Predict the response variable in validation datasets.

Merge the discovery datasets with each validation dataset, train a glmnet model on the samples from the discovery datasets, then predicts the response variable for the samples in the respective validation dataset.

Usage

metapredict(
  ematList,
  studyMetadata,
  sampleMetadata,
  discoveryStudyNames,
  alpha,
  lambda,
  weights,
  batchColname = "study",
  covariateName = NA,
  className = "class",
  type = "response",
  ...
)

Arguments

ematList: Named list of expression matrices.
studyMetadata: data.frame of study metadata.
sampleMetadata: data.frame of sample metadata, with rownames corresponding to sample names.
discoveryStudyNames: vector of study names for training.
alpha: value of alpha for the elastic net mixing parameter.
lambda: value of regularization parameter.
weights: vector of weights for training the glmnet model.
batchColname: column in sampleMetadata containing batch information for sva::ComBat().
covariateName: column in sampleMetadata containing additional covariates for sva::ComBat() besides batch.
className: column in sampleMetadata containing values of the response variable.
type: type of prediction to make, passed to glmnet::predict.glmnet().
...: Other arguments passed to glmnet::glmnet().

Value

A named list of objects from glmnet::predict.glmnet().