MOFA (version 1.3.1)

MOFAmodel: Class to store a Multi-Omics Factor Analysis (MOFA) model

Description

The MOFAmodel is an S4 class used to store all relevant data to analyse a MOFA model.

Slots

InputData

the input data before being parsed to Training Data. Either a MultiAssayExperiment object or a list of matrices, one per view.

TrainData

the parsed data used to fit the MOFA model A list with one matrix per view.

ImputedData

the parsed data with the missing values imputed using the MOFA model. A list with one matrix per view.

Expectations

expected values of the different variables of the model. A list of matrices, one per variable. The most relevant are "W" for weights and "Z" for factors.

TrainStats

list with training statistics such as evidence lower bound (ELBO), number of active factors, etc.

DataOptions

list with the data processing options such as whether to center or scale the data.

TrainOptions

list with the training options such as maximum number of iterations, tolerance for convergence, etc.

ModelOptions

list with the model options such as likelihoods, number of factors, etc.

FeatureIntercepts

list with the feature-wise intercepts. Only used internally.

Dimensions

list with the relevant dimensionalities of the model. N for the number of samples, M for the number of views, D for the number of features of each view and K for the number of infered latent factors.

Status

Auxiliary variable indicating whether the model has been trained.