Changes in version 1.13.1                        

Bug fixes

  - paired_ora(): the combined (paired) analysis now uses genes tested
    for both expression and splicing as its universe; previously genes
    without a splicing result were included. Missing paired results are
    now filled in the paired columns instead of overwriting the splicing
    relative risk and enrichment score. Affects paired ORA results
    since 1.1.0 (Bioconductor 3.18); expression-only and splicing-only
    ORA are unchanged.
  - paired_diff(): test = "Wald" no longer fails; the reduced model is
    only used for the likelihood ratio test.
  - The reduced model for the likelihood ratio test now removes the
    group term by name instead of the first term of the design.
  - paired_diff(): surrogate variable estimation is now seeded
    (sva_seed), so results are reproducible between runs.

Changes

  - paired_diff(): the number of automatically estimated surrogate
    variables is capped so that the model keeps enough residual degrees
    of freedom (sva_min_residual_df). An optional upper limit can be set
    with sva_max_n; when set, the number is also capped at the square
    root of the number of samples. An exact request (sva_n) stops with
    an informative error when it is unsafe instead of being reduced
    silently.
  - paired_diff(): SVA diagnostics (requested, estimated and retained
    surrogate variables, residual degrees of freedom) are stored in
    metadata(result)$pairedGSEA_sva.

                        Changes in version 1.7.0                        

  - Ensure compatibility with updated msigdbr and new R version
  - Fixed rare bug in aggregate_pvalue

                        Changes in version 1.1.2                        

  - Fix bug where DEXSeq fails if there are NAs in the metadata

                        Changes in version 1.1.1                        

  - Improved how surrogate variables are transferred in the design
    formulas

                        Changes in version 1.1.0                        

  - paired_ora now adjusts gene background for analysis to reduce bias
  - paired_ora now also does an ora combining the genes from the two
    analyses
  - Added plotting mode paired = TRUE to plot_ora for the new paired_ora
    analysis type
  - Added baseMean to gene-level aggregation output

                        Changes in version 1.0.0                        

  - Version bump for Bioconductor release
  - Minor edits to documentation

                       Changes in version 0.99.6                        

  - Improved test coverage and depth
  - Coerced paired_ora output to DataFrame from data.table
  - Added value field to data man pages

                       Changes in version 0.99.5                        

  - Remove non-exported man pages

                       Changes in version 0.99.4                        

  - Remove filter_gene_sets option, as it is inherent in fgsea::ora
  - Added test depth and coverage for paired_diff and paired_ora
  - Fixed wrong storing location for splicing intermediate results

                       Changes in version 0.99.3                        

  - Rewrote code base to remove tidyverse dependencies
  - Added \code{...} and \link{...} where appropriate in documentation
  - Added input parameter checks
  - Reduced redundant input parameters from aggregate_pvalue
  - Modularized paired_ora and plot_ora
  - Added filter_gene_sets parameter to help users reduce gene set bias
  - Increased test coverage and depth
  - Moved data scripts to inst/script

                       Changes in version 0.99.2                        

  - Implement limma as alternative analysis method
  - Increased test coverage significantly
  - Removed usage of deprecated purrr::when

                       Changes in version 0.99.1                        

  - Updated vignette with a brief paragraph on the motivation of
    pairedGSEA.
  - Updated NAMESPACE to include all imported packages. Suggested
    packages are added in notes.

                       Changes in version 0.99.0                        

  - Submitted to Bioconductor