filterPeaks-functions: Removes less frequent peaks.

In MALDIquant: Quantitative Analysis of Mass Spectrometry Data

Removes less frequent peaks.

Description

This function removes infrequently occuring peaks in a list of MassPeaks objects.

Usage

filterPeaks(l, minFrequency, minNumber, labels, mergeWhitelists=FALSE)

Arguments

l	list, list of MassPeaks objects.
minFrequency	double, remove all peaks which occur in less than minFrequency*length(l) MassPeaks objects. It is a relative threshold.
minNumber	double, remove all peaks which occur in less than minNumber MassPeaks objects. It is an absolute threshold.
labels	factor, (one for each MassPeaks object) to do groupwise filtering. The levels of the factor label define the groups. If not specified a single group is assumed.
mergeWhitelists	logical, if FALSE the filtering criteria are applied groupwise. If TRUE peaks that survive the filtering in one group (level of labels) these peaks are also kept in other groups even if their frequencies are below minFrequency.

Details

For mergeWhitelists=FALSE the filtering uses a separate peak whitelist for each group specified by labels, and is done independently in each group. For mergeWhitelists=TRUE the peak whitelists are combined, which means that peaks that occur frequently in at least one group are also kept in all other groups.

If both minFrequency and minNumber arguments are specified the more stringent threshold is used.

Value

Returns a list of filtered MassPeaks objects.

Author(s)

Sebastian Gibb [e-mail]

See Also

intensityMatrix, MassPeaks

Website: https://strimmerlab.github.io/software/maldiquant/

Examples

## load package
library("MALDIquant")

## create four MassPeaks objects and add them to the list
p <- list(createMassPeaks(mass=1:2, intensity=1:2),
          createMassPeaks(mass=1:3, intensity=1:3),
          createMassPeaks(mass=1:4, intensity=1:4),
          createMassPeaks(mass=1:5, intensity=1:5))

## only keep peaks which occur in all MassPeaks objects
filteredPeaks <- filterPeaks(p, minFrequency=1)

## compare result
intensities <- intensityMatrix(filteredPeaks)

## peaks at mass 3,4,5 are removed
all(dim(intensities) == c(4, 2)) # TRUE
all(intensities[,1] == 1)        # TRUE
all(intensities[,2] == 2)        # TRUE

## only keep peaks which occur in all MassPeaks objects in a group
## (e.g. useful for technical replicates)
groups <- factor(c("a", "a", "b", "b"), levels=c("a", "b"))
filteredPeaks <- filterPeaks(p, minFrequency=1, labels=groups)

## peaks at mass 3 were removed in group "a"
filteredPeaks[groups == "a"]

## peaks at mass 5 were removed in group "b"
filteredPeaks[groups == "b"]

## only keep peaks which occur at least twice in a group
groups <- factor(c("a", "a", "b", "b", "b"), levels=c("a", "b"))
filteredPeaks <- filterPeaks(c(p, p[[3]]), minNumber=2, labels=groups)

## peaks at mass 3 were removed in group "a"
filteredPeaks[groups == "a"]

## peaks at mass 5 were removed in group "b"
filteredPeaks[groups == "b"]

## apply different minFrequency arguments to each group
groups <- factor(c("a", "a", "b", "b", "b"), levels=c("a", "b"))
filteredPeaks <- filterPeaks(c(p, p[[3]]), minFrequency=c(1, 2/3), labels=groups)
intensityMatrix(filteredPeaks)
#     1 2  3  4
#[1,] 1 2 NA NA
#[2,] 1 2 NA NA
#[3,] 1 2  3  4
#[4,] 1 2  3  4
#[4,] 1 2  3  4


## demonstrate the use of mergeWhitelists
groups <- factor(c("a", "a", "b", "b"), levels=c("a", "b"))

## default behaviour
filteredPeaks <- filterPeaks(p, minNumber=2, labels=groups)
intensityMatrix(filteredPeaks)
#     1 2  3  4
#[1,] 1 2 NA NA
#[2,] 1 2 NA NA
#[3,] 1 2  3  4
#[4,] 1 2  3  4

## use mergeWhitelists=TRUE to keep peaks of group "a" that match all filtering
## criteria in group "b"
## (please note that mass == 3 is not removed in the second MassPeaks object)
filteredPeaks <- filterPeaks(p, minNumber=2, labels=groups,
                             mergeWhitelists=TRUE)
intensityMatrix(filteredPeaks)
#     1 2  3  4
#[1,] 1 2 NA NA
#[2,] 1 2  3 NA
#[3,] 1 2  3  4
#[4,] 1 2  3  4

MALDIquant documentation built on Sept. 11, 2024, 6:15 p.m.