- Added
immunotherapy
geneset
to compute BCS of immunotherapy-related signatures.
- Fixed dependency errors. Beyondcell requires
Seurat
>=4.0.0,<5.0.0 andMatrix
==1.6_1.1. Related to issue rgb(17, 85, 17). - Added Spatial Transcriptomics tutorial example.
- Fixed an issue (#132) where
bcScore
's check for normalized data failed. - Fixed an issue (#135) where
bcRanks
was not computing the residual's mean. - Updated tutorials.
- Removed
bcRanks
"general" statistics computation. - Removed
GetStatistics
function.
- Fixed an error in
bcClusters
andbcSignatures
whenspatial = TRUE
andimages = NULL
. - Fixed title error in
bc4Squares
. The current title matches the correspondingidents
level. - Fixed error in
bc4Squares
when drug names were passed astopnames
. - Updated pre-loaded collections, which have been trimmed to 250 up and 250 down genes.
- Updated SSc collection, which only contains protein-coding genes.
- Added functions to plot spatial transcriptomics data on top of selected images.
- Added a function to load our collection of drug signatures:
GetCollection
. - Added a set of functional pathways related to cancer disease: these are available using
include.pathways
inGetCollection
. - Added two new optional arguments to
bc4Squares
:x.cutoff
andy.cutoff
to set custom quadrant thresholds. - Added two new optional arguments to
bcUMAP
:npcs
andseed
to allow users to both set the number of components to compute in the PCA step and reproduce a given UMAP by specifying the seed. - Changed the KNN imputation methodology in
bcRegressOut
. Beyondcell now relies on DMwR instead of bnstruct. - Fixed multiple issues in the package when only a single signature was being provided by the user.
- Updated the SSc collection. A previous source of drug signatures (CCLE) has been replaced by those from the PRISM study. CTRP and GDSC derived signatures have been updated to the latest releases of the aforementioned projects' data.
- Updated the drug annotation to harmonize the mechanism of action for the compounds in the SSc collection.
- Updated all of our pre-loaded collections, which are stored as geneset objects.
- Updated
GenerateGenesets
to work on user-provided genesets stored in GMT format. - Updated beyondcell's dependencies.
- Updated binned scales in
bcSignatures
to improve visualizations. - Removed the
method
argument frombcUMAP
. Only the UWOT library implementation of UMAP is available.
- Fixed an error in
bcSubset
when filtering by the proportion ofNaN
values. Previously, the beyondcell object wasn't a subset when usingnan.sigs
ornan.cells
. Now those signatures/cells below or equal to theNaN
threshold are kept. Related to issue #21.
- Fixed an issue (#17) where if all the cells have a number of expressed genes <
expr.thres
, all the values for that signature will beNaN
and so will be the switch point.
- Added group annotation to rank tables: high/low/differential sensitivity is now defined for each tested drug.
- Fixed an issue (#14) where
bcSignatures
would fail when used with pathways
- Fixed an issue where
bcMerge
would fail if the background matrix was empty.
- Updated pathways with the latest sources from MSigDB 7.3.
- Added a directory called
data-raw
with both the code and the data used to generate the pathway signatures.
- Added two new arguments in
bcUMAP
:method
andreturn.model
.method
can be"umap-learn"
or"uwot"
(default).return.model
is a logical argument that determines whether the function will return the uwot model (only valid whenmethod = "uwot"
). - Changed tutorials: Typos have been corrected, chunks of code now specify the UMAP method (
"umap-learn"
) and figures have been redrawn (they look the same as the originals).
- Initial public release: subsequent Beyondcell releases will follow Semantic Versioning guidelines.