writeDataFrame() of AromaUnitSignalBinaryFile and AromaUnitTabularBinaryFile.Removed defunct method whatDataType().
The segmentByMPCBS() method does not work in R (>= 4.2.0). This is because the mpcbs package has _R_CHECK_LENGTH_1_CONDITION_ and_R_CHECK_LENGTH_1_LOGIC2_ bugs, which give errors in R (>= 4.2.0). The mpcbs package is a legacy package that is no longer maintained. Because of this, it is likely that segmentByMPCBS() and support for MPCBS segmentation will eventually be dropped from aroma.core.
downloadFile() for AromaRepository is now smarter about *.gz files.Re-export S3 generic process() from R.rsp.
Argument pattern for findByGenome() for AromaGenomeTextFile has an sprintf() mistake causing it to produce a warning in R (>= 4.1.0).
WORKAROUND: Package fails R CMD check with _R_CHECK_LENGTH_1_LOGIC2_=true due to PR#17663 (https://bugs.r-project.org/bugzilla/show_bug.cgi?id=17663). Made a simple code tweak to avoid this bug hitting us.
WORKAROUND: There’s a _R_CHECK_LENGTH_1_LOGIC2_ bug in mpcbs::mpcbs.mbic(). Until fixed, we need to disable this check while testing segmentByMPCBS().
findByChipType() for AromaMicroarrayTabularBinaryFile and TextUnitNamesFile would only recognize Windows Shortcut Links with a filename extension in lower case (*.lnk) but not in upper case (*.LNK).extract() from R.utils instead of R.filesets, which would result in R CMD check errors for aroma.affymetrix with R-devel (>= 3.6.0).getRegions() for CopyNumberSegmentationModel used future strategy eager that is defunct in future (>= 1.6.0) - updated to sequential.Now file sizes are reported using IEC binary prefixes, i.e. bytes, KiB, MiB, GiB, TiB, …, YiB.
Now genomic positions / distances are reported in ‘Mbp’ (was ‘MB’ and ‘Mb’).
Objects no longer report on memory (RAM) usage.
Package used %<=% internally with was deprecated in future (>= 1.4.0).
exportTotalCnRatioSet() for AromaUnitTotalCnBinarySet and exportFracBDiffSet() for AromaUnitFracBCnBinarySet would return files with file names matching .asb, which would incorrectly also include .asb.md5 files. Thanks to Thomas Grombacher at the Merck Group for the report.
Default method whatDataType() is now defunct.
Argument .old of getChipType() for AromaMicroarrayTabularBinaryFile is now defunct (was deprecated in 2008!).
Previously deprecated argument maxNAFraction to getRawCnData() of CopyNumberChromosomalModel and for fit() of CopyNumberSegmentationModel is now defunct.
Removed previously defunct apply() for SampleAnnotationFile.
ROBUSTNESS: Using do.call(fcn) internally instead of do.call("fcn").
ROBUSTNESS: Using getPathname() internally instead of private field.
getOutputDataSet(..., onMissing = "dropall") for AromaTransform would throw error on no applicable method getFullNames() for NULL.apply() for SampleAnnotationFile is now defunct; use applyTo().
Drop defunct methods.
apply() for SampleAnnotationFile to applyTo().ROBUSTNESS: Explicitly importing core R functions.
CLEANUP: One occurance of rowMads(..., centers) to rowMads(..., center).
CLEANUP: Drop prototype code never used.
writeRegions(..., format = "wig") gave an error.Deprecated internal whatDataType() method.
Previously deprecated methods now defunct: nbrOfArrays() for readData() for AromaMicroarrayDataSetTuple, AromaMicroarrayDataSet and AromaUnitTotalCnBinarySet, as well as readData() for SampleAnnotationFile.
Removed previously (Feb 2014) defunct method patchPackage() and patch().
ROBUSTNESS: exportAromaUnitPscnBinarySet() asserts that no duplicated files are returned.
getOutputDataSet() for AromaTransform gives more informative error messages.
getOutputDataSet() for AromaTransform would sometimes given an error due to the same output file being identified more than once, for instance, when the full name of an output file is also the prefix of another file. Thanks to Benilton Carvalho (Universidade Estadual de Campinas, Sao Paulo, Brazil) for reporting on this.
extractRawMirroredAlleleBFractions() for RawAlleleBFractions gave Error in setSignals(res, dh) : object 'res' not found.
display() for Explorer would try to open a pathname in a way that only worked on MS Windows. Thanks to Sunghee Oh (S. Korea) for reporting on this.aromaSettings are loaded when the packages is loaded. Previously the package had to be attached.fit2d() for matrix utilizes use() for aroma.light.writeDataFrame(..., columnNamesPrefix="none") for AromaUnitSignalBinaryFile would give an error.getAromaUflFile() for AromaPlatformInterface.Move byChipType() from classes AromaCellCpgFile and AromaCellPositionFile to superclass AromaCellTabularBinaryFile.
Package does less package reordering.
SampleAnnotationSet$loadAll() would give an error if annotationData/samples/ didn’t exist or did not contain any SAF files.getAromaUgpFile() for AromaUnitSignalBinaryFile.Added SuggestsNote field to DESCRIPTION with list of packages that are recommended for the most common use cases.
Bumped package dependencies.
ds[[idx]] instead of getFile(ds, idx) where possible.<ns>::<fcn>() calls with repetive <fcn>() calls; the :: operator is fairly expensive.Previously deprecated downloadPackagePatch(), patchPackage() and patch() for AromaPackage are now defunct.
Removed defunct methods.
readDataFrame() for AromaGenomeTextFile and AromaUcscGenomeTextFile explicitly passed arguments ... to NextMethod(), which would cause them to be duplicated in certain cases.R.rsp::rfile() for compiling RSP files instead of the to-be-deprecated rspToHtml().getGenericSummary() for RichDataFrame would throw Error in data[i, , drop = FALSE] : subscript out of bounds if it had zero rows.findChangePointsByState() for SegmentedGenomicSignalsInterface.exportAromaUnitPscnBinarySet().onMissing to getOutputDataSet() for AromaTransform. This replaces argument incomplete, which will be dropped in a future version, but is for now supported as well.Minor adjustments of NAMESPACE imports.
Updated package dependencies.
colMedians() for AromaTabularBinaryFile. If needed, there is always colStats().Now explicitly importing only what is needed in NAMESPACE.
BACKWARD COMPATIBILITY: For R (< 3.0.2) only the maximum of 500 S3 methods are declared in NAMESPACE.
Updated package dependencies.
chromosomes to ChromosomalModel() and setChromosomes() for the same class. If not specified, the default is as before to infer the set of chromosomes from the UGP files.maxNAFraction when setting up the model.Now the aroma.core Package object is also available when the package is only loaded (but not attached).
Now only importing only what is needed from the R.utils package.
Now only importing only what is needed from the matrixStats package.
ROBUSTNESS: Forgot to import R.methodsS3::appendVarArgs().
Updated package dependencies.
Backward-compatibility code for obsolete EBImage was no longer needed.
Updated package dependencies.
exportAromaUnitPscnBinarySet() for AromaUnitTotalCnBinarySet would throw an error on unknown argument 'names' to indexOf().Replaced all internal x11() with dev.new() calls.
Moved more packages from Depends to Imports.
Updated package dependencies.
findFilesTodo() for AromaTransform.Now it’s possible to call aroma.core::doCBS() without loading package.
Turned doCBS() into a default method, cf. aroma.affymetrix::doRMA().
help("doCBS") documents all doCBS() methods.rm() calls with NULL assignments.getChecksum() instead of (internal) R.filesets::digest2() with identical results.extractRawCopyNumbers() for CopyNumberChromosomalModel would throw Error in UseMethod("getChecksum") : no applicable method for 'getChecksum' applied to an object of class "list" if (and only if) R.cache package was not attached. Thanks Yadav Sapkota at University of Alberta for reporting on this.CRAN POLICY: Now all Rd \usage{} lines are at most 90 characters long.
CRAN POLICY: Now all Rd example lines are at most 100 characters long.
Bumped version for CRAN submission.
No updates.
Package no longer try to apply package patches, which was only possible when namespaces where not used.
Package no longer depends on R.cache (only imports it).
is(Homo|Hetero)zygote() and getPhysicalPositions() defunct.sqrtsign() transform. This makes it possible to plot log2 PLM residuals (in aroma.affymetrix) on a blue-to-red scale using setColorMaps(.., "log2,sqrtsigned,bluewhitered") where bluewhitered <- colorRampPalette(c("blue", "white", "red")).Bumped the package dependencies.
Added an Authors@R field to DESCRIPTION.
nbrOfArrays() for AromaMicroarrayDataSet and AromaUnitTotalCnBinarySet.No updates.
Bumped version for CRAN submission.
Utilizing new startupMessage() of R.oo.
Bumped up package dependencies.
Now using argument colClasses (was colClassPatterns) for all readDataFrame():s.
getOneFile() internally.inst/reports/templates/html/archive/, inst/reports/templates/rsp/archive/ and inst/archive/, which all contained old/legacy Explorer versions.lapply() for AromaTabularBinaryFile to colApply().Made several file classes implement FileCacheKeyInterface.
Added CacheKeyInterface and FileCacheKeyInterface.
getRam() and setRam() for AromaSettings did not use memory/ram.Bumped version for CRAN submission.
No updates.
Now getParametersAsString() handles sets of parameters as well.
Added getParameterSets() to ParametersInterface.
Dropped getHeaderParameters() for TextUnitNamesFile, which is now done by getHeader() of TabularTextFile.
Package only depends on digest indirectly via R.cache.
Moved internal .assertDigest() to R.cache.
Now using getName() instead of deprecated getLabel().
Properly declared all cached fields, making it possible to remove nearly all clearCache() implementations because the one for Object takes does the job.
The RSP template for ChromosomeExplorer’s setupExplorer.js had a duplicated entry.
seq_along(x) instead of seq(along = x) everywhere. Similarly, seq(ds) where ds is GenericDataFileSet is now replaced by seq_along(ds). Likewise, seq_len(x) replaces seq(length = x), and length(ds) replaces nbrOfFiles(ds).getColumnNames() to getDefaultColumnNames() for all classes inheriting from GenericTabularFile, because of the new ColumnNamesInterface interface.whichVector() with which(), because the latter is now the fastest again.units to readDataFrame() for AromaUnitChromosomeTabularBinaryFile.Arguments$get(Read|Writ)ablePath() instead of filePath(..., expandLinks = "any").Added argument maxNAFraction to CopyNumberChromosomalModel, which is now the prefer place to specify it, instead of to fit() etc.
Now CopyNumberChromosomalModel() accepts references of type "none", "constant(1)", "constant(2)", and "median", where "none" and "constant(1)" are identical, "constant(2)" uses reference signals that are exactly 2, and "median" uses reference signals are equals the robust average across all samples. For backward compatibility, NULL is still supported, which equals "median".
Arguments$getWritablePath() everywhere instead of mkdirs(), because the former will do a better job in creating and asserting directories on slow shared file systems, and when it fails it gives a more informative error message.Some of the HTML and CSS errors in ArrayExplorer and ChromosomeExplorer that were detected by the W3’s online validation services were correct.
A few errors in the JavaScript of ArrayExplorer has been corrected. It also does a better job of inferring the height of the displayed spatial image. Thanks Laurent Malvert (ID Business Solutions Ltd; IDBS) for the report, troubleshooting and suggestions on this.
Object methods that calls “next” methods, utilizes NextMethod(), which became possible with R.oo v1.10.0.NextMethod(), cf. R-devel thread ‘Do not pass’…’ to NextMethod() - it’ll do it for you; missing documentation, a bug or just me?’ on Oct 16,
trim() being overridden by ditto from the IRanges package, iff loaded.exportAromaUnitPscnBinarySet() for a list.exportAromaUnitPscnBinarySet() for AromaUnitTotalCnBinarySet would throw an error if the files of the exported data set was ordered in a non-lexicographic order.downloadUGC() for AromaRepository.interleave() for Image and RasterImage could generate error 'Error in if (abs(vRatio) > abs(hRatio)) { : missing value where TRUE/FALSE needed if odd/even cell intensity columns (rows) contained all NAs.downloadCDF() for AromaRepository downloads CDFs regardless of *.cdf and *.CDF.escape to findAnnotationData(), which causes such symbols that exist in argument pattern to be automatically be escaped.colBinnedSmoothing() is done over zero-length bins, the output for those bins will be NA.colBinnedSmoothing(..., xOut = xOut) would return binned values in the incorrect order, iff xOut was not ordered. Added a systems test for this case.
colBinnedSmoothing(..., xOut = xOut) could generate bins at the ends that did not contain the outer most xOut values.
ROBUSTNESS: Now the package startup message is outputted at the very end. Previously it was done before some additional setup, which could potentially fail.
ROBUSTNESS: fitWRMA() and fitWHRCModel() now call rcModelWPLM() of preprocessCore instead of .Call(..., PACKAGE = "preprocessCore"). Doing so will somewhat slow down the speed, because rcModelWPLM() introduces some non-needed validation of the arguments. This caused their arguments psiCode and psiK to be dropped.
Added exportAromaUnitPscnBinarySet() for AromaUnitTotalCnBinarySet.
Added alpha versions of classes AromaUnitPscnBinary(File|Set).
Added getNumberOfFilesAveraged() to AromaUnitSignalBinaryFile. Formerly only for AromaUnitTotalCnBinaryFile.
byChipType() for several annotation data file classes.calculateRatios to CopyNumberChromosomalModel().CLEANUP: Package now only depends/imports on CRAN package, which simplifies the installation of aroma.core, and indirectly any packages that depends on it. Note, there are still “suggested” non-CRAN packages, which are required by some/optional methods.
CLEANUP: Package no longer depends on aroma.light, but instead it “suggests” it. This is possible, because now weightedMad() of aroma.light has been added to matrixStats v0.5.0.
CLEANUP: Package no longer “Suggests” affxparser. It was needed because of the findFiles() function, which has been added to R.utils v1.13.1.
Turned weightedMad() into a default method, added help documentation with example code.
CLEANUP: Moved (internal) swapXY() and draw() for density objects to R.utils (>= 1.10.0).
[() for RichDataFrame would loose the class attribute, unless argument drop was FALSE.Now colBinnedSmoothing(), colKernelSmoothing() and colGaussianSmoothing() returns a matrix with column names as in argument Y.
For RawGenomicSignals, added argument fields to binnedSmoothing() and binnedSmoothingByField(). Also added argument field to getSignals() and estimateStandardDeviation().
reports/includes/js/ directory installed by ArrayExplorer and ChromosomeExplorer.Added binnedSmoothingByField() for RawGenomicSignals.
Added subset() for RawGenomicSignals.
Added get- and setVirtualField() for RawGenomicSignals.
Added argument sort to as.data.frame() for RawGenomicSignals.
ROBUSTNESS: Now all writeDataFrame() methods will add quotation marks around column names that contain a comment character (#). This will make the written file readable with read.table() even when there are header comments with prefix #. This issue was reported by Yu Song at Oakland University.
Now createImage() for matrix creates images of type "png::array" by default (used to be "EBImage::Image"). If the latter, then at least EBImage v3.9.7 (July 2011) is required.
write() for RasterImage would write truncated intensities, because we forgot to rescale [0,65535] to [0,1] intensities.ROBUSTNESS: Added system tests for RawGenomicSignals classes.
CLEANUP: Now R CMD check no longer reports that the package is using .Internal() function calls. This was due to how we adjusted colSums() and colMeans() to become generic functions.
Preparing for RawGenomicSignals not being reference variables.
Preparing for supporting multiple-chromosome RawGenomicSignals.
GENERALIZATION: Now binnedSmoothing() of RawGenomicSignals default to generate the same target bins as binnedSmoothing() of a numeric vector. Before the bins had to be specified explicitly by the caller.
GENERALIZATION: Now it is possible to call colBinnedSmoothing() with an empty set of input loci, but still requesting a set of output loci, which then will be all missing values.
x of colBinnedSmoothing() would default to the incorrect number of loci.require.js in v2.4.4, which is needed by the ChromosomeExplorer v3.3.setup(ce, force = TRUE) where ce is your ChromosomeExplorer object. Thanks to Keith Ching at ConsultChing for the initial troubleshooting [http://consultching.com/root/?p=64] leading me on the right track on how update ChromosomeExplorer.findPngDevice() tries all the available settings for argument “type” of png(), for the current platform. This will increase the chances for finding a PNG device that really works.findPngDevice() uses isPackageInstalled("Cairo") instead of require("Cairo") to avoid loading Cairo if not really used.findPngDevice() memoizes the results throughout the current session.hg17 to Human,cytobands,hg17,GLADv2.4.0,HB20100219.txt in annotationData/genomes/Human/.drawCytoband() for ChromosomalModel failed to locate the genome annotation data file containing cytoband information, e.g. Human,cytobands,<tags>.txt. Thanks to Kai Wang at Pfizer for reporting on this.base::serialize(), because it was only applied to R (< 2.12.0) anyway and this package now requires R (>= 2.12.0).ROBUSTNESS: Aroma settings are no longer loaded during R CMD check.
Updated package dependencies.
library() would always return an invisible() object, even if base::library() wouldn’t. This caused plain library() to not list installed packaged. Thanks Venkat Seshan at MSKCC for reporting on this.Now colBinnedSmoothing() handles an unordered xOut.
Added argument units to writeDateFrame() for AromaUnitSignalBinarySet to make it possible to write any subset of units and in any order, e.g. genome order.
Now it is possible to fully specify the location and the width of each bin used by colBinnedSmoothing(), which now also returns the bin counts as part of the attributes. Moreover, the bins are now defined to be strictly disjoint using boundaries [x0,x1) instead of [x0,x1] as before.
Added plotCoverageMap() for AromaUgpFile.
getDefaultExtension() for UnitAnnotationDataFile, which guesses the filename extension from the class name, unless overridden by a subclass.Now byChipType() for UnitAnnotationDataFile and derivatives give an error message with more information on which file it failed to locate, e.g. by specifying filename extension it looked for.
CLEANUP: Now getFullNames() for AromaMicroarrayDataSetTuple no longer produces a warning on is.na() applied to non-(list or vector) of type 'NULL'.
exportTotalCnRatioSet() for AromaUnitTotalCnBinarySet and exportFracBDiffSet() for AromaUnitTotalCnBinarySet tried to call cat(verbose, x) with length(x) > 1.process() of ChromosomeExplorer would throw Error in file(file, ifelse(append, "a", "w")) : [...] cannot open file '/ [...] (new Array(': No such file or directory. Importing the R.rsp namespace in the aroma.core namespace solved this. Thanks Qian Liu for reporting on this.extractPSCNArray() for AromaUnitTotalCnBinary{File|Set}.readHeader(), readRawFooter() and writeRawFooter() of AromaTabularBinaryFile would try to read non-signed 4-byte integers, which is not supported and would instead be read as signed integers. From R v2.13.1 this would generated warnings.getSnpPositions() of AromaCellSequenceFile would throw Error: length(pos) == ncol(cells) is not TRUE. However, it was not until aroma.core v2.1.3 (2011-08-01) was release that some people got problem with this. It turns out that the sanity check catches an error in how getSnpPositions() of AromaCellSequenceFile allocates the result vector, a bug that has been there for a very long time. Luckily, this bug has had no effect on the results for anyone. Thanks to David Goode (Stanford) and Irina Ostrovnaya (MSKCC) for reporting on this.cat() and getOption() of R.utils (instead of ‘base’).getSnpNucleotides() for AromaCellSequenceFile.hpaste() internally wherever applicable.allocateFromUnitAnnotationDataFile() for AromaUnitTabularBinaryFile would include chip type tags in the path, e.g. annotationData/chipTypes/GenomeWidesSNP_6,Full.Now ChromosomeExplorer does a better job of listing chromosomes that are specific to the genome/organism used.
Added getChromosomeLabels() for ChromosomeExplorer.
lapplyInChunks(idxs) for numeric did not correctly handle the case when length(idxs) == 0, because of a typo.Updated the default filename patterns used by findByGenome() for AromaGenomeTextFile to "^%s,chromosomes(|,.*)*[.]txt$".
GENERALIZATION: Now findByGenome() for AromaGenomeTextFile follows the new aroma search conventions.
GENERALIZATION: In addition to search <rootPath>/<set>/<name> paths, findAnnotationData() can also search <rootPath>/<set>/ by not specifying argument name (or setting it to NULL). Also, if it cannot locate any files, it falls back to annotation data available in any of the ** aroma.* ** packages.
Now getGenomeFile() for ChromosomalModel utilizes byGenome() for AromaGenomeTextFile to locate the genome annotation file.
Added static loadAll() for SampleAnnotationSet.
getUnitsOnChromosomes() for AromaUnitChromosomeTabularBinaryFile truly returns valid unit indices. Thanks to Emilie Sohier, France for reporting on a problem related to this.STANDARDIZATION: Now the default output path for all allocateFromUnitAnnotationDataFile() is annotationData/chipTypes/<chipType>/. Before it was the same directory as the original annotation data file, which may for instance have been in a deeper subdirectory, or recently also in a sibling root path.
GENERALIZATION: getAverageFile() for AromaUnitTotalCnBinarySet first searches for an existing result file according to the new search conventions. If not found, then it’s created.
GENERALIZATION: Now the default for createImage() for matrix is to test to create images according to aroma settings option output/ImageClasses.
ROBUSTNESS: Now getAverageFile() for AromaUnitTotalCnBinarySet creates the result file atomically by writing to a temporary file which is renamed afterward.
Now write() and read() for RasterImage throws an informative error message explaining that the png package is needed.
colorize() for Image would throw <simpleError in ...: could not find function "colorMode">, because the colorMode() function needs to be explicitly imported after the recent package cleanups.
colorize() for Image tried to call createImage() using a vector instead of a matrix.
createImage() for matrix would not return the first possible image created (when testing different image classes) but instead continue trying to create image for all possible classes. For instance, this meant that although you had the **EBImage** package installed, but not the png package, it would still in the end try to (also) use png package. If writing PNG images to file, say via ArrayExplorer, this would result in Error in loadNamespace(name) : there is no package called 'png'. Thanks Richard Beyer at University of Washington for reporting on this.
findAnnotationData() to be annotationData/ and annotationData,<tags>/.getTags() to Arguments to R.filesets v0.9.3.Deprecated static method importFromTable() for FileMatrix.
Removed several deprecated methods.
Submitted to CRAN.
No updates.
getRawCopyNumbers() for CopyNumberSegmentationModel, when there are too many non-finite signals, gives a more informative error message containing information also on which sample, reference and chromosome the problem was detected on.fit() for CopyNumberSegmentationModel passes down argument maxNAFraction to the internal sanity test as it used to do before aroma.core v1.3.4 (November 2009).GENERALIZATION: Now spatial PNG image files can also be created utilizing the png package as an alternative to the EBImage package. The EBImage package can be tricky to install on some OSes, e.g. MS Windows 7 64-bit.
Added alpha version of an internal RasterImage class.
writeImage() for Image objects as defined by the EBImage package. This made it possible to remove the remaining explicit dependencies on EBImage in the aroma.affymetrix package.Deprecated internal as.TrueColorImage() for Image and for matrices.
Removed deprecated internal rgbTransform() for Image.
digest() uses the faster serialize() of R v2.12.0. The earlier version was orders of magnitude slower on MS Windows, which we have been patching in the aroma framework since v0.9.3.4. That patch wrote to file, which was slower than the recent fix. This Windows-only patch is no longer needed, and hence neither applied, on R v2.12.0 and beyond.flavor to segmentByGLAD() for RawGenomicSignals, which makes it possible to specify whether daglad() or glad() of the GLAD package will be used. The flavor argument can also be passed as for instance GladModel(..., flavor = "daglad").Mouse,chromosomes.txt to system.file("annotationData/genomes/Mouse", package = "aroma.core").ROBUSTNESS: Whenever the GLAD package is tried to be loaded and it fails due to failure of loading a shared library, then this will be remembered for 24 hours while it will not be loaded again.
Added requireWithMemory().
plot() for CopyNumberSegmentationModel would throw exception Cannot infer number of bases in chromosome. No such chromosome: 25 for chromosome 25.
drawCytoband2() would throw an error if argument cytoband was an empty data frame, which could happen if there is no cytoband annotation data for the requested chromosome. Now it returns quietly.
ROBUSTNESS: Now subsample() for BinnedScatter utilizes resample().
CLEANUP: Removed stext(). It is in R.utils.
CLEANUP: Removed some outdated patches for R v2.7.0 and before.
CLEANUP: Package no longer need to “suggest” geneplotter.
Now fit() for CopyNumberSegmentationModel also passed the optional arguments (...) passed to the constructor function. This makes it possible to specify all arguments while initiating the model, e.g. sm <- CbsModel(..., min.width=5, alpha=0.05).
Now optional arguments ... to CopyNumberChromosomalModel are recorded.
drawCytoband2() used to annotate chromosomal plots with cytobands tries to utilize GLAD package, if available. However, even when GLAD is installed it may still be broken due to missing dynamic libraries, e.g. Error in library.dynam(lib, package, package.lib) : DLL GLAD not found: maybe not installed for this architecture?. We now avoid this too.Added drawDensity() for CopyNumberRegions and RawGenomicSignals.
Added getDensity() for CopyNumberRegions which returns the empirical density of the mean levels weighted by the region lengths.
Added getLength() for CopyNumberRegions, which returns a vector of region lengths.
Now the ylim defaults to c(0,5) for plot() for RawCopyNumbers.
Added basic support for operators +, -, and * to RawGenomicSignals.
Added the AromaGenomeTextFile class.
Added genome annotation file Canine,chromosomes,UGP,HB20100822.txt to system.file("annotationData/genomes/Canine", package = "aroma.core").
annotationData/genomes/Human/Human,chromosomes.txt had an extra TAB on the ChrM row.Added abstract classes SegmentationDataFile and SegmentationDataSet to represent segmentation data results. Current subclasses are CbsSegmentationData(File|Set).
Added more utility methods for CopyNumberRegions.
xRange(), prune(), simulateRawCopyNumbers(), +(), -() and *().writeDataFrame() for AromaUnitTotalCnBinarySet and AromaUnitFracBCnBinarySet to get the correct filename extension. Thanks Nicolas Vergne at the Curie Institute for reporting this.Added subset() for CopyNumberRegions.
Now extractRegion() for RawGenomicSignals also accepts a CopyNumberRegions object for argument regions.
Added extractRegions() for RawGenomicSignals.
writeDateFrame() for AromaUnitSignalBinarySet would write the same data chunk over and over.indexOf() for ChromosomalModel would return NA if a search pattern contained parenthesis ( and ). There was a similar issue in indexOf() for GenericDataFileSet/List in R.filesets, which was solved in R.filesets 0.8.3. Now indexOf() for ChromosomalModel utilizes ditto for GenericDataFileSet for its solution.as.GrayscaleImage(..., transforms = NULL) for matrix would throw Exception: Argument transforms contains a non-function: NULL.updateDataColumn() of AromaTabularBinaryFile would censor signed integers incorrectly; it should censor at/to [-(n+1),n], but did it at [-n,(n+1)] (“two’s complement”). This caused it to write too large values as n+1, which then would be read as -(n+1), e.g. writing 130 would be censored to 128 (should be 127), which then would be read as -128. Added more detailed information on how many values were censored. Thanks Robert Ivanek for report on this.Added trial version of fullname translator files.
doCBS() for character:s support data set tuples.
Added doCBS() for CopyNumberDataSetTuple:s.
R CMD check on R v2.11.0 and v2.12.0 devel.path of all writeDataFrame() methods defaults to <rootPath>,txt/<dataSet>/<chipType>/.Added writeDataFrame() for AromaUnitTabularBinaryFile.
Added argument columnNamesPrefix to all writeDataFrame().
AD HOC: Now getUnitAnnotationDataFile() of AromaPlatformInterface load aroma.affymetrix if needed and if installed.
writeDataFrame() for AromaUnitSignalBinary{File|Set}.getFitFunction() of CbsModel would return a function that would give Error in segmentByCBS.RawGenomicSignals(..., seed = seed): formal argument "seed" matched by multiple actual arguments.Added equals() for CopyNumberRegions.
Added argument seed to CbsModel, which will, if specified, set the random seed (temporarily) each time (per sample and chromosome) before calling the segmentation method.
seed to segmentByCBS() for RawGenomicSignals.CHANGE: Argument arrays of doCBS() for CopyNumberDataSet no longer subset the input data set, but instead is passed to the fit() function of the segmentation model. This way all arrays in the input data set are still used for calculating the pooled reference.
Now segmentByMPCBS() can be used to segment data from a single source.
segmentByMPCBS() for RawGenomicSignals.verbose of getOutputSet() of ChromosomalModel.Submitted to CRAN.
No changes since v1.4.7.
Package passes R CMD check on R v2.10.1 patch and R v2.11.0 devel.
Package passes all redundancy tests.
Argument zooms of ChromosomeExplorer now default to 2^(0:6), instead of 2^(0:7), because the images for zoom 2^7=128 would not display in Firefox for the largest chromosomes.
Updated getGenomeFile() for ChromosomalModel such that it can be used to locate other types of genome annotation files as well, files that may be optional (without giving an error).
Added Human cytoband files to annotationData/genomes/.
Added annotationData/genomes/ to aroma.core (from aroma.affymetrix).
Added getOutputSet() for ChromosomalModel.
Added alpha version of a doCBS().
SIMPLIFICATION: Now it is possible to use ChromosomeExplorer without having to install the GLAD package (used before for cytobands).
SIMPLIFICATION: Now it is possible to plot cytobands without having to install the GLAD package.
print() statements in isCompatibleWith().R CMD check and all redundancy tests.getChromosomes() for class AromaUnitChromosomeTabularBinaryFile validating that the file has a chromosome column.Arguments$getTags() where ever possible.ChromosomalModel:s (and and the ChromosomeExplorer) no longer require unit names files.
getListOfAromaUgpFiles() for ChromosomalModel no longer goes via getListOfUnitNamesFiles(). This opens up the possibility to work with data files without unit names files, e.g. smoothed CN data.
Added getAromaUgpFile() for AromaPlatformInterface.
getUnitNamesFile() for AromaPlatformInterface utilizes the generic getUnitAnnotationDataFile() method.defaults to allocate() of AromaTabularBinaryFile.segmentByMPCBS() via the mpcbs package.R CMD check on R v2.10.1 and R v2.11.0 devel and all of its redundancy tests.getDefaultFullName() for AromaMicroarrayDataSet and AromaTabularBinarySet in order to override (parent = 1) the new default of GenericDataFileSet in R.filesets v0.7.0, which would return the chip type (parent = 0) as the name.getPairedNames() to CopyNumberChromosomalModel which returns combined <test>vs<ref> names, e.g. TumorvsNormal. The default behavior is that if <ref> is the average of a pool, then the paired name is <test>, e.g. Tumor.Added argument pattern to byName() for AromaUnitTotalCnBinarySet and AromaUnitFracBCnBinarySet.
getSets() of AromaMicroarrayDataSetTuple overrides the default method by adding the chip types as the names of the returns list.
ROBUSTNESS: Using new Arguments$getInstanceOf() were possible.
ROBUSTNESS: Now CopyNumberChromosomalModel() asserts that none of the test samples have duplicated names.
ROBUSTNESS: Now all index arguments are validated correctly using the new max argument of Arguments$getIndices(). Before the case where max == 0 was not handled correctly.
getSetTuple() from ChromosomeExplorer.translateFullNames() of ChromosomeExplorer would translate the full names, but return the original ones.
getFullNames() of ChromosomeExplorer reported: Error in UseMethod("translateFullNames"): no applicable method for translateFullNames applied to an object of class "character".
extractMatrix() of AromaUnitCallFile did not recognize NoCalls.extractRawCopyNumbers() for RawCopyNumbers would give an error if the internal logBase was NULL.xScale = 1e-6 by default.Added extractRawCopyNumbers(, ...logBase = 2) to RawCopyNumbers, which can be used to change the logarithmic base of CN ratios, if any.
Added getAverageFile() for AromaUnitTotalCnBinarySet.
Moved ChromosomeExplorer to aroma.core from aroma.affymetrix.
Moved several method for profileCGH objects to aroma.core from aroma.affymetrix.
ALPHA: First successful run of segmentation with CbsModel with AromaUnitTotalCnBinarySet data sets.
Added isAverageFile() for AromaUnitSignalBinaryFile.
Added class AromaUnitTotalCnBinarySetTuple.
Added interfaces CopyNumberDataFile, CopyNumberDataSet, and CopyNumberDataSetTuple (internal for now).
GENERALIZATION: Moved CbsModel, GladModel and HaarSegModel (all segmentation models) of aroma.affymetrix to here.
GENERALIZATION: Moved all (generalized) ChromosomalModel, CopyNumberChromosomalModel, and CopyNumberSegmentationModel from aroma.affymetrix to here.
getAromaUflFile() to UnitAnnotationDataFile.Now AromaUflFile$allocateFromUnitNamesFile() works.
importFrom(ufl, dat, ...) where ufl was an AromaUflFile and dat a GenericTabularFile gave an error reporting that importFromGenericTabularFile() is abstract/not defined for AromaUflFile:s. Thanks Paolo Guarnieri for reporting this.
New public release.
More recent dependencies on Bioconductor packages.
R CMD check on R v2.10.0 and all redundancy tests.getExtensionPattern() to most GenericDataFile classes.setName() for RawGenomicSignals.CLEANUP: Updated to use byPath() instead fromFiles().
CLEANUP: Move the Interface class to the R.oo package.
R CMD check on R v2.9.2 and all redundancy tests.getUnitsOnChromosomes() for AromaUnitChromosomeTabularBinaryFile returns a vector by default (unlist = TRUE).Added yRange(), yMin() and yMax() for RawGenomicSignals.
Added extractRawCopyNumbers() for RawSequenceReads.
smoothScatter(), which is used in for instance aroma.affymetrix, is in the graphics package from R v2.9.0. It was previously in geneplotter v1.2.4 and before(!). The code imported it from geneplotter would therefore give an error of a missing function with geneplotter v1.2.5. We are now assuming R v2.9.0 by default, but if not available, but geneplotter v1.2.4 or earlier is, geneplotter is loaded. If neither are available, a dummy smoothScatter() is setup reporting an error.totalAndFracBData/ to the search path of byName() for AromaUnit(FracB|Total)CnBinarySet.Added allocateFromUnitAnnotationDataFile() for AromaUnitSignalBinaryFile and AromaUnitTabularBinaryFile.
Added getUnitTypesFile() for AromaPlatform and AromaPlatformInterface.
Added UnitAnnotationDataFile, which now the UnitNamesFile and the new UnitTypesFile inherits from.
Now uses() for Interface takes multiple Interface:s.
Added class RawSequenceReads.
Added support for argument byCount to binnedSmoothingByState() of SegmentedCopyNumbers. It is rather complex how it works, but we tried to immitate how it works with byCount = FALSE.
Now binnedSmoothing() of RawGenomicSignals demark locus fields that were not binned to be regular fields. Ideally all locus fields (including custom ones) should be binned, but we leave that for a future implementation.
Added get-, setStateColorMap(), and getStateColors() for class SegmentedGenomicSignalsInterface.
Added argument keepUnits = FALSE to extractRawGenomicSignals() of AromaUnitSignalBinaryFile.
Now RawGenomicSignals(y = rgs) sets all locus fields in rgs if it is a RawGenomicSignals object.
examples() that requires “suggested” packages to be ran conditionally, so that it does not throw an error (in R CMD check) if the packages is not available. This is required for a package to be put on CRAN.exportTotalCnRatioSet() would return a AromaUnitFracBCnBinarySet.is(Homo|Hetero)zygous(), not is(Homo|Hetero)zygote().ADDED: Added support for “birdseed” encoding in extractGenotypes() and updateGenotypes() of AromaUnitGenotypeCallFile.
Added SegmentedGenomicSignalsInterface, which implements all the methods for SegmentedCopyNumbers and the new SegmentedAlleleBFractions.
Added getFields() to Interface as an ad-hoc solutions to avoid print(<Interface>) throwing Error in UseMethod("getFields") : no applicable method for "getFields".
updateGenotypes() of AromaUnitGenotypeCallFile is now much faster in counting A:s and B:s.isHomozygote() of AromaUnitGenotypeCallFile was not correct.
getOutputDataSet() of AromaTransform failed to identify the output files if (and only if) a filename translator was applied to the input data set.
New public release.
No updates.
R CMD check on R v2.9.0 and all redundancy tests.Added extractRawAlleleBFractions() for AromaUnitFracBCnBinaryFile.
Added exportFracBDiffSet() for AromaUnitFracBCnBinarySet.
ADDED: Now all segmentByNnn() for RawGenomicSignals also returns so called aromaEstimates, which are additional parameter estimates, e.g. robust estimates of the standard deviation of all full-resolution CNs as well as per identified region.
ROBUSTNESS: Now allocate() for AromaTabularBinaryFile first allocates a temporary file which is then renamed. This makes the file allocation more atomic, and therefore less error prone to interrupts etc.
ROBUSTNESS: Now indexOf() for UnitNamesFile assert that exactly one of the pattern and names arguments is given. It also gives an informative error message if pattern is a vector.
ROBUSTNESS: Now all arguments that are expected to be exactly one character string (not an empty vector or a vector of length two) are asserted to be that too. This closes some potential bugs.
EXCEPTION HANDLING: All methods that modifies an existing file, are now asserting that the file permissions allow modifying it, and if not gives a clear error message that this is the case.
Arguments$getNumerics(), which, in R.utils v1.1.5 and new, corces to doubles only if the input is not already doubles. This will lower the memory usage in some cases since an integers use 4 bytes and doubles 8 bytes. This also save disk space if object is saved to disk, e.g. a DNAcopy fit object with integer genomic positions instead of double ones, save approx 5-7% disk space.Moved the Explorer class and its support files under inst/ from aroma.affymetrix to aroma.core.
Moved the AromaCellPositionFile class from aroma.affymetrix to aroma.core.
CLEANUP: Moved all file set specific classes and methods to the new R.filesets package.
allocateFromUnitNamesFile() for AromaUnitSignalBinaryFile would not call generic allocate() but the one for this class.
extractMergedRawCopyNumbers(..., unshift = TRUE) would estimate the relative shifts between platforms using smoothed CNs over the genomic region defined by the first data set. Now it is done over the region defined by the union of all data sets. The impact of this bug should be neglectable or zero.
binnedSmoothing() of RawGenomicSignals uses weighted estimates (by default) if weights exists.Added RawAlleleBFractions() extending RawGenomicSignals.
Added extractCopyNumberRegions() for profileCGH, DNAcopy, and HaarSeg segmentation object. Before they were in aroma.affymetrix.
Added segmentByCBS(), segmentByGLAD(), and segmentByHarSeeg() to RawGenomicSignals, which all support weights if the underlying segmentation method does. To date, it is only CBS and HaarSeg that supports weights. These methods also support memoization.
Added extractRegion() to RawGenomicSignals.
Added getWeights() and hasWeights() to RawGenomicSignals.
Added {add,subtract,multiply,divide}By() for RawGenomicSignals all utilizing support method applyBinaryOperator().
ALPHA: Added classes GenericDataFile{Set}List.
ALPHA: Added class AromaUnitTotalCnBinaryFileList with methods extractRawCopyNumbers() and extractMergedRawCopyNumbers().
Now aroma.core works with the IRanges and grid packages, regardless on the order they were loaded.
getPlatform() of TextUnitNamesFile would sometimes return a list of length one, instead of an single character string.
extractSubset() of RawGenomicSignals did not recognize all locus fields.
Now getOutputDataSet() of AromaTransform returns a data set with files ordered such that the fullnames are ordered the same way as the input data set. Suggested by Xin (Victoria) Wang at UC Berkeley.
Now getOutputDataSet() of AromaTransform scans for the output data files with fullnames matching those of the input data set. This is one step closer to being able to batch process subsets of data sets when the method is, say, a single-array method.
UPDATE: Replace argument robust of colBinnedSmoothing() with a more generic FUN argument. The default arguments give identical results.
Now static fromFiles() of GenericDataFileSet supports empty data sets.
Now readDataFrame() of AromaUgpFile and AromaUnitTabularBinaryFile accepts rows = integer(0).
Added abstract class AromaUnitChromosomeTabularBinaryFile to make it easier to setup up new annotation file formats based on chromosomes. AromaUgpFile is now inheriting from this class. getUnitsOnChromosomes() now returns a list stratified by chromosome.
Added extractByChromosomes() for AromaUgpFile.
Added argument translate to getColumnNames() of TabularTextFile.
R CMD check on R v2.9.0.Our internal interleave() for Image (defined by EBImage) gave Error in z[idxOdd,, ] : incorrect number of dimensions. This was because internal image structure of the Image class changed (back) to being a 2-dim array. Now interleave() handles both 2d and 3d arrays.
lines() of RawGenomicSignals did not recognize xScale and yScale.
as.character() of GenericDataFile would throw Error in sprintf("%d", getFileSize(db, "numeric")) : use format %f, %e, %g or %a for numeric objects whenever the file size is returned as a double, which happens for very large files (> 2^31-1 bytes).
as.character() of GenericDataFileSet would throw an error if the data set was empty, because then there was no path.
getUnitNames(..., units = NULL) of TextUnitNamesFile would make the object believe there are zero units in the file. Thanks Shermane Teo at the National University of Singapore for reporting this.
binnedSmoothing(..., byCount = TRUE) of RawGenomicSignals would give error [...] object "ys" not found.
When passing a single data points to colBinnedSmoothing(), it would throw the exception: Range of argument 'by' is out of range [0,0]: [<by>,<by>].
Added argument units to extractRawCopyNumbers() of AromaUnitSignalBinaryFile.
Added RawGenomicSignals which RawCopyNumbers now inherits from.
Added readDataFrame(..., units = NULL) to AromaUnitSignalBinaryFile.
Added optional argument/field name to RawCopyNumbers, which is also used by SegmentedCopyNumbers when querying the “truth” function(s) for the copy-number state. This makes it possible to use one (conditional) truth functions for all samples.
Added hasBeenModified() to GenericDataFile.
Now hasTags(..., tags) of GenericData{File|Set} splits the tags argument.
Now RawCopyNumbers() takes RawCopyNumbers objects as input.
Now as.character() of GenericDataFile also reports the exact file size in case the file size is reported in kB, MB, etc. It also tries to report the relative pathname rather than the absolute.
getChecksum() of GenericDataFile caches results unless the file has been modified since last time.byName() of AromaUnit(FracB|Total)CnBinarySet searches rawCnData/ then cnData/.Added SegmentedCopyNumbers class.
Added various smoothing methods to RawCopyNumbers.
Added colBinnedSmoothing() and colKernelSmoothing().
Added getAromaUgpFile() to UnitNamesFile.
Added shakyText().
Added static byChipType() to UnitNamesFile.
Added argument fullname to getChipType() of AromaUnitSignalBinaryFile.
Now argument files in extract() of GenericDataFileSet can also be a vector of string.
ALPHA: Added support for “smart” (e.g. "*") subdirectories in static findByName() of GenericDataFileSet.
ROBUSTNESS: Added support for optional validation/selection by the number of units/cells to all static byChipType() methods. This is done by specifying argument nbrOfUnits/nbrOfCells. This is intend for internal use only.
ROBUSTNESS: Added a sanity check to getAromaUgpFile() of UnitNamesFile and AromaUnitSignalBinaryFile, which asserts that the number of units in the located UGP file match that the number of units in the data file.
ROBUSTNESS: Now findAnnotationData() always returns pathnames ordered by the length of their fullnames. Before this was only done if firstOnly = TRUE.
Added getMaxLengthRepeats() to AromaCellSequenceFile.
Added AromaUnitSignalBinary{File|Set}, AromaUnitTotalCnBinary{File|Set}, AromaUnitFracBCnFinary{File|Set}, AromaUnitCall{File|Set}, and AromaUnitGenotypeCall{File|Set}.
readFooter() of AromaTabularBinaryFile did not return the correct list for nested structures.
groupBySnpNucleotides() of AromaCellSequenceFile would return an empty element missing for some chip types, e.g. Mapping10K_Xba142. Now such empty elements are dropped.
getAttributes() for GenericDataFile:s would give an error if there were no attributes.
SPEEDUP: Now predict() of ProbePositionEffects is 6-7 times faster.
SPEEDUP: Now the result of isMissing() for AromaCellSequenceFile can be cached.
readDataFrame() of AromaTabularBinaryFile would forget to close the connection if verbose output was activated. When R later would close such connections, a warning would be generated.R CMD check on R v2.7.2 and R v2.8.0rc.as.GrayscaleImage() in argument transforms to a list using as.list() and not list().skip to writeChecksum() of GenericDataFile.getSnpPositions(), getSnpShifts(), getSnpNucleotides(), and groupBySnpNucleotides() to AromaCellSequenceFile.fitGenotypeCone() to support both flavors "sfit" (old) and "expectile" (new).Added support for argument positions to countBases() for AromaCellSequenceFile.
ALPHA: Added fitMultiDimensionalCone().
readSequences() for AromaCellSequenceFile translated raw values to incorrect nucleotides.R CMD check on R v2.7.1 and R v2.8.0 devel on MS Windows XP.getCreatedOn(), getLastModifiedOn(), and getLastAccessedOn() to GenericDataFile. These a just wrapper accessing file.info() fields.SPEEDUP: Added patch for base::serialize() on MS Windows, which in turn will speed up digest::digest() and all methods that use the latter to generate hashcodes.
SPEEDUP: Now as.character() for GenericTabularFile:s and TabularTextFile:s reports the number of data rows and the number of text lines as NA if the files are too large and cached results are not already available.
SPEEDUP: Replaced all which() with faster whichVector().
updateDataColumn() coerce values to doubles before censoring them for raw and integer columns.Now countBases() of AromaCellSequenceFile returns “raw” counts if argument mode = "raw".
ALPHA: Added setFullName() and setName() to GenericDataFile and GenericDataFileSet. The plan is to have these replace what getAlias() and setAlias() do today.
BETA: Added setFullNameTranslator() to GenericDataFileSet.
findByName() assert that the data set name is not empty.BETA: Added setFullNamesTranslator() (for the files) to AromaGenericFileSet.
ALPHA: Added protected update2() to GenericDataFileSet.
ALPHA: Added private ProbePositionEffects, etc.
Added a general importFrom() for AromaTabularBinaryFile, which calls matching importFrom<ClassName>(), if found.
Added support for raw data columns in AromaTabularBinaryFile.
BETA: Added readTableHeader().
BETA: Added AromaCellSequenceFile.
BETA: Added classes AromaMicroarrayTabularBinaryFile and AromaCellTabularBinaryFile.
readDataFrame():
R CMD check on R v2.7.0 patched.getOutputFiles() of AromaTransform.Added AromaTransform (extracted from aroma.affymetrix::Transform).
Added nbrOfArrays() to AromaMicroarrayDataSet.
Now private xmlToList() handles more complex XML strings.
Added member chromosome to RawCopyNumbers.
Added gaussianSmoothing() to RawCopyNumbers.
Added argument xOut to gaussianSmoothing() in order to specify at what loci the smoothed signals should be calculated.
Added gzip()/gunzip() to GenericDataFile.
Added static findByName() and byName() to GenericDataFileSet.
Added static getFileClass() to GenericDataFileSet. This makes it possible to remove most fromFiles() from subclasses but also the validation of argument files of their constructors.
Added validate() to GenericDataFileSet, which is called at the end of fromFiles(..., .validate = TRUE).
Added extractMatrix to GenericTabular(File|Set) and AromaTabularBinary(File|Set).
Added equals() and hasFile() to GenericDataFileSet.
Added classes GenericTabularFileSet and AromaTabularBinarySet.
Added stextChipType() for strings.
ALPHA: Now equals() for GenericDataFile:s uses the class, pathname, file size and file-contents checksum to do the comparison. Similarly, for GenericDataFileSet:s, the class, path, number of files, pathnames, and finally equals() for each file pair is used for the comparison.
ALPHA: Added full name translation for GenericDataFile:s.
ALPHA: Added UnitNamesFile.
Package passes R CMD check on R v2.7.0 patched.
Now findPngDevice() search for a working PNG device in the following order: png(..., type = "cairo"), png(..., type = "cairo1"), Cairo::CairoPNG(), R.utils::png2(), and finally plain png().
stextNnn() methods from AffymetrixFile to AffymetrixMicroarrayDataSet. Created superclass AromaMicroarrayDataSetTuple of AffymetrixCelSetTuple.BUG FIX: readDataFrame() did not read the first data row if there was no column header; it was eaten up by a preceeding readHeader().
BUG FIX: findAnnotationDataByChipType(chipType = "GenomeWideSNP_6", pattern = "^GenomeWideSNP_6.*[.]ugp$") would find file GenomeWideSNP_6,Full,na24.ugp before GenomeWideSNP_6,na24.ugp. Now we return the one with the shortest full name.
Package created.
Took all core classes and core methods in aroma.affymetrix that are independent of the Affymetrix platform and placed the here.
The version is numbered starting from the current version of aroma.affymetrix.
interleave() for Image gave Error in z[idxOdd, ] : incorrect number of dimensions. The internal image structure is a 3-dim array.
When searching with firstOnly = FALSE, findAnnotationData() was identifying files that are in “private” directory. This is how affxparser::findFiles() works. Such files are now filtered out.