Pathway insights and predictive modeling for type 2 diabetes using polygenic risk scores

Abstract

Type 2 diabetes (T2D) poses a significant global health burden. We developed a polygenic risk score (PRS) model based on genome-wide association study (GWAS) findings and integrated it with clinical data to predict T2D risk. This study analyzed electronic medical records from a major medical center in Taiwan, comprising 315,424 T2D cases and 141,484 controls. Fourteen genome-wide significant SNPs were identified and used to construct the T2D PRS. The integrated predictive model showed high accuracy (AUROC 0.842) and was validated in the Taiwan Biobank. A risk score ranging from 0 to 19 was established for clinical use. Phenome-wide association study (PheWAS) revealed links between PRSs and T2D-related complications, such as diabetic retinopathy and hypertension. Pathway analysis highlighted biological processes including IL-15 production and WNT/β-catenin signaling. Our findings support the use of PRSs in personalized T2D risk assessment and early prevention strategies.

Keywords: Pathway analysis; PheWAS; Polygenic risk score; Score system; Type 2 diabetes.

Fig. 1

Flowchart of the study design.

Fig. 2

Distribution of the polygenic risk scores, corresponding strata plot, and the receiver operating characteristic curve of the polygenic risk scores for the prediction of type 2 diabetes in the (A, B, C) target and (D, E, F) validation groups. Clinical variables included T2D_Diag_Age, gender, waist circumference, blood pressure, and diabetic family history.

Fig. 3

Phenome-wide association study analysis of the correlation between polygenic risk scores and human diseases (SNPs gene loci, denoting a P < 5 × 10^–8).