mc-ASTRA (mc_astra, commonly imported as mca) is a Python package for building interpretable maps of tissue and sample variability from single-cell and spatial omics data.
It integrates molecular, compositional, and spatial tissue descriptors together with sample-level information, such as clinical variables or technical covariates, to identify the main sources of variation across a collection of tissues. Using flexible factor models and downstream biological interpretation, mc-ASTRA connects these differences to coordinated multicellular programs and changes in tissue organization.
The package provides modular workflows for preprocessing, constructing multi-view tissue representations, fitting and exploring tissue-state maps, and interpreting the multicellular processes underlying them. It integrates with the scverse ecosystem, uses MOFA-FLEX for flexible factor modeling, and supports the incorporation of biological and technical prior knowledge.
mc-ASTRA currently targets Python 3.12 and 3.13. mc-ASTRA currently targets the dev branch of MOFA-FLEX: https://github.com/bioFAM/mofaflex.git@main.
Install mc-ASTRA latest development version:
pip install git+https://github.com/saezlab/mc-astra.git@mainPyPI installation coming soon!
Import the package as:
import mc_astra as mcaRicardo Omar Ramirez Flores, Jan David Lanzer, Daniel Dimitrov, Britta Velten, Julio Saez-Rodriguez (2023) Multicellular factor analysis of single-cell data for a tissue-centric understanding of disease eLife 12:e93161