peaklist-class: "Peaklist"

In IPPD: Isotopic peak pattern deconvolution for Protein Mass Spectrometry by template matching

Description

Object returned from a call to getPeaklist.

Slots

peaklist

A matrix with rows equal to the number of peak patterns used in fitting and five columns. For each row (pattern), the first column contains the initial position of the peak pattern (loc_init, in Thomson), the second one contains the position of the most intense peak within a peak pattern (loc_most_intense, in Thomson), the third one contains the charge state (charge), the fourth one the quantification of the overall intensity of the pattern (quant) and the last one contains the amplitude at the most intense peak (amplitude).

peaklistprocessed

The result after applying the postprocessing procedure to the raw list peaklist. It is a matrix of the same structure as peaklist, with two up to four additional columns: the fifth column contains the local noise level at the position, and the last one contains the ratio of the amplitude to local noise level (signal-to-noise ratio), hence quantitatively describing the significance of the peak pattern. If goodnessoffit has been set to TRUE, there are two additional columns labelled goodness_of_fit and ratio_adj. The column goodness_of_fit contains a local assessment of the goodness-of-fit at the respective positions in the spectrum, while the column ratio_adj contains a goodness-of-fit adjusted signal-to-noise ratio. If postprocessed has been set to FALSE when calling getPeaklist, this slot is an empty matrix.

model

One of "Gaussian" or "EMG".

averagine.table

Averagine table used, cf.getPeaklist.

loss

One of "L2" or "L1".

alpha

Function for the parameter "alpha" (model = "EMG").

sigma

Function for the parameter "sigma" (both models).

mu

Function for the parameter "mu" (model = "EMG").

charges

A numeric vector containing the set of charge states used.

basis

The matrix of basis functions if returnbasis has been set to TRUE when calling getPeaklist. Otherwise, this slot is an empty matrix.

book

A matrix annotating each column of basis. Each row corresponds to a template/peak pattern characterized by the initial position of the peak pattern (column initial), the position of the most intense peak (column most_intense) and the charge state (column charge).

beta

A numeric vector storing the amplitudes for each peak pattern in book/each column of basis, obtained by non-negative least squares estimation (if loss = "L2") or by non-negative least absolute deviation estimation (if loss = "L1"). Hence, together with book, cbind(book, beta) constitutes a predecessor of peaklist.

locnoise

A matrix of local quantiles of the intensities computed along the spectrum. Each row corresponds to a m/z value of the spectrum (ordered increasingly), and each column to a quantile as indicated by the column names. For instance, slot(object, "locnoise")[,"0.5"] would give the vector of local medians.

noiselevel

A numeric vector storing the noise level used for computation of the signal-to-noise ratios appearing in peaklistprocessed.

goodnessoffit

A numeric vector storing the local goodness-of-fit criterion. An empty vector unless control.postprocessing$goodnessfittrue was TRUE when getPeaklist was called.

data

A list storing the spectrum used to generate the object.

Methods

show

Use show(object) for brief information about the object.

threshold

Only applicable if postprocessed has been set to TRUE when calling getPeaklist. A function to threshold peaklistprocessed, yielding a peaklist maintaining only those peaks exceeding a specified threshold. Optionally, re-fitting of the spectrum using only those templates that have passed the thresholding step can be performed. The argument list is given by

object

An object of class peaklist.

threshold

The value to be used as threshold.

ratio

Whether to use the signal-to-noise ratio "ratio" or the goodness-of-fit adjusted signal-to-noise ratio "ratio_adj"

refit

Whether re-fitting of the spectrum should be done, using only the templates above the threshold.

trace

A logical indicating whether information tracing the different steps of the fitting process should be displayed. Only used if refit = TRUE.

eps

Function values below eps are set equal to precisely zero in order to make the basis function matrix sparse. Only used if refit = TRUE.

.

visualize

A function to displaying the results of the fitting procedure, in particular significant peak patterns, in selected m/z ranges. The description of the arguments is as follows:

object

An object of class peaklist.

mz

The vector of m/z values used to obtain object.

intensities

The vector of intensities used to obtain object.

lower, upper

The m/z range for which the result of the fitting procedure should be visualized (lower < upper). Hint: the difference upper - lower should be between 0-30. If it is significantly larger, computation times increase considerably. Moreover, this will lead to a reduced quality of the plot.

truth

May be set to TRUE if one has a precise knowledge of the true underlying signal (normally only occuring for simulated data) and one wants to compare the results with the gold standard. If truth = TRUE, booktrue (s. below) has to be specified. In this case, the true signal is displayed as upper panel in the plot.

signal

A logical indicating whether the raw spectrum in the range [lower,upper] should be plotted. If TRUE, it is displayed as upper panel or as one of the middle panels.

fitted

A logical indicating whether basically all (with restrictions, s. cutoff.functions and cutoff.eps below) fitted templates should be plotted. If TRUE, the result is displayed as a middle panel or as lower panel.

postprocessed

A logical indicating whether fitted templates after postprocessing should be plotted. If TRUE, the result is displayed as lower panel.

fittedfunction

A logical indicating whether the fitted function should be drawn as lines in the panel displaying the raw spectrum (requires signal = TRUE). Note that the fitted function is the result of summing up all fitted templates multiplied by their amplitude. Setting fittedfunction = TRUE requires that object contains the matrix of basis functions.

fittedfunction.cut

A logical indicating whether the fitted function after cutting peak patterns falling below a prespecified cutoff should be drawn as lines in the panel displaying the raw spectrum (requires signal = TRUE). Setting fittedfunction.cut = TRUE requires that object contains the matrix of basis functions.

quantile

Optional argument. If quantile is specified, it has to be one of the values 0.1,...,0.9, and the corresponding local quantile as contained in the slot locnoise (s. above) will be added to the plot as dotted line.

booktrue

Required if and only if truth = TRUE. In this case, booktrue has to be a matrix having the same structure as cbind(slot(object, "book"), slot(object, "beta")), i.e. the first two columns of booktrue contain initial- and most intense peak locations within a peak pattern, the third column contains the charge state and the fourth one the amplitude of the most intense peak. Note: the 'truth' is computed according to slot(object, model). This excludes the following scenario: assuming that the true peak shapes is "EMG", but the chosen model is "Gaussian". Then the true model in this function is computed according to the latter model.

cutoff.eps, cutoff.functions

Control arguments reducing the number of templates to be plotted if fitted = TRUE. Specifying cutoff.eps removes all templates with intensity smaller than cutoff.eps. Specifying cutoff.functions as a positive integer only plots the cutoff 'most intense' templates.

...

Additional options passed to plot.

Note that the colours in the plot have no meaning. They are only used to distinguish between different patterns within a panel. Further note that the colour of a pattern appearing in several panels may vary from panel to panel.

IPPD documentation built on April 28, 2020, 6:58 p.m.