% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/IntegratedLearner.R
\name{IntegratedLearner}
\alias{IntegratedLearner}
\title{Integrated machine learning for multi-omics prediction and classification}
\usage{
IntegratedLearner(
  feature_table,
  sample_metadata,
  feature_metadata,
  feature_table_valid = NULL,
  sample_metadata_valid = NULL,
  folds = 5,
  seed = 1234,
  base_learner = "SL.BART",
  base_screener = "All",
  meta_learner = "SL.nnls.auc",
  run_concat = TRUE,
  run_stacked = TRUE,
  verbose = FALSE,
  print_learner = TRUE,
  refit.stack = FALSE,
  family = gaussian(),
  ...
)
}
\arguments{
\item{feature_table}{An R data frame containing multiview features (in rows) and samples (in columns). 
Column names of \code{feature_metadata} must match the row names of \code{sample_metadata}.}

\item{sample_metadata}{An R data frame of metadata variables (in columns). 
Must have a column named \code{subjectID} describing per-subject unique identifiers. 
For longitudinal designs, this variable is expected to have non-unique values. 
Additionally, a column named \code{Y} must be present which is the outcome of interest (can be binary or continuous). 
Row names of \code{sample_metadata} must match the column names of \code{feature_table}.}

\item{feature_metadata}{An R data frame of feature-specific metadata across views (in columns) and features (in rows).
Must have a column named \code{featureID} describing per-feature unique identifiers. 
Additionally, a column named \code{featureType} should describe the corresponding source layers.
Row names of \code{feature_metadata} must match the row names of \code{feature_table}.}

\item{feature_table_valid}{Feature table from validation set for which prediction is desired. 
Must have the exact same structure as \code{feature_table}. If missing, uses \code{feature_table} for \code{feature_table_valid}.}

\item{sample_metadata_valid}{Sample-specific metadata table from independent validation set when available. 
Must have the exact same structure as \code{sample_metadata}.}

\item{folds}{How many folds in the V-fold nested cross-validation? Default is 10.}

\item{seed}{Specify the arbitrary seed value for reproducibility. Default is 1234.}

\item{base_learner}{Base learner for late fusion and early fusion. 
Check out the \href{https://cran.r-project.org/web/packages/SuperLearner/vignettes/Guide-to-SuperLearner.html}{SuperLearner user manual} for all available options. Default is \code{`SL.BART`}.}

\item{base_screener}{Whether to screen variables before fitting base models? \code{All} means no screening which is the default.
Check out the \href{https://cran.r-project.org/web/packages/SuperLearner/vignettes/Guide-to-SuperLearner.html}{SuperLearner user manual} for all available options.}

\item{meta_learner}{Meta-learner for late fusion (stacked generalization). Defaults to \code{`SL.nnls.auc`}.
Check out the \href{https://cran.r-project.org/web/packages/SuperLearner/vignettes/Guide-to-SuperLearner.html}{SuperLearner user manual} for all available options.}

\item{run_concat}{Should early fusion be run? Default is TRUE. Uses the specified \code{base_learner} as the learning algorithm.}

\item{run_stacked}{Should stacked model (late fusion) be run? Default is TRUE.}

\item{verbose}{logical; TRUE for \code{SuperLearner} printing progress (helpful for debugging). Default is FALSE.}

\item{print_learner}{logical; Should a detailed summary be printed? Default is TRUE.}

\item{refit.stack}{logical; For late fusion, post-refit predictions on the entire data is returned if specified. Default is FALSE.}

\item{family}{Currently allows \code{`gaussian()`} for continuous or \code{`binomial()`} for binary outcomes.}

\item{...}{Additional arguments. Not used currently.}
}
\value{
A \code{SuperLearner} object containing the trained model fits.
}
\description{
Performs integrated machine learning to predict a binary or continuous outcome based on two or more omics layers (views). 
The \code{IntegratedLearner} function takes a training set (Y, X1, X2,...,Xn) and returns the predicted values based on a validation set.
It also performs V-fold nested cross-validation to estimate the prediction accuracy of various fusion algorithms. 
Three types of integration paradigms are supported: early, late, and intermediate. 
The software includes multiple ML models based on the \code{\link[SuperLearner]{SuperLearner}} R package as well as several data exploration capabilities and visualization modules in a unified estimation framework.
}
\author{
Himel Mallick, \email{him4004@med.cornell.edu}
}
\keyword{metagenomics,}
\keyword{microbiome,}
\keyword{multiomics,}
\keyword{scRNASeq,}
\keyword{singlecell}
\keyword{tweedie,}