MEAD Paper Accepted in JRSS-B

Our paper, Statistical Inference for Cell Type Deconvolution, has been accepted for publication in the Journal of the Royal Statistical Society Series B.

👉 preprint: ArXiv: 2202.06420

MEAD is a statistical framework for estimating cell-type proportions in bulk RNA-seq samples using single-cell RNA-seq reference data. Two main methodological contributions are:

• Identifiability under arbitrary gene-specific cross-platform scaling differences, which arise naturally when combining bulk and single-cell expression measurements from different technologies.

• Valid statistical inference for cell-type proportions and their comparisons across individuals, accounting for gene–gene correlation and shared uncertainty from the deconvolution step.

Simulation studies and real-data analyses demonstrate competitive estimation accuracy and reliable statistical inference in realistic settings.