PDMBySvmWeightVector: Compute phenotypic distance matrix by SVM weight vector

Description

This function performs an SVM classification between a given sample and the negative control, calculates the weight vector, and then computes the phenotypic distance matrix based on the weight vectors.

Usage

PDMBySvmWeightVector(x, unames, neg='rluc', selectedCellFtrs, distMethod=c('manhattan','euclidean', 'correlation','mahalanobis'), verbose=FALSE, kernel='linear', ...)

Arguments

An imageHTS object.

unames

A character vector, containing the well names from where to collect the cell features. See getUnames for details.

neg

A character string to identify the negative controls.

selectedCellFtrs

A character vector for cell features to be used in the calculation. If missing, all features are used.

distMethod

A character string indicating which distance method should be used. This must be (an abbreviation of) one of the strings 'manhattan', 'euclidean', 'correlation' or 'mahalanobis'.

verbose

A logical scalar indicating whether progress should be reported.

kernel

The kernel argument for the svm function of the e1071 package.

...

Additional arguments to be passed to the svm function of the e1071 package.

Value

A symmetric distance matrix with dimensions equaling to the length of unames.

Details

For each well, this function collects features of all cells from the well and all cells from the negative control wells, and performs a bi-class classification using Support Vector Machine (SVM). The classification weight vectors are calculated for all wells passed to PDMByWellAvg to compute the phenotypic distance matrix.

Examples

Run this code


  library('phenoDist')

  ## load the imageHTS object
  load(system.file('kimorph', 'kimorph.rda', package='phenoDist'))
  x@localPath <- file.path(tempdir(), 'kimorph')

  ## calculate pair-wise svm distance matrix
  load(system.file('kimorph', 'selectedFtrs.rda', package='phenoDist'))
  pdm <- PDMBySvmWeightVector(x, unames=getUnames(x,plate=1, row=2:3, col=3), neg='rluc', selectedCellFtrs=selectedCellFtrs, distMethod='euclidean', verbose=FALSE, cost=1, kernel='linear')