findPeaks-MSW: Single-spectrum non-chromatography MS data peak detection

In sneumann/xcms: LC-MS and GC-MS Data Analysis

Single-spectrum non-chromatography MS data peak detection

Description

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 peakDetectionCWT functions from the MassSpecWavelet package.

Usage

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'
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

Arguments

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 Savitzky-Golay filter is used (in the MassSpecWavelet) package 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 peakDetectionCWT and identifyMajorPeaks 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.

Details

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.

Value

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.

Slots

snthresh,verboseColumns,scales,nearbyPeak,peakScaleRange,ampTh,minNoiseLevel,ridgeLength,peakThr,tuneIn,addParams

See corresponding parameter above.

Note

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.

Author(s)

Joachim Kutzera, Steffen Neumann, Johannes Rainer

See Also

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: findChromPeaks(), findChromPeaks-centWave, findChromPeaks-centWaveWithPredIsoROIs, findChromPeaks-massifquant, findChromPeaks-matchedFilter

Examples

library(MSnbase)
## 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-mzML", package = "msdata"),
                    recursive = TRUE, full.names = TRUE)
fticr <- readMSData(fticrf[1], 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))

sneumann/xcms documentation built on Nov. 23, 2024, 6:53 p.m.