A landmark study published in Nature Medicine reports strong evidence that polygenic scores (PGSs), a measure of inherited genetic risk, can effectively predict body mass index (BMI) and risk of obesity across developmental stages and diverse ancestral backgrounds.
Obesity remains a critical global health crisis. Based on projections from WHO data, more than half of the world’s population will be affected by obesity by 2035. While interventions range from lifestyle modifications to bariatric surgery, effective and widely accessible options remain limited. Due to the early onset of obesity and its continued presence into adulthood, obesity researchers have long sought predictive tools to assess obesity risk during childhood.
A widely used PGS for obesity developed by Khera and colleagues in 2016 was based on a BMI genome-wide association study (GWAS) of over 339,000 individuals, primarily of European ancestry. It explained around 8.5% of the variance in adult BMI. However, its predictive power diminished in non-European populations, highlighting the urgent need for PGSs developed from more diverse populations to promote equitable healthcare outcomes.
Researchers had previously utilized genetic data from over 5.1 million individuals, including participants from the GIANT consortium and 23andMe. The ancestry represented in the sample was diverse, comprising 71.1% European, 14.4% American (often admixed), 8.4% East Asian, 4.6% African, and 1.5% South Asian.
Using PRS-CS(x), the researchers developed both ancestry-specific and multi-ancestry PGSs. The best-performing score, PGSLC, was a linear combination of five ancestry-specific scores optimized using data from the UK Biobank (UKBB). It was validated across multiple cohorts, including the UKBB, the Million Veteran Program (MVP), BioMe Biobank, Uganda General Population Cohort, and the Avon Longitudinal Study of Parents and Children (ALSPAC) study.
Polygenic score – linear combination (PGSLC) explained 17.6% of the variance in BMI among European-ancestry participants in the UKBB and up to 16% in East Asian Americans, but only 2.2% in rural Ugandans, underscoring disparities caused by underrepresentation in GWAS datasets.
In the ALSPAC cohort, higher PGSs were associated with a rapid increase in BMI gains at the age of 2.5 years and an earlier adiposity rebound. Including PGS almost doubled the prediction accuracy of BMI from age 5 years onward (for example, from 11% to 21% at age 8 years), and improved the prediction of BMI at age 18 years based on age 5 data (for example, from 22% to 35% at age 5 years).
Higher PGSs were also associated with greater weight gain in adulthood. In short-term weight loss interventions, individuals with higher PGSs initially lost slightly more weight (0.55 kg per standard deviation [SD]) but subsequently regained more weight over time. These results highlight the potential of early-life PGSs in predicting obesity.
The PGS predicted both BMI and the incidence of obesity in adults. For example, among European participants, 69.5% of the top 1% of PGSLC developed obesity, compared to 54.9% in the PGSKhera group. In the lifestyle interventions trials Action for Health in Diabetes (Look AHEAD) and Diabetes Prevention Program (DPP), those higher in PGSLC had greater initial weight loss (0.55 kg per SD; 95% confidence interval [CI]-0.94 to -0.16) but also had showed higher weight regain (0.48 kg per SD; 95% CI 0.00 to 0.95).
This study demonstrated that PGSs, particularly multi-ancestry scores such as PGSLC, can improve the prediction of obesity risk across the life course and among people of diverse ancestries. It highlights the urgent need to increase diversity in genetic research. The findings suggest that early-life genetic screening may be instrumental in identifying individuals at high risk and guiding targeted interventions.
References: Smit RAJ, Wade KH, Hui Q, et al. Polygenic prediction of body mass index and obesity through the life course and across ancestries. Nat Med. 2025. doi:10.1038/s41591-025-03827-z
