AmbiDose: per-cell removal budgets for ambient RNA contamination

AmbiDose removes 92.0% of cross-species UMIs in human-mouse mixture while retaining 99.6% of same-species UMIs, reducing ambient RNA contamination in single-cell RNA sequencing.

• Why it matters: Ambient RNA contamination hampers accurate gene expression analysis in droplet-based single-cell and single-nucleus RNA sequencing, creating a need for improved removal methods.
• What they did: The approach combines an ambient profile from empty droplets with cell-specific removal budgets and expression-based gene protection, tested on human-mouse GEM-X mixtures and kidney-tumor and fetal-liver cohorts, with simulations supporting its effectiveness.
• The result: AmbiDose reduces off-target marker expression and marker leakage, enabling more accurate expression analysis, and is available as a Python package compatible with Scanpy.