Abstract
Variations of cell-type proportions within tissues could be informative of biological aging and disease risk. Single-cell RNA-sequencing offers the opportunity to detect such differential abundance (DA) patterns, yet this task can be statistically challenging due to the noise in single-cell data, inter-sample variability and because DA patterns are often of small effect size. Here we present a DA-testing paradigm called ELVAR that uses cell attribute aware clustering when inferring differentially enriched communities within the single-cell manifold. Using simulated and real single-cell and single-nucleus RNA-Seq datasets, we benchmark ELVAR against an analogous DA algorithm that uses Louvain for clustering, as well as local neighborhood-based DA-testing methods, demonstrating that ELVAR improves the sensitivity to detect DA-shifts in relation to aging, precancerous states and Covid-19 phenotypes. In effect, leveraging cell attribute information when inferring cell communities can denoise single-cell data, avoid the need for batch correction and help retrieve more robust cell states for subsequent DA-testing. ELVAR is available as an open-source R-package.
Competing Interest Statement
The authors have declared no competing interest.