Complex traits arise from the combined effects of rare and common genetic variation, development and environment, but resolving their joint contributions has been limited by statistical power. Here, we meta-analyze effects of recurrent copy number variants (CNVs), polygenic scores, sex, age and medications on height and body mass index in 1,447,001 individuals across 6 biobanks and clinical cohorts. CNVs show largely mirror dose-dependent effects of deletions and duplications on both traits, but a subset of loci exhibit asymmetric dose-responses on adult height, consistent with buffering of one allele but not the other. Polygenic background and medications combine with CNVs in ways broadly consistent with additivity. However, detailed analyses of loci at 16p11.2 and 22q11.2 reveal context-dependent effects that vary across development, physiology and sex. At 22q11.2, the net effect of a CNV reflects opposing and reinforcing contributions of multiple genes, providing a potential mechanism for buffering of dosage effects. These results indicate that genetic effects follow additive patterns in aggregate, while context-dependent deviations are widespread for specific loci.

Background: Polygenic risk scores (PRS) are increasingly recognised as a promising tool for stratifying breast cancer risk, thus contributing to personalised screening strategies. However, their clinical utility, cost-effectiveness, and potential for equitable implementation across European health systems remain uncertain. This study aimed to develop evidence-based, context-sensitive recommendations for the integration of PRS into breast cancer screening using the Grading of Recommendations Assessment, Development and Evaluation Evidence to Decision (GRADE EtD) framework. Methods: A multidisciplinary panel of nine experts was convened under the EU4Health-funded CAN.HEAL project. We synthesised findings from a systematic review of decision-analytic models evaluating the clinical utility of PRS-enhanced screening for breast cancer compared to standard age-based protocols, alongside evidence collected on feasibility, acceptability, cost-effectiveness, and equity. Panellists rated the evidence and issued recommendations via a structured two-round consensus process. Results: Four decision-analytic models demonstrated potential benefits of PRS-enhanced screening in terms of life-years gained and reduced mortality. However, these benefits were partially offset by increases in false positives and overdiagnosis. Evidence on implementation revealed mixed acceptability, particularly regarding psychological impact and data privacy concerns, whilst feasibility was constrained by limited genomic infrastructure and workforce readiness. Cost-effectiveness varied by setting and model assumptions. Equity concerns were substantial, with existing PRS models primarily calibrated to populations of European ancestry. Conclusions: In light of the very low certainty of evidence and context-dependent trade-offs, the panel conditionally recommends against the routine adoption of PRS in organised breast cancer screening at this stage. Instead, national and regional systems should support implementation studies to evaluate clinical utility, encompassing equity and system readiness issues. Responsible implementation will require iterative learning, stakeholder co-design, and built-in evaluation to ensure that future scale-up is evidence-informed, ethically sound, and system-adapted.

Genetic predisposition and alcohol consumption are risk factors for increased blood pressure (BP), but their interactions influencing BP remain understudied. We conducted population-specific and cross-population meta-analyses of genome-wide gene-alcohol (GxAlc) interactions affecting BP in >1.1M individuals from multiple populations. We identified 46 GxAlc interaction loci for BP, including 21 from one-degree-of-freedom interaction tests (PGxAlc<5×10-8; or <0.05/Meff, Meff independent BP associations at P<10-5), and 25 from two-degree-of-freedom tests of main and interaction effects (PGxAlc<0.05/M2df, M2df independent 2df-associations at P2df<5×10-8), including 7 novel and 39 known BP loci. The 12q24 locus highlights the genetic effect of BRAP-rs11066001 on BP, being ~6 times larger in current drinkers than in non-drinkers. Gene prioritization with 46 GxAlc loci identified 15 genes with ≥3 lines of evidence (location, literature, druggability, functional/regulatory annotation, or pathway analyses). Several loci showed sex- and population-specific effects and revealed biological pathways of alcohol's influence on BP, suggesting mechanisms underlying alcohol-induced hypertension.

Introduction: Despite identifying hundreds of susceptibility loci for coronary artery disease (CAD), the genetic factors that contribute to risk more strongly or specifically in one sex remain poorly understood. Based on sexually dimorphic patterns observed in prior metabolomics studies, we tested the hypothesis that genetically decreased activity of carbamoyl-phosphate synthetase 1 ( CPS1 ), the rate-limiting enzyme of the urea cycle, was associated with atherosclerosis differentially in male and female humans and mice. Methods: Logistic regression was used to test a functional genetic variant (rs715) in CPS1 for association with risk of CAD separately in men and women from multiple multi-ancestry large cohorts (n=1,803,558), followed by fixed-effects meta-analyses across all subjects, ancestries, and datasets. A genetic mouse model of Cps1 deficiency was characterized for sex-specific differences in metabolite levels, cardiometabolic traits, and atherosclerotic lesion formation. Sex differences in hepatic CPS1 expression were evaluated in subjects from the STARNET cohort and a panel of inbred mouse strains. Results: The CPS1 activity-decreasing allele of rs715 exhibited a significant sexually dimorphic association with decreased risk of CAD (P-heterogeneity=1.3x10 -4 ) in women (OR=0.95, 95% CI 0.94-0.96; P=9.5x10 -13 ; n=1,016,646) compared to men (OR=0.99, 95% CI 0.98-1.00; P=5.3x10 -3 ; n=786,912). Heterozygous Cps1 deficiency decreased aortic lesion formation in male mice compared to wildtype male littermates (452,855±23,550 vs. 343,481±29,752µm 2 /section; P=8.7x10 -3 ) but not in female mice. However, Cps1 deficiency in mice did not lead to sexually dimorphic differences in levels of plasma amino acids or atherogenic metabolites. Regardless of rs715 genotype, hepatic CPS1 expression was ~19% lower in women compared to men (11.5±1.3 vs. 11.8±1.1 log 2 CPM; P=1.1x10 -5 ) while the opposite profile was observed among a panel of inbred mouse strains with male mice having ~28% lower expression than female mice (12.2±0.13 vs. 12.5±0.09 log 2 units; P=2.7x10 -51 ). Gonadectomy experiments did not provide evidence that sex hormones played a role in regulating Cps1 expression in the liver. Conclusions: These results provide evidence for the atheroprotective properties of genetically decreased CPS1 activity in humans and mice through sexually dimorphic pattens. Additional studies will be needed to elucidate the underlying biological mechanisms for these differential effects.

Summary Background Amiodarone is a widely used antiarrhythmic which frequently induces thyroid dysfunction, including both amiodarone-induced hypothyroidism (AIH) and thyrotoxicosis (AIT). Whether genetic factors contribute to these adverse drug reactions is unknown. In this study, we aimed to identify genetic variants that influence the risk of amiodarone-induced thyroid dysfunction and to evaluate their potential to support genotype-guided risk screening. Methods This pharmacogenetic study comprised two genome-wide meta-analyses of AIH and AIT using five datasets (Copenhagen Hospital Biobank, The Danish Blood Donor Study, Estonian Biobank, deCODE genetics, and Mass General Brigham Biobank). Key measures included the odds ratio (OR) per risk allele, the variants’ effects on spontaneous thyroid disease and biomarkers, and their clinical predictive ability, assessed by the area under the receiver operating curve (AUC), positive and negative predictive values (PPV and NPV). Findings The AIH meta-analysis (880 cases, 4,031 controls) identified three genome-wide significant loci in: FOXE1 (rs36052460; OR 2.58, allele frequency [AF] = 64.3%, P = 2.55 × 10 −44 ), FOXA2 (rs2424459; OR 1.67, AF = 71.3%, P = 2.59 × 10 −14 ), and ADAM32 (rs12681571; OR 1.49, AF = 61.8%, P = 3.05 × 10 −9 ). The AIT meta-analysis (385 cases, 4,936 controls) identified one locus in CAPZB (rs867355; OR 1.63, AF = 66.1%, P = 3.49 × 10 −8 ). In risk prediction models, a polygenic risk score (PRS) of the AIH variants increased the AUC by 9.2% (95% CI 6.6 – 11.9%), which outperformed a genome-wide hypothyroidism PRS (1.5% AUC increase, 95%CI 0.0 – 2.9%). Similarly, the CAPZB variant improved AIT prediction (AUC increase of 4.0%, 95% CI 0.4 – 7.5%) beyond a hyperthyroidism PRS (0.2% AUC increase, 95%CI -0.8 – 1.2%). Genotype-guided screening would identify individuals at low risk (NPVs ranging from 90-95% and PPVs 2-20%). Interpretation We identified genetic variants that influence the risk of developing amiodarone-induced thyroid dysfunction. Genotype-guided screening offers a potential complement to current strategies and personalize pre-treatment risk assessment for patients initiating amiodarone therapy.

Background:Atrioventricular node ablation (AVNA) is an effective rate-control strategy for patients with AF refractory to medical therapy. However, data on long-term survival outcomes and prognostic factors in this population remain limited. Thus, we aimed to evaluate long-term mortality in patients with refractory AF who underwent AVNA in a single country. Methods:We conducted a nationwide retrospective cohort study including all patients who underwent AVNA followed by permanent pacing (n=435) in Estonia from 2012 to 2022. The primary endpoint was all-cause mortality. Comparisons were made between CRT and right ventricular (RV) pacing, AVNA timing (≤6 months and later following device implantation) and by patient sex. Results:Acute success was achieved in 418 (96.1%); mean age was 73.2 ± 10.2 years and 56.5% were female. Median follow-up was 4.2 years (range 2 days to 12.9 years), with median survival of 7.8 years. No significant survival difference was observed between CRT and RV pacing overall nor in ad hoc AVNA and pacemaker implantation cases. Among CRT recipients, early AVNA (≤6 months post-implantation) significantly reduced mortality risk compared with later AVNA (HR 0.48; 95% CI [0.24-0.96]; p=0.038). In RV-paced patients, timing showed no survival impact, but renal function independently predicted mortality (p<0.001). Survival was similar between sexes. Conclusion:In this 10-year nationwide cohort, AVNA with permanent pacing resulted in survival comparable to local population norms. CRT provided no universal survival benefit over RV pacing, but early AVNA was associated with significantly improved outcomes in CRT recipients.

Background/Objectives: Metabolomics, in combination with genetic data, is a powerful approach to study the biochemical consequences of genetic variation. We assessed the impact of human gene knockouts (KOs) on the metabolite levels of Estonia Biobank (EstBB) participants and integrated the results with electronic health record data. Methods: In 150,000 EstBB genotyped participants, we identified 723 KOs with 152 different predicted loss of function (pLoF) variants in 115 genes. For those KOs and 258 controls, 1387 metabolites were profiled using ultra-high-performance liquid chromatography-tandem mass spectrometry. Results: We identified 48 associations linking rare pLoF variants in 22 genes to 43 metabolites. Out of 48 associations, 27 (56%) were found in genes that cause inborn errors of metabolism. The top associations identified in our analysis included genes and metabolites involved in the degradation pathway of the pyrimidine bases uracil and thymine (DPYD and UPB1). We found DPYD gene KOs to be associated with elevated levels of Uracil, confirming that DPD-deficiency is a leading cause of severe 5-Fluorouracil toxicity. Overall, 54% of reported associations are gene targets of approved drugs or bioactive drug-like compounds. Conclusions: Our findings contribute to assessing the impact of human KOs on metabolite levels and offer insights into gene functions, disease mechanism, and drug target validation.

BACKGROUND AND AIMS:Deviations from the population mean in sleep duration have been associated with increased risk for developing dyslipidemia and atherosclerotic cardiovascular disease, but the mechanism of effect is poorly characterized. We performed large-scale genome-wide gene-sleep interaction analyses of lipid levels to identify genetic variants underpinning the biomolecular pathways of sleep-associated lipid disturbances and to suggest possible druggable targets. METHODS:We collected data from 55 cohorts with a combined sample size of 732,564 participants (87 % European ancestry) with data on lipid traits (high-density lipoprotein [HDL-c] and low-density lipoprotein [LDL-c] cholesterol and triglycerides [TG]). Short (STST) and long (LTST) total sleep time were defined by the extreme 20 % of the age- and sex-standardized values within each cohort. Based on cohort-level summary statistics data, we performed meta-analyses for one-degree of freedom tests of interaction and two-degree of freedom joint tests of the SNP-main and -interaction effect on lipid levels. RESULTS:The one-degree of freedom variant-sleep interaction test identified 10 novel loci (Pint<5.0e-9), and we additionally identify 7 loci within the two-degree of freedom analyses (Pjoint<5.0e-9 in combination with Pint<6.6e-6). Multiple loci, including those mapped to APSH (target for aspartic and succinic acid) and SLC8A1 showed biological plausibility and druggability potential based on literature. CONCLUSIONS:Collectively, the 17 (9 with short and 8 with long sleep) loci provided evidence into the biomolecular mechanisms underlying sleep-associated lipid changes, including potential involvement of the vitamin D receptor pathway. Collectively, these findings may contribute developing novel interventions for treating dyslipidemia in people with sleep disturbances.

Large biobanks have set a new standard for research and innovation in human genomics and implementation of personalized medicine. The Estonian Biobank was founded a quarter of a century ago, and its biological specimens, clinical, health, omics, and lifestyle data have been included in over 800 publications to date. What makes the biobank unique internationally is its translational focus, with active efforts to conduct clinical studies based on genetic findings, and to explore the effects of return of results on participants. In this review, we provide an overview of the Estonian Biobank, highlight its strengths for studying the effects of genetic variation and quantitative phenotypes on health-related traits, development of methods and frameworks for bringing genomics into the clinic, and its role as a driving force for implementing personalized medicine on a national level and beyond.

Over 85% of the population experience acne at some point in their lives, with its severity spanning a quantitative spectrum, from mild, transient outbreaks to more persistent, severe forms of the condition. Moderate to severe disease poses a substantial global burden arising from both the physical and psychological impacts of this highly visible condition. The analytical approach taken in this study aimed to address the impact of variation in the dichotomisation of acne case control status, driven by ascertainment and study design, on effect size estimates across independent genetic association studies of acne. Through a fixed intercept meta-regression framework, we combined evidence genome-wide for association with acne across studies in which case-control status had been ascertained in different settings, allowing for different severity threshold definitions. Across a combined sample of 73,997 cases and 1,103,940 controls of European, South Asian and African American ancestry we identify genetic variation at 165 genomic loci that influence acne risk. There is evidence for both shared and ancestry specific components to the genetic susceptibility to acne and for sex differences in the magnitude of effect of risk alleles at three loci. We observe that common genetic variation explains 13.4% of acne heritability on the liability scale. Consistent with the hypothesis that genetic risk primarily operates at the level of individual pilosebaceous units, a polygenic score derived from this case-control study of acne susceptibility is associated with both self-reported and clinically assessed acne severity in adolescence, further strengthening the link between genetic risk and disease severity. Prioritisation of causal genes at the identified acne risk loci, provides genetic validation of the targets of established and emerging acne therapies, including retinoid treatments. The identified acne risk loci are enriched for genes encoding downstream effectors of RXRA signalling, including SOX9 and components of the WNT and p53 pathways. Illustrating that the control of stem cell lineage plasticity and cellular fate are important mechanisms through which genetic variation influences acne susceptibility within the pilosebaceous unit.

BACKGROUND:Dopaminergic neuron depletion in the substantia nigra (SN) and the pathological aggregation of α-synuclein are the neuropathological hallmarks of Parkinson's disease (PD). OBJECTIVES:This study aimed to investigate the association between the polygenic risk score for PD (PD-PRS) and transcranial sonography (TCS)-measured SN hyperechogenicity to enhance the accuracy of PD susceptibility prediction. METHODS:PD-PRSs were calculated for over 41,000 Estonian Biobank participants age 55+ years without a PD diagnosis. Participants in the highest and lowest PD-PRS percentiles (n = 222) underwent TCS measurements and Sniffin' sticks olfactory testing. A multivariable logistic regression model was used to examine the associations between PD-PRS, risk and prodromal markers, and SN hyperechogenicity. RESULTS:Data from 204 participants with TCS measurements were analyzed, including 107 individuals in the high-risk PD-PRS group and 97 in the low-risk PD-PRS group. Incorporating PD-PRS group assignment improved the explained variance in SN hyperechogenicity from 17.2% to 31.9%. Participants in the low-risk PD-PRS group had 0.16 times lower odds (95% confidence interval (CI) = 0.07-0.35, P < 0.001) of developing SN hyperechogenicity compared to high-risk PD-PRS individuals. Each unit increase in the Sniffin' sticks olfactory test score was significantly associated with reduced odds of SN hyperechogenicity (adjusted odds ratio = 0.60, 95% CI = 0.47-0.78, P = 0.002). CONCLUSIONS:Our findings indicate that TCS-measured SN hyperechogenicity is associated with PD-PRS and olfactory impairment. This combined assessment may improve early diagnosis of prodromal PD by pinpointing individuals at increased risk.

OBJECTIVE:The clinical efficacy of remote ischaemic preconditioning (RIPC) remains unclear. This pilot study, a substudy of a randomised controlled trial, tested whether repeated RIPC reduces arterial stiffness, end organ damage, and oxidative stress in patients with intermittent claudication (IC). METHODS:In a single centre, randomised, sham controlled, double blind trial, 42 males with Fontaine stage IIa or IIb IC were allocated at a ratio of 1:1 to receive RIPC or sham using an automated device for 28 days in an outpatient setting (ClinicalTrials.gov ID NCT05084066). Secondary outcomes included changes in augmentation index (AIx), heart rate corrected AIx, carotid-femoral pulse wave velocity (cf-PWV), and biomarkers for cardiac (high sensitivity troponin T [hs-TnT], N-terminal pro B-type natriuretic peptide [NT-proBNP]), renal (creatinine, urea, cystatin C, β2 microglobulin, neutrophil gelatinase associated lipocalin, kidney injury molecule 1), and oxidative stress (high sensitivity C reactive protein, interleukin-6, interleukin-18, myeloperoxidase, adiponectin, oxidised low density lipoprotein). RESULTS:Data from 41 patients (RIPC n = 23, sham n = 18) aged 64.9 ± 7.4 years were analysed. The median change in cf-PWV was 0.2 m/s (interquartile range [IQR] -0.6, 0.6) in the RIPC group vs. 0.2 m/s (IQR -0.3, 0.8) in the sham group (p = .54). The median change in hs-TnT was 0 ng/L (IQR -1, 2) in the RIPC group vs. 1 ng/L (IQR -1, 2) in the sham group (p = .54). NT-proBNP showed a median change of -7 ng/L (IQR -32, 18) in the RIPC group vs. 6 ng/L (IQR -14, 35) in the sham group (p = .14). No statistically significant differences were observed between groups for arterial stiffness, oxidative stress, or renal biomarkers. CONCLUSION:Repeated RIPC did not significantly alter arterial stiffness, end organ damage, or oxidative stress biomarkers compared with sham treatment. It is possible that patients with IC already experience repeated RIPC from their ischaemic legs, thereby attenuating any additional effects from arm induced RIPC.

Although both short and long sleep duration are associated with elevated hypertension risk, our understanding of their interplay with biological pathways governing blood pressure remains limited. To address this, we carried out genome-wide cross-population gene-by-short-sleep and long-sleep duration interaction analyses for three blood pressure traits (systolic, diastolic, and pulse pressure) in 811,405 individuals from diverse population groups. We discovered 22 novel gene-sleep duration interaction loci for blood pressure, mapped to 23 genes. Investigating these genes’ functional implications shed light on neurological, thyroidal, bone metabolism, and hematopoietic pathways that necessitate future investigation for blood pressure management that caters to sleep health lifestyle. Non-overlap between short sleep (12) and long sleep (10) interactions underscores the plausible nature of distinct influences of both sleep duration extremes in cardiovascular health. Several of our loci are specific towards a particular population background or sex, emphasizing the importance of addressing heterogeneity entangled in gene-environment interactions, when considering precision medicine design approaches for blood pressure management.

Acne vulgaris is highly prevalent, with 85% of the population experiencing it at some point in their lives. Disease severity spans a quantitative spectrum, from mild, transient outbreaks to more persistent, severe forms of the condition. Moderate-to-severe disease poses a substantial global burden arising from both the physical and psychological impacts of this highly visible condition (Layton AM, Thiboutot D, Tan J. Reviewing the global burden of acne: how could we improve care to reduce the burden? Br J Dermatol 2021; 184: 219–25). Acne pathogenesis is due to factors contributing to comedogenesis and inflammation within the pilosebaceous unit. This study aims to characterize the genetic contribution to acne through a genome-wide association metaregression. Through a fixed-intercept metaregression framework, we combined evidence from genome-wide studies encompassing 73 997 cases and 1 103 940 controls of European, South Asian and African American ancestry. Our approach addressed the impact of variation in the dichotomization of acne case–control status, enabling generation of a more accurate polygenic risk score (PRS) for acne. We evaluated the performance of our PRS using two independent cohorts totalling 4540 individuals with varying acne severity. Our study also incorporated identification of ancestry-specific components to genetic susceptibility, as well as sex-stratified analysis. We identified genetic variation at 165 genomic loci that influence acne risk and found that common genetic variation explains 13.4% of acne heritability on the liability scale. We established evidence for both shared and ancestry-specific components to the genetic susceptibility to acne, and identified sex differences in the magnitude of effect of risk alleles at three loci. PRS performance was initially evaluated using an independent cohort of 2058 individuals with self-reported acne (severity classed as none, mild, moderate, severe). We found that the mean of the PRS distribution significantly increased with self-reported disease severity, with one SD increase in PRS corresponding to 1.63-fold increase in the likelihood of reporting more severe disease. To further validate this tool, we then investigated the distribution of PRS in a cross-sectional cohort of 2482 teenage patients. Within this cohort acne severity was categorized (clear/almost clear, mild, moderate/severe) by trained dermatologists from three-dimensional facial photographs taken at age ∼13 years. Consistently with the findings for self-reported acne, a single SD increase in polygenic risk corresponded to a 1.53-fold increase in the odds of being one severity category higher. A PRS model derived from a genome-wide association metaregression enables prediction of acne severity, demonstrating proof-of-principle of a genomics-based precision medicine tool.

The proteome holds great potential as an intermediate layer between the genome and phenome. Previous protein quantitative trait locus studies have focused mainly on describing the effects of common genetic variations on the proteome. Here, we assessed the impact of the common and rare genetic variations as well as the copy number variants (CNVs) on 326 plasma proteins measured in up to 500 individuals. We identified 184 cis and 94 trans signals for 157 protein traits, which were further fine-mapped to credible sets for 101 cis and 87 trans signals for 151 proteins. Rare genetic variation contributed to the levels of 7 proteins, with 5 cis and 14 trans associations. CNVs were associated with the levels of 11 proteins (7 cis and 5 trans ), examples including a 3q12.1 deletion acting as a hub for multiple trans associations; and a CNV overlapping NAIP , a sensor component of the NAIP-NLRC4 inflammasome which is affecting pro-inflammatory cytokine interleukin 18 levels. In summary, this work presents a comprehensive resource of genetic variation affecting the plasma protein levels and provides the interpretation of identified effects.

We performed large-scale genome-wide gene-sleep interaction analyses of lipid levels to identify novel genetic variants underpinning the biomolecular pathways of sleep-associated lipid disturbances and to suggest possible druggable targets. We collected data from 55 cohorts with a combined sample size of 732,564 participants (87% European ancestry) with data on lipid traits (high-density lipoprotein [HDL-c] and low-density lipoprotein [LDL-c] cholesterol and triglycerides [TG]). Short (STST) and long (LTST) total sleep time were defined by the extreme 20% of the age- and sex-standardized values within each cohort. Based on cohort-level summary statistics data, we performed meta-analyses for the one-degree of freedom tests of interaction and two-degree of freedom joint tests of the main and interaction effect. In the cross-population meta-analyses, the one-degree of freedom variant-sleep interaction test identified 10 loci (Pint<5.0e-9) not previously observed for lipids. Of interest, the ASPH locus (TG, LTST) is a target for aspartic and succinic acid metabolism previously shown to improve sleep and cardiovascular risk. The two-degree of freedom analyses identified an additional 7 loci that showed evidence for variant-sleep interaction (Pjoint<5.0e-9 in combination with Pint<6.6e-6). Of these, the SLC8A1 locus (TG, STST) has been considered a potential treatment target for reduction of ischemic damage after acute myocardial infarction. Collectively, the 17 (9 with STST; 8 with LTST) loci identified in this large-scale initiative provides evidence into the biomolecular mechanisms underpinning sleep-duration-associated changes in lipid levels. The identified druggable targets may contribute to the development of novel therapies for dyslipidemia in people with sleep disturbances.

Type 2 diabetes (T2D) is a heterogeneous disease that develops through diverse pathophysiological processes1,2 and molecular mechanisms that are often specific to cell type3,4. Here, to characterize the genetic contribution to these processes across ancestry groups, we aggregate genome-wide association study data from 2,535,601 individuals (39.7% not of European ancestry), including 428,452 cases of T2D. We identify 1,289 independent association signals at genome-wide significance (P < 5 × 10-8) that map to 611 loci, of which 145 loci are, to our knowledge, previously unreported. We define eight non-overlapping clusters of T2D signals that are characterized by distinct profiles of cardiometabolic trait associations. These clusters are differentially enriched for cell-type-specific regions of open chromatin, including pancreatic islets, adipocytes, endothelial cells and enteroendocrine cells. We build cluster-specific partitioned polygenic scores5 in a further 279,552 individuals of diverse ancestry, including 30,288 cases of T2D, and test their association with T2D-related vascular outcomes. Cluster-specific partitioned polygenic scores are associated with coronary artery disease, peripheral artery disease and end-stage diabetic nephropathy across ancestry groups, highlighting the importance of obesity-related processes in the development of vascular outcomes. Our findings show the value of integrating multi-ancestry genome-wide association study data with single-cell epigenomics to disentangle the aetiological heterogeneity that drives the development and progression of T2D. This might offer a route to optimize global access to genetically informed diabetes care.