Connecting polygenic disease risk to cell states and regulatory programs through single-cell chromatin accessibility

Abstract Single-cell ATAC-sequencing enables high-resolution mapping of cell states and their putative regulatory elements. The enrichment of genome-wide association study (GWAS) signals within context-specific regulatory regions can reveal disease-relevant cell populations. We present Single-Cell ATAC-seq Disease Score (SCADS), a computational framework that integrates GWAS summary statistics with scATAC-seq data to prioritize disease-relevant cells. SCADS produces calibrated, cell-level scores that are comparable across datasets and traits by: (i) identifying chromatin co-regulatory regions via topic modeling, (ii) quantifying their polygenic GWAS enrichment, and (iii) calculating cell-specific disease scores from topic weights and enrichment levels. Across extensive simulations, SCADS outperforms existing methods in power while controlling false positives. Applied to multiple autoimmune traits, SCADS reveals marked heterogeneity in disease relevance within canonical immune cell types. We specifically studied the differential disease relevance within CD8 + T and colon epithelial cells for inflammatory bowel disease, and pinpointed underlying gene programs and variants. SCADS is scalable, interpretable, and modular, providing a general framework to connect noncoding regulatory variation to cellular identity and function.