*Updated NCI Policies on Server UI
*Minor change to ensure compatibility with multiple datasets.
*Require R-4.1.0 in order to make the package CRAN compliant (fixes a single rtracklayer check server install issue)
*Following CRAN guidance, removed a command to close all connections after opening a file.
*Bug fixes related to the suggested plotly dependency.
*Moved the plotly package to suggests because it has been temporarily orphaned.
*Made package check successfully without suggested dependencies.
*Made conditional BSgenome.Hsapiens.UCSC.hg19 to improve OSX installation.
*Made conditional InteractionSet to make OSX installation more seamless.
*We have now added smoothie::kernel2d smooth as an alternative smoother to spatialfil (now disabled) and now our package should once again be CRAN ready.
*With spatialfil removed from CRAN, we have disabled the spatialfil requirement to run the additional examples. *This algorithm is neat and we'll look at other ways to calculate interaction probability, but for the moment, we'll just disable things to keep the package on CRAN.
*Fixed the gene search in the shiny application *Added gene listing for row and column genes to the left of the plot. **These gene data table outputs are searchable and the size is adjustable. *Disabled automatic switching to the sample data tab on click so that users can properly use the new gene column feature. *CRAN fixes to keep the package publicly available
*Reduced imports to 20 and moved the remainder of packages into suggests. *Fixed whole genome map. *Optimized load time to plot a whole genome map (should be roughly a 50% reduction, according to profvis).
*Removed Biostrings dependency to clear NOTE (although packages that CNVScope depends on require it). The prior Biostrings issue has been fixed.
*Added an explicit dependency, Biostrings to keep CRAN check errors on debian from occurring. *Made a small fix to rebinGenomicInteractions, fixing a check at the end to handle a vector input **This will make the package compatible with R 4.0.
*Removed all blockseg references in code. It was a nice feature to have a third algorithm, but blockseg will not be on CRAN shortly.
*Removed blockseg as a strong dependency, per CRAN.
*Updated documentation (some new images have been added to show the effect of bin sizes and resolution). *Fixed a broken URL that was causing a CRAN note.
*Updated documentation for finely binned matrices (see the input matrix vignette). *Several sizes of NBL whole genome matrices are available within the package (1e5,2.5e5,1e6,1e7,1e8). **These are for package use only. The HD matrices are not speed-optimal for app use. **Functionality of basic package functions remains the same as 2.7.2.5. *passes CRAN checks again with flying colors (no notes, warnings, or errors).
*Added P-value filter *Added correlatoin options *Added additional app visualization options. *Added material on how to make higher resolution maps.
*Minor changes to package dependency CNVScopePublicData.
*Another minor cran fix to CNVScopeserver.
*Small changes to CNVScopeserver to fix a CRAN check.
*Small changes in 2.4.9 to the README, for publicity. *Major changes to many functions to pass CRAN checks (global variable issues). CNVScope v2.4.8 (Release date: 2019-07-17)
*Small changes to dependencies (including pwr package) and vignettes.
*Minor fix to freadGDCfile documentation
*More information about BLCA, SKCM, and AML in additional examples vignette **Their processed files have been placed within the package **These are available only on the github site as the CRAN size limitations forbid it. *Added the ability to specify chrM and chrY in formSampleMatrixFromRawGDCData function.
Changes:
- Fixed vignettes not properly linking to the package. CNVScope v2.4.4 (Release date: 2019-07-09) ==============
Changes:
- Added a power analysis vignette
- Fixed issues with formSampleMatrixFromRawGDCData ** These issues arose from changing types with underlying Bioconductor dependencies from character to factor Rle.
- Updated the build for R-base v3.6.1 *prior to this, this package now works with CNVScopePublicData, a large package containing *all that is requisite to run the server locally.