v1.3.4: Minor bugfixes: * Manual updated to reflect changes in the most recent version of QoRTs: the mergeNovelSplices function now automatically calculates size factors, if size factors are not set explicitly. * Fixed typos in the manual. * Bugfix: Basic functionality now works without a GFF file being specified (as per specification) * Iterated version number to match Bioconductor versioning. * Added the ability to suppress the creation of a DESeq2 count container object (for advanced users). v1.1.14: Minor bugfixes: (thanks to Warren McGee) * When making html summary files with "use.multigene.aggregates" genes set to TRUE, JunctionSeq will use shortened html file names in order to avoid filename length limitations on certain filesystems when there are numerous genes overlapping with one another. Aggregate gene names will be truncated to only include the first and last member genes iff there are more than 2 member genes, with a "..." in between. * If minimalImageFilenames is FALSE, then Image files will similarly be renamed. Otherwise they will be reduced in size even further, as per the v1.1.10 update. You can control the naming of the png files using the options: "minimalImageFilenames", "name.files.with.geneID", and "number.plots" * Fixed a bug in the phenotype table on the HTML index page. Various minor "quality-of-life" improvements: * When available, gene names will be included in all tables. * The main index data table now includes mouseover text describing each column. * Added additional information to each plot page. * Added word wrapping to aggregate genes in the main index tables (thanks to Warren McGee). v1.1.12: * Moved NEWS file to base directory * Updated CITATION file with arxiv preprint. v1.1.10: * Changed naming scheme for gene-profile plots, to reduce the length of the filenames. Some web hosts are not compatible with excessively long file names. When writeHTMLresults is TRUE, JunctionSeq will now default to a shorter naming scheme in which the image files are named by number rather than by gene. This behavior can be eliminated setting the new minimalImageFilenames parameter to FALSE. * Minor adjustments to the vignette. * JunctionSeq is now built and tested on SL6 in R 3.2.3 v1.1.3: * Very minor typographical fixes and clarifications in the vignette. v1.1.1: * Release version. Minor cosmetic changes. v0.99.8: * Very minor changes, mostly just fixing typos in the documentation and testing the functionality of the Bioconductor svn server. v0.99.6: * JunctionSeq is now pre-approved by Bioconductor, and will probably be included in the next Bioconductor release. * Running in verbose mode, readJunctionSeqCounts now reports the number of gene aggregates found, and the total number of genes that belong to them. * Iterated version number to 0.99.6, to sync with bioconductor submission version numbering. v0.6.46: * Minor update. Added references to the arxiv preprint manuscript. Fixed bad links in the vignette. Fixed a minor bug with the browser track creator. v0.6.42: * Minor update. No real functional changes, but several changes to code style. * Removed all trailing semicolons * Used non-vectorized logical functions in conditional statements. * Removed old commented-out code. * Renamed the example dataset package to "JctSeqData", rather than "JctSeqExData2". v0.6.30: * Very minor bugfix, fixing an issue when JunctionSeq is installed with the most recent pre-release version of the SummarizedExperiments bioconductor package (in the development version of R). * Also fixed some typos in the vignette. v0.6.25: Major structural changes: * DESeq2 code is now called as foreign function calls and calls to exported DESeq2 functions. Copy-over of DESeq2 code greatly reduced. * As a consequence: JunctionSeq no longer requires Rcpp, RcppArmadillo. * JunctionSeq no longer requires compilation, it no longer includes any C++ code. In addition, a few minor bugfixes: * Fixed chromosome labels in gene profile plots. * Fixed behavior when label.p.vals is FALSE v0.6.8: Several bugfixes, along with some error checking. * Fixed one place where I used an "T" instead of "TRUE", so BiocCheck stops complaining. * Local dispersion fit setting now propagates properly from runJunctionSeqAnalysis to fitJunctionSeqDispersionFunction. * Local dispersion fit now works with the standard "DESeq2-style" method.GLM. * Manual setting of inclusion thresholds now works with the "DESeq2-style" method.GLM. * keep.hypothesisTest.fit now prints a warning when attempted with the "DESeq2-style" method.GLM, as saving the fit statistics cannot easily be implemented using this method (since all the model fitting happens in C++). * If you attempt to plot a non-existent gene, it now prints an informative error. * Fixed other output-compatibility issues with the legacy "simpleML" mode. * Tested all sorts of combinations of method modes. Added warnings to at least some of the incompatible combinations. * Reduced extraneous messages in debug mode.