When rownames are numeric characters e.g., “1”, ensure they stay character when converting MultiAssayExperiment to longFormat.
Changes in version 1.28.0
New features
Dropped experiments are no longer kept in the metadata slot. They can be seen with drops() (@LTLA, #323).
Bug fixes and minor improvements
Checking colnames in sampleMap vs ExperimentList is more robust by only comparing unique and sorted values in each.
Changes in version 1.26.0
New features
showReplicated displays the actual colnames of technical replicates by assay and biological unit.
The bracket replacement method [<- for MultiAssayExperiment now also replaces the names with those from the right-hand side of the operation, if any (@DarioS, #319)
Bug fixes and minor improvements
During single assay replacement [[<-, the re-ordering of assays based on the value input was invalid when empty assays present (@danielinteractive, #322).
Permuting assays also updates the order of names in the MultiAssayExperiment and assays in the sampleMap
Changes in version 1.24.0
New features
replicates provides the actual colnames identified as replicate observations for a particular biological unit in the sampleMap
Bug fixes and minor improvements
Added an assay<- replacement method for robustifying saveHDF5MultiAssayExperiment with plain matrices
Resolve issue when colData has one column when merging two MultiAssayExperiment objects, i.e., using the c method (@cvanderaa, #315)
Increase efficiency in colnames and rownames methods (@cvanderaa, #314)
Make ‘prefix’ inputs consistent in saveHDF5MultiAssayExperiment and loadHDF5MultiAssayExperiment (@asiyeka, #313)
Improve performance for replicated method
Update wideFormat documentation, when replicates present additional sets of columns will be appended to the produced DataFrame (@DarioS, #312)
Changes in version 1.22.0
Bug fixes and minor improvements
Add data("miniACC") to examples after removing lazy loading.
Class definition prototypes defined for cleaner extensibility (@hpages, #306).
Doc and internal improvments to MultiAssayExperimentToMAF
synAssay and nonSynAssay now require exact assay names in MultiAssayExperimentToMAF
Changes in version 1.20.0
Bug fixes and minor improvements
Avoid dropping experiments with repeated calls to subsetByColData and remove harmonization (@cvanderaa, #302)
getWithColData suppresses messages from natural subsetting operations by default with verbose = FALSE (@bhagwataditya, #301)
getWithColData was using the old default (drop = TRUE) and causing an error when the experiment is empty (@danielinteractive, #300).
Calls to the internal .harmonize operation are reduced to increase memory efficiency, when identical experiment colnames present (@LTLA, #299).
subsetByColData now errors on subscript vectors longer than the nrow of the colData (previously a warning).
colData<- includes a check for identical row names. If so, direct replacement of the colData occurs without harmonization.
Added a warning when an empty sampleMap is provided in the constructor function which may cause unexpected behavior. Documentation is updated to include more details on the sampleMap input.
Changes in version 1.18.0
New features
saveHDF5MultiAssayExperiment allows users to save data from most classes (excluding RaggedExperiment) into a single H5 file (ctb @hpages)
Support for maftools conversion has been added as MultiAssayExperimentToMAF (ctb @PoisonAlien)
renameColname and renamePrimary provide renaming facilities for column names in experiments and rownames in the colData, respectively
Users can now rename some or all the column names in experiments using colnames(mae) <- value
When replacing colData or experiments (including [[<-), new rownames and colnames (respectively) are checked against existing values and an error is given when none match
Using List objects to replace the data in the ExperimentList is now supported
splitAssay allows users to separate / split columns across assays
makeHitList is a facilitator function to create the logical lists that are required as input to splitAssay
drop argument when subsetting a MultiAssayExperiment is now FALSE by default
Bug fixes and minor improvements
Updated the constructor function to auto-populate rownames in colData when it is missing (@LTLA, #287)
The metadata now includes names of dropped experiments
Updated validity checks to support array-like classes
Dropped experiments are tracked in the metadata
Changes in version 1.16.0
New features
Coercion methods from list/List to MultiAssayExperiment method now available.
Bug fixes and minor improvements
Provide more details in documentation for mergeReplicates
Improved documentation for accessor function return values, helper function examples (@llrs, #281)
Fixed bug when using longFormat with character assay matrices (@jonocarroll, #282)
Changes in version 1.14.0
New features
exportClass creates a number of .csv data files for exporting data
Allow vector input i for selecting assays in longFormat (@lgatto, #266)
Updates to ‘Using MultiAssayExperiment with DelayedMatrix’ vignette
Bug fixes and minor improvements
Warn when colData rownames and ExperimentList colnames are empty (@LTLA#262)
Add informative error message for ExperimentList (@lgatto, #265)
Informative warning when dropping ExperimentList element columns (@lwaldron)
Fixes to constructor functions, MultiAssayExperiment and MatchedAssayExperiment (@lgatto, #267#268, @lwaldron)
Add warning when j in mae[i, j, k] is longer than colData rows
Strict argument matching between generic and methods
Updates due to class(matrix())
UpsetSamples more robust to differences in names between split sampleMap and names(ExperimentList) (@jonocarroll, #269)
Refactored and improved UpsetSamples
ExperimentList propagation of mcols and metadata (@vobencha, #270)
Enforcement of validObject with replacement methods colData and sampleMap (@vobencha, #271)
Changes in version 1.12.0
Bug fixes and minor improvements
Improvements to the main vignette, MultiAssayExperiment class schematic now included (@mtmorgan, #261)
Updated documentation for the upsetSamples function
Update code to use splitAsList from S4Vectors (@hpages)
Fixed bug with metadata disappearing from ExperimentList when replacing it inside a MultiAssayExperiment object (@lawremi, #259)
Fixed the formatting of the NEWS file
Changes in version 1.10.0
New features
getWithColData now allows easy extraction of experiments (such as SummarizedExperiment) with associated colData data
Single bracket replace method implemented for MultiAssayExperiment assays
Bug fixes and minor improvements
isEmpty method fixed for ExperimentLists containing matrices
MultiAssayExperiment now inherits from the standard Annotated virtual class
c method better distinguishes between list and ExperimentList inputs
Improvements on .getHits internal method for obtaining correct queries on row metadata
Subsetting mechanism tweaked to do nothing when subsetting by NULL rows compared to empty rows (i.e., character(0L))
Improved README.md
Changes in version 1.8.0
New features
The single bracket replacement method [<- added to support assignment of assay datasets
Users can now rename experiments in a MultiAssayExperiment with names(x) <- value
replicated and mergeReplicates functions have been refactored and improved
combining MultiAssayExperiments now possible with c function
Subsetting MultiAssayExperiment by a list or List type class now returns experiments in the input order for rows, columns, and assays
Bug fixes and minor improvements
wideFormat function improvements and bug fixes with name indicator subsetting
BiocGenerics:::replaceSlots used instead of replace methods
Added tests for anyReplicated, c, and names<- functions
Unit tests added for replacement method testing
Renamed objects in examples for brevity and descriptiveness
Updated importFrom directives
Internal sample names check now only works on non-empty colnames
Various documentation improvements
listToMap re-written for efficiency
Various improvements to subsetting mechanism
subsetByAssay bug fixed when using an integer index (@vjcitn, #)
c is smarter at matching colnames with primary names and creating a sampleMap
The MatchedAssayExperiment constructor function now works either a single MultiAssayExperiment or the essential components of one.
Renamed duplicated function to replicated
Added coercion functions from List and list to ExperimentList
Improve speed of reshape functions from previous change (wideFormat)
Remove clusterOn function and move to Bioconductor/MultiOmicQC package on GitHub
duplicated has been deprecated, use replicated and anyReplicated
removed dependencies on tidyr and reshape2
Moved the API shiny function to waldronlab/MultiAssayShiny package
Reduced imports (removed shinydashboard and shiny)
Method requirement checks for classes are practical using try()
Deprecated methods removed: pData
Deprecated class removed: RangedRaggedAssay
Assay-selective subsetting implemented via list/List class subsettors
Bug fixes and minor improvements
Added an isEmpty method for ExperimentList to account for an empty matrix
Documentation improvements to MultiAssayExperiment-class and MultiAssayExperiment-helpers
c internals improved
Explicitly check for DataFrame in ExperimentList
Fixed documentation warnings for inexact links
Fix subsetting order in bracket method ([)
Minor vignette changes
Supply a collapse character for wideFormat column names
upsetSamples does not munge experiment names with special characters when check.names = FALSE (by default keeps hyphens, underscores, etc.). A nameFilter functional argument allows operations such as substr on the experiment names. (@vjcitn, #231)
Updates to prepMultiAssay
Enhancements to the main vignette
New format for NEWS section
updated duplicated function now returns FALSE for non-duplicated samples
Improved ExperimentList constructor now handles multiple key = value entries