@@ -3,15 +3,15 @@ Type: Package

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

        applications to differential variability and differential

        distribution testing

-Version: 3.1.23

-Date: 2022-10-14

+Version: 3.1.25

+Date: 2022-10-18

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

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

 VignetteBuilder: knitr

 Encoding: UTF-8

 biocViews: Classification, Survival

 Depends: R (>= 4.1.0), generics, methods, S4Vectors, MultiAssayExperiment, BiocParallel, survival

-Imports: grid, genefilter, utils, dplyr, tidyr, rlang, randomForest

+Imports: grid, genefilter, utils, dplyr, tidyr, rlang, ranger

 Suggests: limma, edgeR, car, Rmixmod, ggplot2 (>= 3.0.0), gridExtra (>= 2.0.0), cowplot,

         BiocStyle, pamr, PoiClaClu, parathyroidSE, knitr, htmltools, gtable,

         scales, e1071, rmarkdown, IRanges, robustbase, glmnet, class, randomForestSRC,

@@ -10,8 +10,8 @@

 #' @param outcome A vector of class labels of class \code{\link{factor}} of the

 #' same length as the number of samples in \code{measurements} or a character vector of length 1 containing the

 #' column name in \code{measurements} if it is a \code{\link{DataFrame}}. Or a \code{\link{Surv}} object or a character vector of

-#' length 2 or 3 specifying the time and event columns in \code{measurements} for survival outcome.

-#' @param outcomeColumns If \code{measurements} is a \code{\link{MultiAssayExperiment}}, the column name(s) in \code{colData(measurements)} representing the outcome.

+#' length 2 or 3 specifying the time and event columns in \code{measurements} for survival outcome. If \code{measurements} is a

+#' \code{\link{MultiAssayExperiment}}, the column name(s) in \code{colData(measurements)} representing the outcome.

 #' @param outcomeTrain For the \code{train} function, either a factor vector of classes, a \code{\link{Surv}} object, or

 #' a character string, or vector of such strings, containing column name(s) of column(s)

 #' containing either classes or time and event information about survival.

@@ -118,8 +118,6 @@ setMethod("crossValidate", "DataFrame",

               # Which data-types or data-views are present?

               assayIDs <- unique(mcols(measurements)$assay)

               if(is.null(assayIDs)) assayIDs <- 1

-              

-              checkData(measurements, outcome)

               # Check that other variables are in the right format and fix

               nFeatures <- cleanNFeatures(nFeatures = nFeatures,

@@ -184,17 +182,12 @@ Using an ordinary GLM instead.")

                                       characteristicsLabel = characteristicsLabel

                                   )

                               },

-

                               simplify = FALSE)

                           },

-

                           simplify = FALSE)

                       },

-

                       simplify = FALSE)

-

                   result <- unlist(unlist(resClassifier))

-

               }

               ################################

@@ -208,7 +201,6 @@ Using an ordinary GLM instead.")

                   # The below loops over different combinations of assays and merges them together.

                   # This allows someone to answer which combinations of the assays might be most useful.

-

                   if(!is.list(assayCombinations) && assayCombinations == "all") assayCombinations <- do.call("c", sapply(seq_along(assayIDs), function(nChoose) combn(assayIDs, nChoose, simplify = FALSE)))

                   result <- sapply(assayCombinations, function(assayIndex){

@@ -244,7 +236,6 @@ Using an ordinary GLM instead.")

                       if(length(assayCombinations) == 0) stop("No assayCombinations with \"clinical\" data")

                   }

-

                   result <- sapply(assayCombinations, function(assayIndex){

                       CV(measurements = measurements[, mcols(measurements)[["assay"]] %in% assayIndex],

                          outcome = outcome, assayIDs = assayIndex,

@@ -306,7 +297,7 @@ Using an ordinary GLM instead.")

 # One or more omics data sets, possibly with clinical data.

 setMethod("crossValidate", "MultiAssayExperiment",

           function(measurements,

-                   outcomeColumns, 

+                   outcome, 

                    nFeatures = 20,

                    selectionMethod = "t-test",

                    selectionOptimisation = "Resubstitution",

@@ -319,7 +310,7 @@ setMethod("crossValidate", "MultiAssayExperiment",

                    nCores = 1,

                    characteristicsLabel = NULL, ...)

           {

-              measurementsAndOutcome <- prepareData(measurements, outcomeColumns, ...)

+              measurementsAndOutcome <- prepareData(measurements, outcome, ...)

               crossValidate(measurements = measurementsAndOutcome[["measurements"]],

                             outcome = measurementsAndOutcome[["outcome"]], 

@@ -558,24 +549,6 @@ generateCrossValParams <- function(nRepeats, nFolds, nCores, selectionOptimisati

 }

 ######################################

-

-

-######################################

-######################################

-checkData <- function(measurements, outcome){

-    if(is.null(rownames(measurements)))

-        stop("'measurements' DataFrame must have sample identifiers as its row names.")

-    if(any(is.na(measurements)))

-        stop("Some data elements are missing and classifiers don't work with missing data. Consider imputation or filtering.")

-

-    # !!!  Need to check mcols has assay NUm

-

-}

-######################################

-

-

-

-######################################

 ######################################

 #' A function to generate a ModellingParams object

 #'

@@ -643,9 +616,10 @@ generateModellingParams <- function(assayIDs,

     knownClassifiers <- .ClassifyRenvir[["classifyKeywords"]][, "classifier Keyword"]

     if(!classifier %in% knownClassifiers)

         stop(paste("Classifier must exactly match of these (be careful of case):", paste(knownClassifiers, collapse = ", ")))

-    

+

     classifierParams <- .classifierKeywordToParams(classifier)

-    classifierParams$trainParams@tuneParams <- c(classifierParams$trainParams@tuneParams, performanceType = performanceType)

+    if(!is.null(classifierParams$trainParams@tuneParams))

+      classifierParams$trainParams@tuneParams <- c(classifierParams$trainParams@tuneParams, performanceType = performanceType)

     selectionMethod <- unlist(selectionMethod)

@@ -715,6 +689,7 @@ generateMultiviewParams <- function(assayIDs,

                                           nFeatures = nFeatures,

                                           selectionMethod = selectionMethod,

                                           selectionOptimisation = "none",

+                                          performanceType = performanceType,

                                           classifier = classifier,

                                           multiViewMethod = "none")

@@ -833,11 +808,6 @@ CV <- function(measurements = NULL,

                characteristicsLabel = NULL)

 {

-    # Check that data is in the right format

-    if(!is.null(measurements))

-      checkData(measurements, outcome)

-    else

-      checkData(x, x)

     # Check that other variables are in the right format and fix

     nFeatures <- cleanNFeatures(nFeatures = nFeatures,

                                 measurements = measurements)

@@ -1072,9 +1042,9 @@ train.list <- function(x, outcomeTrain, ...)

 #' @rdname crossValidate

 #' @method train MultiAssayExperiment

 #' @export

-train.MultiAssayExperiment <- function(x, outcomeColumns, ...)

+train.MultiAssayExperiment <- function(x, outcome, ...)

           {

-              prepArgs <- list(x, outcomeColumns)

+              prepArgs <- list(x, outcome)

               extraInputs <- list(...)

               prepExtras <- trainExtras <- numeric()

               if(length(extraInputs) > 0)

@@ -11,6 +11,7 @@ GLMtrainInterface <- function(measurementsTrain, classesTrain, ..., verbose = 3)

     } else {fitData <- measurementsTrain}

   glm(class ~ . + 0, family = binomial, data = fitData, ...)

 }

+attr(GLMtrainInterface, "name") <- "GLMtrainInterface"

 # model is of class glm.

 GLMpredictInterface <- function(model, measurementsTest, returnType = c("both", "class", "score"),

@@ -51,11 +51,17 @@ setMethod("prepareData", "matrix",

 setMethod("prepareData", "DataFrame",

   function(measurements, outcome, useFeatures = "all", maxMissingProp = 0.0, topNvariance = NULL)

 {

+  if(is.null(rownames(measurements)))

+  {

+    warning("'measurements' DataFrame must have sample identifiers as its row names. Generating generic ones.")

+    rownames(measurements) <- paste("Sample", seq_len(nrow(measurements)))

+  }      

+            

   if(useFeatures != "all") # Subset to only the desired ones.

     measurements <- measurements[, useFeatures]

   # Won't ever be true if input data was MultiAssayExperiment because wideFormat already produces valid names.  

-  if(all.equal(colnames(measurements), make.names(colnames(measurements))) != TRUE)

+  if(!all(colnames(measurements) == make.names(colnames(measurements))))

   {

     warning("Unsafe feature names in input data. Converted into safe names.")

     mcols(measurements)$feature <- colnames(measurements) # Save the originals.

@@ -114,7 +120,7 @@ setMethod("prepareData", "DataFrame",

     else # Three columns. Therefore, counting process data.

       outcome <- survival::Surv(outcome[, 1], outcome[, 2], outcome[, 3])

   }

-  

+

   # Remove samples with indeterminate outcome.

   dropSamples <- which(is.na(outcome) | is.null(outcome))

   if(length(dropSamples) > 0)

@@ -125,8 +131,9 @@ setMethod("prepareData", "DataFrame",

   # Remove features with more missingness than allowed.

   nSamples <- nrow(measurements)

-  dropFeatures <- which(apply(measurements, 2, function(featureMeasurements) sum(is.na(featureMeasurements)))

-                        / nrow(measurements) > maxMissingProp)

+  measurementsMatrix <- as.matrix(measurements) # For speed of calculation.

+  dropFeatures <- which(apply(measurementsMatrix, 2, function(featureMeasurements) sum(is.na(featureMeasurements)))

+                        / nrow(measurementsMatrix) > maxMissingProp)

   if(length(dropFeatures) > 0)

     measurements <- measurements[, -dropFeatures]

@@ -158,11 +165,13 @@ setMethod("prepareData", "MultiAssayExperiment",

   if(!all(useFeatures[, "assay"] %in% c(names(measurements), "clinical")))

     stop("Some assay names in first column of 'useFeatures' are not assay names in 'measurements' or \"clinical\".")

-  clinicalColumns <- colnames(MultiAssayExperiment::colData(measurements))

+  clinicalColumnsDataset <- colnames(MultiAssayExperiment::colData(measurements))

   if("clinical" %in% useFeatures[, "assay"])

   {

     clinicalRows <- useFeatures[, "assay"] == "clinical"      

     clinicalColumns <- useFeatures[clinicalRows, "feature"]

+    if(length(clinicalColumns) == 1 && clinicalColumns == "all")

+      clinicalColumns <- setdiff(clinicalColumnsDataset, outcomeColumns)

     useFeatures <- useFeatures[!clinicalRows, ]

   } else {

     clinicalColumns <- NULL

@@ -171,7 +180,6 @@ setMethod("prepareData", "MultiAssayExperiment",

   if(nrow(useFeatures) > 0)

   {

     measurements <- measurements[, , unique(useFeatures[, "assay"])]

-  

     # Get all desired measurements tables and clinical columns (other than the columns representing outcome).

     # These form the independent variables to be used for making predictions with.

     # Variable names will have names like RNA_BRAF for traceability.

@@ -13,7 +13,10 @@

 #' are features.

 #' @param outcome Either a factor vector of classes, a \code{\link{Surv}} object, or

 #' a character string, or vector of such strings, containing column name(s) of column(s)

-#' containing either classes or time and event information about survival.

+#' containing either classes or time and event information about survival. If

+#' \code{measurements} is a \code{MultiAssayExperiment}, the names of the column (class) or

+#' columns (survival) in the table extracted by \code{colData(data)} that contain(s) the samples'

+#' outcome to use for prediction.

 #' @param crossValParams An object of class \code{\link{CrossValParams}},

 #' specifying the kind of cross-validation to be done.

 #' @param modellingParams An object of class \code{\link{ModellingParams}},

@@ -26,9 +29,6 @@

 #' package.  Transformation, selection and prediction functions provided by

 #' this package will cause the characteristics to be automatically determined

 #' and this can be left blank.

-#' @param outcomeColumns If \code{measurementsTrain} is a \code{MultiAssayExperiment}, the

-#' names of the column (class) or columns (survival) in the table extracted by \code{colData(data)}

-#' that contain(s)s the samples' outcome to use for prediction.

 #' @param ... Variables not used by the \code{matrix} nor the \code{MultiAssayExperiment} method which

 #' are passed into and used by the \code{DataFrame} method or passed onwards to \code{\link{prepareData}}.

 #' @param verbose Default: 1. A number between 0 and 3 for the amount of

@@ -70,9 +70,11 @@ setMethod("runTests", c("matrix"), function(measurements, outcome, ...) # Matrix

 setMethod("runTests", "DataFrame", function(measurements, outcome, crossValParams = CrossValParams(), modellingParams = ModellingParams(),

            characteristics = S4Vectors::DataFrame(), ..., verbose = 1)

 {

-  # Get out the outcome if inside of data table.           

   if(is.null(rownames(measurements)))

-    stop("'measurements' DataFrame must have sample identifiers as its row names.")

+  {

+    warning("'measurements' DataFrame must have sample identifiers as its row names. Generating generic ones.")

+    rownames(measurements) <- paste("Sample", seq_len(nrow(measurements)))

+  }

   if(any(is.na(measurements)))

     stop("Some data elements are missing and classifiers don't work with missing data. Consider imputation or filtering.")            

@@ -93,7 +95,7 @@ input data. Autmomatically reducing to smaller number.")

   # Element names of the list returned by runTest, in order.

   resultTypes <- c("ranked", "selected", "models", "testSet", "predictions", "tune", "importance")

-  

+

   # Create all partitions of training and testing sets.

   samplesSplits <- .samplesSplits(crossValParams, outcome)

   splitsTestInfo <- .splitsTestInfo(crossValParams, samplesSplits)

@@ -112,6 +114,7 @@ input data. Autmomatically reducing to smaller number.")

       message("Processing sample set ", setNumber, '.')

     # crossValParams is needed at least for nested feature tuning.

+    

     runTest(measurements[trainingSamples, , drop = FALSE], outcome[trainingSamples],

             measurements[testSamples, , drop = FALSE], outcome[testSamples],

             crossValParams, modellingParams, characteristics, verbose,

@@ -180,9 +183,9 @@ input data. Autmomatically reducing to smaller number.")

 #' @import MultiAssayExperiment methods

 #' @export

 setMethod("runTests", c("MultiAssayExperiment"),

-          function(measurements, outcomeColumns, ...)

+          function(measurements, outcome, ...)

 {

-  prepArgs <- list(measurements, outcomeColumns)              

+  prepArgs <- list(measurements, outcome)              

   extraInputs <- list(...)

   prepExtras <- numeric()

   if(length(extraInputs) > 0)

@@ -139,6 +139,12 @@

         return(list(NULL, rankings[[1]], NULL))

     tuneParamsTrain <- list(topN = topNfeatures)

+    performanceIndex <- match("performanceType", names(modellingParams@trainParams@tuneParams))

+    if(!is.na(performanceIndex))

+    {

+      performanceType <- modellingParams@trainParams@tuneParams[["performanceType"]]

+      modellingParams@trainParams@tuneParams <- modellingParams@trainParams@tuneParams[-performanceIndex]

+    }

     tuneParamsTrain <- append(tuneParamsTrain, modellingParams@trainParams@tuneParams)

     tuneCombosTrain <- expand.grid(tuneParamsTrain, stringsAsFactors = FALSE)  

     modellingParams@trainParams@tuneParams <- NULL

@@ -37,7 +37,7 @@ crossValidate(measurements, outcome, ...)

 \S4method{crossValidate}{MultiAssayExperiment}(

   measurements,

-  outcomeColumns,

+  outcome,

   nFeatures = 20,

   selectionMethod = "t-test",

   selectionOptimisation = "Resubstitution",

@@ -119,7 +119,7 @@ crossValidate(measurements, outcome, ...)

 \method{train}{list}(x, outcomeTrain, ...)

-\method{train}{MultiAssayExperiment}(x, outcomeColumns, ...)

+\method{train}{MultiAssayExperiment}(x, outcome, ...)

 \method{predict}{trainedByClassifyR}(object, newData, ...)

 }

@@ -130,7 +130,8 @@ or a list of these objects containing the data.}

 \item{outcome}{A vector of class labels of class \code{\link{factor}} of the

 same length as the number of samples in \code{measurements} or a character vector of length 1 containing the

 column name in \code{measurements} if it is a \code{\link{DataFrame}}. Or a \code{\link{Surv}} object or a character vector of

-length 2 or 3 specifying the time and event columns in \code{measurements} for survival outcome.}

+length 2 or 3 specifying the time and event columns in \code{measurements} for survival outcome. If \code{measurements} is a

+\code{\link{MultiAssayExperiment}}, the column name(s) in \code{colData(measurements)} representing the outcome.}

 \item{...}{Parameters passed into \code{\link{prepareData}} which control subsetting and filtering of input data.}

@@ -162,8 +163,6 @@ with each element being a vector of assays to combine. Special value \code{"all"

 \item{characteristicsLabel}{A character specifying an additional label for the cross-validation run.}

-\item{outcomeColumns}{If \code{measurements} is a \code{\link{MultiAssayExperiment}}, the column name(s) in \code{colData(measurements)} representing the outcome.}

-

 \item{x}{Same as \code{measurements} but only training samples.}

 \item{outcomeTrain}{For the \code{train} function, either a factor vector of classes, a \code{\link{Surv}} object, or

@@ -19,7 +19,7 @@

   verbose = 1

 )

-\S4method{runTests}{MultiAssayExperiment}(measurements, outcomeColumns, ...)

+\S4method{runTests}{MultiAssayExperiment}(measurements, outcome, ...)

 }

 \arguments{

 \item{measurements}{Either a \code{\link{matrix}}, \code{\link{DataFrame}}

@@ -32,7 +32,10 @@ are passed into and used by the \code{DataFrame} method or passed onwards to \co

 \item{outcome}{Either a factor vector of classes, a \code{\link{Surv}} object, or

 a character string, or vector of such strings, containing column name(s) of column(s)

-containing either classes or time and event information about survival.}

+containing either classes or time and event information about survival. If

+\code{measurements} is a \code{MultiAssayExperiment}, the names of the column (class) or

+columns (survival) in the table extracted by \code{colData(data)} that contain(s) the samples'

+outcome to use for prediction.}

 \item{crossValParams}{An object of class \code{\link{CrossValParams}},

 specifying the kind of cross-validation to be done.}

@@ -52,10 +55,6 @@ and this can be left blank.}

 \item{verbose}{Default: 1. A number between 0 and 3 for the amount of

 progress messages to give.  A higher number will produce more messages as

 more lower-level functions print messages.}

-

-\item{outcomeColumns}{If \code{measurementsTrain} is a \code{MultiAssayExperiment}, the

-names of the column (class) or columns (survival) in the table extracted by \code{colData(data)}

-that contain(s)s the samples' outcome to use for prediction.}

 }

 \value{

 An object of class \code{\link{ClassifyResult}}.