2023-08-19 Scott Ritchie * DESCRIPTION: Release 1.2.7 * Package overview help file is now correctly documented as requested by CRAN following breaking changes in Roxygen 7.0.0 that changed the way this help file was internally tagged in the source code. * Resolves CRAN NOTE requesting removal of system requirements of C++11 or greater, as R has required this since R 3.6. To pass CRAN checks, NetRep will no longer support R < 3.6 going forward. 2023-01-06 Scott Ritchie * DESCRIPTION: Release 1.2.6 * DESCRIPTION: Date field removed 2022-12-28 Scott Ritchie * DESCRIPTION: Release 1.2.5 * Replaced sprintf with snprintf when creating progress bar, fixing compiler warning on macOS 13 and above. 2020-10-06 Scott Ritchie * DESCRIPTION: Release 1.2.4 * Added short form citation to Description field in package DESCRIPTION. A full citation can still be found via citation("NetRep"). * Removed GitHub page README.md from package bundle. 2020-10-06 Scott Ritchie * DESCRIPTION: Release 1.2.3 * Minor patch to fix new package build checks introduce by CRAN from R 3.6 onwards. 2020-10-06 Scott Ritchie * DESCRIPTION: Release 1.2.2 * Minor patch to remove deprecated Armadillo C++ macro which will allow RcppArmadillo to update to the latest version of armadillo. As a side effect, users of old compilers may see a compiler warning from RcppArmadillo when building this package (this will only affect you if your compiler is too old to support R >= 4.0.0). 2018-06-11 Scott Ritchie * DESCRIPTION: Release 1.2.1 * Updated Date Field in DESCRIPTION. 2018-06-11 Scott Ritchie * DESCRIPTION: Release 1.2.0 * Removed dependency on the ref package, which is being archived. The internal representation of the data, correlation, and network matrix for the dataset loaded into RAM has been changed from a matrix (previously passed by reference between some functions using the ref package) to an environment that contains a variable storing the matrix. By passing these environments between functions, the loaded dataset can be changed within an internal function without the callee function creating a copy of the previously loaded dataset to be restored on the completion of the function call. * Enhanced code clarity by removing use of with(), which in many instances made it unclear what scope variables were being operated on due to conflicts between process user input in an environment variable and unprocessed user input in function arguments. * Added some minor optimisations to memory usage. The garbage collector is now called at the end of each user exported function. This ensures that any large matrices that have been loaded from disk are correctly unloaded, even when a function returns an error due to malformed user input. * Reduced vignette file size by changing images from png to jpeg. 2017-08-15 Scott Ritchie * DESCRIPTION: Release 1.1.5 * Bugfix to data heatmap and module summary bar charts: sample labels previously appeared in reverse order (i.e. did not correspond to the shown data). This has now been fixed. * Enhancement to data heatmap and module summary bar charts: Previously, when specifying 'orderSampleBy' only samples present in both the plotted 'test' dataset and the 'orderSamplesBy' dataset were shown. Now, samples present in the the plotted 'test' dataset but not present in the 'orderSamplesBy' dataset are also shown on the data heatmap and module summary bar charts. These are shown below the samples present in both dataset on both the data heatmap and module summary bar charts, separated from the samples present in both datasets by a horizontal black line. These samples appear ordered by their module summary values in the 'test' dataset. * R/plot-utils.R: plotProps() now returns an additional value indicating the number of samples present in the 'test' dataset not present in the 'orderSamplesBy' dataset. * R/plot-functions.R: plotSquareHeatmap() and plotMultiBar() now have an additional argument, 'yLine', that tells both functions where to draw a horizontal black line (i.e. to indicate separation of samples present in both the 'orderSamplesBy' and 'test' dataset from those present only in the 'test' dataset). * R/plot-functions.R: plotMultiBar() now shows sample labels in the correct order. * R/plot-module.R: plotModule() now keeps track of number of samples present only in the 'test' dataset to pass to plotSquareHeatamp() and plotMultiBar(). * R/plot-module.R: plotModule() now keeps track of number of samples present only in the 'test' dataset to pass to plotSquareHeatamp() and plotMultiBar(). * R/plot-topology.R plotData() and plotSummary() now keep track of number of samples present in the 'test' dataset to pass to plotSquareHeatmap() and plotMultiBar() respectively. 2017-08-11 Scott Ritchie * DESCRIPTION: Release 1.1.4 * R/plot-functions.R: fixed bug in plotSquareHeatmap() that caused 'plotModule' and 'plotData' to crash when specifying 'orderSamplesBy' when not all samples from the 'orderSamplesBy' dataset were present in the dataset to be rendered. Similarly, this fixes bugs in 'plotCorrelation' and 'plotNetwork' arising when setting 'symmetric=TRUE' when nodes from the discovery network were not present in the test dataset to be rendered. 2017-08-04 Scott Ritchie * DESCRIPTION: Release 1.1.3 * Fixed bug in module data heatmap legends: the range of values present on the legend is now restricted to the range of values present in the data heatmap when the data contains both positive and negative values. Previously 'plotModule' and 'plotData' would show legends for a range of values centered at 0. * Fixed bug in module data heatmap legends: the range of values shown now respects the values passed by the user in the 'dataRange' argument. * Fixed bug in gradient legends: the axis ticks at the lower or upper end of the range specified by the user now have add "<" and ">" to the labels to reflect values of that color may be outside the range shown. * R/plot-functions.R: 'addGradientLegend' now keeps separate track of the range of values present in the data, the range of values to be shown on the legend, and the range of values occupied by the color gradient used in the heatmap. * R/plot-functions.R: 'addGradientLegend' an axis tick will always be placed at 0 on the legend if the range of values on the corresponding heatmap can be both negative and positive. * R/plot-functions.R: extra argument add to 'plotSquareHeatmap' and 'plotTriangleHeatmap' to reflect separation of heatmap legend range from the range of values occupied by the color palette used in the heatmap. Calls to these functions from 'plotModule', 'plotData', 'plotCorrelation', and 'plotNetwork' have been updated accordingly. * R/plot-module.R: added additional logic to 'plotModule' to separately determine the range of values present in the data, those to be shown on the data heatmap legend, and the range of values the heatmap palette occupies. * R/plot-topology.R: as above but for the 'plotData' function. 2017-08-03 Scott Ritchie * DESCRIPTION: Release 1.1.2 * Clarified documentation for "backgroundLabel" argument used in many package functions to describe how to disable the behaviour. 2017-08-03 Scott Ritchie * DESCRIPTION: Release 1.1.1 * R/plot-module.R: Fixed bug where 'plotModule' would crash when attempting to plot a single module where only 1 node was present in the test dataset (e.g. where only 1 of the module's genes is expressed in the test cell-line). * R/plot-topology.R: Fixed same bug as above for 'plotCorrelation', 'plotNetwork', and 'plotData' 2017-07-03 Scott Ritchie * DESCRIPTION: Release 1.1.0 * R/modulePreservation.R: 'nPerm' argument may now be set to 0 to calculate the module preservation statistics across two datasets while skipping the permutation test procedure. * src/permutations.cpp: reorganised code to return the observed test statistics without performing the permutation test procedure when the 'nPerm' argument in modulePreservation() is set to 0. * src/permutationsNoData.cpp: reorganised code to return the observed test statistics without performing the permutation test procedure when the 'nPerm' argument in modulePreservation() is set to 0. * man/modulePreservation.Rd: added description detailing the use case for 'nPerm=0'. * man/*: minor changes in all man files due to generation with roxygen 6.0.1 instead of roxygen 5.0.1. 2016-11-15 Scott Ritchie * DESCRIPTION: Release 1.0.4 * Restructured and curated documentation: Rd files are no longer generated for internal functions, and functions intended to extend basic package usage have been removed from the package manual. 2016-10-20 Scott Ritchie * DESCRIPTION: Release 1.0.3 * README: updated to reflect publication in Cell Systems and package publication on CRAN (as of version 1.0.2). * R/NetRep.R: fixed spelling mistake in package documentation. * man/NetRep.Rd: compilation of R/NetRep.R spelling fix. * R/check-user-input.R: in processInput() improved handling of user input arugments 'modules', 'moduleAssignments', 'backgroundLabel' and 'test' to handle cases where user code has overwritten a necessary list element with 'NULL'. * R/check-user-input.R: numeric labels in 'modules' are now returned as "character", fixing a bug where the incorrect modules would be reported in the 'contigency' table list element returned by contigencyTable() in R/contingency.R * R/contingency.R: fixed bug where ratios returned in the 'propVarsExplained' list element by contingencyTable() were greater than 1 due to mismatch in in module order of the vectors when calculating the ratios (one ordered as text, one numerically). * R/contingency.R: improved the 'contingency' table list element returned by contigencyTable() so that it now reports module sizes in their discovery datasets, the number of nodes present in the other dataset in the contigency table comparison, and includes modules with no nodes present in the other dataset. * R/networkProperties.R: in netPropsInternal() improved progress reporting messages so that the module discovery dataset is listed as well as the dataset in which the module properties are being calculated.