Nick Martin was a doctoral student of mine at the University of Birmingham in the mid 1970s. In this review, I discuss two of Nick’s earliest and most seminal contributions to the field of behavior genetics. First, Martin and Eaves’ (1977) extension of the model-fitting approach to multivariate data, which laid the theoretical groundwork for a generation of multivariate behavior genetic studies. Second, the Martin et al.’s (1978) manuscript on the power of the classical twin design, which showed that thousands of twin pairs would be required in order to reliably estimate components of variance, and has served as impetus for the formation of large-scale twin registries across the world. I discuss these contributions against the historical backdrop of a time when we and others were struggling with the challenge of figuring out how to incorporate gene-by-environment interaction, gene–environment correlation, mate selection and cultural transmission into more complex genetic models of human behavior.

Background: Considerable evidence from twin and adoption studies indicates that genetic and shared environmental factors play a role in the initiation of smoking behavior. Although twin and adoption designs are powerful to detect genetic and environmental influences, they do not provide information on the processes of assortative mating and parent–offspring transmission and their contribution to the variability explained by genetic and/or environmental factors. Methods: We examined the role of genetic and environmental factors in individual differences for smoking initiation (SI) using an extended kinship design. This design allows the simultaneous testing of additive and non-additive genetic, shared and individual-specific environmental factors, as well as sex differences in the expression of genes and environment in the presence of assortative mating and combined genetic and cultural transmission, while also estimating the regression of the prevalence of SI on age. A dichotomous lifetime ‘ever’ smoking measure was obtained from twins and relatives in the ‘Virginia 30,000’ sample and the ‘Australian 25,000’. Results: Results demonstrate that both genetic and environmental factors play a significant role in the liability to SI. Major influences on individual differences appeared to be additive genetic and unique environmental effects, with smaller contributions from assortative mating, shared sibling environment, twin environment, cultural transmission, and resulting genotype-environment covariance. Age regression of the prevalence of SI was significant. The finding of negative cultural transmission without dominance led us to investigate more closely two possible mechanisms for the lower parent–offspring correlations compared to the sibling and DZ twin correlations in subsets of the data: (1) age × gene interaction, and (2) social homogamy. Neither of the mechanism provided a significantly better explanation of the data. Conclusions: This study showed significant heritability, partly due to assortment, and significant effects of primarily non-parental shared environment on liability to SI.

Drinking alcohol is a normal behavior in many societies, and prior studies have demonstrated it has both genetic and environmental sources of variation. Using two very large samples of twins and their first-degree relatives (Australia ≈ 20,000 individuals from 8,019 families; Virginia ≈ 23,000 from 6,042 families), we examine whether there are differences: (1) in the genetic and environmental factors that influence four interrelated drinking behaviors (quantity, frequency, age of initiation, and number of drinks in the last week), (2) between the twin-only design and the extended twin design, and (3) the Australian and Virginia samples. We find that while drinking behaviors are interrelated, there are substantial differences in the genetic and environmental architectures across phenotypes. Specifically, drinking quantity, frequency, and number of drinks in the past week have large broad genetic variance components, and smaller but significant environmental variance components, while age of onset is driven exclusively by environmental factors. Further, the twin-only design and the extended twin design come to similar conclusions regarding broad-sense heritability and environmental transmission, but the extended twin models provide a more nuanced perspective. Finally, we find a high level of similarity between the Australian and Virginian samples, especially for the genetic factors. The observed differences, when present, tend to be at the environmental level. Implications for the extended twin model and future directions are discussed.

The genetic and social causes of individual differences in attitudes to gun control are estimated in a sample of senior male and female twin pairs in the United States. Genetic and environmental parameters were estimated by weighted least squares applied to polychoric correlations for monozygotic (MZ) and dizygotic (DZ) twins of both sexes. The analysis suggests twin similarity for attitudes to gun control in men is entirely genetic while that in women is purely social. Although the volunteer sample is small, the analysis illustrates how the well-tested concepts and methods of genetic epidemiology may be a fertile resource for deepening our scientific understanding of biological and social pathways that affect individual risk to gun violence.

Whether monozygotic (MZ) and dizygotic (DZ) twins differ from each other in a variety of phenotypes is important for genetic twin modeling and for inferences made from twin studies in general. We analyzed whether there were differences in individual, maternal and paternal education between MZ and DZ twins in a large pooled dataset. Information was gathered on individual education for 218,362 adult twins from 27 twin cohorts (53% females; 39% MZ twins), and on maternal and paternal education for 147,315 and 143,056 twins respectively, from 28 twin cohorts (52% females; 38% MZ twins). Together, we had information on individual or parental education from 42 twin cohorts representing 19 countries. The original education classifications were transformed to education years and analyzed using linear regression models. Overall, MZ males had 0.26 (95% CI [0.21, 0.31]) years and MZ females 0.17 (95% CI [0.12, 0.21]) years longer education than DZ twins. The zygosity difference became smaller in more recent birth cohorts for both males and females. Parental education was somewhat longer for fathers of DZ twins in cohorts born in 1990–1999 (0.16 years, 95% CI [0.08, 0.25]) and 2000 or later (0.11 years, 95% CI [0.00, 0.22]), compared with fathers of MZ twins. The results show that the years of both individual and parental education are largely similar in MZ and DZ twins. We suggest that the socio-economic differences between MZ and DZ twins are so small that inferences based upon genetic modeling of twin data are not affected.

Background: There continues to be significant investment in the detection of genotype × environment interaction (G × E) in psychiatric genetics. The implications of the method of assessment for the genetic analysis of psychiatric disorders are examined for simulated twin data on symptom scores and environmental covariates. Methods: Additive and independent genetic and environmental risks were simulated for 10,000 monozygotic (MZ) and 10,000 dizygotic (DZ) twin pairs and the ‘subjects’ administered typical simulated checklists of clinical symptoms and environmental factors. A variety of standard tests for G × E were applied to the simulated additive risk scores, sum scores derived from the checklists and transformed sum scores. Results: All analyses revealed no evidence for G × E for latent risk but marked evidence for G × E and other effects of modulation in the sum scores. These effects were all removed by transformation. An integrated genetic and psychometric model, accounting for both the causes of latent liability and a theory of measurement, was fitted to a sample of the simulated sum-score data and showed that there was no significant modulation of the parameters of the genetic model by environmental covariates (i.e., no G × E). Conclusions: Claims to detect G × E based on analytical methods that ignore the theory of measurement must be subjected to greater scrutiny prior to publication.

A trend toward greater body size in dizygotic (DZ) than in monozygotic (MZ) twins has been suggested by some but not all studies, and this difference may also vary by age. We analyzed zygosity differences in mean values and variances of height and body mass index (BMI) among male and female twins from infancy to old age. Data were derived from an international database of 54 twin cohorts participating in the COllaborative project of Development of Anthropometrical measures in Twins (CODATwins), and included 842,951 height and BMI measurements from twins aged 1 to 102 years. The results showed that DZ twins were consistently taller than MZ twins, with differences of up to 2.0 cm in childhood and adolescence and up to 0.9 cm in adulthood. Similarly, a greater mean BMI of up to 0.3 kg/m2 in childhood and adolescence and up to 0.2 kg/m2 in adulthood was observed in DZ twins, although the pattern was less consistent. DZ twins presented up to 1.7% greater height and 1.9% greater BMI than MZ twins; these percentage differences were largest in middle and late childhood and decreased with age in both sexes. The variance of height was similar in MZ and DZ twins at most ages. In contrast, the variance of BMI was significantly higher in DZ than in MZ twins, particularly in childhood. In conclusion, DZ twins were generally taller and had greater BMI than MZ twins, but the differences decreased with age in both sexes.

The public health burden of alcohol is unevenly distributed across the life course, with levels of use, abuse, and dependence increasing across adolescence and peaking in early adulthood. Here, we leverage this temporal patterning to search for common genetic variants predicting developmental trajectories of alcohol consumption. Comparable psychiatric evaluations measuring alcohol consumption were collected in three longitudinal community samples (N = 2,126, obs = 12,166). Consumption-repeated measurements spanning adolescence and early adulthood were analyzed using linear mixed models, estimating individual consumption trajectories, which were then tested for association with Illumina 660W-Quad genotype data (866,099 SNPs after imputation and QC). Association results were combined across samples using standard meta-analysis methods. Four meta-analysis associations satisfied our pre-determined genome-wide significance criterion (FDR < 0.1) and six others met our ‘suggestive’ criterion (FDR <0.2). Genome-wide significant associations were highly biological plausible, including associations within GABA transporter 1, SLC6A1 (solute carrier family 6, member 1), and exonic hits in LOC100129340 (mitofusin-1-like). Pathway analyses elaborated single marker results, indicating significant enriched associations to intuitive biological mechanisms, including neurotransmission, xenobiotic pharmacodynamics, and nuclear hormone receptors (NHR). These findings underscore the value of combining longitudinal behavioral data and genome-wide genotype information in order to study developmental patterns and improve statistical power in genomic studies.

For over 100 years, the genetics of human anthropometric traits has attracted scientific interest. In particular, height and body mass index (BMI, calculated as kg/m2) have been under intensive genetic research. However, it is still largely unknown whether and how heritability estimates vary between human populations. Opportunities to address this question have increased recently because of the establishment of many new twin cohorts and the increasing accumulation of data in established twin cohorts. We started a new research project to analyze systematically (1) the variation of heritability estimates of height, BMI and their trajectories over the life course between birth cohorts, ethnicities and countries, and (2) to study the effects of birth-related factors, education and smoking on these anthropometric traits and whether these effects vary between twin cohorts. We identified 67 twin projects, including both monozygotic (MZ) and dizygotic (DZ) twins, using various sources. We asked for individual level data on height and weight including repeated measurements, birth related traits, background variables, education and smoking. By the end of 2014, 48 projects participated. Together, we have 893,458 height and weight measures (52% females) from 434,723 twin individuals, including 201,192 complete twin pairs (40% monozygotic, 40% same-sex dizygotic and 20% opposite-sex dizygotic) representing 22 countries. This project demonstrates that large-scale international twin studies are feasible and can promote the use of existing data for novel research purposes.

Objective: Despite an increasing recognition that psychiatric disorders can be diagnosed as early as preschool, little is known how early genetic and environmental risk factors contribute to the development of psychiatric disorders during this very early period of development. Method: We assessed infant temperament at age 1, and attention deficit hyperactivity disorder (ADHD), oppositional defiant disorder (ODD), and separation anxiety disorder (SAD) at ages 3 through 5 years in a sample of Hispanic twins. Genetic, shared, and non-shared environmental effects were estimated for each temperamental construct and psychiatric disorder using the statistical program MX. Multivariate genetic models were fitted to determine whether the same or different sets of genes and environments account for the co-occurrence between early temperament and preschool psychiatric disorders. Results: Additive genetic factors accounted for 61% of the variance in ADHD, 21% in ODD, and 28% in SAD. Shared environmental factors accounted for 34% of the variance in ODD and 15% of SAD. The genetic influence on difficult temperament was significantly associated with preschool ADHD, SAD, and ODD. The association between ODD and SAD was due to both genetic and family environmental factors. The temperamental trait of resistance to control was entirely accounted for by the shared family environment. Conclusions: There are different genetic and family environmental pathways between infant temperament and psychiatric diagnoses in this sample of Puerto Rican preschool age children.

Little is known regarding the underlying relationship between smoking initiation and current quantity smoked during adolescence into young adulthood. It is possible that the influences of genetic and environmental factors on this relationship vary across sex and age. To investigate this further, the current study applied a common causal contingency model to data from a Virginia-based twin study to determine: (1) if the same genetic and environmental factors are contributing to smoking initiation and current quantity smoked; (2) whether the magnitude of genetic and environmental factor contributions are the same across adolescence and young adulthood; and (3) if qualitative and quantitative differences in the sources of variance between males and females exist. Study results found no qualitative or quantitative sex differences in the relationship between smoking initiation and current quantity smoked, though relative contributions of genetic and environmental factors changed across adolescence and young adulthood. More specifically, smoking initiation and current quantity smoked remain separate constructs until young adulthood, when liabilities are correlated. Smoking initiation is explained by genetic, shared, and unique environmental factors in early adolescence and by genetic and unique environmental factors in young adulthood; while current quantity smoked is explained by shared environmental and unique environmental factors until young adulthood, when genetic and unique environmental factors play a larger role.

The importance of including developmental and environmental measures in genetic studies of human pathology is widely acknowledged, but few empirical studies have been published. Barriers include the need for longitudinal studies that cover relevant developmental stages and for samples large enough to deal with the challenge of testing gene–environment–development interaction. A solution to some of these problems is to bring together existing data sets that have the necessary characteristics. As part of the National Institute on Drug Abuse-funded Gene-Environment-Development Initiative, our goal is to identify exactly which genes, which environments, and which developmental transitions together predict the development of drug use and misuse. Four data sets were used of which common characteristics include (1) general population samples, including males and females; (2) repeated measures across adolescence and young adulthood; (3) assessment of nicotine, alcohol, and cannabis use and addiction; (4) measures of family and environmental risk; and (5) consent for genotyping DNA from blood or saliva. After quality controls, 2,962 individuals provided over 15,000 total observations. In the first gene–environment analyses, of alcohol misuse and stressful life events, some significant gene–environment and gene–development effects were identified. We conclude that in some circumstances, already collected data sets can be combined for gene–environment and gene–development analyses. This greatly reduces the cost and time needed for this type of research. However, care must be taken to ensure careful matching across studies and variables.

The analysis of genetic and environmental contributions to preterm birth is not straightforward in family studies, as etiology could involve both maternal and fetal genes. Markov Chain Monte Carlo (MCMC) methods are presented as a flexible approach for defining user-specified covariance structures to handle multiple random effects and hierarchical dependencies inherent in children of twin (COT) studies of pregnancy outcomes. The proposed method is easily modified to allow for the study of gestational age as a continuous trait and as a binary outcome reflecting the presence or absence of preterm birth. Estimation of fetal and maternal genetic factors and the effect of the environment are demonstrated using MCMC methods implemented in WinBUGS and maximum likelihood methods in a Virginia COT sample comprising 7,061 births. In summary, although the contribution of maternal and fetal genetic factors was supported using both outcomes, additional births and/or extended relationships are required to precisely estimate both genetic effects simultaneously. We anticipate the flexibility of MCMC methods to handle increasingly complex models to be of particular relevance for the study of birth outcomes.

Amulticategory item-response theory model was developed to characterize developmental changes in three items relating to the assessment of puberty in adolescent twin girls and boys. The model allowed for the fixed effects of age on probability of endorsing the responses and for the random effects of individual differences on the timing of pubertal changes relative to chronological age. In girls, the model was applied three-wave data on twin pairs (N = 414 female monozygotic [MZ] and 197 female dizygotic [DZ] pairs) and female twins from boy–girl pairs (N = 300 twins) from the Virginia Twin Study of Adolescent Behavioral Development. In boys, the data comprised 318 MZ and 185 DZ pairs and 297 male twins from boy–girl pairs. A total of 3172 and 2790 individual twin assessments were available in girls and boys, respectively, spanning ages 8–17 years. The availability of twin data allows the contributions of genes, the shared environment and individual unique environmental experiences to be resolved in the relative timing of pubertal changes. Parameters of the mixed model including fixed effects of age and random effects of genes and environment were estimated by Markov Chain Monte Carlo simulations using the BUGS algorithm for Gibbs sampling. The estimated standard deviation of random differences in the timing of puberty relative to age was 0.96 years in girls and 1.01 years in boys. The estimated intraclass correlations for the relative timing of pubertal changes were 0.99 +/−0.01 in MZ girls, 0.52 +/−0.02 in DZ girls, 0.88 +/−0.04 in MZ boys and 0.44+/−0.02 in DZ boys, indicating a very large contribution of genetic factors to the relative timing of pubertal change in both sexes. Additive genetic factors account for an estimated 96.3+/−3.3% of the total variance in random effects in girls and 88.0+/−3.6% in boys. Shared environmental influences account for 3.6+/−3.4% in girls and 0% in boys. In girls, nonshared environmental effects explain 0.1+/−0.1% of the total residual variance. The comparable figure in boys is 12.0+/−;3.6%.

Arecurring criticism of the twin method for quantifying genetic and environmental components of human differences is the necessity of the so-called “equal environments assumption” (EEA) (i.e., that monozygotic and dizygotic twins experience equally correlated environments). It has been proposed to test the EEA by stratifying twin correlations by indices of the amount of shared environment. However, relevant environments may also be influenced by genetic differences. We present a model for the role of genetic factors in niche selection by twins that may account for variation in indices of the shared twin environment (e.g., contact between members of twin pairs). Simulations reveal that stratification of twin correlations by amount of contact can yield spurious evidence of large shared environmental effects in some strata and even give false indications of genotype x environment interaction. The stratification approach to testing the equal environments assumption may be misleading and the results of such tests may actually be consistent with a simpler theory of the role of genetic factors in niche selection.

Considerable evidence from twin and adoption studies indicates that both genetic and shared environmental factors play a substantial role in the liability to antisocial behavior. Although twin and adoption designs can resolve genetic and environmental influences, they do not provide information about assortative mating, parent–offspring transmission, or the contribution of these factors to trait variation. We examined the role of genetic and environmental factors for conduct disorder (CD) using a twin–parent design. This design allows the simultaneous estimation of additive genetic, shared and individual-specific environmental effects, as well as sex differences in the expression of genes and environment in the presence of assortative mating and combined genetic and cultural transmission. A retrospective measure of CD was obtained from twins and their parents or guardians in the Virginia Twin Study of Adolescent Behavior Development and its Young Adult Follow up sample. Both genetic and environmental factors play a significant role in the liability to CD. Major influences on individual differences appeared to be additive genetic (38%–40%) and unique environmental (39%–42%) effects, with smaller contributions from the shared environment (18%–23%), assortative mating (~2%), cultural transmission (~2%) and resulting genotype-environment covariance. This study showed significant heritability, which is slightly increased by assortative mating, and significant effects of primarily nonparental shared environment on CD.

Few papers have attempted to address the potential of extended kinships of twins to resolve the complexities of biological and cultural inheritance in humans. Since the mid-80s these issues have largely been buried beneath the quest for specific genes that contribute to individual differences in complex traits and liability to disease. The articles of Keller et al. and Medland and Keller (current issue) revisit these issues. History will decide whether these new papers represent the final gasp of a paradigm superceded by the ‘new’ genetics or contain the timely seeds of new birth in the face of a 30-year genetic research program otherwise poised on the cusp of degeneration (Lakatos & Musgrave, 1970).

The classical twin design uses data on the variation of and covariation between monozygotic and dizygotic twins to infer underlying genetic and environmental causes of phenotypic variation in the population. By using data from additional relative classes, such as parents, extended twin family designs more comprehensively describe the causes of phenotypic variation. This article introduces an extension of previous extended twin family models, the Cascade model, which uses information on twins as well as their siblings, spouses, parents, and children to differentiate two genetic and six environmental sources of phenotypic variation. The Cascade also relaxes assumptions regarding mating and cultural transmission that existed in previous extended twin family designs. The estimation of additional parameters and relaxation of assumptions is potentially important, not only because it allows more fine-grained descriptions of the causes of phenotypic variation, but more importantly, because it can reduce the biases in parameter estimates that exist in earlier designs.

Research has consistently shown that religiousness is associated with lower levels of alcohol and drug use, but little is known about the nature of adolescent religiousness or the mechanisms through which it influences problem behavior in this age group. This paper presents preliminary results from the Mid-Atlantic School Age Twin Study, a prospective, population-based study of 6–18-year-old twins and their mothers. Factor analysis of a scale developed to characterize adolescent religiousness, the Religious Attitudes and Practices Inventory (RAPI), revealed three factors: theism, religious/spiritual practices, and peer religiousness. Twin correlations and univariate behavior-genetic models for these factors and a measure of belief that drug use is sinful reveal in 357 twin pairs that common environmental factors significantly influence these traits, but a minor influence of genetic factors could not be discounted. Correlations between the multiple factors of adolescent religiousness and substance use, comorbid problem behavior, mood disorders, and selected risk factors for substance involvement are also presented. Structural equation modeling illustrates that specific religious beliefs about the sinfulness of drugs and level of peer religiousness mediate the relationship between theistic beliefs and religious/spiritual practices on substance use. Limitations and future analyses are discussed.