Incorporating cell states for mapping eQTLs in single-cell studies

Single-cell expression quantitative trait loci (eQTL) studies have revealed genetic regulation of gene expression with cell-type specificity. However, methods for genome-wide eQTL mapping that account for cell-state-dependent effects remain limited. To fill this gap, we propose scarf-QTL, which maps eQTLs across cell states and implements retrospective association tests for robust statistical inference. Simulations showed that scarf-QTL controls type I error rates and improves statistical power compared with pseudobulk methods. Applied to the OneK1K dataset, scarf-QTL identified 30% more cell-type-specific eGenes (genes with eQTLs) than the pseudobulk analysis and revealed distinct cell-state-dependent eQTL patterns in dendritic cells.