makePeptideTable: Create peptide table from evidence data

In bartongroup/Proteus: Downstream analysis of the MaxQuant output

Create peptide table from evidence data

Description

makePeptideTable computes a peptide table and related data. Peptide table is a matrix with columns corresponding to conditions and rows corresponding to peptide sequences.

Usage

makePeptideTable(
  evi,
  meta,
  sequence.col = c("sequence", "modified_sequence"),
  protein.col = c("protein", "protein_group"),
  measure.cols = measureColumns,
  aggregate.fun = aggregateSum,
  ...,
  experiment.type = c("label-free", "TMT", "SILAC"),
  ncores = 4
)

Arguments

evi	Evidence table created with readEvidenceFile.
meta	Data frame with metadata. As a minimum, it should contain "sample" and "condition" columns.
sequence.col	A column name to identify peptides. Can be either "sequence" or "modified_sequence".
protein.col	A column name to use for peptide-to-protein reference. Can be either "protein" or "protein_group".
measure.cols	A named list of measure columns; should be the same as used in readEvidenceFile
aggregate.fun	A function to aggregate pepetides with the same sequence/sample.
...	Additional parameters passed to the aggregate function
experiment.type	Type of the experiment, "label-free", "TMT" or "SILAC".
ncores	Number of cores for parallel processing

Details

The evidence file contains a column called "Experiment" and one or more columns with measure values. In case of label-free experiment there is only one measure column: "Intensity". In case of TMT experiment there are several measure columns, usually called "Reporter intensity 0", "Reporter intensity 1", and so on. makePeptideTable will combine "Experiment" and measure columns in a way defined by the metadata (parameter meta). The name of the combined column will be named using "sample" column in metadata.

The result is a proteusData object containing a table with rows corresponding to peptides and columns corresponding to samples (as defined in metadata). Each cell of the table is an aggregated measure values over all evidence entries corresponding to the given sequence and experiment. How these measure values are aggregated is controlled by the parameter aggregate.fun.

There are two aggregation functions provided in this package: aggregateMedian and aggregateSum. Depending on the needs, the user can provide any arbitrary function to perform aggregation. We recommend using aggregateSum for label-free and TMT experiments and aggregateMedian for SILAC experiments.

The aggregate function should be in form: function(wp, ...). wp is a matrix containing measurements from for a given peptide. Rows are evidence file entries, columns are samples. ... are additional parameters for the function that are passed from makePeptideTable. The aggregate function should return a vector of values for each sample. That is the length of the vector should be the same as the number of columns in wp. For example, the default aggregate function aggregateSum calculates sums in each column of wp.

This function, apart from aggregating evidence data into peptides, builds a peptide-to-protein reference and stores is in the returned object ($pep2prot field). It can be used at peptide level, to identify which protein peptides belong to. It is then passed to makeProteinTable function and used to aggregate peptides to proteins. Here we can decide how peptides are aggregated to proteins. This is controlled by the protein.col parameter, which indicates which evidence column should be used to build peptide-to-protein relation. If "protein" is used (default) the peptite-to-protein relation is build on leading razor proteins. If "protein_group" is used, then peptide-to-protein relation is based on protein groups.

Only samples from metadata are used, regardless of the content of the evidence data. This makes selection of samples for downstream processing easy: select only required rows in the metadata data frame.

Value

A proteusData object, containing peptide intensities and metadata.

Examples

## Not run: 
library(proteusLabelFree)
data(proteusLabelFree)
pepdat <- makePeptideTable(evi, meta, ncores=2)

## End(Not run)

bartongroup/Proteus documentation built on April 22, 2023, 5:33 a.m.