pepmatch: Identify peak pairs.

In goat-anti-rabbit/labelpepmatch.R: Tools For Interpreting Mass Spectra With Labeled Peptides

Description

Identifies pairs of labelled peptides in an lpm_input object.

Usage

pepmatch(lpm_input, elutionthresh = 0.2, elutionunit = "minutes",
  labelthresh = 0.1, labelcountmax = 6, labellightmass, labelheavymass,
  label = "unspecified", minmolweight = 132, quantmin = 0, FDR = T,
  iterations = 10, cores = 1, verbose = FALSE)

Arguments

lpm_input	Input object of class "LPM_input", generated by the read functions of labelpepmatch, or example data object present in package.
elutionthresh	Numeric. Threshold for elution time difference between two peaks to be considered a peakpair. Default is 0.2 (minutes)
elutionunit	Character. Default is "minutes", alternative is "seconds". This is sketchy, only just for FDR estimation.
labelthresh	Numeric. Threshold for molecular weight difference (in Dalton) between to peaks to differ from theoretical mass difference between two labelled peptides. Regardless of number of labels. Default is 0.1.
labelcountmax	Integer. Maximal number of labels allowed. Default is 6.
labellightmass	Numeric. Mass of light label. Default is 128.1177, the mass of light TMAB
labelheavymass	Numeric. Mass of heavy label. Default is 137.1728, the mass of heavy TMAB
label	Character. Optional argument. If it is "TMAB", automatically the right values for light and heavy TMAB are used. Can be extended in the future with newer labels.
minmolweight	Numeric. Minimal molecular weight for a peptide to be retained. Default is 132, the smallest possible peptide.
quantmin	Numeric vector of one or two elements. Minimal untransformed quantity for a feature to be retained. The highest value of the vector is the cutoff for the most abundant peak in a pair, the lowest for the least abundant. If you want to make sure you pick up extreme up-or downregulations, the vector should contain one zero or very small number. If you want to discart a peak pair once one of the peaks is below a quantity threshold, you can set the threshold with one minimal value (a vector with one element). This is equivalent to a vector with two equal elements.
FDR	Logical. Generates mock data according to a restricted randomization procedure that produces mock data with a very comparable structure to the original data. Structure elements that are retained are the dispersion of features in the retention time - m/z space and the structure of decimals for m/z. This is important because decimals are non random in m/z data. For more information see lpm_mockdata.
iterations	Integer: the number of iterations to estimate the FDR. Careful: for large datasets this can take a long time! A modest number (3 to 5) of iterations will already give a good indication of FDR.
cores	Interger. Number of cores that can be used on the computer for calculation. When >1, the packages foreach and doParallel will be used for multithreading. Optimally, the number of cores is a multiple of the number of runs. Since every run makes up a single thread, it makes no sense to use more cores than the number of runs.
verbose	Logical. Gives verbose output.

Details

This is the central function of the labelpepmatch package. Candidate matching features are those features that have the same charge, that co-elute within a given time interval and that have a mass difference of an integer multiple (the number of labels) of the mass difference between two labels. However, only m/z values are given, and masses have to be calculated first (deconvolution). Note that in the case one or more labels are present, the deconvoluted mass will not be correct because an unknown number of charges originate from labels rather than from protons. However, this error is systematic and would by definition be the same for two features within a peak pair. In other words, the mass difference within a peak pair is still calculated correctly. The function follows a simple algorithm where features are first sorted based on their preliminarily deconvoluted masses and retention time. Every feature is then matched to a set of equally charged features that fall within the same retention time window. Once a peak pair is detected, the number of labels is inferred from the mass difference and the correct mass of the peptide is recalculated.

Value

An object of class pepmatched without mass matchings. This object can serve as an imput to the pep.id function, or in case of no mass matching, can go directly in make.statlist. If FDR is TRUE, it also contains a summary of the false discovery rate: the mean, median, minimal and maximal number of false positive hits per run over all iterations, along with their proportion to the presumed real positives. Finally, if FDR is TRUE, the pepmatched object contains all the mass match precisions of the false positive hits along with the actual mockdata used for FDR estimation.

Author(s)

Rik Verdonck & Gerben Menschaert

See Also

lpm_mockdata

goat-anti-rabbit/labelpepmatch.R documentation built on May 17, 2019, 7:29 a.m.