Finding a causal gene is a fundamental problem in genomic medicine. We present a causal inference framework, CoCoA-diff, that prioritizes disease genes by adjusting confounders without prior knowledge of control variables in single-cell RNA-seq data. |