Minor changes:
EnsemblToEntrez
: Updated unit test to reflect breaking change in pending AcidGenomes package update.
Minor changes:
melt
: Needed to ensure that join column is factor, to avoid unwanted type switch duringleftJoin
step.
Minor changes:
- Removed
export
method forMultiAssayExperiment
, which has migrated to cBioPortalAnalysis package instead. The MultiAssayExperiment package introduces too many unwanted dependencies.
Minor changes:
convertSampleIdsToNames
: Quiet down this function when no sample name conversion is necessary.
Major changes:
- Starting a new release series, based on breaking changes in other Acid Genomics dependency packages.
- Now enforcing strict camel case for all function names.
- Replaced usage of
"2"
with"To"
in relevant generics. export
: Updated method to use AcidGenerics conventions rather than BiocIO.- Replaced usage of stringi with AcidBase variants.
Minor changes:
AcidExperimentTestsUrl
: No longer uses specific version.
Minor changes:
- Now consistently using
DFrame
instead ofDataFrame
;GRanges
instead ofGenomicRanges
in S4 coercion and method definitions. - Updated package to use new
Ensembl2Ncbi
andNcbi2Ensembl
generics now defined in AcidGenomes, instead of previousEnsembl2Entrez
andEntrez2Ensembl
. We are now using "NCBI" consistently across packages instead of the legacy "Entrez" name.
Minor changes:
- Updated dependencies to Bioconductor 3.16.
- Migrated
requireNamespaces
import from AcidBase to goalie.
Major changes:
export
: Added initial method support forMultiAssayExperiment
.
Minor changes:
export
: Hardened method to only usecon
, dropping support for legacydir
approach. These methods have been restricted based on the pending pipette v0.10 release series.- Hardened dependencies to match Bioconductor 3.15 release series more clearly.
Minor changes:
- Hardened
tempdir
andunlink
calls in unit tests and working examples, to improve compatibility in Windows testing environment.
Minor changes:
export
: HardenedSummarizedExperiment
against edge cases where noassays
,colData
, orrowData
are defined.
Minor changes:
- Updated lintr checks and testthat unit tests.
Major changes:
- Now requiring R 4.2 / Bioconductor 3.15.
- Lightened package by eliminating strong dependencies on scales, sessioninfo, and stringr package. Also removed dependency on tibble package via pipette.
makeSummarizedExperiment
will conditionally slotsession_info
from sessioninfo package if installed, otherwise will fall back tosessionInfo
from utils package.
Minor changes:
- Reworking factor level support as
droplevels2
instead ofdroplevels
, to avoid method collisions with Bioconductor 3.15. - Reformatted package code using styler conventions.
Major changes:
export
: ReworkedSummarizedExperiment
to use new BiocIO generic approach. This generic now usescon
instead ofdir
to define the directory path. Note that previously, the object name would be defined as a subdirectory when usingdir
to define the path. We are currently keeping this legacy approach when usingdir
, but no attempt to parse the object name and create a corresponding subdirectory is employed when using the newcon
argument.convertGenesToSymbols
andconvertSymbolsToGenes
: Simplified the number of exported methods to now simply dispatch onSummarizedExperiment
. Added support for newstrict
mode check, which intentionally does not allow propagation of anyNA
values during interconversion steps.
Minor changes:
selectSamples
: The function now automatically relevels factors defined incolData
.- Reworked internal code to no longer depend on
formalsList
andgetNameInParent
from AcidBase package. Note thatgetNameInParent
is now only exported in goalie package. Also removed import of now defunctas.data.frame
methods previously defined in pipette package. - S4 methods are now defined alphabetically in package documentation.
- Hardened internal
mapGenes
code to no longer usetryCatch
call and silently error on map failure. selectSamples
: Factors defined incolData
are now automatically releveled via internaldroplevels
call forSummarizedExperiment
method.
Major changes:
aggregate
: Added support viaMARGIN
argument to work either row-wise (1
; default) or column-wise (2
).SummarizedExperiment
method has been updated to support these conventions. CorrespondingaggregateCols
andaggregateRows
functions have been simplified to merely wrap anaggregate
call withMARGIN
argument predefined.convertTranscriptsToGenes
: Now all assays are aggregated, rather than simplify returning primarycounts
assay.
Minor changes:
calculateMetrics
: Updated to supportassay
argument.convertGenesToSymbols
: Removed unnecessarystrict
argument forSummarizedExperiment
method, and simplified handling of mismatchedGene2Symbol
internally.estimateSizeFactors
now supportsassay
argument.integerCounts
now supportsassay
argument.aggregate
keeps track of mapping factor used internally forSummarizedExperiment
method. This is defined asaggregateRows
oraggregateCols
in the object metadata.
Minor changes:
- Now importing
bapply
from goalie instead of AcidBase. - Updated internal gene mapping code to better handle mismatched Gene2Symbol input, which can happen in pointillism package.
Major changes:
- Reworked and improved unit tests to achieve 100% code coverage.
calculateMetrics
: Simplifed internal assert checks. Thematrix
andMatrix
methods (handed off fromSummarizedExperiment
method) now work primarily onrowData
instead ofrowRanges
, since we're not using any genomic coordinates here in checks. Prefiltering (disabled by default, but can be enabled withprefilter = TRUE
) now removes column elements (e.g. cells and/or samples) with all mito counts (mitoRatio >= 1
) and an artifically highlog10FeaturesPerCount >= 1
value. These settings help make 100% code coverage possible on minimal test datasets.makeSummarizedExperiment
: Improved and fixed CLI messages on detected of mismatched features to be slotted intorowRanges
. This can happen when incorrect values are defined inensemblRelease
orgffFile
. This should help resolve the cryptic CLI error message seen in bcbioRNASeq issue #170.
Minor changes:
- Made unused
matchesGene2Symbol
andmatchesInterestingGroups
functions defunct. These are still reexported in basejump and DESeqAnalysis, which will be cleaned up in future releases. - Improved and hardened CLI messages. Now using
abort
instead ofstop
internally, to support better stylized error messages. - Now calling
toStringInline
instead of basetoString
internally to properly handle string handoff to cli package. autopadZeros
:SummarizedExperiment
method now hassampleNames
argument, which allows the user to selectively pad the sample names columnsampleNames
defined incolData
. Enabled by default, matching previous internal behavior.estimateSizeFactors
: Simplified internal assert checks that look for any zerolibSizes
. Also improved code coverage to check for specific values returned bytype
argument.
Major changes:
convertGenesToSymbols
: Addedstrict
andquiet
arguments. Improved and hardened internal code, improving the verbosity of calls toGene2Symbol
generator. Added additional asserts to ensure that mismatches (e.g. row flipping) don't occur.Ensembl2Entrez
,Gene2Symbol
: Updated formals to passthrough to other primary methods defined in AcidGenomes package.
Minor changes:
- Hardened and expanded code coverage for expected gene-to-symbol mapping
behavior in main
Gene2Symbol
generator, focusing on theformat
argument.
Minor changes:
- Updated package dependency version cutoffs.
Gene2Symbol
: Tightened up internalmatch.arg
handling forformat
argument, in definedSummarizedExperiment
method.
Minor changes:
mapGenesToRownames
,mapGenesToIDs
,mapGenesToSymbols
: Improved the flexibility of gene identifier matching forSummarizedExperiment
method. Now also attempts to match againstgeneSynonyms
when defined inrowData
, which is incredibly useful for easy matching of deprecated gene names.- Removed unnecessary
Gene2Symbol
methods formapGenesToRownames
,mapGenesToIDs
, andmapGenesToSymbols
. matchesGene2Symbol
no longer attempts to use the now removedGene2Symbol
method formapGenesToRownames
internally. We simplified this check using the internal.mapGenes
function to locate index position instead.
Minor changes:
- Updated the documentation to provide compatibility with R 4.1 and Bioconductor 3.13 release.
- Updated dependency versions.
Minor changes:
makeSummarizedExperiment
: Relaxed the validity check on syntactically valid names in rownames and colnames of the object. Now the function informs the user with a CLI message instead of erroring.
Minor changes:
- Renamed all internal instances of "blacklist" to "denylist".
- Added
denylist
override option tomakeSummarizedExperiment
(recommended to be left enabled by default), which will be used in cBioPortalAnalysis package update.
Minor changes:
- Now enforcing lower camel case for
sampleData
return. - Simplified generic handling for Matrix methods.
Minor changes:
- More internal NAMESPACE reworks. IRanges classes now defined in AcidGenerics.
Minor changes:
- Internal NAMESPACE rework, using pipette and AcidGenerics.
Minor changes:
- Reworked the list of reexported functions. Now reexporting only from SummarizedExperiment and sessioninfo.
Minor changes:
- Now reexporting some commonly used functions from SummarizedExperiment.
Minor changes:
sampleData
: Bug fix for unwanted early return on minimal colData defined in SummarizedExperiment. This handling is required for downstream automatic plotting functions defined in AcidPlots, that make use of "sampleName" metadata column.
Minor changes:
- Added
ignoreVersion
option to gene mapping functions.
Major changes:
- Ensuring we're enforcing lower camel case naming conventions for metadata
describing the columns slotted in
colData
.
Minor changes:
- Imported coercion methods from
SummarizedExperiment
. - Added
package.R
, describing package info.
Initial release, with code migrated from basejump package.