Malva

Documentation | Web Platform

Malva is a nucleotide sequence indexer that enables sequence search at single-cell and spatial resolution: query any sequence across millions of single cells, in seconds.

Malva powers Malva Index, a large collection of datasets spanning 75+ million human single cells. You can use the Malva Client API to connect to and query the Malva Index.

Code and binaries are freely available for academic non-profit use. See LICENSE for details.

What Malva Can Do

• Search any nucleotide sequence across single cells from your own datasets
• In silico probe-based detection of viral/bacterial transcripts, circular RNAs, splice variants, and mutations

Installation

Malva can be installed from pip.

pip install malva

Linux note: installing liburing is recommended for optimal I/O performance.

sudo apt-get install liburing2      # Debian/Ubuntu
sudo dnf install liburing           # RHEL/Fedora

Malva will build and run without it, but I/O will fall back to standard pread.

Building from Source

Requires Python ≥ 3.9, a C++17 compiler, and Cython ≥ 3.0.

pip install poetry cython>=3.0 numpy setuptools wheel pkgconfig
pip install .

See the documentation for full build instructions, usage, and examples.

Citation

If you use Malva in your research, please cite our paper in Nature:

@article{LenPerin2026,
  title = {Ultrafast and reference-free sequence discovery in single-cell data},
  ISSN = {1476-4687},
  url = {http://dx.doi.org/10.1038/s41586-026-10975-w},
  DOI = {10.1038/s41586-026-10975-w},
  journal = {Nature},
  publisher = {Springer Science and Business Media LLC},
  author = {León-Periñán,  Daniel and Karaiskos,  Nikos and Rajewsky,  Nikolaus},
  year = {2026},
  month = Aug 
}

Contact

• Issues: GitHub Issues
• General inquiries: Rajewsky Lab @ MDC Berlin