Ancestry-Adjusted Polygenic Risk Scores for Predicting Obesity Risk in the Indonesian Population

TL;DR

This study develops an ancestry-adjusted polygenic risk score for obesity tailored to Indonesians, improving prediction accuracy by accounting for genetic diversity and population-specific factors.

Contribution

It introduces a novel method for adjusting PRS using principal components to enhance obesity risk prediction in the Indonesian population.

Findings

Ancestry-adjusted PRS increased AUC by 5% over unadjusted scores.

The method improves obesity risk classification accuracy.

Population-specific adjustments are crucial for effective genetic risk assessment.

Abstract

Obesity prevalence in Indonesian adults increased from 10.5% in 2007 to 23.4% in 2023. Studies showed that genetic predisposition significantly influences obesity susceptibility. To aid this, polygenic risk scores (PRS) help aggregate the effects of numerous genetic variants to assess genetic risk. However, 91% of genome-wide association studies (GWAS) involve European populations, limiting their applicability to Indonesians due to genetic diversity. This study aims to develop and validate an ancestry adjusted PRS for obesity in the Indonesian population using principal component analysis (PCA) method constructed from the 1000 Genomes Project data and our own genomic data from approximately 2,800 Indonesians. We calculate PRS for obesity using all races, then determine the first four principal components using ancestry-informative SNPs and develop a linear regression model to predict PRS based on these principal components. The raw PRS is adjusted by subtracting the predicted score to obtain an ancestry adjusted PRS for the Indonesian population. Our results indicate that the ancestry-adjusted PRS improves obesity risk prediction. Compared to the unadjusted PRS, the adjusted score improved classification performance with a 5% increase in area under the ROC curve (AUC). This approach underscores the importance of population-specific adjustments in genetic risk assessments to enable more effective personalized healthcare and targeted intervention strategies for diverse populations.