A groundbreaking study published in Nature Medicine presents a new polygenic risk score (PGS) that significantly improves the prediction of obesity from early childhood. Developed through a collaboration between the GIANT consortium and 23andMe, the score draws on genetic data from more than five million people.

Led by Roelof Smit and senior author Ruth Loos of the Novo Nordisk Foundation Center for Basic Metabolic Research (CBMR) at the University of Copenhagen, the multi-ancestry PGS is twice as effective as previous scores and explains up to 17.6% of BMI variation in European populations. Importantly, it can predict future obesity risk as early as age five, before weight-related changes typically begin to manifest, enabling earlier intervention strategies.

While highly predictive in European populations, performance was lower in individuals of African ancestry—a gap researchers attribute to underrepresentation in the training data.

"This new polygenic score is a dramatic improvement in predictive power and a leap forward in the genetic prediction of obesity risk, which brings us much closer to clinically useful genetic testing," said Professor Loos.

Beyond prediction, the study also found that individuals with higher genetic risk responded more strongly to lifestyle interventions but were also more likely to regain weight—underscoring both the promise and complexity of using genetics in obesity prevention and care.

For more information, please visit the Novo Nordisk Foundation Center for Basic Metabolic Research.