VERSION 1.3.5
- Add gam approximation of NES computation explanation to man files.
- Add clear user messages to gsea.
VERSION 1.3.4
- gsea now warns when the provided score is not centered. We also added a new paremeter that allows to force data centering.
VERSION 1.3.3
- htmlpage has been replaced by write.html which is much more versatile.
- New function eset2genelevel has been added.
- ExpressionPhenoTest's survival analysis bayesian approach now use's the implementation from BMA package to compute posterior probabilities.
VERSION 1.3.2
- Added a function to export gseaData objects to html files with plots and links to online databases.
- Added functions gsea.selGsets and gsea.selVars to subset gseaData objects.
- ExpressionPhenoTest now allows to compute posterior probabilities instead of pvalues.
- Function to easily perform gene set enrichment analysis with gene ontology and kegg databases have been added. Their names are gsea.go and gsea.kegg.
- Functions to obtain Gene ontolology and Kegg gene sets have been added. Their names are getGo and getKegg.
- The plot method for gsea objects now alows to plot a selection of gene sets (or one specific gene set). In previous versions it allways ploted all gene sets.
- The package works with a new object to store gsea data. The object is called gseaData.
- smoothCoxph function did not calculate the confidence interval's lower bound when values were negative. Now it does. It also adds the possibility to do the calculations using logrisk.
VERSION 1.1.2
- Use R instead of C when computing enrichment scores.
VERSION 1.1.1
- Use R's memory instead of C memory in C code.
- Import functions from mgcv library that are required by findCopyNumber.
- Add getNes method.
- Allocate vectors of non R variables in C code. This seemed to give problems under windows.
- When producing the GSEA plot the option to plot only NES did not work.
- Fixed bug in export2CSV which caused it to ignore the row.names argument
- Fixed bug in ExpressionPhenoTest which caused to incorrectly compute P-values when adjustment variables were specified and the variable of interest was categorical or ordinal
VERSION 1.0.1
-findCopyNumber has been added to package. This function is very similar to ACE. Finds regions with copy number variation.
-GSEA's wilcoxon plot did not take into account signatures with only one gene.
-Wilcoxon test implemented in GSEA now accounts for a gene set's mean to be different from 0 and from all genes' mean value.
-When running GSEA with a non-centered score big gene sets may produce problems. The software accounts for the maximum number of genes allowed in a gene set. If some gene set is bigger than that number wilcoxon test is used instead of permutation test.
VERSION 1.0.0
-First stable version.
