scverse/anndataR: Version 1.0.0 (Bioconductor 3.22)

{anndataR} brings the AnnData object to R. It enables native reading and writing of H5AD files and conversion to/from SingleCellExperiment and Seurat objects.

Features

• An **{R6}(( class to work with AnnData objects in R (either in-memory or on-disk)
• Natively read/write H5AD files without needing a Python environment
• Convert to/from SingleCellExperiment objects
• Convert to/from Seurat objects

Changes from v0.99.0

New features

• Implemented an AnnDataView class, which provides a lazy view of an AnnData object without copying data (PR #324)
• Implemented S3 methods for AbstractAnnData objects: dim, nrow, ncol, dimnames, rownames, colnames, and [ (PR #324)
• Add a ReticulateAnnData class for seamless Python integration via {reticulate} (PR #322, PR #372)
• Add a get_generator_types() function that returns allowed/example types for generate_dataset() (PR #354)

Major changes

• Refactor obs_names/var_names handling for improved data consistency (PR #328)
  • InMemoryAnnData now stores obs_names and var_names as separate private fields instead of relying on rownames of obs/var `data frames
  • HDF5AnnData maintains separate obs/var names management to ensure consistency between obs/var data frames and dim names
  • All matrix data (X, layers, obsm, varm, obsp, varp) is now stored internally without dim names for consistency
  • Dim names are added on-the-fly when users access data, ensuring proper obs/var` name display

Minor changes

• Handle slots that may have incomplete dimensions when converting from Seurat. These are now skipped with a warning instead of indirectly raising an error. (PR #369)
• Add a compression parameter to additional write operations in HDF5AnnData for consistency (PR #328)
• Refactor setter methods in HDF5AnnData and InMemoryAnnData to use pipe operators for cleaner code (PR #328)
• Generalise the layers created by generate_dataset() when format = "Seurat" (PR #354)
• Add checks for type arguments to generate_dataset() (PR #354)

Bug fixes

• Avoid writing character datasets to H5AD files with LZF compression as it causes R to crash (PR #356)
• Fix Seurat conversion for PCA loadings with variable feature subsets (PR #328)
  • Seurat PCA loadings only contain variable features, not all genes
  • {anndataR} now properly expands loadings matrix to include all genes with zeros for non-variable features
  • Adds a warning when rownames don't match var_names during conversion
• Directly use obs_names and var_names properties instead of corresponding indirect S3 methods rownames and colnames inside the package (PR #328)
• Fix the error message variable name in .validate_aligned_array() method (expected_colnames → expected_rownames) (PR #328)

Documentation

• Add a citation file, citation("anndataR") now returns details of the {anndataR} preprint (PR #351)
• Update vignettes to clarify and expand text and improve formatting, including adding mapping figures to conversion vignettes (PR #360, PR #372)
• Minor updates to function documentation (PR #319)
• Add explanatory comments for matrix generation alignment with Python dummy-anndata (PR #328)
• Replace \donttrun with \donttest in man pages (PR #371)

Development

• Add an AGENTS.md with instructions for AI agents used during development (PR #367)

Full Changelog: https://github.com/scverse/anndataR/compare/v0.99.0...v1.0.0