J Comput Biol. 2025 Oct 9. doi: 10.1177/15578666251383561. Online ahead of print.
ABSTRACT
Metacell partitioning is a common preprocessing step in single-cell data analysis, used to reduce sparsity by aggregating similar cells. However, existing metacell partitioning algorithms may inadvertently group heterogeneous cells, potentially biasing downstream analyses. The resulting metacell partitions can vary substantially with different hyperparameter settings, leaving users uncertain about which result to trust. The mcRigor R package offers a statistical method for evaluating and optimizing metacell partitioning in single-cell data analysis. This article provides instructions for installing and using mcRigor to support more rigorous and interpretable metacell-based workflows.
PMID:41066085 | DOI:10.1177/15578666251383561
