metabolyseR 0.15.4

• Fixed various tidyverse warnings.

• Fixed an error when calculating the mds dimensions for multiple class comparisons with differing numbers of observations.

• Added the transformPercent() method for the AnalysisData S4 class to scale as a percentage of feature maximum intensity.

• Feature intensities are now displayed as relative percent intensities in heat maps generated by plotExplanatoryHeatmap().

• Reduced the gap between the dendrogram and heat map in outputs of plotExplanatoryHeatmap().

metabolyseR 0.15.3

• Fixed the margin value displayed in plots from plotSupervisedRF()

• The plotExplanatoryHeatmap() method for the Analysis S4 class now returns a warning and skips plotting if an error is encountered whilst trying to plot a heat map.

metabolyseR 0.15.2

• Added the argument refactor to the method transformTICnorm() to enable the feature intensities of total ion count (TIC) normalised data to be refactored back to a range consistent with the original data by multiplying the normalised values by the median TIC.

• Removed the permutation cap when the perm argument of randomForest() is less than the total number of permutations possible.

metabolyseR 0.15.1

• The class occupancy methods now throw a helpful error message if no features are available on which to compute class occupancy.

• Fixed a bug where it was not possible to customize the order of class labels in the legend of plotLDA().

metabolyseR 0.15.0

• It is now possible to specify multiple cls arguments to the aggregation methods.

• Correlation thresholds are now available for both coefficient and total number using the minCoef and maxCor arguments in the correlations() method.

• Added the predictions() accessor method for the RandomForest S4 class to enable the retrieval of the out-of-bag model response predictions.

• The occupancy filtering methods now error if the value supplied to argument occupancy is non-numeric.

• Memory usage and performance improvements for the randomForest() method.

• Added type() and response() methods for the Univariate S4 class.

• plotLDA() now returns a warning and skips plotting if an error is encountered during PC-LDA.

• The value pre-treated is now the default for argument type in the Analysis S4 class accessor methods.

• Added the value argument to the explanatoryFeatures() method to allow the specification of on which importance value to apply the specified threshold.

• The specified cls argument is now correctly displayed on the x-axis title of the resulting plots from the plotFeature() method.

metabolyseR 0.14.10

• Added the method predict() for the RandomForest S4 class to predict model response values.

• Added the method mtry() for the AnalysisData S4 class to return the default mtry random forest parameter for a given response variable.

• Added the method tune() for the AnalysisData S4 class to tune the random forest parameters mtry and ntree for a given response variable.

metabolyseR 0.14.9

• Suppressed name repair console message encountered during random forest permutation testing.

• Added the proximity() method for extracting sample proximities from the RandomForest S4 class.

• Added the mds() method to perform multidimensional scaling on sample proximities from the RandomForest S4 class.

• Added the roc() method to calculate receiver-operator characteristic curves from the RandomForest S4 class.

metabolyseR 0.14.8

• An error is now thrown during random forest classification when less than two classes are specified.

• plotSupervisedRF() now skips plotting if errors are encountered during random forest training.

metabolyseR 0.14.7

• Single replicate classes now automatically removed by plotLDA().

metabolyseR 0.14.6

• plotExplanatoryHeatmap method for the Analysis class now returns the plot only if the number of plots is equal to 1.

• Removed reference to the nCores parameter from the documentation example of metabolyse().