A groundbreaking new study has revealed that your DNA may predict whether you’re likely to develop obesity—years before symptoms emerge.
Using genetic data from over five million people, scientists created a highly accurate polygenic risk score that signals obesity risk from early childhood. The score is so predictive that signs often appear before age five, offering a crucial window for early intervention. While the test performs best in people of European ancestry, it opens up powerful possibilities for personalized health strategies, especially for children.
Genetic Clues to Obesity Risk in Childhood
According to projections from the World Obesity Federation, more than half of the global population could be overweight or obese by the year 2035. While there are treatment options like medication, lifestyle changes, and surgery, these approaches are not consistently effective or accessible to everyone.
To explore new ways of tackling this growing issue, researchers from around the world analyzed genetic data from over five million individuals. They developed a tool called a polygenic risk score (PGS), which is closely linked to adult obesity and shows noticeable patterns as early as childhood. This score could be used to pinpoint children and teens who are genetically more likely to become obese later in life, offering a chance to introduce preventive measures like healthy lifestyle guidance before weight issues begin.
Why Early Intervention Could Be Key
“What makes the score so powerful is the consistency of associations between the genetic score and body mass index before the age of five and through to adulthood – timing that starts well before other risk factors start to shape their weight later in childhood. Intervening at this point could theoretically make a huge impact,” said Assistant Professor Roelof Smit at the University of Copenhagen and lead author of the research published in the journal Nature Medicine.
Even small differences in a person’s genes can significantly influence health when they work together. Thousands of genetic variants have been linked to obesity, including ones that affect the brain and appetite regulation. The polygenic risk score functions like a calculator, adding up the combined effect of those genetic variants to estimate a person’s overall risk. In this study, the PGS was able to account for nearly 17% of the variation in body mass index among individuals—a much greater predictive power than earlier attempts.
The Largest Genetic Dataset Ever Used
To create these PGS, the scientists drew on the genetic data of more than five million people – the largest and most diverse genetic dataset ever – including genetic data from the Genetic Investigation of ANthropometric Traits (GIANT) consortium and consumer DNA testing firm, 23andMe.
The researchers then tested whether their new PGS was associated with obesity using datasets of the physical and genetic characteristics of more than 500,000 people, including BMI data tracked over time from the Children of the 90s study. They found that their new PGS was twice as effective as the previous best method at predicting a person’s risk of developing obesity.
Implications for Public Health
Dr. Kaitlin Wade, Associate Professor in Epidemiology at the University of Bristol and second author on this paper, said: “Obesity is a major public health issue, with many factors contributing to its development, including genetics, environment, lifestyle, and behavior. These factors likely vary across a person’s life, and we believe that some of these originate in childhood.
“We were delighted to contribute data from the Children of the 90s study to this exceptional and insightful research into the genetic architecture of obesity. We hope this work will contribute to detecting individuals at high risk of developing obesity at an earlier age, which could have a vast clinical and public health impact in the future.”
Genetics and Lifestyle Interventions
The research team also investigated the relationship between a person’s genetic risk of obesity and the impact of lifestyle weight loss interventions, such as diet and exercise. They discovered that people with a higher genetic risk of obesity were more responsive to interventions but also regained weight more quickly when the interventions ended.
Despite drawing on the genomes of a wider population, the new PGS has its limitations. For example, it was far better at predicting obesity in people with European-like ancestry than in people with African ancestry. This flags the need for work like this in more representative groups.
Reference: “Polygenic prediction of body mass index and obesity through the life course and across ancestries” by Roelof A. J. Smit, Kaitlin H. Wade, Qin Hui, Joshua D. Arias, Xianyong Yin, Malene R. Christiansen, Loic Yengo, Michael H. Preuss, Mariam Nakabuye, Ghislain Rocheleau, Sarah E. Graham, Victoria L. Buchanan, Geetha Chittoor, Marielisa Graff, Marta Guindo-Martínez, Yingchang Lu, Eirini Marouli, Saori Sakaue, Cassandra N. Spracklen, Sailaja Vedantam, Emma P. Wilson, Shyh-Huei Chen, Teresa Ferreira, Yingjie Ji, Tugce Karaderi, Kreete Lüll, Moara Machado, Deborah E. Malden, Carolina Medina-Gomez, Amy Moore, Sina Rüeger, Masato Akiyama, Matthew A. Allison, Marcus Alvarez, Mette K. Andersen, Vivek Appadurai, Liubov Arbeeva, Eric Bartell, Seema Bhaskar, Lawrence F. Bielak, Joshua C. Bis, Sailalitha Bollepalli, Jette Bork-Jensen, Jonathan P. Bradfield, Yuki Bradford, Caroline Brandl, Peter S. Braund, Jennifer A. Brody, Ulrich Broeckel, Kristoffer S. Burgdorf, Brian E. Cade, Qiuyin Cai, Silvia Camarda, Archie Campbell, Marisa Cañadas-Garre, Jin-Fang Chai, Alessandra Chesi, Seung Hoan Choi, Paraskevi Christofidou, Christian Couture, Gabriel Cuellar-Partida, Rebecca Danning, Frauke Degenhardt, Graciela E. Delgado, Alessandro Delitala, Ayşe Demirkan, Xuan Deng, Alexander Dietl, Maria Dimitriou, Latchezar Dimitrov, Rajkumar Dorajoo, …, Cristen J. Willer, Kristin L. Young, Segun Fatumo, Jeanne M. McCaffery, Nicholas J. Timpson, Joel N. Hirschhorn, Yan V. Sun, Sonja I. Berndt and Ruth J. F. Loos, 21 July 2025, Nature Medicine.
DOI: 10.1038/s41591-025-03827-z
Never miss a breakthrough: Join the SciTechDaily newsletter.
Follow us on Google and Google News.