- Add feature scaling options for PCA and UMAP
- SCE/SPE-compatible
- Fix compatibility with SeuratWrappers
- Implemented SmoothLabels for k-nearest neighbors cluster label smoothing
- Parallel clustering for Leiden graph-based clustering
- Version depedency on leidenAlg (>= 1.1.0) for compatibility with igraph (>= 1.5.0)
- Seed setting for clustering
- Neighborhood sampling for computing neighborhood feature matrices. See arguments
sample_size
,sample_renorm
andseed
in function ComputeBanksy
- Implemented Azimuthal Gabor filters in ComputeBanksy, with the number of
harmonics determined by the
M
argument. To obtain similar results as version 0.1.3, useM=0
- RunPCA and RunUMAP renamed to RunBanksyPCA and RunBanksyUMAP respectively to avoid namespace collisions with other packages
- Changed some function argument names (e.g.
spatialMode
tospatial_mode
)
- Interoperability with SingleCellExperiment with asBanksyObject
- Passing Bioc and R CMD checks
- BanksyObject: dimensionality reduction slot name changed to reduction
- RunPCA and RunUMAP
- Plotting is generalised (plotReduction instead of plotUMAP / plotPCA)
- ConnectClusters improved and returns BanksyObject
- ARI computation and plotting
- Legacy version