Allele-specific expression (ASE) in single-cell resolution can reveal stochastic and dynamic features of gene expression in greater detail. However, analyzing single-cell ASE poses unique analytical challenges due to the extremely low depth of sequencing coverage per cell. In addition, allelic proportions often form U-shaped or W-shaped distribution because of the frequent occurrence of random monoallelic bursts. We propose a new method, scBASE, a “soft” zero-and-one inflated model for cellular allelic proportions in which we tackle data sparsity and variability by harnessing information derived from cell's population context.
- Free software: MIT license
- Documentation: https://churchill-lab.github.io/scBASE/