Abstract 3816: Fine-mapping candidate causal cell types in solid cancers using cis-regulatory elements

Abstract Background: Genome-wide association studies (GWAS) have identified numerous genetic variants linked to cancer susceptibility, but pinpointing the specific causal cell types and mechanisms remains challenging. We integrated GWAS with functional genomics data to uncover candidate causal cell types for ten solid cancers, including breast (overall and estrogen receptor [ER]-stratified), endometrial, ovarian, prostate, colorectal, esophageal, lung, and kidney. Methods: We analyzed GWAS summary statistics from individuals of European ancestry for each cancer (average cases = 51,290) alongside over 1,000 cell-type-specific (CTS) cis-regulatory element single nucleotide polymorphism (SNP) annotations. Causal cell type fine-mapping was performed using CT-FM and CT-FM-SNP. CT-FM used genome-wide CTS heritability enrichment Z-scores from stratified linkage disequilibrium score regression (SLDSC), a co-regulation matrix of CTS SNP annotations, and a 1000 Genomes Project LD reference panel, outputting posterior inclusion probabilities (PIPs) indicating the likelihood of each cell type driving genetic associations. We used SuSiE to identify candidate causal SNPs (PIPs ≥ 0.5) across the genome for each cancer. CT-FM-SNP then restricted fine-mapping to SNP-annotations overlapping these candidate SNPs, providing additional SNP-level evidence of candidate causal cell types. High-confidence causal cell types were defined as those with PIPs ≥ 0.50, and candidate causal cell types as those with PIPs ≥ 0.05. Results: SLDSC identified 69 cell types with significant heritability enrichment (q-values ≤ 0.05) across five cancers. CT-FM refined these results, revealing mammary luminal epithelial cells as high-confidence causal cell types for overall (PIP = 0.76) and ER+ breast cancer (PIP = 0.86), prostate epithelial cells (VCaP) for prostate cancer (PIP = 1.00), and colon epithelial cells (LoVo) for colorectal cancer (PIP = 0.56). Additional candidate cell types included mammary basal epithelial cells for overall/ER+ breast cancer; mesenchymal stem cells, colon goblet cells, colon epithelial cells and small intestine goblet cells for colorectal cancer; naïve B cells, proximal tubule epithelial cells, Jurkat cells and mesodermal cells for renal cancer. Based on 555 candidate causal SNPs across all cancers, CT-FM-SNP identified mammary luminal epithelial cells as the most frequently detected high-confidence causal cell type for overall (11/103 SNPs) and ER+ breast cancer (9/73 SNPs), prostate epithelial cells (VCaP) for prostate cancer (34/209 SNPs), naïve B cells for renal cancer (5/81 SNPs), and prostate epithelial cells for one SNP in colorectal cancer. Conclusion: Our findings provide insights into the cellular origins of solid cancers. Larger, more diverse GWAS datasets and comprehensive CTS SNP annotations are needed to better link genetic associations with cancer etiology. Citation Format: Xueyao Wu, Artem Kim, Sara Lindström, Nicholas Mancuso, Peter Kraft. Fine-mapping candidate causal cell types in solid cancers using cis-regulatory elements [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 3816.