@@ -3,8 +3,8 @@ Type: Package

 Title: A framework for cross-validated classification problems, with

        applications to differential variability and differential

        distribution testing

-Version: 3.1.16

-Date: 2022-09-12

+Version: 3.1.17

+Date: 2022-09-15

 Author: Dario Strbenac, Ellis Patrick, John Ormerod, Graham Mann, Jean Yang

 Maintainer: Dario Strbenac <dario.strbenac@sydney.edu.au>

 VignetteBuilder: knitr

@@ -26,6 +26,7 @@ export(crossValidate)

 export(distribution)

 export(edgesToHubNetworks)

 export(featureSetSummary)

+export(finalModel)

 export(generateCrossValParams)

 export(generateModellingParams)

 export(interactorDifferences)

@@ -68,6 +69,7 @@ exportMethods(chosenFeatureNames)

 exportMethods(crossValidate)

 exportMethods(distribution)

 exportMethods(featureSetSummary)

+exportMethods(finalModel)

 exportMethods(interactorDifferences)

 exportMethods(length)

 exportMethods(models)

@@ -43,6 +43,8 @@

 #' @param yLabel Label to be used for the y-axis of true positive rate.

 #' @param showAUC Logical. If \code{TRUE}, the AUC value of each result is

 #' added to its legend text.

+#' @param ... Parameters not used by the \code{ClassifyResult} method but passed to

+#' the \code{list} method.

 #' @return An object of class \code{ggplot} and a plot on the current graphics

 #' device, if \code{plot} is \code{TRUE}.

 #' @author Dario Strbenac

@@ -862,8 +862,8 @@ setClassUnion("ModellingParamsOrNULL", c("ModellingParams", "NULL"))

 #' show,ClassifyResult-method sampleNames sampleNames,ClassifyResult-method

 #' predictions predictions,ClassifyResult-method actualOutcome

 #' actualOutcome,ClassifyResult-method features features,ClassifyResult-method

-#' models models,ClassifyResult-method performance

-#' performance,ClassifyResult-method tunedParameters

+#' models models,ClassifyResult-method finalModel finalModel,ClassifyResult-method

+#' performance performance,ClassifyResult-method tunedParameters

 #' tunedParameters,ClassifyResult-method totalPredictions

 #' totalPredictions,ClassifyResult-method

 #' @docType class

@@ -911,6 +911,8 @@ setClassUnion("ModellingParamsOrNULL", c("ModellingParams", "NULL"))

 #' \describe{

 #' \item{\code{models(result)}}{A \code{list} of the models fitted for each training.}}

 #' \describe{

+#' \item{\code{finalModel(result)}}{A deployable model fitted on all of the data for use on future data.}}

+#' \describe{

 #' \item{\code{chosenFeatureNames(result)}}{A \code{list} of the features selected for each training.}}

 #' \describe{

 #' \item{\code{predictions(result)}}{Returns a \code{DataFrame} which has columns with test sample,

@@ -1054,6 +1056,20 @@ setMethod("models", "ClassifyResult",

             object@models

           })

+#' @export

+#' @usage NULL

+setGeneric("finalModel", function(object, ...)

+standardGeneric("finalModel"))

+

+#' @rdname ClassifyResult-class

+#' @usage NULL

+#' @export

+setMethod("finalModel", "ClassifyResult",

+          function(object)

+          {

+            object@finalModel

+          })

+

 #' @export

 #' @usage NULL

 setGeneric("predictions", function(object, ...)

@@ -252,7 +252,7 @@ input data. Autmomatically reducing to smaller number.")

         predictedOutcome <- predictedOutcome[, na.omit(match(c("class", "risk"), colnames(predictedOutcome)))]

     performanceChanges <- round(performancesWithoutEach - calcExternalPerformance(outcomeTest, predictedOutcome, performanceType), 2)

-    if(is.null(S4Vectors::mcols(measurementsTrain)))

+    if(is.null(S4Vectors::mcols(measurementsTrain)) || !any(c("assay", "feature") %in% colnames(S4Vectors::mcols(measurementsTrain))))

     {

       selectedFeatures <- colnames(measurementsTrain)[selectedFeaturesIndices]

     } else {

@@ -272,22 +272,22 @@ input data. Autmomatically reducing to smaller number.")

   {

     if(!is.null(rankedFeaturesIndices))

     {

-      if(is.null(S4Vectors::mcols(measurementsTrain)) || !"assay" %in% colnames(S4Vectors::mcols(measurementsTrain)))

+      if(is.null(S4Vectors::mcols(measurementsTrain)) || !any(c("assay", "feature") %in% colnames(S4Vectors::mcols(measurementsTrain))))

       {

-        rankedFeatures <- originalFeatures[rankedFeaturesIndices]

+        rankedFeatures <- colnames(measurementsTrain)[rankedFeaturesIndices]

       } else {

         featureColumns <- na.omit(match(c("assay", "feature"), colnames(S4Vectors::mcols(measurementsTrain))))          

-        rankedFeatures <- originalFeatures[rankedFeaturesIndices, featureColumns]

+        rankedFeatures <- S4Vectors::mcols(measurementsTrain)[rankedFeaturesIndices, featureColumns]

       }

     } else { rankedFeatures <- NULL}

     if(!is.null(selectedFeaturesIndices))

     {

-      if(is.null(S4Vectors::mcols(measurementsTrain)) || !"assay" %in% colnames(S4Vectors::mcols(measurementsTrain)))

+      if(is.null(S4Vectors::mcols(measurementsTrain)) || !any(c("assay", "feature") %in% colnames(S4Vectors::mcols(measurementsTrain))))

       {

-        selectedFeatures <- originalFeatures[selectedFeaturesIndices]

+        selectedFeatures <- colnames(measurementsTrain)[selectedFeaturesIndices]

       } else {

         featureColumns <- na.omit(match(c("assay", "feature"), colnames(S4Vectors::mcols(measurementsTrain))))  

-        selectedFeatures <- originalFeatures[selectedFeaturesIndices, ]

+        selectedFeatures <- S4Vectors::mcols(measurementsTrain)[selectedFeaturesIndices, ]

       }

     } else { selectedFeatures <- NULL}

   } else { # Nested use in feature selection. No feature selection in inner execution, so ignore features. 

@@ -16,6 +16,8 @@

 \alias{features,ClassifyResult-method}

 \alias{models}

 \alias{models,ClassifyResult-method}

+\alias{finalModel}

+\alias{finalModel,ClassifyResult-method}

 \alias{performance}

 \alias{performance,ClassifyResult-method}

 \alias{tunedParameters}

@@ -83,6 +85,8 @@ most popular value of the parameter in cross-validation is used.}

 \describe{

 \item{\code{models(result)}}{A \code{list} of the models fitted for each training.}}

 \describe{

+\item{\code{finalModel(result)}}{A deployable model fitted on all of the data for use on future data.}}

+\describe{

 \item{\code{chosenFeatureNames(result)}}{A \code{list} of the features selected for each training.}}

 \describe{

 \item{\code{predictions(result)}}{Returns a \code{DataFrame} which has columns with test sample,

@@ -27,6 +27,9 @@

 \arguments{

 \item{results}{A list of \code{\link{ClassifyResult}} objects.}

+\item{...}{Parameters not used by the \code{ClassifyResult} method but passed to

+the \code{list} method.}

+

 \item{mode}{Default: "merge". Whether to merge all predictions of all

 iterations of cross-validation into one set or keep them separate. Keeping

 them separate will cause separate ROC curves to be computed for each