R/filterMQ.R

In jchitpin/lipidMS: Annotating and barcoding the lipidome using targeted mass spectrometry

## Basic loading and filtering of MultiQuant output.
## This may or may not become a helper function in the future.
## This function will likely call other functions to filter the MultiQuant
## output like deisotoping, removing dehydrations and dimers, etc.
filterMQ <- function(nameMQ){
    # Load MultiQuant output file
    MQ <- fread(nameMQ, sep = "\t", header = T,
                na.strings = "N/A", select = c(2,4,21,56),
                strip.white = T)

    # Proper column names to be scanned in
    properNames <- c("Sample Index", "Sample Name", "Mass Info", "Retention Time")
    properNospace <- c("Sample.Index", "Sample.Name", "Mass.Info", "Retention.Time")

    # Check MultiQuant column headers; return error if they don't match
    if (any(colnames(MQ) != properNames) == TRUE){
        stop("The column names or orders do not match the following:
            \"Sample Index\",
            \"Sample Name\",
            \"Mass Info\",
            \"Retention Time\"")
    }

    # Rename columns to remove whitespace which interferes with Data.Table package
    setnames(MQ, properNospace)

    ## Do we need to do any checks to clean the data or warn the user about
    ## potential issues?
    ## Ex. Retention Time is < 5 minutes?

    ## Remove peaks that were not found (identified by a 'Retention.Time' of NA)
    MQ <- MQ[is.na(Retention.Time) == F]

    ## Split "Mass.Info" into Q1 and Q3 as numerics
    MQ <- MQ[, c("MQ.Q1", "Q3") := tstrsplit(Mass.Info, " / ")]
    MQ[, MQ.Q1 := as.numeric(MQ.Q1)]
    MQ[, Q3 := as.numeric(Q3)]

    ## Remove "Mass.Info" column and re-order columns
    MQ[,"Mass.Info":=NULL]
    setcolorder(MQ, c("Sample.Index", "MQ.Q1", "Q3", "Retention.Time"))

    ## Create an index of filtered MultiQuant peaks
    MQ[, ID := .I]

    ## Call |.cleanMQ| function
    MQ <- .isotopeMQ(MQ)

    return(MQ)
}