peaksAlignment: Data Structure for pairwise alignment of 2 GCMS samples

In flagme: Analysis of Metabolomics GC/MS Data

Description

Store the raw data and optionally, information regarding signal peaks for a number of GCMS runs

Usage

peaksAlignment(d1, d2, t1, t2, gap=0.5, D=50, timedf=NULL, df=30,
               verbose=TRUE, usePeaks=TRUE, compress=TRUE, metric=2,
               type=2, penality=0.2)

Arguments

d1	matrix of MS intensities for 1st sample (if doing a peak alignment, this contains peak apexes/areas; if doing a profile alignment, this contains scan intensities. Rows are m/z bins, columns are peaks/scans.
d2	matrix of MS intensities for 2nd sample
t1	vector of retention times for 1st sample
t2	vector of retention times for 2nd sample
gap	gap penalty for dynamic programming algorithm. Not used if type=2
D	time window (on same scale as retention time differences, t1 and t2. Default scale is seconds.)

timedf	list (length = the number of pairwise alignments) of matrices giving the expected time differences expected at each pair of peaks used with usePeaks=TRUE.
df	integer, how far from the diagonal to go to calculate the similarity of peaks. Smaller value should run faster, but be careful not to choose too low.
verbose	logical, whether to print out info.
usePeaks	logical, TRUE uses peakdata list, FALSE uses rawdata list for computing similarity.
compress	logical, whether to compress the similarity matrix into a sparse format.
metric	numeric, different algorithm to calculate the similarity matrix between two mass spectrum. metric=1 call normDotProduct(); metric=2 call ndpRT(); metric=3 call corPrt()
type	numeric, two different type of alignment function
penality	penalization applied to the matching between two mass spectra if (t1-t2)>D

Details

peaksAlignment is a hold-all data structure of the raw and peak detection data.

Value

peaksAlignment object

Author(s)

Mark Robinson, Riccardo Romoli

References

Mark D Robinson (2008). Methods for the analysis of gas chromatography - mass spectrometry data PhD dissertation University of Melbourne.

See Also

peaksDataset, clusterAlignment

Examples

## see clusterAlignment, it calls peaksAlignment

## Not Run:
gcmsPath <- paste(find.package("gcspikelite"), "data", sep="/")
cdfFiles <- dir(gcmsPath,"CDF", full=TRUE)

# read data, peak detection results
pd <- peaksDataset(cdfFiles[1:3], mz=seq(50,550), rtrange=c(7.5,10.5))
pd <- addXCMSPeaks(files=cdfFiles[1:3], object=pd, peakPicking=c('mF'),
                   snthresh=3, fwhm=10,  step=0.1, steps=2, mzdiff=0.5,
                   sleep=0)
## review peak picking
plot(pd, rtrange=c(7.5, 10.5), runs=c(1:3))

## align two chromatogram
pA <- peaksAlignment(pd@peaksdata[[1]], pd@peaksdata[[2]],
                     pd@peaksrt[[1]], pd@peaksrt[[2]], D=50,
                     metric=3, compress=FALSE, type=2, penality=0.2)

plot(pA)
pA@v$match

par(mfrow=c(2,1))
plot(pd@peaksdata[[1]][,15], type='h', main=paste(pd@peaksrt[[1]][[15]]))
plot(pd@peaksdata[[2]][,17], type='h',
     main=paste(pd@peaksrt[[2]][[17]]))
## End (Not Run)

flagme documentation built on Nov. 8, 2020, 5:24 p.m.