interactorDifferences: Convert Individual Features into Differences Between Binary...

In DarioS/ClassifyR: A framework for cross-validated classification problems, with applications to differential variability and differential distribution testing

Convert Individual Features into Differences Between Binary Interactors Based on Known Sub-networks

Description

This conversion is useful for creating a meta-feature table for classifier training and prediction based on sub-networks that were selected based on their differential correlation between classes.

Usage

## S4 method for signature 'matrix'
interactorDifferences(measurements, ...)

## S4 method for signature 'DataFrame'
interactorDifferences(
  measurements,
  featurePairs = NULL,
  absolute = FALSE,
  verbose = 3
)

## S4 method for signature 'MultiAssayExperiment'
interactorDifferences(measurements, useFeatures = "all", ...)

Arguments

measurements	Either a matrix, DataFrame or MultiAssayExperiment containing the training data. For a matrix, the rows are samples, and the columns are features.
...	Variables not used by the matrix nor the MultiAssayExperiment method which are passed into and used by the DataFrame method.
featurePairs	A object of type Pairs.
absolute	If TRUE, then the absolute values of the differences are returned.
verbose	Default: 3. A number between 0 and 3 for the amount of progress messages to give. This function only prints progress messages if the value is 3.
useFeatures	If measurements is a MultiAssayExperiment, "all" or a two-column table of features to use. If a table, the first column must have assay names and the second column must have feature names found for that assay. "clinical" is also a valid assay name and refers to the clinical data table.

Details

The pairs of features known to interact with each other are specified by networkSets.

Value

An object of class DataFrame with one column for each interactor pair difference and one row for each sample. Additionally, mcols(resultTable) prodvides a DataFrame with a column named "original" containing the name of the sub-network each meta-feature belongs to.

Author(s)

Dario Strbenac

References

Dynamic modularity in protein interaction networks predicts breast cancer outcome, Ian W Taylor, Rune Linding, David Warde-Farley, Yongmei Liu, Catia Pesquita, Daniel Faria, Shelley Bull, Tony Pawson, Quaid Morris and Jeffrey L Wrana, 2009, Nature Biotechnology, Volume 27 Issue 2, https://www.nature.com/articles/nbt.1522.

Examples

  pairs <- Pairs(rep(c('A', 'G'), each = 3), c('B', 'C', 'D', 'H', 'I', 'J'))
                           
  # Consistent differences for interactors of A.                                           
  measurements <- matrix(c(5.7, 10.1, 6.9, 7.7, 8.8, 9.1, 11.2, 6.4, 7.0, 5.5,
                           3.6, 7.6, 4.0, 4.4, 5.8, 6.2, 8.1, 3.7, 4.4, 2.1,
                           8.5, 13.0, 9.9, 10.0, 10.3, 11.9, 13.8, 9.9, 10.7, 8.5,
                           8.1, 10.6, 7.4, 10.7, 10.8, 11.1, 13.3, 9.7, 11.0, 9.1,
                           round(rnorm(60, 8, 0.3), 1)), nrow = 10)
                         
  rownames(measurements) <- paste("Patient", 1:10)
  colnames(measurements) <- LETTERS[1:10]
  
  interactorDifferences(measurements, pairs)

DarioS/ClassifyR documentation built on Nov. 4, 2024, 1:06 p.m.