feature-grouping: Compounding of LC-MS features

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

Compounding of LC-MS features

Description

Feature compounding aims at identifying and grouping LC-MS features representing different ions or adducts (including isotopes) of the same originating compound. The MsFeatures package provides a general framework and functionality to group features based on different properties. The groupFeatures methods for XcmsExperiment() or XCMSnExp objects implemented in xcms extend these to enable the compounding of LC-MS data considering also e.g. feature peak shaped. Note that these functions simply define feature groups but don't actually aggregate or combine the features.

See MsFeatures::groupFeatures() for an overview on the general feature grouping concept as well as details on the individual settings and parameters.

The available options for groupFeatures on xcms preprocessing results (i.e. on XcmsExperiment or XCMSnExp objects after correspondence analysis with groupChromPeaks()) are:

• Grouping by similar retention times: groupFeatures-similar-rtime().

• Grouping by similar feature values across samples: AbundanceSimilarityParam().

• Grouping by similar peak shape of extracted ion chromatograms: EicSimilarityParam().

An ideal workflow grouping features should sequentially perform the above methods (in the listed order).

Compounded feature groups can be accessed with the featureGroups function.

Usage

## S4 method for signature 'XcmsResult'
featureGroups(object)

## S4 replacement method for signature 'XcmsResult'
featureGroups(object) <- value

Arguments

object	an XcmsExperiment() or XCMSnExp() object with LC-MS pre-processing results.
value	for ⁠featureGroups<-⁠: replacement for the feature groups in object. Has to be of length 1 or length equal to the number of features in object.

Author(s)

Johannes Rainer, Mar Garcia-Aloy, Vinicius Veri Hernandes

See Also

plotFeatureGroups() for visualization of grouped features.

sneumann/xcms documentation built on Nov. 3, 2024, 10:33 p.m.