How Taiwan’s Giant Genomics Project Is Rewriting the Future of Disease Prediction

A sweeping genomic effort in Taiwan has revealed something that global precision medicine has long overlooked, that the best way to predict disease is to study the people who will be living with its consequences. Researchers at Academia Sinica have now shown that building genetic risk tools tailored to Han Chinese populations can transform how common illnesses are forecast and understood.

In work published in Nature on October 15, 2025, scientists analyzed genomic and clinical data from more than half a million participants in the Taiwan Precision Medicine Initiative. By conducting the largest genome wide association analysis of Han Chinese individuals to date, they developed the first population specific polygenic risk score models for diseases ranging from type 2 diabetes to autoimmune disorders to heart disease, achieving markedly stronger accuracy than tools based on European data. “This project marks a milestone for precision medicine in East Asia,” said Dr. Cathy S. J. Fann, senior corresponding author at Academia Sinica. “By integrating large scale genomic and clinical data, we are building predictive models that truly reflect the real genetic architecture of our population.”

What Population Specific Genetics Reveals

The research team combed through 695 diseases and 24 quantitative traits in 463,447 Han Chinese participants and identified 2,656 independent genetic associations, including 95 previously unknown variants. Those signals, which collectively explain up to 10.3 percent of health variation in the Taiwanese population, illuminate shared genetic roots across major disease clusters, including cardiometabolic traits, autoimmune and infectious disorders, and kidney related conditions. The ability to detect associations that are rare or invisible in European cohorts underscores why ancestry specific analyses matter for clinical accuracy and health equity.

“Our results show how locally developed risk prediction can substantially enhance healthcare precision,” said Dr. Hung Hsin Chen, the first and corresponding author at Academia Sinica.

The team then built polygenic risk score models with tools such as LDpred2 and PRSmix Plus, reaching area under the curve values above 0.8 for ankylosing spondylitis and about 0.7 for several other high impact conditions, including psoriasis, atrial fibrillation, rheumatoid arthritis and type 2 diabetes. When these scores were tested in external cohorts from the Taiwan Biobank, UK Biobank and the All of Us Project, the Taiwan specific models consistently outperformed European derived ones for people of East Asian ancestry.

New Disease Insights and Public Health Stakes

Some of the most striking findings emerged from Taiwan’s unique hepatitis B landscape. With more than 23,000 hepatitis B cases in the dataset, researchers uncovered 19 new genetic loci tied to infection and identified genetic correlations that run opposite to expectations, including inverse relationships between hepatitis B and autoimmune disorders such as psoriasis and systemic lupus erythematosus. These patterns show how local pathogen exposure interacts with human genomes in ways that global datasets often fail to capture.

“Precision medicine should serve everyone, not just the populations that have been most studied,” said Dr. Chen. “Our work offers a blueprint for other countries to develop population specific risk prediction frameworks.”

The broader implication is clear. Nearly a quarter of the world’s people are of East Asian ancestry, yet they represent less than 4 percent of participants in past genome wide studies. That imbalance carries real clinical costs, from miscalibrated risk estimates to missed variants that matter only in certain populations. By releasing their data and methods, the TPMI team hopes to inspire large scale genomic efforts in other understudied groups and ultimately build a fairer foundation for global precision medicine.

Nature: 10.1038/s41586-025-09350-y