e_data. See details of as.peakData for specifying column names.", sep = ""))
if(!(edata_cname %in% names(e_meta))) stop(paste("Peak/Mass column ", edata_cname," not found in e_meta. Column names for peak/mass identifiers must match for e_data and e_meta. See details of as.peakData for specifying column names.", sep = ""))
# check that e_data has unique rows #
if(nrow(e_data) != length(unique(dplyr::pull(e_data, edata_cname)))) stop("The 'edata_cname' identifier is non-unique.")
# check that f_data has unique rows #
if(nrow(f_data) != length(unique(dplyr::pull(f_data, fdata_cname)))) stop("The 'fdata_cname' identifier is non-unique.")
# check that instrument_type is a valid string #
if(!(instrument_type %in% c("21T", "12T"))) stop("Instrument type is not valid. See details of as.peakData for valid options.")
# check that mass_cname is not null #
if(is.null(mass_cname)) stop("'mass_cname' must be specified")
# check that either mf_cname or elemental cnames are not null #
if(is.null(mf_cname) & any(c(is.null(element_col_names$C), is.null(element_col_names$H)))) {
stop("Either 'mf_cname' or both 'C' and 'H' must be specified in 'element_col_names'")
}
# check that cname arguments are found #
if(!(mass_cname %in% names(e_meta))) {
stop(paste("Mass column", mass_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = ""))
}
if(!is.null(element_col_names$C)){
if(!(element_col_names$C %in% names(e_meta))) stop(paste("Carbon column", element_col_names$C, " not found in e_meta. See details of as.peakData for specifying column names.", sep = ""))
}
if(!is.null(element_col_names$H)){
if(!(element_col_names$H %in% names(e_meta))) stop(paste("Hydrogen column", element_col_names$H, " not found in e_meta. See details of as.peakData for specifying column names.", sep = ""))
}
# Check that the elements/isotopes in element_col-names are within coreMS element/isotope superset.
if(!all(names(element_col_names) %in% element_names)){
stop(paste("The following elements in 'element_col_names' are not supported: ", paste(names(element_col_names[!(names(element_col_names) %in% element_names)]), collapse = ','), ". See details of as.peakData for specifying column names.", sep = ""))
}
# Check that non-NULL elements in element_col_names are present within e_meta
for(element in names(element_col_names)){
if(!is.null(element_col_names[[element]]) && ! element %in% names(e_meta)){
stop(paste("The following element in 'element_col_names' is not found in e_meta: ", element))
}
}
if(!is.null(isotopic_cname)){
if(!(isotopic_cname %in% names(e_meta))) stop(paste("Isotopic column", isotopic_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = ""))
}
if(!is.null(ratio_cnames$o2c_cname)){
if(!(ratio_cnames$o2c_cname %in% names(e_meta))) stop(paste("Oxygen:Carbon Ratio column ", ratio_cnames$o2c_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(ratio_cnames$h2c_cname)){
if(!(ratio_cnames$h2c_cname %in% names(e_meta))) stop(paste("Hydrogen:Carbon Ratio column ", ratio_cnames$h2c_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(kmass_cname)){
if(!is.null(names(kmass_cname))){
if(!all(names(kmass_cname)) %in% c('CH2', 'CO2', 'H2', 'H2O', 'CHO', '')) stop("Names (not the values) of kmass_cname must be in 'CH2', 'CO2', 'H2', 'H2O', 'CHO'")
}
if(!all(kmass_cname %in% names(e_meta))) stop(paste("Kendrick Mass column ", kmass_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(kdefect_cname)){
if(!is.null(names(kdefect_cname))){
if(!all(names(kdefect_cname)) %in% c('CH2', 'CO2', 'H2', 'H2O', 'CHO', '')) stop("Names (not the values) of kdefect_cname must be in 'CH2', 'CO2', 'H2', 'H2O', 'CHO'")
}
if(!all(kdefect_cname %in% names(e_meta))) stop(paste("Kendrick Defect column ", kdefect_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(nosc_cname)){
if(!(nosc_cname %in% names(e_meta))) stop(paste("NOSC column ", nosc_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(gfe_cname)){
if(!(gfe_cname %in% names(e_meta))) stop(paste("Gibbs Free Energy column ", gfe_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(mfname_cname)){
if(!(mfname_cname %in% names(e_meta))) stop(paste("Molecular formula name/description column ", mfname_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(aroma_cname)){
if(!(aroma_cname %in% names(e_meta))) stop(paste("Aromaticity column ", aroma_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(modaroma_cname)){
if(!(modaroma_cname %in% names(e_meta))) stop(paste("Modified aromaticity column ", modaroma_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(mf_cname)){
if(!(mf_cname %in% names(e_meta))) stop(paste("Molecular formula column ", mf_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(dbe_cname)){
if(!(dbe_cname %in% names(e_meta))) stop(paste("Double-bond equivalent column ", dbe_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(dbeo_cname)){
if(!(dbeo_cname %in% names(e_meta))) stop(paste("Double-bond equivalent (minus Oxygen) column ", dbeo_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(dbeai_cname)){
if(!(dbeai_cname %in% names(e_meta))) stop(paste("Double-bond equivalent (aromaticity index) column ", dbeai_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(elcomp_cname)){
if(!(elcomp_cname %in% names(e_meta))) stop(paste("Elemental composition column ", elcomp_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(extraction_cname)){
if(!(extraction_cname %in% names(f_data))) stop(paste("Extraction Method column ", extraction_cname, " not found in f_data. See details of as.peakData for specifying column names.", sep = "") )
}
# check that the Sample column name is in f_data column names #
if(!(fdata_cname %in% names(f_data))) stop(paste("Sample column ", fdata_cname, " not found in f_data. See details of as.pepData for specifying column names.", sep = ""))
# check that all samples in e_data are present in f_data #
edat_sampid = which(names(e_data) == edata_cname)
samps.miss = sum(!(names(e_data[,-edat_sampid]) %in% dplyr::pull(f_data,fdata_cname)))
if( samps.miss > 0) stop(paste( samps.miss, " samples from e_data not found in f_data", sep = ""))
# check for any extra samples in f_data not in e_data - necessary to remove before group_designation function #
if(any(!(dplyr::pull(f_data, fdata_cname) %in% names(e_data)))){
f_data <- f_data[-which(!(dplyr::pull(f_data,fdata_cname) %in% names(e_data))),]
warning("Extra samples were found in f_data that were not in e_data. These have been removed from f_data.")
}
# check that f_data has at least 2 columns #
# if(ncol(f_data) < 2) stop("f_data must contain at least 2 columns")
# if e_meta is provided, check that all peaks in e_data occur in e_meta #
if(!is.null(e_meta)){
if(sum(!(dplyr::pull(e_data,edata_cname) %in% dplyr::pull(e_meta,edata_cname))) > 0 ) stop("Not all peaks in e_data are present in e_meta")
}
# if e_meta is provided, remove any extra features that were provided #
if(!is.null(e_meta)){
if(any(!(dplyr::pull(e_meta,edata_cname) %in% dplyr::pull(e_data,edata_cname)))){
e_meta <- e_meta[-which(!(dplyr::pull(e_meta, edata_cname) %in% dplyr::pull(e_data, edata_cname))),]
warning("Extra peaks were found in e_meta that were not in e_data. These have been removed from e_meta.")
}
}
# convert the unique identifier column to character, to avoid factors #
e_data[, edata_cname] = as.character(dplyr::pull(e_data, edata_cname))
e_meta[, edata_cname] = as.character(dplyr::pull(e_meta, edata_cname))
# check that if isotopic_cname is non-NULL then isotopic_notation is also non-NULL #
if(!is.null(isotopic_cname)){
if(is.null(isotopic_notation)) stop("You must specify 'isotopic_notation', if 'isotopic_cname' is non-NULL")
}
# if isotopic_cname is not NULL #
if(!is.null(isotopic_cname)){
# filter out peaks where isotopic_cname = TRUE #
ids_rmv = e_meta[which(as.character(e_meta[,isotopic_cname]) == isotopic_notation),edata_cname]
if(length(ids_rmv) > 0){
e_meta = e_meta[-which(dplyr::pull(e_meta, edata_cname) %in% ids_rmv), ]
e_data = e_data[-which(dplyr::pull(e_data, edata_cname) %in% ids_rmv), ]
}
}
# if mf_cname is NULL and any of o_cname, n_cname, s_cname or p_cname are NULL, create columns of zeroes for them #
if (is.null(mf_cname) & is.null(element_col_names$O)) {
element_col_names$O <- tail(make.unique(c(colnames(e_meta), "O")), 1) #unique column name
e_meta[[element_col_names$O]] <- 0
}
if (is.null(mf_cname) & is.null(element_col_names$N)) {
element_col_names$N <- tail(make.unique(c(colnames(e_meta), "N")), 1) #unique column name
e_meta[[element_col_names$N]] <- 0
}
if (is.null(mf_cname) & is.null(element_col_names$S)) {
element_col_names$S <- tail(make.unique(c(colnames(e_meta), "S")), 1) #unique column name
e_meta[[element_col_names$S]] <- 0
}
if (is.null(mf_cname) & is.null(element_col_names$P)) {
element_col_names$P <- tail(make.unique(c(colnames(e_meta), "P")), 1) #unique column name
e_meta[[element_col_names$P]] <- 0
}
# store results #
res = list(e_data = e_data, f_data = f_data, e_meta = e_meta)
# set column name attributes #
attr(res, "cnames") = list(edata_cname = edata_cname, fdata_cname = fdata_cname, mass_cname = mass_cname,
extraction_cname = extraction_cname, mf_cname = mf_cname,
element_col_names = element_col_names, isotopic_cname = isotopic_cname,
ratio_cnames = ratio_cnames, kmass_cname = kmass_cname, kdefect_cname = kdefect_cname,
nosc_cname = nosc_cname, gfe_cname = gfe_cname, mfname_cname = mfname_cname,
aroma_cname = aroma_cname, modaroma_cname = modaroma_cname, dbe_cname = dbe_cname,
dbeo_cname = dbeo_cname, dbeai_cname = dbeai_cname, elcomp_cname = elcomp_cname
)
attr(res, "data_info") = list(data_scale = data_scale, instrument_type = instrument_type)
# set group dataframe attribute to NULL, will be filled in after running group_designation function #
attr(res, "group_DF") = NULL
# set class of list #
class(res) = c("peakData","ftmsData")
# check for empty rows and remove them with molfilt
if(check_rows){
molfilt <- molecule_filter(res)
if(any(molfilt$Num_Observations == 0)) res <- applyFilt(molfilt, res, min_num = 1)
}
# set filters attributes #
attr(res, "filters") = NULL
#if mf_cname is NULL and elemental columns are non-NULL, construct formulae #
if(is.null(mf_cname) & all(!sapply(element_col_names, is.null))){
res = assign_mf(res)
}
# # if mf_cname is non-NULL and any elements within element_cols are NULL, parse formulae #
if(!is.null(mf_cname) & any(sapply(element_col_names, is.null))){
res = parse_mf(res)
}
return(res)
}
#' Convert Data to compoundData Class
#'
#' Converts a list object or several data.frames of FT-MS data to an object of the class 'compoundData'. Objects of the class 'compoundData' are lists with three obligatory components \code{e_data}, \code{f_data}, and \code{e_meta}.
#'
#' @param e_data a \eqn{p \times n + 1} data.frame of expression data, where \eqn{p} is the number of observed compounds and \eqn{n} is the number of samples. Each row corresponds to data for each peak. One column specifying a unique identifier for each peak/mass (row) must be present.
#' @param f_data a data.frame with \eqn{n} rows. Each row corresponds to a sample with one column giving the unique sample identifiers found in e_data column names and other columns providing qualitative and/or quantitative traits of each sample.
#' @param e_meta a data.frame with \eqn{p} rows. Each row corresponds to a compound with one column giving a unique peak/identifier (must be named the same as the column in \code{e_data}) and other columns giving meta information. At a minimum a column giving the mass of each peak and a column giving molecular formulae or columns giving elemental counts must be present.
#' @param edata_cname character string specifying the name of the column containing a unique identifier for each peak/mass in \code{e_data} and \code{e_meta}.
#' @param fdata_cname character string specifying the name of the column containing the sample identifiers in \code{f_data}.
#' @param mass_cname character string specifying the name of the column containing the peak/mass identifiers in \code{e_meta}. Note: this is often the same as \code{edata_cname} for cases where mass is used as a unique identifier.
#' @param compound_cname character string specifying the name of the column containing the compound identifier in \code{e_meta}. This is a compound identifier related to a database (e.g. MetaCyc)
#' @param ... further arguments (see \code{\link{as.peakData}})
#'
#' @details \code{as.compoundData} constructs a compoundData object which is an ftmsData object where the rows of \code{e_data} correspond to compounds.
#' @rdname as.compoundData
as.compoundData <- function(e_data, f_data, e_meta, edata_cname, fdata_cname, mass_cname, compound_cname, ...){
res <- as.peakData(e_data, f_data, e_meta, edata_cname, fdata_cname, mass_cname, ...)
class(res) <- c("compoundData", "ftmsData")
# if e_meta is provided, check that reaction_cname is in it
if (!is.null(e_meta)) {
if(!(compound_cname %in% names(e_meta))) stop(paste("Compound column", compound_cname, " not found in e_meta.", sep = ""))
}
res <- setCompoundColName(res, compound_cname)
return(res)
}
#' Convert Data to reactionData Class
#'
#' Converts a list object or several data.frames of FT-MS data to an object of the class 'reactionData'. Objects of the class 'reactionData' are lists with two obligatory components \code{e_data} and \code{f_data} and one optional compounent, \code{e_meta}.
#'
#' @param e_data a \eqn{p \times n + 1} data.frame of expression data, where \eqn{p} is the number of observed reactions and \eqn{n} is the number of samples. Each row corresponds to data for each peak. One column specifying a unique identifier for each peak/mass (row) must be present.
#' @param f_data a data.frame with \eqn{n} rows. Each row corresponds to a sample with one column giving the unique sample identifiers found in e_data column names and other columns providing qualitative and/or quantitative traits of each sample.
#' @param e_meta a data.frame with \eqn{p} rows. Each row corresponds to a reaction with one column giving a unique reaction identifier (must be named the same as the column in \code{e_data}) and other columns giving meta information. At a minimum a column giving the mass of each peak and a column giving molecular formulae or columns giving elemental counts must be present.
#' @param edata_cname character string specifying the name of the column containing a unique identifier for each reaction in \code{e_data} and \code{e_meta}.
#' @param fdata_cname character string specifying the name of the column containing the sample identifiers in \code{f_data}.
#' @param reaction_cname character string specifying the name of the column containing the reaction identifiers in \code{e_meta}.
#' @param instrument_type character string giving the type of FT-MS instrument data was generated by. Valid options are: "12T" and "21T". Defaults to "12T". This information is used to determine appropriate plotting functions for Van Krevelen, Kendrick, etc. plots.
#' @param db character string specifying the database from which reaction information is drawn
#' @param ... further arguments
#'
#' @details \code{as.reactionData} constructs a reactionData object which is an ftmsData object where the rows of \code{e_data} correspond to reactions.
#' @rdname as.reactionData
as.reactionData <- function(e_data, f_data, e_meta = NULL, edata_cname, fdata_cname, reaction_cname, instrument_type = "12T", db=NA, ...){
# initial checks #
# check that e_data, f_data, and e_meta are data.frames #
if(!inherits(e_data, "data.frame")) stop("e_data must be of the class 'data.frame'")
if(!inherits(f_data, "data.frame")) stop("f_data must be of the class 'data.frame'")
if(!inherits(e_meta, "data.frame")) stop("e_meta must be of the class 'data.frame'")
# check that the peak column exists in e_data and e_meta (if applicable) #
if(!(edata_cname %in% names(e_data))) stop(paste("Column ", edata_cname," not found in e_data. See details of as.reactionData for specifying column names.", sep = ""))
if(!(edata_cname %in% names(e_meta))) stop(paste("Column ", edata_cname," not found in e_meta. Column names for dentifiers must match for e_data and e_meta. See details of as.reactionData for specifying column names.", sep = ""))
# check that e_data has unique rows #
if(nrow(e_data) != length(unique(dplyr::pull(e_data, edata_cname)))) stop("The 'edata_cname' identifier is non-unique.")
# check that f_data has unique rows #
if(nrow(f_data) != length(unique(dplyr::pull(f_data, fdata_cname)))) stop("The 'fdata_cname' identifier is non-unique.")
# check that the Sample column name is in f_data column names #
if(!(fdata_cname %in% names(f_data))) stop(paste("Sample column ", fdata_cname, " not found in f_data. See details of as.pepData for specifying column names.", sep = ""))
# check that all samples in e_data are present in f_data #
edat_sampid = which(names(e_data) == edata_cname)
samps.miss = sum(!(names(e_data[,-edat_sampid]) %in% dplyr::pull(f_data, fdata_cname)))
if( samps.miss > 0) stop(paste( samps.miss, " samples from e_data not found in f_data", sep = ""))
# check for any extra samples in f_data not in e_data - necessary to remove before group_designation function #
if(any(!(dplyr::pull(f_data, fdata_cname) %in% names(e_data)))){
f_data <- f_data[-which(!(dplyr::pull(f_data, fdata_cname) %in% names(e_data))),]
warning("Extra samples were found in f_data that were not in e_data. These have been removed from f_data.")
}
# check that f_data has at least 2 columns #
if(ncol(f_data) < 2) stop("f_data must contain at least 2 columns")
# if e_meta is provided, check that all peaks in e_data occur in e_meta #
if(!is.null(e_meta)){
if(sum(!(dplyr::pull(e_data, edata_cname) %in% dplyr::pull(e_meta, edata_cname))) > 0 ) stop("Not all identifiers in e_data are present in e_meta")
}
# if e_meta is provided, remove any extra features that were provided #
if(!is.null(e_meta)){
if(any(!(dplyr::pull(e_meta, edata_cname) %in% dplyr::pull(e_data, edata_cname)))){
e_meta <- e_meta[-which(!(dplyr::pull(e_meta, edata_cname) %in% dplyr::pull(e_data, edata_cname))),]
warning("Extra rows were found in e_meta that were not in e_data. These have been removed from e_meta.")
}
}
# if e_meta is provided, check that reaction_cname is in it
if (!is.null(e_meta)) {
if(!(reaction_cname %in% names(e_meta))) stop(paste("Reaction column", reaction_cname, " not found in e_meta.", sep = ""))
}
# convert the unique identifier column to character, to avoid factors #
e_data[, edata_cname] = as.character(dplyr::pull(e_data, edata_cname))
e_meta[, edata_cname] = as.character(dplyr::pull(e_meta, edata_cname))
# store results #
res = list(e_data = e_data, f_data = f_data, e_meta = e_meta)
# set column name attributes #
attr(res, "cnames") = list(edata_cname = edata_cname, fdata_cname = fdata_cname, reaction_cname=reaction_cname)
# set group dataframe attribute to NULL, will be filled in after running group_designation function #
attr(res, "group_DF") = NULL
# set filters attributes #
attr(res, "filters") = NULL
attr(res, "data_info") <- list(data_scale=NA, instrument_type=instrument_type)
attr(res, "DB") <- db
class(res) <- c("reactionData", "ftmsData")
return(res)
}
#' Convert Data to moduleData Class
#'
#' Converts a list object or several data.frames of FT-MS data to an object of the class 'moduleData'. Objects of the class 'moduleData' are lists with two obligatory components \code{e_data} and \code{f_data} and one optional component, \code{e_meta}.
#'
#' @param e_data a \eqn{p \times n + 1} data.frame of expression data, where \eqn{p} is the number of observed module nodes and \eqn{n} is the number of samples. Each row corresponds to data for each peak. One column specifying a unique identifier for each peak/mass (row) must be present.
#' @param f_data a data.frame with \eqn{n} rows. Each row corresponds to a sample with one column giving the unique sample identifiers found in e_data column names and other columns providing qualitative and/or quantitative traits of each sample.
#' @param e_meta a data.frame with \eqn{p} rows. Each row corresponds to a module node with one column giving a unique module node identifier (must be named the same as the column in \code{e_data}) and other columns giving meta information. At a minimum a column giving the mass of each peak and a column giving molecular formulae or columns giving elemental counts must be present.
#' @param edata_cname character string specifying the name of the column containing a unique identifier for each module node in \code{e_data} and \code{e_meta}.
#' @param fdata_cname character string specifying the name of the column containing the sample identifiers in \code{f_data}.
#' @param module_cname character string specifying the name of the column containing the module identifiers in \code{e_meta}.
#' @param module_node_cname character string specifying the name of the column containing the module node identifiers in \code{e_meta}
#' @param instrument_type character string giving the type of FT-MS instrument data was generated by. Valid options are: "12T" and "21T". Defaults to "12T". This information is used to determine appropriate plotting functions for Van Krevelen, Kendrick, etc. plots.
#' @param db character string specifying the database from which reaction information is drawn
#' @param ... further arguments
#'
#' @details \code{as.moduleData} constructs a moduleData object which is an ftmsData object where the rows of \code{e_data} correspond to unique module nodes.
#' @rdname as.moduleData
# @param node_label_cname character string specifying the name of the column containing the display name for each module node, in \code{e_meta}
as.moduleData <- function(e_data, f_data, e_meta = NULL, edata_cname, fdata_cname, module_cname, module_node_cname,
instrument_type = "12T", db=NA, ...){
# initial checks #
# check that e_data, f_data, and e_meta are data.frames #
if(!inherits(e_data, "data.frame")) stop("e_data must be of the class 'data.frame'")
if(!inherits(f_data, "data.frame")) stop("f_data must be of the class 'data.frame'")
if(!inherits(e_meta, "data.frame")) stop("e_meta must be of the class 'data.frame'")
# check that the peak column exists in e_data and e_meta (if applicable) #
if(!(edata_cname %in% names(e_data))) stop(paste("Column ", edata_cname," not found in e_data. See details of as.moduleData for specifying column names.", sep = ""))
if(!(edata_cname %in% names(e_meta))) stop(paste("Column ", edata_cname," not found in e_meta. Column names for identifiers must match for e_data and e_meta. See details of as.moduleData for specifying column names.", sep = ""))
# check that e_data has unique rows #
if(nrow(e_data) != length(unique(dplyr::pull(e_data, edata_cname)))) stop("The 'edata_cname' identifier is non-unique.")
# check that f_data has unique rows #
if(nrow(f_data) != length(unique(dplyr::pull(f_data, fdata_cname)))) stop("The 'fdata_cname' identifier is non-unique.")
# check that the Sample column name is in f_data column names #
if(!(fdata_cname %in% names(f_data))) stop(paste("Sample column ", fdata_cname, " not found in f_data. See details of as.pepData for specifying column names.", sep = ""))
# check that all samples in e_data are present in f_data #
edat_sampid = which(names(e_data) == edata_cname)
samps.miss = sum(!(names(e_data[,-edat_sampid]) %in% dplyr::pull(f_data, fdata_cname)))
if( samps.miss > 0) stop(paste( samps.miss, " samples from e_data not found in f_data", sep = ""))
# check for any extra samples in f_data not in e_data - necessary to remove before group_designation function #
if(any(!(dplyr::pull(f_data, fdata_cname) %in% names(e_data)))){
f_data <- f_data[-which(!(dplyr::pull(f_data, fdata_cname) %in% names(e_data))),]
warning("Extra samples were found in f_data that were not in e_data. These have been removed from f_data.")
}
# check that f_data has at least 2 columns #
if(ncol(f_data) < 2) stop("f_data must contain at least 2 columns")
# if e_meta is provided, check that all peaks in e_data occur in e_meta #
if(!is.null(e_meta)){
if(sum(!(dplyr::pull(e_data, edata_cname) %in% dplyr::pull(e_meta, edata_cname))) > 0 ) stop("Not all identifiers in e_data are present in e_meta")
}
# if e_meta is provided, remove any extra features that were provided #
if(!is.null(e_meta)){
if(any(!(dplyr::pull(e_meta, edata_cname) %in% dplyr::pull(e_data, edata_cname)))){
e_meta <- e_meta[-which(!(dplyr::pull(e_meta, edata_cname) %in% dplyr::pull(e_data, edata_cname))),]
warning("Extra rows were found in e_meta that were not in e_data. These have been removed from e_meta.")
}
}
# if e_meta is provided, check that reaction_cname is in it
if (!is.null(e_meta)) {
if(!(module_cname %in% names(e_meta))) stop(paste("Module column", module_cname, " not found in e_meta.", sep = ""))
if(!(module_node_cname %in% names(e_meta))) stop(paste("Module node column", module_node_cname, " not found in e_meta.", sep = ""))
# if(!is.null(node_label_cname))
# if (!(node_label_cname %in% names(e_meta))) stop(paste("Node label column", node_label_cname, " not found in e_meta.", sep = ""))
}
# convert the unique identifier column to character, to avoid factors #
e_data[, edata_cname] = as.character(dplyr::pull(e_data, edata_cname))
e_meta[, edata_cname] = as.character(dplyr::pull(e_meta, edata_cname))
# store results #
res = list(e_data = e_data, f_data = f_data, e_meta = e_meta)
# set column name attributes #
attr(res, "cnames") <- list(edata_cname = edata_cname, fdata_cname = fdata_cname, module_cname=module_cname, module_node_cname=module_node_cname)
# set group dataframe attribute to NULL, will be filled in after running group_designation function #
attr(res, "group_DF") = NULL
# set filters attributes #
attr(res, "filters") = NULL
attr(res, "data_info") <- list(data_scale=NA, instrument_type=instrument_type)
attr(res, "DB") <- db
class(res) <- c("moduleData", "ftmsData")
return(res)
}
EMSL-Computing/fticRanalysis documentation built on March 23, 2024, 8:36 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions as.moduleData as.reactionData as.compoundData .as.peakData as.peakData
Documented in as.compoundData as.moduleData as.peakData as.reactionData
#' Convert Data to peakData Class
#'
#' Converts a list object or several data.frames of FT-MS data to an object of the class 'peakData'. Objects of the class 'peakData' are lists with three obligatory components \code{e_data}, \code{f_data}, and \code{e_meta}.
#'
#' @param e_data a \eqn{p \times n + 1} data.frame of expression data, where \eqn{p} is the number of observed peaks and \eqn{n} is the number of samples. Each row corresponds to data for each peak. One column specifying a unique identifier for each peak/mass (row) must be present.
#' @param f_data a data.frame with \eqn{n} rows. Each row corresponds to a sample with one column giving the unique sample identifiers found in e_data column names and other columns providing qualitative and/or quantitative traits of each sample.
#' @param e_meta a data.frame with \eqn{p} rows. Each row corresponds to a peak/mass with one column giving a unique peak/identifier (must be named the same as the column in \code{e_data}) and other columns giving meta information. At a minimum a column giving the mass of each peak and a column giving molecular formulae or columns giving elemental counts must be present.
#' @param edata_cname character string specifying the name of the column containing a unique identifier for each peak/mass in \code{e_data} and \code{e_meta}.
#' @param fdata_cname character string specifying the name of the column containing the sample identifiers in \code{f_data}.
#' @param mass_cname character string specifying the name of the column containing the peak/mass identifiers in \code{e_meta}. Note: this is often the same as \code{edata_cname} for cases where mass is used as a unique identifier.
#' @param ... further arguments
#'
#' @details Objects of class 'peakData' contain some attributes that are referenced by downstream functions. These attributes must be specified (or added using available functions) to reference downstream functions for: Kendrick plots, Van Krevelen plots, and functions involving databases.
#'
#' If your data contains information about isotopic peaks (e.g. C13), you should specify the attribute \code{isotopic_cname} which gives the column in \code{e_meta} that contains an indicator of yes/no for each peak. Additionally, you must specify the attribute \code{isotopic_notation} which is a character string indicating the value in column \code{isotopic_cname} which indicates that a peak is isotopic.
#' Currently, any peaks that are isotopic are removed from the dataset, as available methods (e.g. Van Krevelen plot) are not applicable to these peaks.
#'
#' Attributes giving general information about the data object:
#' \tabular{ll}{
#' data_scale \tab character string giving the scale that the data is on. Valid options include 'log2', 'log10', 'log' (for natural log), 'pres' (for 0/1 presence/absence data), and 'abundance'. Default value is 'abundance'. \cr
#' \tab \cr
#' instrument_type \tab character string giving the type of FT-MS instrument data was generated by. Valid options are: "12T" and "21T". Defaults to "12T". This information is used to determine appropriate plotting functions for Van Krevelen, Kendrick, etc. plots. \cr
#' }
#' Attributes giving extra information in \code{f_data}:
#' \tabular{ll}{
#' extraction_cname \tab character string specifying the name of the column, in \code{f_data}, containing information as to what extraction method was used for a sample. Only necessary if \code{e_data} contains samples from multiple extraction methods. \cr
#' }
#' Attributes giving extra information in \code{e_meta}:
#' \tabular{ll}{
#' mass_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the mass information for each peak. \cr
#' \tab \cr
#' mf_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the mass (empirical) formula for a peak/mass. \cr
#' \tab \cr
#' element_col_names \tab named list of character strings specifying element/isotope column names, in \code{e_meta}, containing the respective count for each peak/mass. \cr
#' \tab \cr
#' isotopic_cname \tab character string specifying the name of the column, in \code{e_meta}, containing information about whether each peak is isotopic or not. \cr
#' \tab \cr
#' isotopic_notation \tab character string specifying the value used in column \code{isotopic_cname} which indicates that a peak is isotopic. \cr
#' \tab \cr
#' ratio_cnames \tab named list of character strings specifying the name of the column, in \code{e_meta}, containing the respective ratio of two elements or isotopes for each peak/mass. \cr
#' \tab \cr
#' kmass_cname \tab a possibly named character vector specifying the name of the columns, in \code{e_meta}, containing the Kendrick Mass for each peak/mass. Names should be any of 'CH2', 'CO2', 'H2', 'H2O', 'CHO' and correspond to the base compounds used to calculate each of the Kendrick Masses \cr
#' \tab \cr
#' kdefect_cname \tab a possibly named character vector specifying the name of the column, in \code{e_meta}, containing the Kendrick Defect for each peak/mass. Names should be any of 'CH2', 'CO2', 'H2', 'H2O', 'CHO' and correspond to the base compounds used to calculate each of the Kendrick Masses\cr
#' \tab \cr
#' nosc_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the NOSC value for each peak/mass \cr
#' \tab \cr
#' gfe_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the Gibb's Free Energy value for each peak/mass \cr
#' \tab \cr
#' mfname_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the name/description for each peak/mass \cr
#' \tab \cr
#' aroma_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the aromaticity value for each peak/mass \cr
#' \tab \cr
#' modaroma_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the modified aromaticity value for each peak/mass \cr
#' \tab \cr
#' dbe_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the double-bond equivalent values for each peak/mass \cr
#' \tab \cr
#' dbeo_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the double-bond equivalent minus oxygen value for each peak/mass \cr
#' \tab \cr
#' dbeai_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the double-bond equivalent aromaticity index value for each peak/mass \cr
#' \tab \cr
#' elcomp_cname \tab character string specifying the name of the column, in \code{e_meta}, containing the general elemental composition of each peak/mass \cr
#' \tab \cr
#' check_rows \tab logical indicating whether to remove peaks with no nonzero entries. Defaults to FALSE \cr
#' }
#'
#'@author Lisa Bramer
#'
#'@export
as.peakData <- function(e_data, f_data, e_meta, edata_cname, fdata_cname, mass_cname, ...){
.as.peakData(e_data, f_data, e_meta, edata_cname, fdata_cname, mass_cname, ...)
}
# Need to change the col names param here
.as.peakData <- function(e_data, f_data, e_meta, edata_cname, fdata_cname, mass_cname,
extraction_cname = NULL, mf_cname = NULL,
element_col_names = list("C"="C", "H"="H", "O"=NULL, "N"=NULL, "S"=NULL, "P"=NULL),
isotopic_cname = NULL, isotopic_notation = NULL,
ratio_cnames = list("O:C" = NULL, "H:C" = NULL), kmass_cname = NULL,
kdefect_cname = NULL, nosc_cname = NULL, gfe_cname = NULL, mfname_cname = NULL,
aroma_cname = NULL, modaroma_cname = NULL, dbe_cname = NULL, dbeo_cname = NULL, dbeai_cname = NULL,
elcomp_cname = NULL, instrument_type = "12T", data_scale = "abundance", check_rows = FALSE){
# check that e_data, f_data, and e_meta are data.frames #
# if(!inherits(e_data, "data.frame")) stop("e_data must be of the class 'data.frame'")
# if(!inherits(f_data, "data.frame")) stop("f_data must be of the class 'data.frame'")
# if(!inherits(e_meta, "data.frame")) stop("e_meta must be of the class 'data.frame'")
# make sure e_data, f_data, and e_meta are data.frames #
e_data <- as.data.frame(e_data)
f_data <- as.data.frame(f_data)
e_meta <- as.data.frame(e_meta)
# check that the peak column exists in e_data and e_meta (if applicable) #
if(!(edata_cname %in% names(e_data))) stop(paste("Peak/Mass column ", edata_cname," not found in e_data. See details of as.peakData for specifying column names.", sep = ""))
if(!(edata_cname %in% names(e_meta))) stop(paste("Peak/Mass column ", edata_cname," not found in e_meta. Column names for peak/mass identifiers must match for e_data and e_meta. See details of as.peakData for specifying column names.", sep = ""))
# check that e_data has unique rows #
if(nrow(e_data) != length(unique(dplyr::pull(e_data, edata_cname)))) stop("The 'edata_cname' identifier is non-unique.")
# check that f_data has unique rows #
if(nrow(f_data) != length(unique(dplyr::pull(f_data, fdata_cname)))) stop("The 'fdata_cname' identifier is non-unique.")
# check that instrument_type is a valid string #
if(!(instrument_type %in% c("21T", "12T"))) stop("Instrument type is not valid. See details of as.peakData for valid options.")
# check that mass_cname is not null #
if(is.null(mass_cname)) stop("'mass_cname' must be specified")
# check that either mf_cname or elemental cnames are not null #
if(is.null(mf_cname) & any(c(is.null(element_col_names$C), is.null(element_col_names$H)))) {
stop("Either 'mf_cname' or both 'C' and 'H' must be specified in 'element_col_names'")
}
# check that cname arguments are found #
if(!(mass_cname %in% names(e_meta))) {
stop(paste("Mass column", mass_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = ""))
}
if(!is.null(element_col_names$C)){
if(!(element_col_names$C %in% names(e_meta))) stop(paste("Carbon column", element_col_names$C, " not found in e_meta. See details of as.peakData for specifying column names.", sep = ""))
}
if(!is.null(element_col_names$H)){
if(!(element_col_names$H %in% names(e_meta))) stop(paste("Hydrogen column", element_col_names$H, " not found in e_meta. See details of as.peakData for specifying column names.", sep = ""))
}
# Check that the elements/isotopes in element_col-names are within coreMS element/isotope superset.
if(!all(names(element_col_names) %in% element_names)){
stop(paste("The following elements in 'element_col_names' are not supported: ", paste(names(element_col_names[!(names(element_col_names) %in% element_names)]), collapse = ','), ". See details of as.peakData for specifying column names.", sep = ""))
}
# Check that non-NULL elements in element_col_names are present within e_meta
for(element in names(element_col_names)){
if(!is.null(element_col_names[[element]]) && ! element %in% names(e_meta)){
stop(paste("The following element in 'element_col_names' is not found in e_meta: ", element))
}
}
if(!is.null(isotopic_cname)){
if(!(isotopic_cname %in% names(e_meta))) stop(paste("Isotopic column", isotopic_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = ""))
}
if(!is.null(ratio_cnames$o2c_cname)){
if(!(ratio_cnames$o2c_cname %in% names(e_meta))) stop(paste("Oxygen:Carbon Ratio column ", ratio_cnames$o2c_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(ratio_cnames$h2c_cname)){
if(!(ratio_cnames$h2c_cname %in% names(e_meta))) stop(paste("Hydrogen:Carbon Ratio column ", ratio_cnames$h2c_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(kmass_cname)){
if(!is.null(names(kmass_cname))){
if(!all(names(kmass_cname)) %in% c('CH2', 'CO2', 'H2', 'H2O', 'CHO', '')) stop("Names (not the values) of kmass_cname must be in 'CH2', 'CO2', 'H2', 'H2O', 'CHO'")
}
if(!all(kmass_cname %in% names(e_meta))) stop(paste("Kendrick Mass column ", kmass_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(kdefect_cname)){
if(!is.null(names(kdefect_cname))){
if(!all(names(kdefect_cname)) %in% c('CH2', 'CO2', 'H2', 'H2O', 'CHO', '')) stop("Names (not the values) of kdefect_cname must be in 'CH2', 'CO2', 'H2', 'H2O', 'CHO'")
}
if(!all(kdefect_cname %in% names(e_meta))) stop(paste("Kendrick Defect column ", kdefect_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(nosc_cname)){
if(!(nosc_cname %in% names(e_meta))) stop(paste("NOSC column ", nosc_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(gfe_cname)){
if(!(gfe_cname %in% names(e_meta))) stop(paste("Gibbs Free Energy column ", gfe_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(mfname_cname)){
if(!(mfname_cname %in% names(e_meta))) stop(paste("Molecular formula name/description column ", mfname_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(aroma_cname)){
if(!(aroma_cname %in% names(e_meta))) stop(paste("Aromaticity column ", aroma_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(modaroma_cname)){
if(!(modaroma_cname %in% names(e_meta))) stop(paste("Modified aromaticity column ", modaroma_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(mf_cname)){
if(!(mf_cname %in% names(e_meta))) stop(paste("Molecular formula column ", mf_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(dbe_cname)){
if(!(dbe_cname %in% names(e_meta))) stop(paste("Double-bond equivalent column ", dbe_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(dbeo_cname)){
if(!(dbeo_cname %in% names(e_meta))) stop(paste("Double-bond equivalent (minus Oxygen) column ", dbeo_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(dbeai_cname)){
if(!(dbeai_cname %in% names(e_meta))) stop(paste("Double-bond equivalent (aromaticity index) column ", dbeai_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(elcomp_cname)){
if(!(elcomp_cname %in% names(e_meta))) stop(paste("Elemental composition column ", elcomp_cname, " not found in e_meta. See details of as.peakData for specifying column names.", sep = "") )
}
if(!is.null(extraction_cname)){
if(!(extraction_cname %in% names(f_data))) stop(paste("Extraction Method column ", extraction_cname, " not found in f_data. See details of as.peakData for specifying column names.", sep = "") )
}
# check that the Sample column name is in f_data column names #
if(!(fdata_cname %in% names(f_data))) stop(paste("Sample column ", fdata_cname, " not found in f_data. See details of as.pepData for specifying column names.", sep = ""))
# check that all samples in e_data are present in f_data #
edat_sampid = which(names(e_data) == edata_cname)
samps.miss = sum(!(names(e_data[,-edat_sampid]) %in% dplyr::pull(f_data,fdata_cname)))
if( samps.miss > 0) stop(paste( samps.miss, " samples from e_data not found in f_data", sep = ""))
# check for any extra samples in f_data not in e_data - necessary to remove before group_designation function #
if(any(!(dplyr::pull(f_data, fdata_cname) %in% names(e_data)))){
f_data <- f_data[-which(!(dplyr::pull(f_data,fdata_cname) %in% names(e_data))),]
warning("Extra samples were found in f_data that were not in e_data. These have been removed from f_data.")
}
# check that f_data has at least 2 columns #
# if(ncol(f_data) < 2) stop("f_data must contain at least 2 columns")
# if e_meta is provided, check that all peaks in e_data occur in e_meta #
if(!is.null(e_meta)){
if(sum(!(dplyr::pull(e_data,edata_cname) %in% dplyr::pull(e_meta,edata_cname))) > 0 ) stop("Not all peaks in e_data are present in e_meta")
}
# if e_meta is provided, remove any extra features that were provided #
if(!is.null(e_meta)){
if(any(!(dplyr::pull(e_meta,edata_cname) %in% dplyr::pull(e_data,edata_cname)))){
e_meta <- e_meta[-which(!(dplyr::pull(e_meta, edata_cname) %in% dplyr::pull(e_data, edata_cname))),]
warning("Extra peaks were found in e_meta that were not in e_data. These have been removed from e_meta.")
}
}
# convert the unique identifier column to character, to avoid factors #
e_data[, edata_cname] = as.character(dplyr::pull(e_data, edata_cname))
e_meta[, edata_cname] = as.character(dplyr::pull(e_meta, edata_cname))
# check that if isotopic_cname is non-NULL then isotopic_notation is also non-NULL #
if(!is.null(isotopic_cname)){
if(is.null(isotopic_notation)) stop("You must specify 'isotopic_notation', if 'isotopic_cname' is non-NULL")
}
# if isotopic_cname is not NULL #
if(!is.null(isotopic_cname)){
# filter out peaks where isotopic_cname = TRUE #
ids_rmv = e_meta[which(as.character(e_meta[,isotopic_cname]) == isotopic_notation),edata_cname]
if(length(ids_rmv) > 0){
e_meta = e_meta[-which(dplyr::pull(e_meta, edata_cname) %in% ids_rmv), ]
e_data = e_data[-which(dplyr::pull(e_data, edata_cname) %in% ids_rmv), ]
}
}
# if mf_cname is NULL and any of o_cname, n_cname, s_cname or p_cname are NULL, create columns of zeroes for them #
if (is.null(mf_cname) & is.null(element_col_names$O)) {
element_col_names$O <- tail(make.unique(c(colnames(e_meta), "O")), 1) #unique column name
e_meta[[element_col_names$O]] <- 0
}
if (is.null(mf_cname) & is.null(element_col_names$N)) {
element_col_names$N <- tail(make.unique(c(colnames(e_meta), "N")), 1) #unique column name
e_meta[[element_col_names$N]] <- 0
}
if (is.null(mf_cname) & is.null(element_col_names$S)) {
element_col_names$S <- tail(make.unique(c(colnames(e_meta), "S")), 1) #unique column name
e_meta[[element_col_names$S]] <- 0
}
if (is.null(mf_cname) & is.null(element_col_names$P)) {
element_col_names$P <- tail(make.unique(c(colnames(e_meta), "P")), 1) #unique column name
e_meta[[element_col_names$P]] <- 0
}
# store results #
res = list(e_data = e_data, f_data = f_data, e_meta = e_meta)
# set column name attributes #
attr(res, "cnames") = list(edata_cname = edata_cname, fdata_cname = fdata_cname, mass_cname = mass_cname,
extraction_cname = extraction_cname, mf_cname = mf_cname,
element_col_names = element_col_names, isotopic_cname = isotopic_cname,
ratio_cnames = ratio_cnames, kmass_cname = kmass_cname, kdefect_cname = kdefect_cname,
nosc_cname = nosc_cname, gfe_cname = gfe_cname, mfname_cname = mfname_cname,
aroma_cname = aroma_cname, modaroma_cname = modaroma_cname, dbe_cname = dbe_cname,
dbeo_cname = dbeo_cname, dbeai_cname = dbeai_cname, elcomp_cname = elcomp_cname
)
attr(res, "data_info") = list(data_scale = data_scale, instrument_type = instrument_type)
# set group dataframe attribute to NULL, will be filled in after running group_designation function #
attr(res, "group_DF") = NULL
# set class of list #
class(res) = c("peakData","ftmsData")
# check for empty rows and remove them with molfilt
if(check_rows){
molfilt <- molecule_filter(res)
if(any(molfilt$Num_Observations == 0)) res <- applyFilt(molfilt, res, min_num = 1)
}
# set filters attributes #
attr(res, "filters") = NULL
#if mf_cname is NULL and elemental columns are non-NULL, construct formulae #
if(is.null(mf_cname) & all(!sapply(element_col_names, is.null))){
res = assign_mf(res)
}
# # if mf_cname is non-NULL and any elements within element_cols are NULL, parse formulae #
if(!is.null(mf_cname) & any(sapply(element_col_names, is.null))){
res = parse_mf(res)
}
return(res)
}
#' Convert Data to compoundData Class
#'
#' Converts a list object or several data.frames of FT-MS data to an object of the class 'compoundData'. Objects of the class 'compoundData' are lists with three obligatory components \code{e_data}, \code{f_data}, and \code{e_meta}.
#'
#' @param e_data a \eqn{p \times n + 1} data.frame of expression data, where \eqn{p} is the number of observed compounds and \eqn{n} is the number of samples. Each row corresponds to data for each peak. One column specifying a unique identifier for each peak/mass (row) must be present.
#' @param f_data a data.frame with \eqn{n} rows. Each row corresponds to a sample with one column giving the unique sample identifiers found in e_data column names and other columns providing qualitative and/or quantitative traits of each sample.
#' @param e_meta a data.frame with \eqn{p} rows. Each row corresponds to a compound with one column giving a unique peak/identifier (must be named the same as the column in \code{e_data}) and other columns giving meta information. At a minimum a column giving the mass of each peak and a column giving molecular formulae or columns giving elemental counts must be present.
#' @param edata_cname character string specifying the name of the column containing a unique identifier for each peak/mass in \code{e_data} and \code{e_meta}.
#' @param fdata_cname character string specifying the name of the column containing the sample identifiers in \code{f_data}.
#' @param mass_cname character string specifying the name of the column containing the peak/mass identifiers in \code{e_meta}. Note: this is often the same as \code{edata_cname} for cases where mass is used as a unique identifier.
#' @param compound_cname character string specifying the name of the column containing the compound identifier in \code{e_meta}. This is a compound identifier related to a database (e.g. MetaCyc)
#' @param ... further arguments (see \code{\link{as.peakData}})
#'
#' @details \code{as.compoundData} constructs a compoundData object which is an ftmsData object where the rows of \code{e_data} correspond to compounds.
#' @rdname as.compoundData
as.compoundData <- function(e_data, f_data, e_meta, edata_cname, fdata_cname, mass_cname, compound_cname, ...){
res <- as.peakData(e_data, f_data, e_meta, edata_cname, fdata_cname, mass_cname, ...)
class(res) <- c("compoundData", "ftmsData")
# if e_meta is provided, check that reaction_cname is in it
if (!is.null(e_meta)) {
if(!(compound_cname %in% names(e_meta))) stop(paste("Compound column", compound_cname, " not found in e_meta.", sep = ""))
}
res <- setCompoundColName(res, compound_cname)
return(res)
}
#' Convert Data to reactionData Class
#'
#' Converts a list object or several data.frames of FT-MS data to an object of the class 'reactionData'. Objects of the class 'reactionData' are lists with two obligatory components \code{e_data} and \code{f_data} and one optional compounent, \code{e_meta}.
#'
#' @param e_data a \eqn{p \times n + 1} data.frame of expression data, where \eqn{p} is the number of observed reactions and \eqn{n} is the number of samples. Each row corresponds to data for each peak. One column specifying a unique identifier for each peak/mass (row) must be present.
#' @param f_data a data.frame with \eqn{n} rows. Each row corresponds to a sample with one column giving the unique sample identifiers found in e_data column names and other columns providing qualitative and/or quantitative traits of each sample.
#' @param e_meta a data.frame with \eqn{p} rows. Each row corresponds to a reaction with one column giving a unique reaction identifier (must be named the same as the column in \code{e_data}) and other columns giving meta information. At a minimum a column giving the mass of each peak and a column giving molecular formulae or columns giving elemental counts must be present.
#' @param edata_cname character string specifying the name of the column containing a unique identifier for each reaction in \code{e_data} and \code{e_meta}.
#' @param fdata_cname character string specifying the name of the column containing the sample identifiers in \code{f_data}.
#' @param reaction_cname character string specifying the name of the column containing the reaction identifiers in \code{e_meta}.
#' @param instrument_type character string giving the type of FT-MS instrument data was generated by. Valid options are: "12T" and "21T". Defaults to "12T". This information is used to determine appropriate plotting functions for Van Krevelen, Kendrick, etc. plots.
#' @param db character string specifying the database from which reaction information is drawn
#' @param ... further arguments
#'
#' @details \code{as.reactionData} constructs a reactionData object which is an ftmsData object where the rows of \code{e_data} correspond to reactions.
#' @rdname as.reactionData
as.reactionData <- function(e_data, f_data, e_meta = NULL, edata_cname, fdata_cname, reaction_cname, instrument_type = "12T", db=NA, ...){
# initial checks #
# check that e_data, f_data, and e_meta are data.frames #
if(!inherits(e_data, "data.frame")) stop("e_data must be of the class 'data.frame'")
if(!inherits(f_data, "data.frame")) stop("f_data must be of the class 'data.frame'")
if(!inherits(e_meta, "data.frame")) stop("e_meta must be of the class 'data.frame'")
# check that the peak column exists in e_data and e_meta (if applicable) #
if(!(edata_cname %in% names(e_data))) stop(paste("Column ", edata_cname," not found in e_data. See details of as.reactionData for specifying column names.", sep = ""))
if(!(edata_cname %in% names(e_meta))) stop(paste("Column ", edata_cname," not found in e_meta. Column names for dentifiers must match for e_data and e_meta. See details of as.reactionData for specifying column names.", sep = ""))
# check that e_data has unique rows #
if(nrow(e_data) != length(unique(dplyr::pull(e_data, edata_cname)))) stop("The 'edata_cname' identifier is non-unique.")
# check that f_data has unique rows #
if(nrow(f_data) != length(unique(dplyr::pull(f_data, fdata_cname)))) stop("The 'fdata_cname' identifier is non-unique.")
# check that the Sample column name is in f_data column names #
if(!(fdata_cname %in% names(f_data))) stop(paste("Sample column ", fdata_cname, " not found in f_data. See details of as.pepData for specifying column names.", sep = ""))
# check that all samples in e_data are present in f_data #
edat_sampid = which(names(e_data) == edata_cname)
samps.miss = sum(!(names(e_data[,-edat_sampid]) %in% dplyr::pull(f_data, fdata_cname)))
if( samps.miss > 0) stop(paste( samps.miss, " samples from e_data not found in f_data", sep = ""))
# check for any extra samples in f_data not in e_data - necessary to remove before group_designation function #
if(any(!(dplyr::pull(f_data, fdata_cname) %in% names(e_data)))){
f_data <- f_data[-which(!(dplyr::pull(f_data, fdata_cname) %in% names(e_data))),]
warning("Extra samples were found in f_data that were not in e_data. These have been removed from f_data.")
}
# check that f_data has at least 2 columns #
if(ncol(f_data) < 2) stop("f_data must contain at least 2 columns")
# if e_meta is provided, check that all peaks in e_data occur in e_meta #
if(!is.null(e_meta)){
if(sum(!(dplyr::pull(e_data, edata_cname) %in% dplyr::pull(e_meta, edata_cname))) > 0 ) stop("Not all identifiers in e_data are present in e_meta")
}
# if e_meta is provided, remove any extra features that were provided #
if(!is.null(e_meta)){
if(any(!(dplyr::pull(e_meta, edata_cname) %in% dplyr::pull(e_data, edata_cname)))){
e_meta <- e_meta[-which(!(dplyr::pull(e_meta, edata_cname) %in% dplyr::pull(e_data, edata_cname))),]
warning("Extra rows were found in e_meta that were not in e_data. These have been removed from e_meta.")
}
}
# if e_meta is provided, check that reaction_cname is in it
if (!is.null(e_meta)) {
if(!(reaction_cname %in% names(e_meta))) stop(paste("Reaction column", reaction_cname, " not found in e_meta.", sep = ""))
}
# convert the unique identifier column to character, to avoid factors #
e_data[, edata_cname] = as.character(dplyr::pull(e_data, edata_cname))
e_meta[, edata_cname] = as.character(dplyr::pull(e_meta, edata_cname))
# store results #
res = list(e_data = e_data, f_data = f_data, e_meta = e_meta)
# set column name attributes #
attr(res, "cnames") = list(edata_cname = edata_cname, fdata_cname = fdata_cname, reaction_cname=reaction_cname)
# set group dataframe attribute to NULL, will be filled in after running group_designation function #
attr(res, "group_DF") = NULL
# set filters attributes #
attr(res, "filters") = NULL
attr(res, "data_info") <- list(data_scale=NA, instrument_type=instrument_type)
attr(res, "DB") <- db
class(res) <- c("reactionData", "ftmsData")
return(res)
}
#' Convert Data to moduleData Class
#'
#' Converts a list object or several data.frames of FT-MS data to an object of the class 'moduleData'. Objects of the class 'moduleData' are lists with two obligatory components \code{e_data} and \code{f_data} and one optional component, \code{e_meta}.
#'
#' @param e_data a \eqn{p \times n + 1} data.frame of expression data, where \eqn{p} is the number of observed module nodes and \eqn{n} is the number of samples. Each row corresponds to data for each peak. One column specifying a unique identifier for each peak/mass (row) must be present.
#' @param f_data a data.frame with \eqn{n} rows. Each row corresponds to a sample with one column giving the unique sample identifiers found in e_data column names and other columns providing qualitative and/or quantitative traits of each sample.
#' @param e_meta a data.frame with \eqn{p} rows. Each row corresponds to a module node with one column giving a unique module node identifier (must be named the same as the column in \code{e_data}) and other columns giving meta information. At a minimum a column giving the mass of each peak and a column giving molecular formulae or columns giving elemental counts must be present.
#' @param edata_cname character string specifying the name of the column containing a unique identifier for each module node in \code{e_data} and \code{e_meta}.
#' @param fdata_cname character string specifying the name of the column containing the sample identifiers in \code{f_data}.
#' @param module_cname character string specifying the name of the column containing the module identifiers in \code{e_meta}.
#' @param module_node_cname character string specifying the name of the column containing the module node identifiers in \code{e_meta}
#' @param instrument_type character string giving the type of FT-MS instrument data was generated by. Valid options are: "12T" and "21T". Defaults to "12T". This information is used to determine appropriate plotting functions for Van Krevelen, Kendrick, etc. plots.
#' @param db character string specifying the database from which reaction information is drawn
#' @param ... further arguments
#'
#' @details \code{as.moduleData} constructs a moduleData object which is an ftmsData object where the rows of \code{e_data} correspond to unique module nodes.
#' @rdname as.moduleData
# @param node_label_cname character string specifying the name of the column containing the display name for each module node, in \code{e_meta}
as.moduleData <- function(e_data, f_data, e_meta = NULL, edata_cname, fdata_cname, module_cname, module_node_cname,
instrument_type = "12T", db=NA, ...){
# initial checks #
# check that e_data, f_data, and e_meta are data.frames #
if(!inherits(e_data, "data.frame")) stop("e_data must be of the class 'data.frame'")
if(!inherits(f_data, "data.frame")) stop("f_data must be of the class 'data.frame'")
if(!inherits(e_meta, "data.frame")) stop("e_meta must be of the class 'data.frame'")
# check that the peak column exists in e_data and e_meta (if applicable) #
if(!(edata_cname %in% names(e_data))) stop(paste("Column ", edata_cname," not found in e_data. See details of as.moduleData for specifying column names.", sep = ""))
if(!(edata_cname %in% names(e_meta))) stop(paste("Column ", edata_cname," not found in e_meta. Column names for identifiers must match for e_data and e_meta. See details of as.moduleData for specifying column names.", sep = ""))
# check that e_data has unique rows #
if(nrow(e_data) != length(unique(dplyr::pull(e_data, edata_cname)))) stop("The 'edata_cname' identifier is non-unique.")
# check that f_data has unique rows #
if(nrow(f_data) != length(unique(dplyr::pull(f_data, fdata_cname)))) stop("The 'fdata_cname' identifier is non-unique.")
# check that the Sample column name is in f_data column names #
if(!(fdata_cname %in% names(f_data))) stop(paste("Sample column ", fdata_cname, " not found in f_data. See details of as.pepData for specifying column names.", sep = ""))
# check that all samples in e_data are present in f_data #
edat_sampid = which(names(e_data) == edata_cname)
samps.miss = sum(!(names(e_data[,-edat_sampid]) %in% dplyr::pull(f_data, fdata_cname)))
if( samps.miss > 0) stop(paste( samps.miss, " samples from e_data not found in f_data", sep = ""))
# check for any extra samples in f_data not in e_data - necessary to remove before group_designation function #
if(any(!(dplyr::pull(f_data, fdata_cname) %in% names(e_data)))){
f_data <- f_data[-which(!(dplyr::pull(f_data, fdata_cname) %in% names(e_data))),]
warning("Extra samples were found in f_data that were not in e_data. These have been removed from f_data.")
}
# check that f_data has at least 2 columns #
if(ncol(f_data) < 2) stop("f_data must contain at least 2 columns")
# if e_meta is provided, check that all peaks in e_data occur in e_meta #
if(!is.null(e_meta)){
if(sum(!(dplyr::pull(e_data, edata_cname) %in% dplyr::pull(e_meta, edata_cname))) > 0 ) stop("Not all identifiers in e_data are present in e_meta")
}
# if e_meta is provided, remove any extra features that were provided #
if(!is.null(e_meta)){
if(any(!(dplyr::pull(e_meta, edata_cname) %in% dplyr::pull(e_data, edata_cname)))){
e_meta <- e_meta[-which(!(dplyr::pull(e_meta, edata_cname) %in% dplyr::pull(e_data, edata_cname))),]
warning("Extra rows were found in e_meta that were not in e_data. These have been removed from e_meta.")
}
}
# if e_meta is provided, check that reaction_cname is in it
if (!is.null(e_meta)) {
if(!(module_cname %in% names(e_meta))) stop(paste("Module column", module_cname, " not found in e_meta.", sep = ""))
if(!(module_node_cname %in% names(e_meta))) stop(paste("Module node column", module_node_cname, " not found in e_meta.", sep = ""))
# if(!is.null(node_label_cname))
# if (!(node_label_cname %in% names(e_meta))) stop(paste("Node label column", node_label_cname, " not found in e_meta.", sep = ""))
}
# convert the unique identifier column to character, to avoid factors #
e_data[, edata_cname] = as.character(dplyr::pull(e_data, edata_cname))
e_meta[, edata_cname] = as.character(dplyr::pull(e_meta, edata_cname))
# store results #
res = list(e_data = e_data, f_data = f_data, e_meta = e_meta)
# set column name attributes #
attr(res, "cnames") <- list(edata_cname = edata_cname, fdata_cname = fdata_cname, module_cname=module_cname, module_node_cname=module_node_cname)
# set group dataframe attribute to NULL, will be filled in after running group_designation function #
attr(res, "group_DF") = NULL
# set filters attributes #
attr(res, "filters") = NULL
attr(res, "data_info") <- list(data_scale=NA, instrument_type=instrument_type)
attr(res, "DB") <- db
class(res) <- c("moduleData", "ftmsData")
return(res)
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.