Genetics and stem cells

Smoking 'is down to your genes'

“Genes determine why some smoke,” according to The Independent , which said that smokers may be able to blame their genes for their addiction.

The news is based on research that identified genetic variants linked to the likelihood of smoking and the number of cigarettes people smoke a day. The well-conducted analysis of several studies has confirmed the associations between smoking behaviour and three genetic variants, two of which had not been identified before. The research also highlighted a complicated causal relationship between the presence of these variants, smoking behaviour and lung cancer, which the study authors say needs untangling.

Overall, the work increases our understanding of the genetics that contribute to smoking behaviour. It is important to highlight that having the genetic variations does not mean that someone will be a smoker, but that the variants are linked with an increased likelihood of smoking.

Where did the story come from?

This study was carried out by a large number of researchers, including Dr Thorgeir Thorgeirsson and colleagues affiliated with various research groups and academic institutions across the world. The research was supported in part by the US National Institutes of Health and European Commission. It was published in the peer-reviewed medical journal Nature Genetics .

The Independent reported this research accurately, although it described the gene variants that the researchers discovered as “mutations”, which is somewhat misleading.

What kind of research was this?

This meta-analysis of genome wide association studies investigated the association between smoking behaviour and the presence of particular genetic variants in the DNA of a large number of people. Genome-wide association studies are a form of case-control study that provide a way to examine the link between DNA and disease or condition characteristics. The power of these studies is increased when several are pooled together and analysed, as they were in this meta-analysis.

What did the research involve?

This research included samples from people from a number of groups of smokers (cases) and people who had never smoked (controls). The study had two main parts. The first part looked specifically at the association between DNA and whether a person smoked, and the second assessed the association between DNA and the number of cigarettes smoked a day.

To assess whether a person smoked, the researchers analysed data on a total of 30,431 people who had smoked at some point in their lives (ever-smokers) and 16,050 people who had never smoked (never-smokers), drawn from 12 separate genome wide association studies. The addition of a thirteenth study provided a total of 31,266 smokers for analysis to assess associations with the number of cigarettes they smoked daily.

The purpose of this phase of study was to establish whether particular DNA variants were more common in people who smoked compared to those who did not, and whether DNA variants were linked to the number of cigarettes smoked a day. As is common in these studies, the researchers then began a second “replication phase”, in which they attempted to verify their findings in a separate population.

The researchers retested their initial findings in two separate sets of people: a set of 45,691 smokers and a sample of 9,040 people drawn from the general population. The first sample was tested for the presence of 277 variants in 15 chromosome regions that had the strongest associations with smoking and for 443 variants in 14 regions that had the strongest associations with number of cigarettes smoked a day. The second sample was tested for the presence of three variants with the strongest associations. 

The researchers then performed a range of related analyses, such as testing to see whether the variants they had identified in the first part of the experiment were associated with nicotine dependence (assessed as a score on a known nicotine dependence tool) and lung cancer.

In their paper, the researchers also outlined what is known about the regions in which these variants are found, including nearby genes, and the associations with smoking behaviours, nicotine metabolism and harms.

What were the basic results?

The study found that gene variants in three regions were associated with the number of cigarettes smoked a day. These variants were:

  • rs1051730[A] at 15q25
  • rs4105144[C] at 19q13
  • rs6474412-T at 8p11

The researchers say that the variants rs4105144[C] and rs6474412-T are in close proximity to genes that have known associations with nicotine metabolism.

How did the researchers interpret the results?

The researchers say that they have discovered DNA variants associated with smoking behaviour within regions that also contain genes involved in nicotine metabolism. They note that while their study found nominal associations between these gene variants and the risk of lung cancer, they cannot tell whether this is because they have an effect on smoking behaviour or whether they increase a person’s vulnerability to the harmful effects of smoking. They say that further studies are needed.


This meta-analysis of genome wide association studies was well conducted and well reported. The researchers used recognised methods in this field and replicated their findings in separate populations to check their validity. The researchers say that while the environmental influences of smoking behaviour are well known, some aspects of smoking behaviour also have a high heritability. Their research has confirmed the association between variants on chromosome 15q25 with smoking behaviour and harms, which have been suggested by other studies, and identified two new variants.

About 78% of people in the first part of the study had the active variant at the rs6474412 site of chromosome 8. In the replication sample, 78–83% of patients with this variant had cancer, whereas 76–81% of those without cancer also had the variant. This suggests that, although there may be significant differences between the frequency of these variants in different populations, the presence of the variant may not be that strong a predictor of who will go on to develop lung cancer.

It is important to note that the gene variants discovered were not unique to smokers and can occur in people who do not smoke. Having the variations does not mean that someone will be a smoker, but it is linked with an increased likelihood of being a smoker. While these findings are important and increase what is known about the genetics of smoking, their direct application to smoking prevention is unclear.

NHS Attribution