• Correlations;
  • genes;
  • interactions;
  • social environments

Laboratory evidence of gene–environment interaction for alcohol preference appeared 35 years ago; social modification of alcohol consumption was demonstrated by cross-fostering newborn mice from inbred high- and low-alcohol preference strains to foster mothers of the other strain [1]. Parallel evidence that social environments modify genetic influences on human drinking patterns has been more difficult to obtain, and not only because we control neither genetic background nor social rearing. There are many risk and protective factors for use/abuse of alcohol in humans, and most are inherently confounded. Peer effects on adolescent drinking illustrate this, because peer effects arise from both socialization and selection and the selection is driven, in part, by assortative pairing for behaviors that are heritable and risk-relevant to substance use [2, 3].

Agrawal and colleagues elucidate these complexities in the interplay of genetic dispositions and social environments [4]. Their study is noteworthy, first, because its analysis focuses upon a social environment, partly (perhaps largely) of one's own creation. Early studies of environmental modulation of genetic dispositions for adolescent drinking focused upon demographic features of environments related to access to alcohol and community surveillance of its use—environmental features which adolescents play little role in creating. However, social environments of substance-using or substance-avoiding peers are formed as adolescents seek affiliation actively with those who share their genetic dispositions. Agrawal and colleagues show that genetic influences on substance use among peers overlap the genetic influences on an adolescent's own substance use, an illustration of gene–environment correlation that confirms earlier reports [5–7]. They then address gene–environment interaction, showing that even after accounting for the active role that females play in selecting peers, substance use by peers modulates expression of a woman's genetic disposition to regularly use substances. Greater substance use among one's peers was associated with greater heritability of own substance use 5 years later, novel results that may be attributed both to greater access to substances, as well as to greater social acceptance and increased opportunity to express (and find reinforcement of) predispositions for their use.

Peer substance use is an important environmental domain moderating genetic influences on substance use. Our analyses of Finnish twin data show similar effects across others. Neighborhood and parenting variables have strong moderating effects on the importance of genetic (and environmental) influences on adolescent substance use. Genetic influences are more important in urban than in rural settings [8], a finding replicated in Minnesota twin data [9]. Genetic influences are magnified in neighborhoods with greater migration, those with a larger proportion of older adolescents and those with proportionately higher alcohol sales [10]. Further, genetic influences assume greater importance when parents have less knowledge of their children's activities and associations [11]. These findings, consistent with those of Agrawal et al., probably reflect mechanisms of social control and/or opportunity. Environments characterized by more social control (e.g. neighborhoods with greater stability; families with increased parental knowledge) assume more importance in an adolescent's substance use; but environments characterized by less social control, fewer constraints and more opportunity allow adolescents to express their genetic predispositions more freely [12]. Increasing genetic influence is evident in many behaviors across the life-span [13] and probably reflects similar mechanisms, such as increased control over one's own behavior in transitioning from childhood and adolescence into adulthood. Controlled by law and monitored by parents, religious institutions and community authority, adolescent substance use illustrates more readily than most behaviors the significant environmental modification of genetic influences. The research by Agrawal and colleagues adds importantly to this line of evidence and raises two questions for future study. First, is the effect enhanced among females? Secondly, is the effect developmentally specific?

Are peer effects greater among females? Some research suggests that girls and women are more sensitive to environmental influences relevant to substance abuse. Adolescent girls may be more likely to engage in assortative peer selection [3,14], and girls may also be more susceptible to the influence of friends [15]. Are these effects limited to early development? Many gene–environment interactions are probably specific to developmental periods, and effects of peer interactions in late adolescence may wane with time and age—perhaps replaced by other social effects from partner and children. The research by Agrawal and colleagues finds that influence of peers extends into early adulthood; does it then wane later? The focus of the next round of gene–environment research will add development into the equation and focus upon gene–environment–development interactions. That focus is stimulated by the kind of research Agrawal and colleagues have offered.


  1. Top of page
  2. Declarations of interest
  3. References