A history of research on gene–environment interaction (G × E) is provided in this article, revealing the fact that there have actually been two distinct concepts of G × E since the very origins of this research. R. A. Fisher introduced what I call the biometric concept of G × E (G × EB), whereas Lancelot Hogben introduced what I call the developmental concept of G × E (G × ED). Much of the subsequent history of research on G × E has largely consisted of the separate legacies of these separate concepts, along with the (sometimes acrimonious) disputes that have arisen time and again when employers of each have argued over the appropriate way to conceptualize the phenomenon. With this history in place, more recent attempts to distinguish between different concepts of G × E are considered, paying particular attention to the commonly made distinction between “statistical interaction” and “interactionism,” and Michael Rutter's distinction between statistical interaction and “the biological concept of interaction.” I argue that the history of the separate legacies of G × EB and G × ED better supports Rutter's analysis of the situation and that this analysis best paves the way for an integrative relationship between the various scientists investigating the place of G × E in the etiology of complex traits.

Human studies have suggested an association between a variable length polymorphism in the serotonin transporter gene promoter region and vulnerability to anxiety and depression. Relative to the long (l) allele, the short (s) allele increases the risk of developing depression in individuals exposed to stressful life events. An orthologue of the human variant is present in rhesus macaques and allows for studies in animals exposed to stress. Here, we used an established model of early life stress exposure, in which rhesus macaques are raised without adults in a group of peers (peer-only reared [PR]), or with their mothers. At 6 months of age, animals were subjected to 4-day long social separations for 4 consecutive weeks, with 3 days of reunion in between. Data were collected during both the acute (Day 1) and chronic phases (Days 2–4) of separation. Behavioral factors were separately extracted for each phase of separation. For acute separation, the behavioral factors generated were despair and behavioral pathology and, for the chronic phase despair, agitation, and behavioral pathology. During both phases of social separation, PR l/s animals were more likely to exhibit pathological behaviors, whereas PR l/l monkeys show higher levels of despair compared to the other three groups. These findings indicate that early stress affects the behavioral response to separation differently as a function of recombinant human serotonin transporter linked polymorphic repeat genotype and suggest that carriers of the s allele are not only more anxious but may also be more vulnerable to developing behavioral pathology in the face of chronic adversity.

The study of gene–environment interaction (G × E) constitutes an area of significant social and clinical significance. Different types of research study designs are being used to investigate the contribution of G × E to psychopathology, although the term G × E has also been used and interpreted in different ways. Despite mixed evidence that G × E contributes to psychopathology, some promising and consistent findings are emerging. Evidence is reviewed in relation to depression, antisocial behavior, schizophrenia, and attention-deficit/hyperactivity disorder. Although findings from various research designs have different meaning, interestingly much of the evidence with regard to the contribution of G × E that has arisen from twin and adoption studies has been for antisocial behavior and depression. It is for these same forms of psychopathology that molecular genetic evidence of G × E has also been most convincing. Finally, current and anticipated methodological challenges and implications for future research in this area are considered.

The rapidly moving study of Gene × Environment interaction (G × E) needs interim conceptual tools to track progress, integrate findings, and apply this knowledge to preventive intervention. We define two closely related concepts: the social mediation of the expression of genetic influences and the interaction between the entire genotype and the social environment (G × E). G × E, the primary focus of this report, assesses individual differences in the full genotype using twin, sibling, and adoption designs and, for the most part, employs fine-grained analyses of relational processes in the social environment. In comparison, studies of Allele × Environment interaction assess the influence on development of one or more measured polymorphisms as modified by environmental factors. G × E studies build on work showing how the social environment responds to genetic influences and how genetic influences shape the social environment. Recent G × E research has yielded new insight into variations in the sensitivity of the social environment to genotypic influences and provides clues to the specificity and timing of these environmental responses that can be leveraged to inform preventive interventions aimed at reducing genetic risk for problem behavior.

As studies of measured gene–environment interactions (G × E) in developmental psychopathology gain momentum, methods for systematically and quantitatively summarizing effects across multiple studies are urgently needed. Meta-analyses of G × E findings are critical for evaluating the overall statistical and theoretical significance of any given G × E based on cumulative and systematically combined knowledge. Although meta-analytic methods for the combination of study findings based on single effect measures such as odds ratios and mean differences are well established, equivalent methods for the meta-analysis of studies investigating interactions are not well developed. This article describes one simple approach to the meta-analysis of G × E effects using, as a contemporaneous example, the interaction of the monoamine oxidase A (MAOA) gene and the impact of childhood maltreatment on risk for developing antisocial behavior.

We examined the influence of a common allelic variation in the dopamine receptor D4 (DRD4) gene and caregiver quality on temperament in early childhood. Children 18–21 months of age were genotyped for the DRD4 48 base pair tandem repeat polymorphism, which has been implicated in the development of attention, sensation seeking, and attention-deficit/hyperactivity disorder. The children also interacted with their caregiver for 10 min in a laboratory setting, and these videotaped interactions were coded for parenting quality using an observational rating procedure. The presence of the DRD4 7-repeat allele was associated with differences in the influence of parenting on a measure of temperamental sensation seeking constructed from caregiver reports on children's activity level, impulsivity, and high-intensity pleasure. Children with the 7-repeat allele were influenced by parenting quality, with lower quality parenting associated with higher levels of sensation seeking; children without the 7-repeat allele were uninfluenced by parenting quality. Differences between alleles were not related to the child's self-regulation as assessed by the effortful control measure. Previous studies have indicated that the 7-repeat allele is under positive selective pressure, and our results are consistent with the hypothesis that the DRD4 7-repeat allele increased children's sensitivity to environmental factors such as parenting. This study shows that genes influence the relation between parenting and temperament in ways that are important to normal development and psychopathology.

Few studies have investigated the role of gene × environment interactions (G × E) in speech, language, and literacy disorders. Currently, there are two theoretical models, the diathesis–stress model and the bioecological model, that make opposite predictions about the expected direction of G × E, because environmental risk factors may either strengthen or weaken the effect of genes on phenotypes. The purpose of the current study was to test for G × E at two speech sound disorder and reading disability linkage peaks using a sib-pair linkage design and continuous measures of socioeconomic status, home language/literacy environment, and number of ear infections. The interactions were tested using composite speech, language, and preliteracy phenotypes and previously identified linkage peaks on 6p22 and 15q21. Results showed five G × E at both the 6p22 and 15q21 locations across several phenotypes and environmental measures. Four of the five interactions were consistent with the bioecological model of G × E. Each of these four interactions involved environmental measures of the home language/literacy environment. The only interaction that was consistent with the diathesis–stress model was one involving the number of ear infections as the environmental risk variable. The direction of these interactions and possible interpretations are explored in the discussion.

Family systems theory proposes that an individual's functioning depends on interactive processes within the self and within the context of dyadic family subsystems. Previous research on these processes has focused largely on behavioral, cognitive, and psychophysiological properties of the individual and the dyad. The goals of this study were to explore genetic and environmental interactions within the family system by examining how the dopamine receptor D2 gene (DRD2) A1+ polymorphism in mothers and children relates to maternal sensitivity, how maternal and child characteristics might mediate those effects, and whether maternal sensitivity moderates the association between DRD2 A1+ and child affective problems. Evidence is found for an evocative effect of child polymorphism on parenting behavior, and for a moderating effect of child polymorphism on the association between maternal sensitivity and later child affective problems. Findings are discussed from a family systems perspective, highlighting the role of the family as a context for gene expression in both mothers and children.

It is established that reading and reading-related processes are heritable; genes thus play an important role in the foundation of individual differences in reading. In this article, we focus on one facet of reading–comprehension. Comprehension is a higher order cognitive skill that requires many other cognitive processes for it to unfold completely and successfully. One such process is executive functioning, which has been associated with genetic variation in the catechol-O-methyltransferase (COMT) gene. Genotypes and haplotypes of four single nucleotide polymorphisms in COMT were investigated in 179 incarcerated adolescent delinquents. Four hierarchical logistic regression models predicting the presence/absence of comprehension difficulties were fitted to the data; genetic variation in COMT and the presence/absence of maternal rejection were investigated as main effects and as effects acting interactively. Three out of four interaction terms were found to be important predictors of individual differences in comprehension. These findings were supported by the results of the haplotype analyses, in which the four investigated polymorphisms were considered simultaneously.

Recent work on the molecular genetics of complex traits in typical and atypical human development has focused primarilyon associations of single genes with behavior. Disparate literature suggests that the presence of one or two copies of the short allele of the serotonin transporter (5-HTT) gene and the long allele (7-repeat allele) version of the dopamine receptor D4 (DRD4) gene predicts internalizing- and externalizing-related behaviors, respectively. Apparently for the first time in the extant literature, we report a gene–gene statistical interaction on behavior problems in a group of typically developing children at age 7. DNA was extracted from buccal cells collected from 108 children and genotyped for short and long alleles of the 5-HTT gene and the short (2–5 repeats) versus long (6–8 repeats) allele of the DRD4 gene. Mothers completed the Child Behavior Checklist. As predicted, children with one or two copies of the short allele of the 5-HTT gene and the long allele version of the DRD4 gene exhibited significantly more internalizing and externalizing behaviors at age 7 than children with other combinations of the 5-HTT and DRD4 short and long genotypes. As well, children with the 5-HTT long and DRD4 long genotypes had the lowest reported scores on internalizing and externalizing behaviors at age 7, suggesting that the presence of the 5-HTT long genotype may serve as a protective factor against these behaviors in children with the long DRD4 genotype. Implications of these findings for understanding cumulative biological risk and protective factors in childhood behavior problems and psychopathology are discussed.

Potential candidate Gene × Environment interactions in the etiology of childhood attention-deficit/hyperactivity disorder (ADHD) are examined between the dopamine receptor D2 gene (DRD2) and putative family environmental risk factors that reflect mothers' marital stability. Specifically, interactions were tested between DRD2 and mothers' marital status, number of marriages or cohabiting relationships, and age at first marriage. Moderate relations were found among the marital stability measures, and mother's marital status and number of marriages or cohabiting relationships (but not age at first marriage) were risk factors for their children's ADHD. All three mother's marital stability variables were associated with either the child's or mother's DRD2 genotypes. Gene × Environment interactions were found for children's ADHD diagnoses between children's DRD2 genotypes and mother's marital status and number of marriages or cohabiting relationships. It is of interest that these interactions were strengthened with the progressive addition of sets of covariates intended to control for alternative causal pathways that represent background genetic and environmental context confounds. The present findings highlight the importance of considering both the nexus of putative environmental risk factors and whether their etiology and effects are truly environmental in future Gene × Environment interaction research.

Consistent evidence shows both genetic and stress-related risks on child and adolescent anxiety, yet few studies have considered the degree to which genetic effects are moderated by stress (gene–environment interaction). We used longitudinal data from both a child and adolescent sample of twins to examine three novel issues on the presence of gene–environment interaction on anxiety symptoms. First, we assessed moderation of genetic risks on anxiety symptoms by negative life events in each age group. Second, by distinguishing between “stable” and “age-specific” genetic factors, we explored the continuity of gene–environment interaction across time and/or its emergence at specific ages. Third, we compared the presence of gene–environment interaction across different symptom types (general, panic, social, and separation). Genetic effects on separation anxiety symptoms in childhood (mean age = 8 years, 6 months) and panic anxiety symptoms in adolescence (mean age = 15 years) increased across independent negative life events. Shared environmental effects on separation anxiety symptoms and nonshared environmental effects on general anxiety symptoms in adolescence were also moderated by negative life events. We interpret these preliminary findings tentatively in the context of gene–environment interaction on anxiety in general, and on early separation and later panic anxiety in particular.

The impact of socioeconomic status (SES) and genetic polymorphisms on individual differences for externalized behaviors have often been investigated separately in studies of children and adults. In a general population sample of 607 Italian preadolescents, we examined the independent and joint effects of SES and the dopamine receptor D4 (DRD4) and serotonin transporter linked promoter region (5-HTTLPR) polymorphisms upon rule-breaking and aggressive behaviors measured with the Child Behavior CheckList/6–18. We found evidence, which was based on both one locus and two-loci genotype analyses, that low SES and DRD4 long and 5-HTTLPR long alleles, both alone and in interaction, are associated with higher aggressive behavior scores. The effects were similar but more modest and limited to one locus genotype analyses for rule-breaking behavior. Consistent with studies that showed the effects of societal moderators on the heritability of externalized behaviors across different segments of the population, we suggest that diminished social constraints associated with low parental SES may act as enhancers of the genetic influence of specific DRD4 and 5-HTTLPR alleles over aggressive behaviors in preadolescence.

Child maltreatment and polymorphisms of the serotonin transporter (5-HTT) and monoamine oxidase A (MAOA) genes were examined in relation to depressive symptomatology. Adolescents (M age = 16.7 years) from low socioeconomic backgrounds with a history of child maltreatment (n = 207) or no such history (n = 132) were interviewed and provided buccal cells for genetic analysis. Gene × environment interactions were observed. Heightened depressive symptoms were found only among extensively maltreated youth with low MAOA activity. Among comparably maltreated youth with high MAOA activity, self-coping strategies related to lower symptoms. Sexual abuse and the 5-HTT short/short genotype predicted higher depression, anxiety, and somatic symptoms. This Gene × Environment interaction was further moderated by MAOA activity level. The results highlight the protective functions of genetic polymorphisms and coping strategies in high risk youth and offer direction for understanding resilience and its promotion from a multiple levels of analysis perspective.

To determine if drinking behavior in adolescence provides a “gateway” leading to the misuse of other psychoactive substances and antisocial behavior, we genotyped 180 Asian adolescent adoptees to determine if they inherited the deficient from of the aldehyde dehydrogenase 2 (ALDH2) enzyme that is important in the metabolism of alcohol. Based on the gateway model, we hypothesized that those with normal enzyme activity (70% of the sample) who began to misuse alcohol would also misuse other drugs and display antisocial tendencies. Those with the enzyme deficiency (30%), because they experience unpleasant side effects associated with drinking, were expected to show less evidence of alcohol misuse and thus be less likely to progress to the misuse of other substances or engage in antisocial acts. Consistent with previous research, we found that ALDH2 deficiency was significantly associated with lower rates of drinking and getting drunk but not with ever having tried alcohol. Contrary to the gateway model, we found no evidence that ALDH2 deficiency was associated with lower rates of nonalcohol substance use or antisociality. Finally, in an examination of factors that may moderate the impact of the metabolic protection because of ALDH2 deficiency, we identified siblings rather than parents as the major source of familial environmental effect on adolescent drinking.