Development and validation of a trans-ancestry polygenic risk score for type 1 diabetes

Area of Science:

• Genetics
• Epidemiology
• Biostatistics

Background:

• Type 1 diabetes (T1D) exhibits high heritability, making polygenic risk scores (PRSs) valuable for risk screening.
• Current PRSs for T1D, primarily developed using European data, demonstrate reduced efficacy in non-European populations.
• Developing a trans-ancestry PRS is crucial for equitable and effective genetic risk assessment in diverse populations.

Purpose of the Study:

• To develop and validate a trans-ancestry polygenic risk score (TA-PRS) for type 1 diabetes with comparable performance across multiple ancestries.
• To improve upon existing European-ancestry-based PRSs for T1D in non-European populations.

Main Methods:

• Utilized the PRS-CSx method with genome-wide association study data from European, East Asian, African American, and Hispanic individuals (N=29,469 cases).
• Developed a TA-PRS by combining a non-HLA component (over one million variants) and the HLA component of a European PRS (GRS2x).
• Evaluated PRS performance using area under the receiver operating curve (AUROC), sensitivity, and specificity in multi-ancestry cohorts and validated in independent cohorts.

Main Results:

• The developed TA-PRS achieved a significantly higher AUROC (0.89) compared to the European-based GRS2x (0.85) in a multi-ancestry cohort.
• TA-PRS demonstrated improved sensitivity across various ancestries, including Europeans (0.71) and South Asians (0.77), outperforming GRS2x (0.56 in Europeans).
• Specificity remained acceptable (≥0.83) across all ancestries, with validation confirmed in four independent cohorts.

Conclusions:

• A novel trans-ancestry PRS (TA-PRS) for type 1 diabetes has been successfully developed, outperforming existing European-centric models.
• The TA-PRS exhibits comparable predictive performance across diverse ancestries, supporting its utility in population-wide screening programs.
• This advancement facilitates more equitable genetic risk stratification for type 1 diabetes, enabling targeted surveillance and preventive strategies.