R/detectPeaks-method.R

In wilsontom/sRm: Toolkit for SRM-MS data (MS) data

#' Detect Peaks
#'
#' @rdname detectPeaks
#' @param object a SRM object
#' @param snthresh a numeric value for the signal-to-noise threshold to use
#' @param peakwidth a numeric vector indicating the minimum and maximum tolerated peak width
#' @param parallel logical; if `TRUE` then `future_map` is used for peak detection
#' @param cores a numeric value for the number of parallel workers to enable, if `parallel = TRUE`
#'
#' @return a SRM object
#'
#' @export

setMethod('detectPeaks', signature = 'SRM',
          function(object,
                   snthresh,
                   peakwidth,
                   parallel = FALSE,
                   cores = 4)
          {
            chrom_split  <-
              object@chroms %>% dplyr::group_by(sampleID, filter) %>% dplyr::group_split()

            if (parallel == FALSE) {
              chromPeaks <-
                purrr::map(chrom_split,  ~ {
                  centWave(.$rt, .$int, snthresh = snthresh, peakwidth = peakwidth)
                })
            }

            if (parallel == TRUE) {
              future::plan(future::multisession, workers = cores)

              chromPeaks <-
                furrr::future_map(chrom_split,  ~ {
                  centWave(.$rt, .$int, snthresh = snthresh, peakwidth = peakwidth)
                })

              future::plan(future::sequential)

            }

            for (i in seq_along(chromPeaks)) {
              chromPeaks[[i]] <-
                chromPeaks[[i]] %>% dplyr::mutate(sampleID = chrom_split[[i]]$sampleID[1],
                                                  filter = chrom_split[[i]]$filter[1]) %>%
                dplyr::select(sampleID, filter, rt, rtmin, rtmax, int, area, sn, peakId)
            }

            object@peaks <-
              chromPeaks %>% dplyr::bind_rows() %>% dplyr::filter(peakId != 0) %>%
              dplyr::left_join(., object@transitions, by = 'filter')

            return(object)

          })