CHANGES IN VERSION 1.9.2 [2017-10-25] ----------------------------------- BUG FIXES o Corrected author name in all documentation CHANGES IN VERSION 1.9.1 [2017-10-23] ----------------------------------- BUG FIXES o Fixed bug in package dependency 'matter' affecting the size of datasets that can be processed with 'batchProcess' o Fixed bug where bin sizes for units='ppm' were twice as wide as they should be in 'readImzML' and 'reduceDimension' CHANGES IN VERSION 1.7.2 [2017-3-22] ----------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o In 'image' method, 'superpose = TRUE' now supports multiple LHS arguments in the formula (e.g., formula = a + b ~ x * y) CHANGES IN VERSION 1.7.1 [2016-11-29] ----------------------------------- NEW FEATURES o PCA is now supported for larger-than-memory on-disk datasets o External 'matter' matrices replace 'Binmat' matrices for on-disk support o Added 'image3D' aliases for all 'ResultSet' subclasses SIGNIFICANT USER-VISIBLE CHANGES o Added 'matter' package to Depends list CHANGES IN VERSION 1.5.2 [2016-10-6] ----------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o Updated 'batchProcess' to support reduceDimension and peakAlign o Now 'peakAlign' looks for an existing 'mean' column in featureData o Added 'matter' support to readAnalyze (previously only readImzML) BUG FIXES o Fixed bug when indexing into data cube using 'imageData' method CHANGES IN VERSION 1.5.1 [2016-9-21] ----------------------------------- BUG FIXES o Corrected author and maintainer contact information CHANGES IN VERSION 1.5.0 [2016-5-25] ----------------------------------- NEW FEATURES o Added experimental support for 'matter' on-disk matrices, from package 'matter', hosted at https://github.com/kuwisdelu/matter, as a replacement for 'Binmat' matrices BUG FIXES o Fixed subsetting SImageData objects with variables CHANGES IN VERSION 1.3.3 [2016-4-20] ----------------------------------- BUG FIXES o Subsetting the S4 part of Binmat objects by row is now an error o Providing non-positive m/z values to 'readImzML' is now an error o Elements of 'imageData' that fail to 'combine' or which are missing from one or more of the objects are now dropped from the result with warning rather than failing o Moved unit tests in 'ints/tests' to 'tests/testthat' CHANGES IN VERSION 1.3.2 [2016-3-29] ----------------------------------- NEW FEATURES o Added 'image3D' method for plotting 3D images o Added 'batchProcess' method for batch pre-processing CHANGES IN VERSION 1.3.1 [2016-3-14] ----------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o Added 'mass.accuracy' and 'units.accuracy' arguments for controlling the m/z accuracy when reading 'processed' imzML o Function 'reduceDimension.bin' now takes argument 'units' with value 'ppm' or 'mz', and new corresponding defaults BUG FIXES o Fixed bug in reading 'processed' imzML format that caused mass spectra to be reconstructed in the wrong order o Improved speed accessing columns of Hashmat sparse matrices o In 'pixelApply' and 'featureApply', zero-length return values are no longer returned as a list when '.simplify=FALSE' o Function 'peakAlign.diff' should be more memory efficient now CHANGES IN VERSION 1.3.0 [2015-12-16] ----------------------------------- NEW FEATURES o Added experimental Binmat class for working with on-disk matrices o Added experimental support for 3D files from benchmark datasets o Added experimental support for plotting 3D images o Added experimental support for 'processed' imzML format SIGNIFICANT USER-VISIBLE CHANGES o Added 'attach.only' argument to readImzML and readAnalyze BUG FIXES o Fixed bug in plotting 3D image slices in the z dimension o Fixed bug where large imzML files could not be read due to byte offsets being stored as ints; they are now stored as doubles. o Fixed bug with strip labels in 3D plotting and with mixed labels o Fixed bug with unique m/z feature names for high mass resolutions CHANGES IN VERSION 1.1.0 [2015-10-1] ----------------------------------- BUG FIXES o Fixed bug in formatting m/z labels affecting R 3.2.2 o Removed dependency on 'fields' because 'maps' is broken on Windows CHANGES IN VERSION 0.99.6 [2015-4-5] ----------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o Bioconductor Release Candidate 7 o Added CITATION file for citing Cardinal BUG FIXES o In 'readAnalyze' and 'readMSIData', removed endianness check in Analyze 7.5 headers because some ABSciex data files specify an incorrect header size, thereby fixing a bug where bits would be swapped wrongly and file read incorrectly. CHANGES IN VERSION 0.99.5 [2015-3-3] ------------------------------------ SIGNIFICANT USER-VISIBLE CHANGES o Bioconductor Release Candidate 6 o Added new vignette for Cardinal design and development o Now using ProtGenerics generics for 'spectra', 'peaks', and 'mz' CHANGES IN VERSION 0.99.4 [2015-2-15] ------------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o Bioconductor Release Candidate 5 o Updated vignette with biological examples o Added new citations to vignette CHANGES IN Version 0.99.3 [2015-2-3] ------------------------------------ SIGNIFICANT USER-VISIBLE CHANGES o Bioconductor Release Candidate 4 o In plot and image methods, 'groups' arg now coerced to factor BUG FIXES o Fixed bug in subset arg in select method o Fixed bug in plot and image methods with NA in 'groups' arg CHANGES IN Version 0.99.2 [2015-1-20] ------------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o Bioconductor Release Candidate 3 o Adjusts NIPALS unit tests for Windows build CHANGES IN Version 0.99.1 [2015-1-12] ------------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o Bioconductor Release Candidate 2 o Cleaned up biocViews BUG FIXES o Fixed bug in SImageData coord factor levels not being properly updated when SImageSet is subsetted CHANGES IN Version 0.99.0 [2014-12-22] -------------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o Bioconductor Release Candidate 1 CHANGES IN Version 0.9.0 [2014-12-22] ------------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o Added standardizeSamples method for between-sample normalization CHANGES IN Version 0.8.9 [2014-12-5] ------------------------------------ SIGNIFICANT USER-VISIBLE CHANGES o Improved plot methods for object summaries o Fixed minor issurs in walkthrough vignette CHANGES IN Version 0.8.8 [2014-11-27] ------------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o Changed modelData for ResultSet from facors to original parameters o Added plot method for summaries of ResultSet objects o Added 'normalize.image' argument for image method o Improved plot and image methods for CrossValidated CHANGES IN VERSION 0.8.7 [2014-11-23] ------------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o Completed Cardinal-walkthrough vignette BUG FIXES o Fixed bug in plotting vectors from ResultSet o Fixed bug in PCA summary in reporting explained variance CHANGES IN Version 0.8.6 [2014-11-21] SIGNIFICANT USER-VISIBLE CHANGES o Completed walkthrough vignettes o Added ImagingMassSpectrometry biocView CHANGES IN Version 0.8.5 [2014-11-20] ------------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o Added 'show' and 'print' methods for most objects o Changed 'combine' method for IAnnotatedDataFame, SImageData, and SImageSet o Completed .Rd documentation for all public methods o Cleaned up R CMD check warnings (except C++ .hpp headers) BUG FIXES o Fixed bug in between-cluster SS calculation for spatialKMeans o Fixed bug with pre-processing 1-pixel datasets CHANGES IN Version 0.8.4 [2014-10-16] ------------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o Changed 'topIons' to 'topLabels' method for summarization o Added MSImageData class for peakData and mzData methods o Changed naming scheme for SImageData (.iData => iData) BUG FIXES o Fixed bug in plotting of classes for SpatialShrunkenCentroids o Fixed bug in fopen in readAnalyze and readImzML affecting Windows CHANGES IN Version 0.8.3 [2014-9-18] ------------------------------------ SIGNIFICANT USER-VISIBLE CHANGES o Added 'coregistration' for SpatialKMeans o Added 'summary' methods for all ResultSet subclasses o Added 'topIons' method for summarization o Added plotting methods for 'ResultSet' classes o Added 'cvApply' method for cross-validation o Added additional statistics to SpatialKmeans class results o Changed 'auto.key' argument to 'key' in plotting methods BUG FIXES o Fixed bug in spatialShrunkenCentroids with prior probabilties o Fixed bug in OPLS with proper scaling of data matrices during prediction CHANGES IN Version 0.8.2 [2014-7-23] ------------------------------------ SIGNIFICANT USER-VISIBLE CHANGES o Added 'peakFilter' method o Minor fixes to documentation CHANGES IN Version 0.8.1 [2014-7-10] ------------------------------------ SIGNIFICANT USER-VISIBLE CHANGES o Changed 'parameters' argument to 'model' for ResultSet plot and image o Added 'auto.key' to plot and image for when 'groups' are specified o Added strip labels to plot and image when using base graphics o Added 'coregister' method for 'SpatialShrunkenCentroids' o Method 'image' is now vectorized for argument 'mz' o Method 'plot' is now vectorized for argument 'coord' BUG FIXES o Fixed a bug where conditions are plotted in an incorrect order CHANGES IN Version 0.8.0 [2014-6-16] ------------------------------------ SIGNIFICANT USER-VISIBLE CHANGES o First public beta available for ASMS 2014. CHANGES IN Version 0.7 [2014-6-2] --------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o Refactored version of the package prepared for internal testing. o New processing methods, renamed to 'normalize', 'smoothSignal', 'reduceBaseline', 'peakPick', 'peakAlign', 'reduceDimension'. o New function 'readMSIData' for reading supported formats. o Native support for imzML added (no longer requires Java). o Methods 'pixels' and 'features' replaced with updated versions. o In 'features' method, m/z matching now done using bisection search. o New analysis method, renamed to spatialShrunkenCentroids. o Analysis methods now return a new class: 'ResultSet'. o New method 'select' to replace 'selectROI' and 'selectPixels'. CHANGES IN Version 0.6 [2013-2-19] ---------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o Major revision to use Bioconductor base packages. o New dependencies on BiocGenerics and Biobase. o New virtual class iSet for generic imaging data. o New classes ImageSet, SImageSet, ImageData, and SImageData. o Revised MSImageSet that uses new parent classes. o New class MIAPE-Imaging for Minimum Information About a Protemics Experiment - Imaging. o New class MSImageProcess for pre-processing information. o Revised functions for importing Analyze 7.5 and imzML o New plotting and image methods using lattice-style formula interface. o New pixelApply and featureApply methods. CHANGES IN Version 0.5 [2013-11-8] ---------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o Added image coregistration for 3D sections. o Added 'volume' method for 3D visualization. CHANGES IN Version 0.4 [2013-11-7] ---------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o Added 'summaryPlot' method for segmentation visualization. o Exposed 'calibrateSegmentation' method to calibrate colors between multiple MSImageSegmentation objects. o Added more thorough verbose output for 'spatialSparseCluster' and 'spatialSparseKMeans' methods. o Improved behavior of 'intensities' method when reconstructing complex datacubes and with non-contiguous pixel requests. o Changed 'trellisImage' to plot all coordinates for dimensions omitted from 'fixCoord' by default. o Minor changes to the 'MSImageSegmentation' class structure. o Improved 'likPlot' method and added AIC and BIC results and the resulting number of classes. CHANGES IN Version 0.3 [2013-7-5] --------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o Improved speed of imaging processing (smoothing and interpolation) methods for use with 'image' and 'trellisImage'. o Added option for verbose output and hooks for GUI progress bars. o For prediction on 'MSImageSegmentation' objects, 'autoDimNames' is now automatically passed along from the original call. o Changed 'MSImageSegmentation' to an S4 class derived from 'list'. o Added option for plotting the probabilities of an 'MSImageSegmentation' using overlaid transparency masks for improved visualization. CHANGES IN Version 0.2 [2013-6-20] ---------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o Package renamed from internal code-name "Canary" to "Cardinal". o Split off GUI to a new package called "CardinaliView". o New GUI in "CardinaliView" now based on gWidgets and RGtk2. o Added slot 'peaks' to 'MSImageSet' class. o Changed 'MSPeakFrame' and 'MSPeakList' to S4 classes, and added new class 'MSPeak Alignment'. o Changed how the 'estimateNoise' method estimates noise. o Changed options in the 'detectPeaks' method to be more useful. o Changed the 'alignPeaks' methods to align to the mean spectrum as a reference when aligning a list of peaks. o Added 'interpolate' option for images. o Added new method "selectPeaks" for manual peak selection. o Modified the 'type' arguments in the plotting methods. o Changed 'crossValidate' method to now return predictions for the full dataset for every fold instead of the test set only. BUG FIXES o Fixed bug where plotting mass spectra for multiple pixels fails on datasets with missing mass spectra. o Fixed bug where the 'unload' method did not remove the object. o Fixed bug in the 'crossValidate' method where the dots arguments were not passed to the 'predict' method. CHANGES IN Version 0.1 [2013-6-6] --------------------------------- SIGNIFICANT USER-VISIBLE CHANGES o First internal release for Stanford Canary Center for Cancer Early Detection as primary beta testers. o Implementation of MSImageSet S4 class; mass spectra are stored as a flat matrix with rows as features and columns as pixels; original data cube can be reconstructed dynamically using a position array; spectra stored using environments to prevent redundant data copying. o Visualization that includes plotting ion images and mass spectra, using both base graphics and lattice graphics, allowing trellis display of multiple ion images and multipe mass spectra. o Pre-processing for that includes normalization, baseline correct, peak-picking, peak-alignment, resampling, and binning. o Multivaritate statistical analysis includes PCA, PLS, and OPLS. o Supervised analysis includes a spatially-aware version of nearest shrunken centroids, PLS-DA, and OPLS-DA. o Unsupervised analysis includes spatially-aware clustering based on nearest shrunken centroids, and an implementation based on sparse k-means. o GUI based on tcltk using rpanel that supports all major processing and statistical analysis methods in the package, except for PLS and OPLS.