Fast, flexible analysis of differences in cellular composition with crumblr.

Crumblr enhances detection of cell composition differences by increasing statistical power 2- to 3-fold in single-cell RNA-seq data analysis.

• Why it matters: Understanding shifts in cell type proportions is crucial for insights into health and disease, but current methods lack power and flexibility for complex study designs.
• What they did: The authors developed crumblr, a scalable statistical tool using precision-weighted linear mixed models that tests multiple cell lineage levels simultaneously, validated through simulations and real datasets.
• The result: Crumblr outperforms existing methods by boosting detection power while controlling false positives, enabling more accurate identification of cellular changes across diverse biological conditions.