Researchers have known for decades that genetics may be more powerful than environment in predicting who will develop obesity. Identical twins tend to have the same body mass index, even if they are reared apart. Adopted children tend to have a degree of obesity similar to their birth parents rather than their adoptive ones.

Identifying the genetic roots of obesity could aid with prevention starting in childhood. But finding a genetic footprint for obesity has proved challenging. With rare exceptions, there’s not one gene or even a few that are the culprits. Instead, obesity is spurred by thousands of gene variants acting in concert. Each variant exerts a tiny effect.

Now, using genetic data from five million people, an international group of hundreds of researchers reports that it has developed an obesity risk score, known also as a polygenic risk score. It combines thousands of gene variants to estimate individuals’ predicted body mass indexes, which continue to be used by doctors to anticipate weight-related health dangers.

The researchers showed that the scores can predict which young children are at risk of obesity as adults. And, in another test, they found that overweight and obese adults with high risk scores quickly regain any weight that they lose with lifestyle programs.

Their paper was published on Monday in the journal Nature Medicine.

Dr. Joel Hirschhorn, an author of the paper and a professor of pediatrics and genetics at Boston Children’s Hospital, cautioned that genetics cannot account for the effects of environment and is therefore inherently limited in predicting obesity.

“We will almost never be able to say a child will have a BMI of 38 as an adult,” he said. “Genetics is not that predictive.”