Genetic Code and Postal Code: Both Matter for Children’s Health

A new study disentangles genetic and neighborhood influences, finding location and genetic background influence young people’s health and behavior.

April 8, 2019

Children in worse-off neighborhoods often leave school early and live shorter lives. Improving neighborhood conditions has been proposed as way of improving health and opportunities for millions of children. But if genetic factors influence both where families are able to live and their children’s health and educational success, improving neighborhoods may not be enough. New research from scientists at Columbia Mailman School of Public Health and the University of California at Irvine provides new insights into the highly debated question of whether the neighborhoods that children live in influence their health and life chances.
The study is the first to bring together genetic and geographic data to test links between children’s neighborhood and genetic risk. The findings are published online in Nature Human Behaviour.
The research team led by Daniel Belsky, PhD, assistant professor in the Department of Epidemiology at the Columbia Mailman School, and Candice Odgers, PhD, at the University of California, Irvine, Department of Psychological Science, linked the genomic, geographic, health, and educational data of thousands of children living in Britain and Wales. They found that children growing up in worse-off neighborhoods also carried a higher genetic risk for poor educational outcomes and earlier childbearing, as determined by genome-wide association studies known as polygenic scoring. The authors replicated their findings in the U.S.-based Add Health Study, where they found that gene-neighborhood correlations may accumulate across generations as young people with a higher genetic risk for poor educational attainment and women who gave birth a younger ages were both born into, and subsequently moved into, worse-off neighborhoods.
“We found genetic risk alone was not enough to explain why children from poorer versus more affluent neighborhoods received less education by late adolescence,” said Belsky, who is also with the Columbia Aging Center. “The data on education could explain only a fraction of the link between neighborhood risk and poor educational qualifications, suggesting that there is ample opportunity for neighborhoods to influence these outcomes.”
“Surprisingly, for obesity, one of the most prevalent and costly health problems facing this generation, we found no link between neighborhood and genetic risk,” observed Odgers. “Children who grew up in worse-off neighborhoods were more likely to become obese by age 18, but they did not carry a higher genetic risk for obesity than their peers living in more advantaged neighborhoods.”
Similarly, for mental health problems, children in worse-off neighborhoods experienced more symptoms of mental disorder, but there was little evidence that the reason for this link was due to genetic risk. For physical and mental health problems, postal code and genetic code both predicted children’s futures.  
Analyses were based on data from the Environmental Risk (E-Risk) Longitudinal Twin Study, which has followed 2,232 twins born in England and Wales in 1994-1995 into young adulthood, and the National Longitudinal Study of Adolescent to Adult Health, which followed 15,000 American secondary school students into adulthood. Genetic risk was measured by polygenic scoring, combining information from recent genome-wide association studies of obesity, of schizophrenia, of age-at-first-birth, and of educational attainment. Neighborhood characteristics and mobility were derived from government data, surveys of residents, and virtual assessment method employing Google Street View. 
Odgers, who developed the neighborhood virtual assessments noted that “advances in both genomics and geospatial analyses are rapidly positioning us to make new discoveries. In this case, they allowed us to identify outcomes, like obesity and mental health, where neighborhoods are most likely to have unique impacts.” But, she added, “This is only a first step in answering the really important question of whether changing neighborhoods can improve children’s lives.”

“In our study, polygenic risk scores showed a link between genetics and neighborhoods for teen pregnancy and poor educational outcomes,” said Belsky. This finding suggests that we should consider neighborhoods when interpreting the results of studies searching for genes related to these outcomes, and also that we should consider genes when examining the effects of neighborhoods.” But, he cautioned that “polygenic risk scores are an evolving and still imperfect tool. They can help us test whether genes and neighborhoods are related. But they cannot tell us how.”
Genetic risk accounted for only a fraction of the differences between children living in different types of neighborhoods. According to Belsky and Odgers this provides some reason to hope that “targeting neighborhoods—especially for physical and mental health—will be enough to improve children’s life outcomes.
Co-authors’ institutions are Duke University; Stanford University; University of North Carolina at Chapel Hill; Kings College, UK; and University of Exeter, UK.
The study was supported by the Medical Research Council (UKMRC G1002190), NICHD (HD077482), Google, and the Jacobs Foundation. The Add Health Study was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (HD31921, HD073342, HD060726), with cooperative funding from 23 other federal agencies and foundations.