- Bug fixed in vignette for CRAN %\VignetteIndexEntry{Generating realistic data}
- Bug fixed to pass R CMD check under Solaris
- CRAN submission.
- Minor updates to C code.
- Minor updates to C code to pass R CMD check cleanly under R 3.4.1
- 'acnr' moved from Depends to Imports
- Added more tests
- Fixed Issue #4
- Fixed issue #5: the package does not rely on cghseg:::segmeanCO anymore!
- Added pruned DPA algorithm [GR]
- Added a test that pDPA and classical DP yield identical results for (small) univariate signals
- Data generation vignette is now .Rmd instead of .Rnw.
- Moved documentation from inlinedocs to roxygen2 (as per Issue #2)
- Minor updates to vignette following changes to 'acnr'.
- PSSeg vignette is now .Rmd instead of .Rnw
- Minor bug fix in 'plotSeg' in the default value of parameter binExclPattern.
- In 'PSSeg': added argument 'rankTransform' to allow rank-based segmentation to be performed.
- Bug fixed in vignette (remove GFLars segmentation)
- Bug fixes for submission to CRAN
- Updates to vignette.
- Updates to CITATION.
- In 'estimateSd': SPEED UP for method "Hall": vectorization.
- In 'jointSeg', 'PSseg': it is now possible to run 'modelSelection' on initial segmentation. Default behavior of the functions is unchanged.
- In 'getCopyNumberDataByResampling', if 'regAnnot' is NULL (the default), frequencies of regions (0,1), (0,2), (1,1) and (1,2) (the most common alterations) are set to represent 90% of the regions.
- Cleanups in DESCRIPTION/NAMESPACE.
- Moved 'doCnaStruct' to 'otherMethods/'.
- Packages passes R CMD check --as-cran.
- Replaced 'jointSeg' by 'jointseg' in 'inst' scripts.
- Some changes to the main vignette (now using 'knitr').
- Updates in scripts to perform ASCAT segmentation.
- Added a vignette to illustrate how to add a new method.
- Changed Illumina to GSE11976 and Affymetrix to GSE29172
- Bug fixed for PSCBS, DynamicProgramming and CBS methods.
- Updated scripts
- Major changes in the interface of the main functions 'PSSeg' and 'jointseg'.
- Now 'doGFLars' is a wrapper calling the lower-level 'segmentByGFLars'.
- Now 'doRBS' is a wrapper calling the lower-level 'segmentByRBS'.
- Added 'binMissingValues' so that it may be used not only for GFLars.
- Cleaned up documentation and examples.
- in 'jointSeg': added argument 'segFUN' and flavor "other" to enable a user-defined segmentation method to be used.
- BUG FIX in 'doCBS': replaced start bkp given by CBS by end bkp as we consider ends of segments for all methods. Only makes a difference when there are missing values in the signal.
- Bug fixed in scripts to build table of the paper.
- Updated mapping for d|het segmentations (median).
- Added scripts to 'inst/otherMethods' and 'inst/syntheticProfiles' to address referees' comments.
- Added 'inst/figures' to reproduce the figures of the paper.
- 'plotSeg': Calculation of segment means would fail when a breakpoint was detected between the first and second position.
- Renamed function 'jointSeg' to 'jointseg' to avoid name clash with package name.
- Now all documented functions are exported.
- Updates in vignette.
- Added scripts to reproduce methods comparison of the jointSeg paper to dedicated subdirectories.
- Added scripts to reproduce figures of the jointSeg paper to 'figures'.
- Removed defunct package PSCN from 'Suggests' field, and moved function 'doPSCN' to (newly created) directory 'zzz.defunct'.
- Removed 'R.utils' from 'Depends' field.
- Updates in 'plotSeg'.
- 'segmentByNnn' replaced by 'doNnn'.
- Replaced 'segmentByCghseg' by 'doDynamicProgramming'.
- Wrappers for PSCN and CnaStruct now operate systematically on log-scaled copy numbers (ie LRRs). Thanks to report by Hugues Sicotte.
- Segmentation methods can now be run on numeric vectors or data.frames (when appropriate).
- Added flavor 'DP'.
- Bumped version number to trigger build by R-forge.
- Added function 'getTpFp' so that performance can be evaluated in a scale which does not depend on the signal length. Updated examples and vignette accordingly.
- Added argument 'regNames' to 'plotSeg' so that region labels are plotted if available.
- Some example code now embedded in 'require()' statements to avoid problems in the R CMD check mechanism of R-forge.
- Added argument 'K' to 'pruneByDP'.
- Modification in 'segmentByCnaStruct'
- Added segmentation methods PELT and CnaStruct
- Added flavor "Hall" to estimate standard deviation in 'estimateSd'.
- Added argument "DP" in 'PSSeg' and 'jointSeg'.
- Added argument "relax" in 'getTprTnr' and definition of true negatives.
- Added return value "rse" in 'segmentByRBS'.
- Added argument "connex" in 'getCopyNumberByResampling'.
- In 'PSSeg' and 'jointSeg': added argument 'jitter'.
- In getCopyNumberDataByResampling': added argument 'connex' to forces adjacent regions to be connex.
- In 'getTprTnr': added argument 'relax'.
- 'segmentByGFLars' now handles missing values.
- In 'PSSeg': Flavor "GFLars" can now be run at full resolution
- BUG FIX in 'segmentByPSCBS'.
- flavor = 'PSBCS' added.
- Removed segmentByPairedPSCBS.
- BUG FIX in 'modelSelection': if the dimension is too small, in 'jointSeg' if flavor = 'CBS' or 'PSCN' no using of 'modelSelection'.
- Bumped acnr dependency (>= 0.1.1) to fix minor case sensitivity issue.
- Reduced graphics size in vignette.
- Updated vignette.
- Passes R CMD check with no errors.
- Updated NAMESPACE.
- Moved dilution data sets to new package 'acnr'.
- Cleanups in doc and return values.
- 'jointSeg' now returns 'bestBkp'
- Removed 'position' from I/O in 'getCopyNumberDataByResampling' and 'PSSeg'.
- In 'plotSeg': added arguments 'exclNames', 'ylabs', and 'binExclPattern' so that 'plotSeg' can handle not only copy number signals.
- Improved doc and examples.
- Now using 'ERMajustment' for model selection.
- BUG FIX in 'segmentByRBS': Empty candidate list would give an error. Now returning early when 'minRegionSize' is too large for 'K'.
- Added 'loadCnRegionData'.
- Added Affymetrix dilution data set (from GEO:GSE29172 and GSE26302).
- Removed 'setNormalContamination' as it is not needed anymore.
- In 'getCopyNumberDataByResampling':
- Added argument 'regAnnot', through which theoretical frequencies for each CN regions can be specified.
- Added arguments 'bkp' and 'regions' to allow for bypassing the breakpoint generation step.
- Made the constraints on CN transitions more generic.
- Added 'loadCnRegionData'.
- Replaced all 'jumps' by 'bkp'.
- Added argument 'platform' to 'segmentByPSCN'.
- Added Illumina dilution data set (from GEO:GSE11976).
- BUG FIX in 'segmentByCghseg': index shift when reshaping results.
- Now 'PSSeg' can also run 'CBS' segmentation.
- Now 'PSSeg' can also run 'PSCN' and 'cghseg' segmentations.
- Updated doc and vignette.
- Now 'segmentByRBS' handles missing values. Therefore, 'PSSeg' and 'jointSeg' with the corresponding flavor also do.
- Added Argument 'statistic' to 'PSSeg'.
- Now 'segmentByRBS' scales each dimension to unit variance using 'estimateSd'.
- Some code and doc cleanups.
- Renamed 'binSeg' to 'segmentByRBS'.
- Renamed 'dpSeg' to 'pruneByDP'.
- Renamed 'segmentByGflars' to 'segmentByGFLars'.
- SPEEDUP: removed redundant calls to 'getRSE' in 'binSeg'.
- Added function 'runPSCN' as a convenient wrapper to PSCN + DP.
- Added function 'getTprTnr' to evaluate the performance of segmentation methods.
- Added function 'prof' for optional reporting of CPU and memory usage.
- Added 'setNormalContamination'.
- BUG FIX in 'binSeg': index shift in correspondence b/w breakpoint position and interval.
- Added 'binSeg' for binary segmentation.
- Added wrappers 'jointSeg', 'plotSeg', 'PSGFL'.
- Added argument 'flavor' to 'jointSeg' so that one can run GFLARS or binary segmentation before dynamic programming.
- Added example data files and scripts based on public data set GSE19539.
- Added scripts to generate these data files from raw CEL files in GSE19539.
- Added model selection part to 'modelSelection'.
- Removed model selection part in 'dpseg'.
- Renamed 'randomProfileTCGA' to 'getCopyNumberDataByResampling'.
- Added 'randomProfileTCGA'.
- BUG FIX in randomProfile: 'minLength' would not work as expected.
- Tentative bug fix in gflars: indices in gammaTemp.
- SPEEDUP: removed unnecessary calls to 'complex'.
- Updated documentation
- Some code cleanups.
- SPEEDUP: removed explicit calls to 'matrix()' and 'apply(..., rev)'.
- Package documentation now done using 'inlinedocs'.
- Improved scripts for assessing replication of the MATLAB version.
- Improved scripts for testing speed.
- BUG FIX: in optimizeLARS.R, the index in 'lambda' was off by 1.
- Added 'value' to the return values of 'optimizeLARS'.
- Renamed 'optimizeLARS' to 'gflars'.
- BUG FIX: in optimizeLARS.R, 'Y' should be centered but not scaled.
- Added dpseg.R: finds the best set of change points from a(n over-) segmentation using joint dynamic programming.
- SPEEDUP: replaced 'apply(, 1, sum)' by 'rowSums()'.
- Added test scripts for replicating MATLAB version.
- Added test scripts for speed trials.
- Created.