findPeaks-MSW: Single-spectrum non-chromatography MS data peak detection
In yclement/xcms: LC-MS and GC-MS Data Analysis

Description Usage Arguments Details Value Slots Note Author(s) See Also Examples

Perform peak detection in mass spectrometry direct injection spectrum using a wavelet based algorithm.

The MSWParam class allows to specify all settings for a peak detection using the MSW method. Instances should be created with the MSWParam constructor.

The findChromPeaks,OnDiskMSnExp,MSWParam method performs peak detection in single-spectrum non-chromatography MS data using functionality from the MassSpecWavelet package on all samples from an OnDiskMSnExp object. OnDiskMSnExp objects encapsule all experiment specific data and load the spectra data (mz and intensity values) on the fly from the original files applying also all eventual data manipulations.

snthresh,snthresh<-: getter and setter for the snthresh slot of the object.

verboseColumns,verboseColumns<-: getter and setter for the verboseColumns slot of the object.

scales,scales<-: getter and setter for the scales slot of the object.

nearbyPeak,nearbyPeak<-: getter and setter for the nearbyPeak slot of the object.

peakScaleRange,peakScaleRange<-: getter and setter for the peakScaleRange slot of the object.

ampTh,ampTh<-: getter and setter for the ampTh slot of the object.

minNoiseLevel,minNoiseLevel<-: getter and setter for the minNoiseLevel slot of the object.

ridgeLength,ridgeLength<-: getter and setter for the ridgeLength slot of the object.

peakThr,peakThr<-: getter and setter for the peakThr slot of the object.

tuneIn,tuneIn<-: getter and setter for the tuneIn slot of the object.

addParams,addParams<-: getter and setter for the addParams slot of the object. This slot stores optional additional parameters to be passed to the identifyMajorPeaks and sav.gol functions from the MassSpecWavelet package.

MSWParam(
  snthresh = 3,
  verboseColumns = FALSE,
  scales = c(1, seq(2, 30, 2), seq(32, 64, 4)),
  nearbyPeak = TRUE,
  peakScaleRange = 5,
  ampTh = 0.01,
  minNoiseLevel = ampTh/snthresh,
  ridgeLength = 24,
  peakThr = NULL,
  tuneIn = FALSE,
  ...
)

## S4 method for signature 'OnDiskMSnExp,MSWParam'
findChromPeaks(
  object,
  param,
  BPPARAM = bpparam(),
  return.type = "XCMSnExp",
  msLevel = 1L
)

## S4 method for signature 'MSWParam'
show(object)

## S4 method for signature 'MSWParam'
snthresh(object)

## S4 replacement method for signature 'MSWParam'
snthresh(object) <- value

## S4 method for signature 'MSWParam'
verboseColumns(object)

## S4 replacement method for signature 'MSWParam'
verboseColumns(object) <- value

## S4 method for signature 'MSWParam'
scales(object)

## S4 replacement method for signature 'MSWParam'
scales(object) <- value

## S4 method for signature 'MSWParam'
nearbyPeak(object)

## S4 replacement method for signature 'MSWParam'
nearbyPeak(object) <- value

## S4 method for signature 'MSWParam'
peakScaleRange(object)

## S4 replacement method for signature 'MSWParam'
peakScaleRange(object) <- value

## S4 method for signature 'MSWParam'
ampTh(object)

## S4 replacement method for signature 'MSWParam'
ampTh(object) <- value

## S4 method for signature 'MSWParam'
minNoiseLevel(object)

## S4 replacement method for signature 'MSWParam'
minNoiseLevel(object) <- value

## S4 method for signature 'MSWParam'
ridgeLength(object)

## S4 replacement method for signature 'MSWParam'
ridgeLength(object) <- value

## S4 method for signature 'MSWParam'
peakThr(object)

## S4 replacement method for signature 'MSWParam'
peakThr(object) <- value

## S4 method for signature 'MSWParam'
tuneIn(object)

## S4 replacement method for signature 'MSWParam'
tuneIn(object) <- value

## S4 method for signature 'MSWParam'
addParams(object)

## S4 replacement method for signature 'MSWParam'
addParams(object) <- value

`snthresh`	`numeric(1)` defining the signal to noise ratio cutoff.
`verboseColumns`	`logical(1)` whether additional peak meta data columns should be returned.
`scales`	Numeric defining the scales of the continuous wavelet transform (CWT).
`nearbyPeak`	logical(1) whether to include nearby peaks of major peaks.
`peakScaleRange`	numeric(1) defining the scale range of the peak (larger than 5 by default).
`ampTh`	numeric(1) defining the minimum required relative amplitude of the peak (ratio of the maximum of CWT coefficients).
`minNoiseLevel`	numeric(1) defining the minimum noise level used in computing the SNR.
`ridgeLength`	numeric(1) defining the minimum highest scale of the peak in 2-D CWT coefficient matrix.
`peakThr`	numeric(1) with the minimum absolute intensity (above baseline) of peaks to be picked. If provided, the smoothing function `sav.gol` function (in the `MassSpecWavelet`) package is called to estimate the local intensity.
`tuneIn`	logical(1) whther to tune in the parameter estimation of the detected peaks.
`...`	Additional parameters to be passed to the `identifyMajorPeaks` and `sav.gol` functions from the `MassSpecWavelet` package.
`object`	For `findChromPeaks`: an `OnDiskMSnExp` object containing the MS- and all other experiment-relevant data. For all other methods: a parameter object.
`param`	An `MSWParam` object containing all settings for the algorithm.
`BPPARAM`	A parameter class specifying if and how parallel processing should be performed. It defaults to `bpparam`. See documentation of the `BiocParallel` for more details. If parallel processing is enabled, peak detection is performed in parallel on several of the input samples.
`return.type`	Character specifying what type of object the method should return. Can be either `"XCMSnExp"` (default), `"list"` or `"xcmsSet"`.
`msLevel`	`integer(1)` defining the MS level on which the peak detection should be performed. Defaults to `msLevel = 1`.
`value`	The value for the slot.

This is a wrapper for the peak picker in Bioconductor's MassSpecWavelet package calling peakDetectionCWT and tuneInPeakInfo functions. See the xcmsDirect vignette for more information.

Parallel processing (one process per sample) is supported and can be configured either by the BPPARAM parameter or by globally defining the parallel processing mode using the register method from the BiocParallel package.

The MSWParam function returns a MSWParam class instance with all of the settings specified for peak detection by the MSW method.

For findChromPeaks: if return.type = "XCMSnExp" an XCMSnExp object with the results of the peak detection. If return.type = "list" a list of length equal to the number of samples with matrices specifying the identified peaks. If return.type = "xcmsSet" an xcmsSet object with the results of the detection.

.__classVersion__,snthresh,verboseColumns,scales,nearbyPeak,peakScaleRange,ampTh,minNoiseLevel,ridgeLength,peakThr,tuneIn,addParams: See corresponding parameter above. .__classVersion__ stores the version from the class. Slots values should exclusively be accessed via the corresponding getter and setter methods listed above.

These methods and classes are part of the updated and modernized xcms user interface which will eventually replace the findPeaks methods. It supports peak detection on OnDiskMSnExp objects (defined in the MSnbase package). All of the settings to the algorithm can be passed with a MSWParam object.

Joachim Kutzera, Steffen Neumann, Johannes Rainer

The do_findPeaks_MSW core API function and findPeaks.MSW for the old user interface.

XCMSnExp for the object containing the results of the peak detection.

Other peak detection methods: chromatographic-peak-detection, findChromPeaks-centWaveWithPredIsoROIs, findChromPeaks-centWave, findChromPeaks-massifquant, findChromPeaks-matchedFilter

## Create a MSWParam object
mp <- MSWParam()
## Change snthresh parameter
snthresh(mp) <- 15
mp

## Loading a small subset of direct injection, single spectrum files
library(msdata)
fticrf <- list.files(system.file("fticr", package = "msdata"),
                    recursive = TRUE, full.names = TRUE)
fticr <- readMSData(fticrf[1:2], msLevel. = 1, mode = "onDisk")

## Perform the MSW peak detection on these:
p <- MSWParam(scales = c(1, 7), peakThr = 80000, ampTh = 0.005,
             SNR.method = "data.mean", winSize.noise = 500)
fticr <- findChromPeaks(fticr, param = p)

head(chromPeaks(fticr))