CPAClassification: CPA classification

In mgerault/pRolocExtra: New functions to pRoloc package

CPA classification

Description

Classification using constrained proportionate assignment method

Usage

CPAClassification(
  object,
  scores = c("prediction", "all", "none"),
  fcol = "markers",
  method = 3,
  ...
)

Arguments

object	An instance of class "MSnSet".
scores	One of "prediction", "all" or "none" to report the score for the predicted class only, for all classes or none.
fcol	The feature meta-data containing marker definitions. Default is markers.
method	An integer (1, 2 or 3) specifying wich Barzilai-Borwein steplength to use; default is 3. For more details see documentation from spg function from package BB.
...	Additional parameters passed to spg from package BB.

Details

CPA was originally described by Michel Jadot et al. in 'Accounting for protein subcellular localization: A compartmental map of the rat liver proteome.', 2017. The goal of this method is to assign probabilities to each protein of belonging to an organelle. This way, we can potentially assign multiple organelles to proteins (a protain may have multiple locations). For this, it uses the mean profile of each organelle of the train dataset and then find the best coefficients to each profile to obtain the profile from a given protein. The coefficients are constrained to sum to one and bounded between 0 and 1 (so we can interpret them as probabilities). To solve this constrained optimization problem, we use spg function from BB package, as in the article. However, in this function, any treatments on the mean profiles are done, unlike in the article. So be sure of your reference proteins and use normalized data.

Value

An instance of class "MSnSet" with CPA and CPA.scores feature variables storing the classification results and scores respectively.

Examples

library(pRolocExtra)
data(tan2009r1)
res <- CPAClassification(tan2009r1)
getPredictions(res, fcol = "CPA")
getPredictions(res, fcol = "CPA", t = 0.75)
plot2D(res, fcol = "CPA")

mgerault/pRolocExtra documentation built on Sept. 15, 2022, 9:26 a.m.