Journal: medRxiv

Article Title: Cross-omic dissection reveals locus-specific heterogeneity and antagonistic pleiotropy between Alzheimer’s disease and type 2 diabetes

doi: 10.64898/2026.03.23.26349030

Figure Lengend Snippet: Workflow integrating Alzheimer’s disease (AD) and type 2 diabetes (T2D) GWAS with expression QTL (eQTL) and methylation QTL (mQTL) datasets. Genome-wide overlap was assessed using linkage disequilibrium score regression (LDSC) and SNP effect concordance analysis (SECA). Local genetic architecture was evaluated using local analysis of [co]variant association (LAVA) and bidirectional two-sample Mendelian randomisation (MR). Cross-trait variant and locus-level sharing was examined through RE2 random-effects meta-analysis and GWAS-pairwise (GWAS-PW) colocalisation. Gene-level and regulatory analysis was assessed using MAGMA gene-based tests and summary-data–based MR (SMR) integrating eQTL and mQTL data.

Article Snippet: We first quantified genome-wide polygenic overlap using linkage disequilibrium score regression (LDSC) and assessed SNP effect concordance as well as SNP-level genetic overlap across p-value thresholds using SNP effect concordance analysis (SECA) .

Techniques: Expressing, Methylation, Genome Wide, Variant Assay

Journal: medRxiv

Article Title: Cross-omic dissection reveals locus-specific heterogeneity and antagonistic pleiotropy between Alzheimer’s disease and type 2 diabetes

doi: 10.64898/2026.03.23.26349030

Figure Lengend Snippet: Panels a–d show global genetic correlations between AD and T2D estimated using LDSC. Results are shown, including (a–b) and excluding (c–d) the APOE region. Panels are stratified by AD GWAS: AD(J) denotes the GWAS (combining clinically diagnosed and proxy cases), and AD(L) denotes the GWAS (clinically diagnosed cases only). Points represent the genetic correlation estimate (r G ), horizontal lines indicate 95% confidence intervals, and the vertical dashed line marks r G ⍰=⍰0. Panels e–h show results from SECA for AD–T2D GWAS pairs, again including (e, g) and excluding (f, h) the APOE region. Panels e–f display the percentage of SNP subsets with significant concordant effects between traits (out of 144 tested subsets). Panels g–h display ORs quantifying the association between SNP effect directions in AD and T2D, with 95% confidence intervals and a dashed vertical line indicating OR⍰=⍰1. In the SECA bar plots, asterisks reflect permuted p-values for observing the number of significant concordant SNP subsets. Permuted P is the p-value for observing significant SNP subsets in the observation. SECA calculates the permuted P value for the number of significant associations with adjustment for testing 144 associations (based on permutations of 1000 replicates). When testing for the association between effect direction in Trait 1 SNPs and Trait 2 SNPs at P SNP < 0.05, we used Fisher’s exact test (two-sided). Note: The SNP effects in trait 1 and trait 2 were positively correlated; given that it is significant, the results indicate the presence of allelic effects that increase the risk for both traits. Asterisks denote statistical significance using conventional thresholds (⍰p⍰<⍰0.05; **⍰p⍰<⍰0.01; ***⍰p⍰<⍰0.001). In LDSC analyses (panels a–d), asterisks reflect the p-value testing whether the global genetic correlation (r G ) differs from zero. In SECA analyses (panels e–h), asterisks reflect either permuted p-values for the number of significant concordant SNP subsets (bar plots) or two-sided Fisher’s exact test p-values for association between SNP effect directions (forest plots). Abbreviations: AD(J) : Alzheimer’s disease GWAS by Jansen et⍰ al., AD(L) : Alzheimer’s disease GWAS by Lambert et⍰al., LDSC: linkage disequilibrium score regression, SECA: single nucleotide polymorphism effect concordance analysis, T2D(M) : T2D GWAS by Mahajan etIZal., T2D(BMIadj) : T2D GWAS adjusted for body-mass index, T2D(UKB): UK Biobank T2D Phecode 250.2, T2D(X) : T2D GWAS by Xue et⍰al., T2D(FG) : FinnGen Release⍰10 T2D.

Article Snippet: We first quantified genome-wide polygenic overlap using linkage disequilibrium score regression (LDSC) and assessed SNP effect concordance as well as SNP-level genetic overlap across p-value thresholds using SNP effect concordance analysis (SECA) .

Techniques: