Train GD2 Model — trainGD2model • GD2Viz

The trainGD2model function trains a GD2 model using a specified reaction activity score (RAS) matrix from training data stored as a SummarizedExperiment object.

Usage

trainGD2model(
  train_data,
  adjust_ras = c("ras", "ras_prob", "ras_prob_rec"),
  adjust_input = c("raw", "ranged", "scaled"),
  center = TRUE,
  type_column = "X_study"
)

Arguments

train_data

A SummarizedExperiment object containing training data. It should include:

assays: "ras", "ras_prob", "ras_prob_path", "ras_prob_rec".
colData: Column data information.
metadata(train_data)$geom: A numeric vector of geometric mean data for genes.

adjust_ras

Which assay to use. One of "ras", "ras_prob", "ras_prob_path", or "ras_prob_rec".

adjust_input

How to adjust the input matrix: "raw", "ranged", or "scaled".

center

Logical; whether to mean-center data if using "scaled" adjustment.

type_column

Column name in colData indicating the sample type.

Value

An SVM model trained using kernlab::ksvm.

Examples

if (FALSE) { # \dontrun{
# Train GD2 model with raw input
model <- trainGD2model(train_data, adjust_ras = "ras", adjust_input = "raw")
} # }