{"doi":"10.1002/ctm2.70076","title":"Genetic variants explain ancestry‐related differences in type 2 diabetes risk","abstract":"Type 2 diabetes (T2D) is a global epidemic, affecting over 400 million people around the world.1 T2D causes devasting complications and is a leading risk factor for ischaemic heart disease and stroke, which are among the top causes of global morbidity and mortality.1 Classically, T2D occurs in adulthood in the setting of obesity and insulin resistance. Increasingly, however, T2D is understood to arise from a complex interplay of environmental and genetic factors, leading to heterogeneity in patient clinical presentation and disease course.2, 3 There have been many attempts to define T2D subtypes using a range of analytic methods, but few efforts have shown real-world clinical utility or given insight into disease pathophysiology.4 Over recent years, advances in large-scale genome-wide association studies (GWASs) have uncovered hundreds of genetic variants that modulate T2D risk. This genetic information has the potential to provide insight into disease biology; yet, clinical translation has been limited, often because the strongest genetic associations are not found in protein-coding regions, which makes it more challenging to identify causal genes and pathways. By leveraging the power of GWAS, our laboratory has developed a complex, high-throughput approach to define T2D disease mechanisms, which may help to identify T2D patient subtypes5, 6 (Figure 1). This approach aggregates GWAS results to assess the link between genetic variants and diabetes-related clinical traits, such as glucose, haemoglobin A1c and body mass index (BMI). We then apply a machine learning method called Bayesian non-negative matrix factorisation to group together closely related variants and traits into clusters. By analysing the top-weighted variants and traits in each cluster, we can infer the most likely biological mechanism contributing to that cluster. Notably, this ‘soft’ clustering method allows a given variant or trait to be assigned to more than one cluster. Most prior genetic analyses have focused on European populations, potentially limiting the applicability for other ancestry groups. To address this limitation, we recently applied our high-throughput pipeline to investigate T2D clusters using current large, multi-ancestry genetic studies.7 Through this approach, we confirmed our previously identified T2D genetic clusters and found three new clusters, yielding a total of 12 clusters. Three clusters were associated with beta cell dysfunction and insulin deficiency, while seven were associated with insulin resistance. Among the insulin resistance clusters, certain clusters were associated with obesity and above-average BMI, whereas two other clusters were associated with a ‘lipodystrophy-like’8 abnormal fat distribution and below-average BMI. Furthermore, we demonstrated significant associations between the clusters and clinical phenotypes. For instance, the lipodystrophy-like genetic clusters were associated with above-average risk of fatty liver disease, whereas another cluster defined by reduced cholesterol levels was associated with below-average risk of coronary artery disease. Importantly, we also assessed differences in T2D genetic clusters across ancestry groups. To do this, we aggregated the genetic variants from each cluster to capture each person's cluster-specific genetic risk. We found that certain clusters were more highly weighted in specific ancestry groups. In particular, the two lipodystrophy-like genetic clusters contained alleles more frequently seen in individuals with East Asian ancestry, compared to other populations (Figure 2). We hypothesised that these genetic differences might account for varied clinical presentations of T2D. Notably, it is well established that individuals with East Asian ancestry have a higher risk of T2D at lower BMI levels, compared to those with European ancestry.9 This association may be due to an increased risk of a lipodystrophy-like phenotype, with a greater degree of metabolically unhealthy","journal":"Clinical and Translational Medicine","year":2024,"id":479143,"datarank":0.0,"base_score":0.0,"endowment":0.0,"self_citation_contribution":0.0,"citation_network_contribution":0.0,"self_endowment_contribution":0.0,"citer_contribution":0.0,"corpus_percentile":null,"corpus_rank":null,"citation_count":2,"citer_count":0,"citers_with_citation_signal":0,"citers_with_endowment":0,"datacite_reuse_total":0,"is_dataset":false,"is_dataset_confidence":0.9599,"is_data_producer":false,"deposit_databanks":null,"is_oa":true,"file_count":0,"downloads":0,"has_version_chain":false,"published_date":"2024-01-01","fair_score":null,"fair_percentile":null,"algorithm_id":"datarank_citation_only_1hop_v6","ranking_scope":"data_only","authors":[{"id":852453,"name":"Kirk Smith","orcid":"0000-0002-4338-4812","position":1,"is_corresponding":false},{"id":7372,"name":"Miriam S. Udler","orcid":"0000-0003-3824-9162","position":2,"is_corresponding":false},{"id":988242,"name":"Aaron J. Deutsch","orcid":"0000-0001-6750-5335","position":0,"is_corresponding":true}],"reference_count":11,"raw_metadata":null,"created_at":"2026-07-19T02:06:50.355747Z","pmid":"39500627","pmcid":null,"fwci":null,"citation_percentile":null,"influential_citations":0,"oa_status":null,"license":null,"views":0,"total_file_size_bytes":0,"version_count":0,"fair_f":null,"fair_a":null,"fair_i":null,"fair_r":null,"fair_zscore":null,"fair_rationale":null,"fair_model":null,"fair_agent_version":null,"fair_fulltext_source":null,"fair_has_llm":null,"fair_computed_at":null,"clinical_trials":[],"software_tools":[],"db_accessions":[],"linked_datasets":[],"topics":[]}