R/organize_libraries_MS2.R

In QizhiSu/mspcompiler: Compile Mass Spectral Libraries from Various Sources

#' Remove retention time for MS2 libraries
#'
#' \code{remove_rt} offers a way to remove all retention time for MS2 library.
#'
#' This function supports parallel computing.
#'
#' @param lib The \code{list} generated by \code{read_lib}.
#'
#' @return A \code{list} without retention time
#' @export
#'
#' @import future.apply
remove_rt <- function(lib) {
  future.apply::future_lapply(lib, function(x) {
    x$RetentionTime <- NA

    return(x)
  })
}


#' Separate positive and negative modes in MS2 library
#'
#' \code{separate_polarity} offers a way to separate a MS2 library based on
#' polarity.
#'
#' Some libraries, e.g., NIST and GNPS have both positive and negative MS2
#' records mixed in a singled file. However, in practice, a MS2 library should
#' be either positive or negative based on the polarity used in the experiment.
#' Therefore, this function provides a way to separate positive and negative
#' modes in MS2 library.
#'
#' @param lib A MS2 library mixed with positive and negative modes.
#' @param polarity The polarity, can be either "pos" or "neg"
#'
#' @return A \code{list}, being positive or negative
#' @export
#'
#' @import rlist
#' @importFrom rlang .data
separate_polarity <- function(lib, polarity) {
  if (polarity == "pos") {
    tmp <-
      rlist::list.filter(
        lib, grepl("P", .data$IonMode, ignore.case = TRUE)
      )
  } else {
    tmp <-
      rlist::list.filter(
        lib, grepl("N", .data$IonMode, ignore.case = TRUE)
      )
  }

  return(tmp)
}


#' Add formula to the gnps library and remove wrong SMILES
#'
#' \code{complete_gnps} offers a way to complete the molecular formula filed in
#' the mgf file.
#'
#' The mgf file downloaded from GNPS has no molecular formula (MF). Therefore, this
#' function tries to calculate the MF from the SMILES (if it exists).
#'
#' @param lib The \code{list} generated by \code{read_lib} from mgf library.
#'
#' @return A \code{list} with molecular formula assigned
#' @export
#'
#' @import future.apply
#' @importFrom ChemmineR smiles2sdf MF
#' @importFrom webchem is.smiles
#'
#' @rawNamespace import(ChemmineR, except = c(groups, view))
complete_gnps <- function(lib) {
  future.apply::future_lapply(lib, function(x) {
    x$Smiles <- ifelse(webchem::is.smiles(x$Smiles), x$Smiles, NA)
    x$Formula <-
      tryCatch(ChemmineR::MF(ChemmineR::smiles2sdf(x$Smiles), addH = TRUE)[[1]],
        error = function(e) NULL
      )

    return(x)
  })
}