identifyMajorPeaks: Identify peaks based on the ridges in 2-D CWT coefficient...

In zeehio/MassSpecWavelet: Peak Detection for Mass Spectrometry data using wavelet-based algorithms

Identify peaks based on the ridges in 2-D CWT coefficient matrix

Description

Indentify the peaks based on the ridge list (returned by getRidge()) in 2-D CWT coefficient matrix and estimated Signal to Noise Ratio (SNR)

Usage

identifyMajorPeaks(
  ms,
  ridgeList,
  wCoefs,
  scales = as.numeric(colnames(wCoefs)),
  SNR.Th = 3,
  peakScaleRange = 5,
  ridgeLength = 32,
  nearbyPeak = FALSE,
  nearbyWinSize = ifelse(nearbyPeak, 150, 100),
  winSize.noise = 500,
  SNR.method = "quantile",
  minNoiseLevel = 0.001,
  excludeBoundariesSize = nearbyWinSize/2
)

Arguments

ms	the mass spectrometry spectrum
ridgeList	returned by getRidge()
wCoefs	2-D CWT coefficients
scales	scales of CWT, by default it is the colnames of wCoefs
SNR.Th	threshold of SNR
peakScaleRange	the CWT scale range of the peak.
ridgeLength	the maximum ridge scale of the major peaks.
nearbyPeak	determine whether to include the small peaks close to large major peaks
nearbyWinSize	the window size to determine the nearby peaks. Only effective when nearbyPeak is true.
winSize.noise	the local window size to estimate the noise level.
SNR.method	method to estimate noise level. Currently, only 95 percentage quantile is supported.
minNoiseLevel	the minimum noise level used in calculating SNR, i.e., if the estimated noise level is less than "minNoiseLevel", it will use "minNoiseLevel" instead. If the noise level is less than 0.5, it will be treated as the ratio to the maximum amplitude of the spectrum.
excludeBoundariesSize	number of points at each boundary of the ms signal that will be excluded in search for peaks to avoid boundary effects.

Details

The determination of the peaks is based on three rules: Rule 1: The maximum ridge scale of the peak should larger than a certain threshold Rule 1.1: Based on the scale of the peak (corresponding to the maximum value of the peak ridge) should be within certain range Rule 2: Based on the peak SNR Rule 3: The peak should not appear at the boundaries of the signal.

Value

Return a list with following elements:

peakIndex	the m/z indexes of the identified peaks
peakCenterIndex	the m/z indexes of peak centers, which correspond to the maximum on the ridge. peakCenterIndex includes all the peaks, not just the identified major peaks.
peakCenterValue	the CWT coefficients (the maximum on the ridge) corresponding to peakCenterIndex
peakSNR	the SNR of the peak, which is the ratio of peakCenterValue and noise level
peakScale	the estimated scale of the peak, which corresponds to the peakCenerIndex
potentialPeakIndex	the m/z indexes of all potential peaks, which satisfy all requirements of a peak without considering its SNR. Useful, if you want to change to a lower SNR threshold later.
allPeakIndex	the m/z indexes of all the peaks, whose order is the same as peakCenterIndex, peakCenterValue, peakSNR and peakScale.

All of these return elements have peak names, which are the same as the corresponding peak ridges. see getRidge() for details.

Author(s)

Pan Du, Simon Lin

References

Du, P., Kibbe, W.A. and Lin, S.M. (2006) Improved peak detection in mass spectrum by incorporating continuous wavelet transform-based pattern matching, Bioinformatics, 22, 2059-2065.

See Also

peakDetectionCWT(), tuneInPeakInfo()

Examples

data(exampleMS)
scales <- seq(1, 64, 3)
wCoefs <- cwt(exampleMS, scales = scales, wavelet = "mexh")

localMax <- getLocalMaximumCWT(wCoefs)
ridgeList <- getRidge(localMax)

SNR.Th <- 3
majorPeakInfo <- identifyMajorPeaks(exampleMS, ridgeList, wCoefs, SNR.Th = SNR.Th)
## Plot the identified peaks
peakIndex <- majorPeakInfo$peakIndex
plotPeak(exampleMS, peakIndex, main = paste("Identified peaks with SNR >", SNR.Th))

zeehio/MassSpecWavelet documentation built on May 6, 2023, 1:32 a.m.