train_wbcc: Train a cell count predictor

In ljsinke/wbccPredictor: Functions to predict sample cell count proportions

Description

Creates a predictor for the cell type composition of whole blood samples from methylation information.

Usage

train_wbcc(data, covariates, cellPercentages,
  model = formula(cellPercentages ~ covariates + data), ncomp = 50,
  keep.model = FALSE, ...)

Arguments

data	matrix of beta values
covariates	matrix of covariates
cellPercentages	matrix of cell percentages
model	specification (e.g.log10(cellPercentages+1) ~ data + covariates)
ncomp	number of PLS components to be tested
keep.model	logical. Default FALSE means do not return full PLS model.
...	additional parameters for plsr()

Value

prediction model plsr object

Author(s)

miterson

ljsinke/wbccPredictor documentation built on May 15, 2019, 12:07 a.m.