Defining type 2 diabetes polygenic risk scores through colocalization and network-based clustering of metabolic trait genetic associations

colocalization
disease subtyping
mendelian randomization
pleiotropy
polygenic risk scores
  • Core Principle: This study partitioned the genetic heterogeneity of Type 2 Diabetes (T2D) using a novel colocalization-first approach followed by network-based clustering of T2D and 20 related metabolic traits across 243 loci.
  • Key Finding: The method identified five distinct T2D biological pathways (Obesity, Lipodystrophic insulin resistance, Liver/lipid metabolism, Hepatic glucose metabolism, and Beta-cell dysfunction), successfully isolating genetically distinct disease mechanisms.
  • Clinical Significance: Partitioned Polygenic Risk Scores (PRSs) showed heterogeneous clinical associations in a validation cohort (n=21,742 T2D individuals); notably, the Lipodystrophic insulin resistance PRS and Beta-cell dysfunction PRS were causally associated with lower BMI, providing genetic validation for the clinically important “lean diabetes” sub-type.
  • Methodological Advance: By integrating colocalization and Mendelian Randomization, the framework provided stronger inferences on the causality and directionality of the genetic associations, which is essential for translating genetic discoveries into targeted T2D treatments.
Published

23 January 2026

PubMed: 38200577 DOI: 10.1186/s13073-023-01255-7 Overview generated by: Gemini 2.5 Flash, 26/11/2025

Key Findings: Dissecting T2D Heterogeneity via Pleiotropy

This study addresses the profound genetic and clinical heterogeneity of Type 2 Diabetes (T2D) by developing a refined framework to partition T2D-associated genetic variants into distinct biological pathways. The goal is to move beyond a single, monolithic T2D diagnosis toward stratified risk prediction and targeted therapeutic strategies. The method leverages the pleiotropic nature of genetic variants—their influence on multiple related traits—to define distinct mechanisms of T2D pathogenesis.

Colocalization-First Partitioning and Clustering

The authors improved upon previous clustering approaches by integrating rigorous statistical checks for shared causality, enhancing the mechanistic interpretability of the resulting risk scores.

  1. Colocalization Analysis: They applied colocalization analysis between T2D and 20 related metabolic traits (selected based on established risk factors and genetic correlation) across 243 T2D loci. This step robustly identified 146 T2D loci where the T2D association was likely caused by the same causal variant as the associated metabolic trait.
  2. Network-Based Clustering: A network-based unsupervised hierarchical clustering approach was then performed using the colocalized variant-trait associations. This successfully grouped the T2D risk loci into five distinct clusters, each representing a unique, interconnected set of T2D and metabolic risk factors.
  3. Causality Check (Mendelian Randomization): The study explicitly assessed the causality and directionality of the variant-trait associations using the Mendelian randomization (MR) Steiger’s Z-test. This confirmed that the genetic associations identified in the clusters were largely causal for the corresponding metabolic phenotypes.

Five Distinct T2D Pathophysiological Clusters

The five identified genetic clusters, which align with distinct T2D pathophysiologies, are: 1. Obesity (High BMI-T2D risk) 2. Lipodystrophic insulin resistance (T2D risk associated with fat distribution/dysfunction) 3. Liver and lipid metabolism 4. Hepatic glucose metabolism 5. Beta-cell dysfunction (Impaired insulin secretion)

Heterogeneous Clinical Profiles and “Lean Diabetes”

Partitioned Polygenic Risk Scores (PRSs) were generated for each cluster and externally validated in 21,742 individuals with T2D across three independent cohorts, demonstrating unique associations with metabolic and clinical outcomes:

  • Opposite BMI Associations: The Obesity PRS was strongly associated with a higher BMI, as expected. Critically, the Lipodystrophic insulin resistance PRS and Beta-cell dysfunction PRS were both associated with lower BMI.
  • Support for Lean Diabetes: The MR Steiger analysis provided causal evidence that increased T2D risk in the lipodystrophic insulin resistance and beta-cell dysfunction pathways was causally associated with lower BMI. This provides a genetic foundation for the “lean diabetes” or non-obese T2D sub-type, where risk is driven by dysfunctional fat/insulin-secretion rather than overall fat mass.
  • Comorbidity Stratification: The Lipodystrophic insulin resistance PRS was uniquely and specifically associated with a higher odds of chronic kidney disease (CKD) (Odds Ratio 1.29), suggesting that individuals whose T2D risk is driven by this pathway may require specific monitoring and intervention for renal complications.

Conclusion

The colocalization-first, network-based clustering methodology successfully and robustly partitioned the genetic heterogeneity underlying T2D. The resulting pathway-specific PRSs provide valuable tools for risk stratification, sub-type identification, and the development of targeted therapies based on an individual’s distinct genetic mechanism of disease.