Major Changes

• Make 'cyt_xgb' compatible with the upcoming XGBoost R package changes (parameter renames, changes in predict() output).
• Fix ROC/AUC computation and confusion matrix computation for XGBoost models.
• Deprecate arguments 'eta' and 'gamma' in cyt_xgb() in favor of 'learning_rate' and 'min_split_loss' using the lifecycle package.

Major Changes

• 'cyt_mint_splsda', a new function for conducting multivariate integrative analysis using sparse partial least squares discriminant analysis (sPLS-DA) with multiple datasets, has been added.

• 'cyt_splsda' now has new arguments:

  • 'batch_col' to specify a batch column for batch correction in the sPLS-DA analysis using a Z-score method.
  • 'multilevel_col' to specify a multilevel column for multilevel analysis in sPLS-DA.
  • 'ind_names' to specify custom individual names for the samples in the plots.

• Added new "ExampleData5" dataset which is the same as "ExampleData1" but with a batch column for demonstration of the 'batch_col' argument in 'cyt_mint_splsda'.

• 'cyt_heatmap' now uses pheatmap for plotting, and provides a proper legend along with additional scaling options.

Minor Changes and Bug Fixes

• Updated the vignette and README files to include examples of the new 'cyt_mint_splsda' function.

• Updated 'cyt_splsda' and 'cyt_pca' logic.

Major Changes

• 'cyt_ttest' now conducts shapiro-wilk test and based on the p-value decides to do Welch two sample t-test or Wilcoxon Rank Sum Test.

Minor Changes and Bug Fixes

• Updated 'cyt_pca' and 'cyt_splsda' to use the appropriate pch argument in plotIndiv().
• Updated 'cyt_ttest' examples.
• Updated citation formatting.
• Changed links

Major Changes

• 'cyt_splsda': Fixed an error occurring when only one variable has VIP > 1 leading to not being able to conduct sPLS-DA analysis. Now, the function checks whether the number of VIP > 1 variables is below 2, if that is true the sPLS-DA model of VIP > 1 is skipped.

• Fixed an issue with the 'verbose' argument of 'cyt_splsda' not working in overall analysis due to calling the wrong object.

• 'cyt_errbp' now uses 'ggplot2' with 'facet_wrap' to create multiple error-bar plots where the p-value and effect size labels are based on t-test comparisons between the groups.

Minor Changes and Bug Fixes

• 'cyt_errbp', 'cyt_bp', 'cyt_bp2', and 'cyt_skku' received a minor theme update keeping the background white in scenarios where it may appear transparent due to using 'theme_minimal()'.

• Fixed some grammatical and spelling issues in the README file.

• Vignette for the package now show plots generated that aren't saved to PDF files. A temporary directory is not created anymore.

• Added references in description field of the DESCRIPTION file to the multivariate methods mentioned in the format 'Authors (Year) doi:...'.

• 'cyt_rf', 'cyt_splsda', 'cyt_ttest', 'cyt_dualflashplot', and 'cyt_volc' now have a logical 'verbose' argument to print the output if user wants to. This is to ensure printed output in console can be easily suppressed.

• 'cyt_xgb' has a new logical argument called 'print_results' that works the same as the 'verbose' argument in other functions mentioned above.

• Reverted the previous 'verbose' conditionals on plots from the functions mentioned above as they are essential to the output. For example, after creating a ggplot2 object of the plots, I have left 'print(a)' without a 'verbose' conditional statement.

• Have removed the arguments for changing graphical parameters in 'cyt_bp', and 'cyt_bp2' so it no longer requires changes to the user's graphical parameters.

• Updated examples from graphical parameters changing without resetting properly to now the graphical parameters reverting to original using 'oldpar <- par(no.readonly = TRUE)' in the beginning and par(oldpar) after execution of code.

• Updated 'getting_started.Rmd' vignette where the files created and saved are now saved to a temporary directory using 'tempdir()' to avoid creating files in the user's working directory. Additionally,
  the vignette now uses 'oldpar <- par(no.readonly = TRUE)' in the beginning and par(oldpar) after code execution to revert to original graphical parameters for the 'cyt_errbp' examples shown.

• Added a new 'format_output' argument to 'cyt_anova' and 'cyt_ttest' to format the output as a data frame which can be printed to show neat format, however still dependent on whether 'verbose' equals TRUE or FALSE.

• 'cyt_rf', 'cyt_splsda', and 'cyt_xgb' now has a 'seed' argument to set the seed for reproducibility.