A study surveying more than 3.4 million people has found nearly 4,000 genetic variants related to the use of alcohol and tobacco, scientists reported Wednesday (Dec 7) in Nature. More than 1,900 of the variants had not previously been linked to substance use behaviors, study coauthor and statistical geneticist at the Penn State College of Medicine Dajiang Liu states in a press release. The novel findings were aided by the fact that a fifth of the genomic samples came from individuals of non-European ancestry, Liu says.

“This is a great study. It demonstrates the power of using large samples from multiple ancestry groups in well-designed analyses,” geneticist and neuroscientist Joel Gelernter of Yale University tells New Scientist.

Although scientists have conducted similar genome-wide association studies (GWAS) in the past, Liu tells New Scientist that the majority of those surveys were conducted primarily on people of European descent. Although social situations and policies may influence a person’s inclination towards smoking and drinking, there is substantial evidence that a person’s genetic makeup may also predispose them towards alcohol and tobacco usage. “We’re at a stage where genetic discoveries are being translated into clinical [applications],” Liu tells Nature. “If we can forecast someones risk of developing nicotine or alcohol dependence using this information, we can intervene early and potentially prevent a lot of deaths.”

See “Genes for Alcohol Use Disorder and Alzheimer’s Risk Overlap: Study

Liu collaborated with more than a hundred other scientists to evaluate the millions of genomic datasets, employing machine learning techniques to link genetic variants to smoking and drinking–related factors. These factors ranged from the age at which a person began the habit to how many cigarettes or drinks they had per day or week or how likely they were to give up the habit. The analyses uncovered nearly 2,500 genetic variants linked to regular smoking, as well as an additional 243 variants linked to how many cigarettes a person smokes per day, 206 linked to whether a person gave up smoking, and 39 linked to the age a person started the habit. Meanwhile, 849 variants were associated with how much alcohol a person drank per week. In all, nearly 4,000 alcohol and tobacco use–associated variants were identified, some of which were found in genes that are associated with the secretion and regulation of the hormone dopamine, New Scientist reports.

Intriguingly, 721 of the total variants were only found as a result of the multi-ancestry testing, Nature reports. Liu tells the publication that most of these variants had similar effects across all ancestries, although risk scores generated only for those of European ancestry were not good predictors of risk for other ancestry groups.

“It is promising to see that the same genes are associated with addictive behaviors across ancestries,” Liu states in the press release. “Having more robust and diverse data will help us develop predictive risk factor tools that can be applied to all populations.”

Despite being the largest genetic study of tobacco and alcohol use thus far, experts say there is still room for improvement. Ananyo Choudhury, a geneticist at the University of the Witwatersrand in Johannesburg, South Africa, tells Nature that the similar effects seen across ancestries could be a result of most of the study subjects living in the United States. Other experts add that incorporating analyses of more international populations would improve the study’s outcome, the publication reports. Liu agrees. “In future phases of the study, we will welcome collaborations from other investigators who have access to additional datasets to further expand our studies,” he tells the magazine.