Kai Yu

Background: Transcriptome-wide association studies (TWAS) integrate gene expression and genome-wide association studies (GWAS) to identify disease susceptibility genes. Because gene expression varies substantially across cell types within tissues, cell type–specific prediction models may enhance the power of TWAS. Methods: We conducted cell type–specific TWAS leveraging single-cell RNA sequencing data from the OneK1K cohort (14 immune cell types, 1.27 million cells) and GWAS summary statistics for 7 cancers (>290 000 cases in total). To improve prediction accuracy, we developed a modeling framework that incorporates shared gene expression effects across cell types. Results: At a false discovery rate of 5%, we identified 106 (Bonferroni 5%: 13) previously unreported loci for breast cancer, 51 (4) loci for prostate cancer, 11 (4) loci for lung cancer, 39 (5) loci for melanoma, 9 (1) loci for ovarian cancer, and 2 (1) loci for diffuse large B-cell lymphoma, with most genes exhibiting cell type specificity. Gene set analyses confirmed joint associations of unreported genes with breast and prostate cancer risk in UK Biobank data. Additional lung tissue single-cell RNA sequencing data with 113 individuals validated 18 of 32 (56.3%) statistically significant genes for lung cancer. Across cancers, 139 statistically significant genes were shared by at least 2 cancer types and were primarily enriched in specific immune cell types. Conclusion: Cell type–specific TWAS improve the identification of novel cancer susceptibility loci and provide insights into the immune landscape of cancer etiology.

The ABO locus is associated with pancreatic ductal adenocarcinoma (PDAC). Potential metabolic mechanisms underlying these associations have not been investigated. We examined associations between genotype-derived ABO blood group (rs505922 and rs8176746) and 1478 prediagnostic serum metabolites in 4042 participants from 8 nested case–control studies within the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial and Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study using linear regression and fixed-effect meta-analysis. We then examined associations between the identified ABO-associated metabolites and PDAC in 2 nested case–control studies (493 cases, 640 controls) using logistic regression and evaluated metabolite mediation of the ABO-PDAC association. Non-O and A (vs O) blood groups were associated with 13 and 20 metabolites, respectively, at false discovery rate <0.20, with 9 in common. The ABO-associated metabolites, sphingosine (non-O: β = 0.15), aspartate (A: β = –0.11), and aspartylphenylalanine (A: β = –0.16) were positively, and fibrinopeptide B (1-13) (non-O: β = 0.13; A: β = 0.21) was inversely associated with PDAC (OR = 0.96-1.07 per SD change log10-metabolite, P <.05) after adjustment for blood group. Non-O (OR = 1.50 [95% CI, 1.16-1.94]) and A (OR = 1.46 [95% CI, 1.10-1.92]) (vs O) blood groups were associated with PDAC, however none of the ABO associated metabolites significantly mediated the association between ABO blood group and PDAC. Our results suggest the ABO-associated metabolites are independent risk factors for PDAC. Trial registration: NCT00342992 and NCT00339495.

Supplementary Table S1 shows results of univariate analysis for 77 chronically elevated serum alanine aminotransferase (cALT)-defined nonalcoholic fatty liver disease SNPs in the Pancreatic Cancer Cohort Consortium (PanScan) data

Supplementary Table S2 shows results of univariate analysis for 22 imaging-defined nonalcoholic fatty liver disease SNPs in the PanScan sample

Supplementary Table S3 shows results of univariate analysis for 36 biopsy-confirmed nonalcoholic fatty liver disease SNPs in the PanScan sample

Supplementary Table S8 shows results of univariate analysis for 17 imaging and biopsy validated SNPs in the PanC4 sample

Supplementary Table S6 shows results of univariate analysis for 22 imaging defined nonalcoholic fatty liver disease SNPs in the PanC4 sample

Supplementary Table S9 shows descriptive Statistics for PanScan sample

Supplementary Figure S1 shows forest plot for minimally adjusted model for the PanScan sample.

Supplementary Figure S2 shows forest plot for minimally adjusted model for PanC4 sample