Changes in version 1.37.1 o Hash-based resolution of tied gene sets in the multilevel algorithm (by Nikita Golikov), properly fixes #151 o P-avlues for ES close to 1 are more accurate now o Internally, the gene weights are now converted to integers (with scaling), which might introduce minor result discrepancies in rare cases Changes in version 1.35.7 o fix handling of zero stat values with gseaParam=0 in plotEnrichment Changes in version 1.35.3 o plotGesecaTable shows a legend o plotCoregulationProfileImage for in-situ profiling o change default color scheme for GESECA plots o fix GESECA plots for Seurat 5 Changes in version 1.35.1 o more reproducibility between platforms (fix #170), but the exact results differ from previous versions. Changes in version 1.33.3 o update code for using `msigdbr` in geseca vignette Changes in version 1.31.6 o plotCoregulationSpatial supports multiple samples Changes in version 1.31.3 o check for empty strings in gene names o minor fix in leading edge calculation Changes in version 1.31.2 o fora() provides fold enrichment scores for more effective prioritisation of results Changes in version 1.25.1 o introduced plotEnrichmentData() function for more flexible plotting o update GESECA plot behavior Changes in version 1.23.1 o Introduced GESECA method for multi-conditional gene set enrichment (see geseca-tutorial vignette for details) o Enrichment table plots now use cowplot (issue #101) o Enrichment table plots are more easily fine-tuned (issue #29) Changes in version 1.21.1 o fix a reproducibility problem (https://github.com/ctlab/fgsea/issues/110) Changes in version 1.19.4 o plotGseaTable now accepts units vector for column widths Changes in version 1.19.2 o Fixed fora() failing to run on a single pathway o Fixed problems random gene set generation for large k (issue #94) o Changed default eps to 1e-50 Changes in version 1.15.2 o Faster perturbate thanks to Nikolay Budin o Cleaner P-value and error estimations Changes in version 1.13.4 o mapIdsList for faster conversion of leadingEdges o experimental support of hypergeometric tests with fora() and collapsePathwaysORA() Changes in version 1.13.3 o experimental support for one-tailed tests Changes in version 1.13.2 o fgsea() implementation changed from simple to multilevel algorithm o proper handling of duplicate genes in gene sets for fgseaSimple() and fgseaMultilevel() Changes in version 1.11.2 o Simpler handling of conditional probabilities o Added the exact algorithm to inst folder Changes in version 1.11.1 o Proper absEps handling Changes in version 1.9.6 o Fix in log2err calculation o Small fixes for plotGseaTable Changes in version 1.9.2 o Added fgseaMultilevel function for more accurate logPval estimation Changes in version 1.7.1 o Setting colwidth to zero make column not to be drawn o Changable line width in plotEnrichment Changes in version 1.5.2 o fgsea throws warning for rank ties o fgsea throws warning for duplicate gene names o Leading edge is now ordered by decreasing of absolute statistic value Changes in version 1.5.1 o Reproducibility fixes o Added collapsePathway function to intelligently collapse overlapping pathways o Added fgseaLabel function for label-permuting GSEA Changes in version 1.1.2 o Fixed building issues Changes in version 1.1.1 o Fixed bug: not using Multicore on Windows Changes in version 0.99.8 o Results are now reproducible with set.seed() Changes in version 0.99.7 o Performance improvement: slightly better sampling and better segment tree Changes in version 0.99.6 o Fixed bug with failing when zero pathways are analyzed Changes in version 0.99.5 o Using `BiocParallel` package instead of `parallel` o Implemented leading edge analysis (`leadingEdge` column of fgsea result)