do_groupChromPeaks_density: Core API function for peak density based chromatographic peak...

In xiaodfeng/DynamicXCMS: LC-MS and GC-MS Data Analysis

Core API function for peak density based chromatographic peak grouping

Description

The do_groupChromPeaks_density function performs chromatographic peak grouping based on the density (distribution) of peaks, found in different samples, along the retention time axis in slices of overlapping mz ranges.

Usage

do_groupChromPeaks_density(peaks, sampleGroups, bw = 30,
  minFraction = 0.5, minSamples = 1, binSize = 0.25,
  maxFeatures = 50, sleep = 0)

Arguments

peaks	A matrix or data.frame with the mz values and retention times of the identified chromatographic peaks in all samples of an experiment. Required columns are "mz", "rt" and "sample". The latter should contain numeric values representing the index of the sample in which the peak was found.
sampleGroups	A vector of the same length than samples defining the sample group assignments (i.e. which samples belong to which sample group). This parameter is mandatory for the PeakDensityParam and has to be provided also if there is no sample grouping in the experiment (in which case all samples should be assigned to the same group).
bw	numeric(1) defining the bandwidth (standard deviation ot the smoothing kernel) to be used. This argument is passed to the [density() method.
minFraction	numeric(1) defining the minimum fraction of samples in at least one sample group in which the peaks have to be present to be considered as a peak group (feature).
minSamples	numeric(1) with the minimum number of samples in at least one sample group in which the peaks have to be detected to be considered a peak group (feature).
binSize	numeric(1) defining the size of the overlapping slices in mz dimension.
maxFeatures	numeric(1) with the maximum number of peak groups to be identified in a single mz slice.
sleep	numeric(1) defining the time to sleep between iterations and plot the result from the current iteration.

Details

For overlapping slices along the mz dimension, the function calculates the density distribution of identified peaks along the retention time axis and groups peaks from the same or different samples that are close to each other. See (Smith 2006) for more details.

Value

A data.frame, each row representing a (mz-rt) feature (i.e. a peak group) with columns:

• "mzmed": median of the peaks' apex mz values.

• "mzmin": smallest mz value of all peaks' apex within the feature.

• "mzmax":largest mz value of all peaks' apex within the feature.

• "rtmed": the median of the peaks' retention times.

• "rtmin": the smallest retention time of the peaks in the group.

• "rtmax": the largest retention time of the peaks in the group.

• "npeaks": the total number of peaks assigned to the feature.

• "peakidx": a list with the indices of all peaks in a feature in the peaks input matrix.

Note that this number can be larger than the total number of samples, since multiple peaks from the same sample could be assigned to a feature.

Note

The default settings might not be appropriate for all LC/GC-MS setups, especially the bw and binSize parameter should be adjusted accordingly.

Author(s)

Colin Smith, Johannes Rainer

References

Colin A. Smith, Elizabeth J. Want, Grace O'Maille, Ruben Abagyan and Gary Siuzdak. "XCMS: Processing Mass Spectrometry Data for Metabolite Profiling Using Nonlinear Peak Alignment, Matching, and Identification" Anal. Chem. 2006, 78:779-787.

See Also

Other core peak grouping algorithms: do_groupChromPeaks_nearest, do_groupPeaks_mzClust

Examples

## Load the test data set
library(faahKO)
data(faahko)

## Extract the matrix with the identified peaks from the xcmsSet:
fts <- peaks(faahko)

## Perform the peak grouping with default settings:
res <- do_groupChromPeaks_density(fts, sampleGroups = sampclass(faahko))

## The feature definitions:
head(res$featureDefinitions)

## The assignment of peaks from the input matrix to the features
head(res$peakIndex)

xiaodfeng/DynamicXCMS documentation built on Aug. 6, 2023, 3:02 p.m.