NEWS
ExCluster 0.99.13
- Fixed bugs that would cause ExCluster to either crash or not properly plot results in some cases
ExCluster 0.99.12
- Loosened FDR calculations slightly (ExCluster was a bit too stringent)
- Added plot.Type option to plotExonlog2FC function, which accepts "bitmap" and "PNG"
- Bug-tested plot.Type so machines with at least Ghostscript or X11 forwarding will have minimal issues
- Changed how files/folders are written & how write-permissions are detected, to avoid bugs
- Updated the vignette to use BiocManager instead of biocLite
ExCluster 0.99.11
- Removed some duplicated code (stripping ID numbers, computing p-values)
- removed several instances of cat() and print() and replaced them with message()
- changed apply() code to use matrixStats instead, which is up to 500 times faster (Thanks Lori!)
- this previous change sped the algorithm up from 1 hour+ to only ~ 20 minute runtime
- changed the test dataset so one of the genes has a p-value < 0.05 and plots results
ExCluster 0.99.10
- The GFF_convert function now outputs GFF3 formatted annotations
- Amended other functions (processCounts, ExCluster) to work on GFF3 formatted annotations
- Changed the output of GFF_convert to be a GRanges object of said GFF3 annotations
- Created a separate, internal, load_ExCluster_functions.R script to load helper functions
- Removed many depenencies on variables outside the environment for said functions
- Created a library of error messages in the ExCluster_errors.R script (internal)
- Separated the large ExCluster function into ExCluster.R, ExClust_compute_stats.R, and ExClust_main_function.R
ExCluster 0.99.9
- Added the function rtracklayerGTFtoGFF, which flattens GTF files imported by rtracklayer to GFF format
- Added the function GRangesFromGFF, which converts GFF formatted data to GRanges format
- Added the function GRangesFromExClustResults, which converts ExCluster function results to GRanges format
- Removed repeated code from the GFF_convert function
- Most functions now have checks to ensure GTF, GFF, and ExClustResults data is formatted correctly'
- Made an improvement to the processCounts function, which handles some edge cases better (zero reads in some conditions)
ExCluster 0.99.8
- Fixed build check errors from 0.99.7
ExCluster 0.99.7
- Fixed a number of bugs resulting from ExCluster() function code changes
- Updated the Vignette to correctly reflect changes suggested from the Bioconductor review
- Added several package imports to better keep track of global variables/functions
- Reduced the lengths of code lines in a number of instances
- Cleaned up error messages for ExCluster() function
ExCluster 0.99.6
- Completely re-wrote the GFF_convert() function to have helper functions and take advantage of GenomicRanges
- Changed processCounts, which was incorrectly counting reads as stranded (is now set to unstranded)
- Altered ExCluster null hypothesis simulations to run faster (about 4 times faster now)
- Made numerous small changes to address the issues with first Bioconductor review
ExCluster 0.99.5
- Fixed incorrectly pushed changes, which caused build error
ExCluster 0.99.4
- Reformatted the Vignette & DESCRIPTION to process the Vignette correctly with Sweave
ExCluster 0.99.3
- Removed extraneous files causing errors in the build
ExCluster 0.99.2
- Reformatted the Vignette & DESCRIPTION to process the Vignette correctly with Sweave
ExCluster 0.99.1
- Removed extraneous files causing errors in the build
ExCluster 0.99.0
- ExCluster package now passes R CMD build and check, and BiocCheck!
- takes about 2-3 hours to run on large datasets
- added a "Toy Dataset" for function manual runnable examples within the package
- more code in ExCluster has been turned into functions to reduce repeated code
- still some repeated code (such as parsing EnsIDs and exon bins)