Release notes

Release notes#


  • Compatibility with scanpy 1.10


This release comes with fixes:

  • for handling and saving colour palettes in MuData for categorical and continuous variables in,

  • for using sparse matrices in the MOFA interface to combine modalities with missing samples in,

  • for error messages and mixing metadata and features when plotting across modalities with


This release comes with improvements and fixes:


This release comes with new features and improvements:

  • MOFA can be now run in the stochastic mode (SVI) using the new arguments for

  • MOFA model weights can be visualised with

  • Plotting module has gained new plots such as

  • It is now possible to define layers as {modality: layer} in

  • Improvements to the TF-IDF normalisation interface including view handling.

  • Dependencies are handled better now such as pysam and scikit-learn.



MuData is now provided as a separate package. Since this release, mudata is a hard dependency of muon. Making the codebase more modular, this will also help developing tools around MuData that do not depend on scanpy or muon.

We fixed a few things including reading .h5mu files in backed mode when modalities have .raw attributes (this is live in the mudata library), SNF functionality ( and colouring plots by var_names that are present in .raw but not in the root AnnData object.

We also introduced some new features for the ATAC module including:

  • handling fragments files with barcodes different than obs_names and

  • supporting atac_peak_annotation.tsv files produced by Cell Ranger ARC 2.0.0.


Initial muon release with MuData (muon.MuData), atac and prot submodules (muon.atac and muon.prot), and multi-omics integration with MOFA ( and WNN (muon.pp.neighbors()).