The children of twins (COT) design has been proposed as an alternative to the adoption study to resolve the direct effects of parental treatment from secondary parent–child association due to genetic factors. The basic analytical approach compares the parent–offspring correlation with the correlation between children and the monozygotic (MZ) twins of their parents. We show that a significant difference between these correlations does not imply direct environmental causality when the measured parental treatment in question is dyadic, that is, influenced by both parents even when mating is random. Nongenetic causal effects yield very similar patterns of correlation to secondary genetic effects on dyadic treatment variables. The fact that many candidate environments, such as parental divorce, are dyadic gives reason to question the interpretation of their correlations with behavior in the children of twins.

Human chromosomes terminate in a number of repeats of the sequence TTAGGG. At birth, each chromosome end is equipped with approximately 15 kb of telomere sequence, but this sequence is shortened during each cell division. In cell cultures telomere shortening is associated with senescence, a phenomenon that has also been observed in normal adult tissues, indicating that telomere loss is associated with organismal ageing. Previous work has established that the rate of telomere loss in humans is age dependent, and recent work shows a sex-specific difference in telomere length and shortening in individuals over the age span of 20 to 75 years. Here, terminal restriction fragment lengths on DNA purified from whole blood were measured to examine the mean telomere length in a cross-sectional cohort of 816 Danish individuals of age 73 to 101 years. In this age group, females show a linear correlation between telomere length and age, whereas the pattern tends to be nonlinear (quadratic in age) for males. This difference in telomere length dynamics between the 2 sexes may be caused by several different mechanisms, including differences in selection by mortality, differences in leukocyte population or different telomerase expression pattern.

The design and interpretation of genetic association studies depends on the relationship between the genotyped variants and the underlying functional variant, often parameterized as the squared correlation or r2 measure of linkage disequilibrium between two loci. While it has long been recognized that placing a constraint on the r2 between two loci also places a constraint on the difference in frequencies between the coupled alleles, this constraint has not been quantified. Here, quantification of this severe constraint is presented. For example, for r2 ≥ .8, the maximum difference in allele frequency is ± .06 which occurs when one locus has allele frequency .5. For r2 ≥ .8 and allele frequency at one locus of .1, the maximum difference in allele frequency at the second locus is only ± .02. The impact on the design and interpretation of association studies is discussed.

Several studies report apparent sibling contrast effects in analyses of twin resemblance. In the presence of genetic differences, contrast effects reduce the dizygotic (DZ) twin correlation relative to that in monozygotic (MZ) twins and produce higher DZ than MZ variance. Explanations of contrast effects are typically cast in terms of direct social interaction between twins or an artifact of the process of rating children by their parents. We outline a model for sibling imitation and contrast effects that depends on social interaction between parents and children. In addition to predicting the observed pattern of twin variances and covariances, the parental mediation of child imitation and contrast effects leads to differences in the variance of parents of MZ and DZ twins and differences between the correlations of parents with their MZ and DZ children.

Congenital anomalies are a major cause of fetal and neonatal death and of childhood morbidity. Chromosomal and other genetic abnormalities, environmental teratogens and some nutritional deficiencies account for some congenital anomalies but the majority are of unknown etiology. The hypothesis is here proposed that a significant proportion of congenital anomalies and cerebral palsy of unknown etiology are attributable to a monozygotic multiple conception with monochorionic placentation and that these anomalies, even in singletons, may be explained by early, unrecognized or unrecorded loss of one conceptus in a monochorionic monozygotic conception. The pathological mechanism is hemodynamic instability with episodes of acute feto–fetal transfusion that produce ischemic organ impairment in either or both twins. The resultant clinical abnormality will depend on range of severity (fetal death, infant death, congenital anomaly, normal infant), site or combination of sites (which organ[s] present[s] with the congenital anomaly) and timing (early, middle or late in gestation as shown by variation in brain pathology that is observed).

Information about linkage disequilibrium (LD) is important in understanding the genome structure and has its applications in association studies. Here we present the first genome-wide LD study based on a founder population (East Finland). The LD data consist of 118 unrelated individuals and around 480,000 SNP pairs genotyped with the Affymetrix 100K genotyping assay. Using the minor allele frequency (MAF) limit of .05, the squared correlation coefficient between two loci (r2) was .48, .37, .28, and .20 for distances of 5, 10, 20, and 40 kb respectively. MAF had a significant effect on the mean r2 so that the extent of useful LD (r2 > .3) varied from 17 kb to 80 kb depending on the limit set for the MAF. For D' the effect of MAF was smaller but reflected the possible age of the mutation: SNPs with high MAF had lower D' than those with low MAF. The X chromosome showed higher D' values than autosomes and the extent of useful LD (r2 > .3) was twice as long on the X chromosome than on the autosomes. Based on the results, LD varies across the genome and is correlated to local recombination rate between and within chromosomes. However, the recombination rate does not explain all the variation found in LD. We also report a number of long chromosomal regions where exceptionally high or low LD were detected.

All etiological studies of complex human traits focus on analyzing the causes of variation. Given this complexity, there is a premium on studying those processes that mediate between gene products and cellular or organismal phenotypes. Studies of levels of gene expression could offer insight into these processes and are likely to be especially useful to the extent that the major sources of their variation are known in normal tissues. The classical study of monozygotic (MZ) and dizygotic (DZ) twins was employed to partition the genetic and environmental influences in gene expression for over 6500 human genes measured using microarrays from lymphoblastoid cell lines. Our results indicate that mean expression levels are correlated about .3 in monozygotic (MZ) and .0 in dizygotic (DZ) twins suggesting an overall epistatic regulation of gene expression. Furthermore, the functions of several of the genes whose expression was most affected by environmental effects, after correction for measurement error, were consistent with their known role in mediating sensitivity to environmental influences.

Adisappointing feature of conventional methods for detecting association between DNA variation and a phenotype of interest is that they tell us little about the hidden pattern of linkage disequilibrium (LD) with the functional variant that is actually responsible for the association. This limitation applies to case-control studies and also to the transmission/disequilibrium test (TDT) and other family-based association methods. Here we present a fresh perspective on genetic association based on two novel concepts called ‘LD squares’ and ‘equi- risk alleles’. These describe and characterize the different patterns of gametic LD which underlie genetic association. These concepts lead to a general principle – the Equi-Risk Allele Segregation Principle – which captures the way in which underlying LD patterns affect the transmission patterns of genetic variants associated with a phenotype. This provides a basis for distinguishing the hidden LD patterns and might help to locate the functional variants responsible for the association.

The first classical twin studies, recognizing the potential of comparing findings in identical twins, have previously been reported to be those by Siemens and by Merriman, both published in 1924. However, we would like to bring to attention a study performed by Walter Jablonski, 2 years earlier (1922), investigating the contribution of heredity to refraction in human eyes. Jablonski examined the eyes of 52 twin pairs and by comparing the size of within-pair differences between identical and non- identical twins was able to infer the heritability of a trait. Therefore, this is likely to be the first reported classical twin study.

It has been argued that culture significantly influences the developmental basis of self-concept. The goal of the present study is to compare the relative importance of genetic and environmental factors to explain individual differences in various dimensions of self-concept in female preadoles- cents of Minnesota in the United States and Seoul in South Korea. Two hundred and eighteen monozygotic (MZ) and 137 dizygotic (DZ) twin pairs from the Minnesota Twin Family Study (MTFS) and 74 MZ and 42 DZ twin pairs from the Seoul Twin Family Study (STFS) completed the 6 cluster scales of the Piers–Harris Children's Self-Concept Scale (P–H). The 6 cluster scales of the P-H include Popularity, Physical Appearance and Attributes, Behavior, Intellectual Competence and School Status, Anxiety, and Happiness and Satisfaction. Univariate model- fitting analyses were performed. In both samples, a model incorporating shared and nonshared environmental influences fitted the data best for Popularity, Anxiety, and Intellectual Competence and School Status, whereas a model including additive genetic and nonshared environmental factors provided the best fit for Physical Appearance and Attributes, and Behavior. The univariate model did not yield an adequate fit for Happiness and Satisfaction. For Physical Appearance and Attributes, and Intellectual Competence and School Status, estimates of additive genetic and environmental factors were significantly different between the MTFS and the STFS samples. For Popularity, Anxiety, and Behavior, however, the genetic and environmental estimates were comparable between the two samples.

A tight link exists between telomere length and both population doublings of a cell culture and age of a given organism. The more population doublings of the cell culture or the higher the age of the organism, the shorter the telomeres. The proposed model for telomere shortening, called the end replication problem, explains why the telomere erodes at each cellular turnover. Telomere length is regulated by a number of associated proteins through a number of different signaling pathways. The determinants of telomere length were studied using whole blood samples from 287 twin pairs aged 73 to 95 years. Structural equation models revealed that a model including additive genetic effects and non- shared environment was the best fitting model and that telomere length was moderately heritable, with an estimate that was sensitive to the telomere length standardization procedure. Sex-specific analyses showed lower heritability in males, although not statistically significant, which is in line with our earlier finding of a sex difference in telomere dynamics among the elderly and oldest-old.

Commentary on Cherkas et al. (2004). Genetic Influences on Female Infidelity and Number of Sexual Partners in Humans: A Linkage and Association Study on the Role of the Vasopressin Receptor Gene (AVPR1A). Twin Research, 7, 649–658.

Questionnaire-based dimensional measures are often employed in epidemiological studies to predict the presence of psychiatric disorders. The present study sought to determine how accurately 4 dimensional mental health measures, the 12-item General Health Questionnaire (GHQ-12), Neuroticism (EPQ-N), the high positive affect and anxious arousal scales from the Mood and Anxiety Symptoms Questionnaire (MASQ-HPA and MASQ-AA) and a composite of all 4, predicted psychiatric caseness as diagnosed by the University of Michigan Composite International Diagnostic Interview (UM-CIDI). Community subjects were recruited through general practitioners; those who agreed to participate were sent a questionnaire containing the above measures. Subsequently, the UM-CIDI was administered by telephone to 469 subjects consisting of sibling pairs who scored most discordantly or concordantly on a composite index of the 4 measures. Logistic Regression and Receiver Operating Characteristic (ROC) curve analyses were carried out to assess the predictive accuracy of the dimensional measures on UM-CIDI diagnosis. A total of 179 subjects, 62 men and 117 women with an average age of 42 years, were diagnosed with at least one of the following psychiatric disorders: depression, dysthymia, generalized anxiety disorder (GAD), social phobia, agoraphobia and panic attack. The six disorders showed high comorbidity. EPQ-N and the Composite Index were found to be very strong and accurate predictors of psychiatric caseness; they were however unable to differentiate between specific disorders. The results from the present study therefore validated the four mental health measures as being predictive of psychiatric caseness.

Genome-wide linkage analysis using multiple traits and statistical software packages is a tedious process which requires a significant amount of manual file manipulation. Different linkage analysis programs require different input file formats, making the task of analyzing data with multiple methods even more time-consuming. We have developed a software tool, AUTOGSCAN, that automates file formatting, the running of statistical analyses, and the summarizing of resulting statistics for whole genome scans with a push of a button, using several independent, and often idiosyncratic, statistical software packages such as MERLIN, SOLAR and GENEHUNTER. We also describe a program, ANALYZE, designed to run qualitative linkage analysis with several different statistical strategies and programs to efficiently screen for linkage and linkage disequilibrium for a given discrete trait. The ANALYZE program can also be used by AUTOGSCAN in a genome-wide sense.

The relationship between subjective wellbeing (SWB) and self-reported sleep problems was investigated in a cohort of Norwegian twins aged 18 to 31 years. Questionnaire data from 8045 same- and opposite-sex twins were analyzed using structural equation modeling to explore the relative effects of genetic and environmental influences on phenotypic variance and covariance. Special attention was paid to sex-specific effects. The correlation between the phenotypes was estimated to be −.43. Univariate analyses indicated considerable genetic influences for both SWB and sleep problems, for male and female twins alike. The best fitting bivariate model specified additive genetic and individual environmental factors for both phenotypes, and nonadditive genetic effects for sleep problems, with no sex-specific effects. Genetic and environmental effects accounted for 60% and 40% of the phenotypic correlation, respectively. Additive genetic factors affecting the two phenotypes were correlated (−.85), suggesting that part of the genetic effects that positively influence SWB also protect against sleep problems. In conclusion, the results indicate considerable overlap in genetic etiology for SWB and sleep problems, for males and females alike.

We examined whether effortful control (EC), a temperament proposed by Rothbart and Bates (1998), has genetically coherent structure. A self- report measure of EC was administered to 450 Japanese twins (151 males and 299 females, ages 17 to 32 years) including 152 monozygotic and 73 dizygotic pairs. Univariate genetic analysis revealed that AE model fit best for the total EC as well as its subscales. The heritability estimate for total EC was 49%, and the estimates for subscales ranged between 32% and 45%. Multivariate genetic analysis revealed that the subscales of EC were genetically correlated to a high degree and environmentally correlated to a moderate degree. These results suggest that EC has substantial genetic basis and genetically coherent structure, supporting the validity of the construct. The implications to molecular genetic study and study of psychopathology were discussed.

The classical twin design (CTD) is the most common method used to infer genetic and environmental causes of phenotypic variation. As has long been acknowledged, different combinations of the common environment/assortative mating, and additive, dominant, and epistatic genetic effects can lead to the same observed covariation between twin pairs, meaning that there is an inherent indeterminacy in parameter estimates arising from the CTD. The CTD circumvents this indeterminacy by assuming that higher-order epistasis is negligible and that the effects of either dominant genetic variation or the common environment are nonexistent. These assumptions, however, lead to consistent biases in parameter estimation. The current paper quantifies these biases and discusses alternative strategies for dealing with parameter indeterminacy in twin designs. One strategy is to model the similarity among other relatives in addition to twins (extended twin-family designs), which reduces but does not eliminate indeterminacy in parameter estimates. A more general strategy, applicable to all twin designs, is to present the parameter indeterminacy explicitly, as in a graph. Presenting the space of mathematically equally likely parameter values is important, not only because it aids the proper interpretation of twin design findings, but also because it keeps behavioral geneticists themselves mindful of methodological assumptions that can easily go unexamined.

Twin studies of complex traits, such as behavior or psychiatric diagnoses, frequently involve univariate analysis of a sum score derived from multiple items. In this article, we show that absence of measurement invariance across zygosity can bias estimates of genetic and environmental components of variance. Specifically, if the item responses are considered as multiple indicators of a latent factor, and the aim is to partition the variance in the latent factor, then the factor loadings relating the items to the factor should be equal for monozygotic (MZ) and dizygotic (DZ) twins. While it seems unlikely, a priori, that these loadings should differ as a function of zygosity, certain special measurement situations are cause for concern. Ratings by parents, or self-ratings of pheno- types which are more easily observed in others than via introspection, may be tainted by the co-twin's phenotype to a greater extent in MZ than DZ pairs. We also show that the analysis of sum scores typically biases both MZ and DZ correlations compared to the true latent trait correlation. These two sources of bias are quantified for a range of values and are shown to be especially acute for sum scores based on binary items. Solutions to these problems include formal tests for measurement invariance across zygosity prior to analysis of the sum or scale scores, and multivariate genetic analysis at the individual item or symptom level.

Genetic factors have historically been thought of as important in the development of obsessive–compulsive disorder (OCD). For the estimation of the relative importance of genetic and environmental factors, twin studies are an obvious approach. Twin studies of OCD have a long history, starting in 1929. In this review, over 70 years of twin research of OCD is presented, using four different approaches that represent the steps in the twin research of OCD from past to present. These steps include (1) case-studies of twins with OCD from the old literature; (2) twin studies of OCD using Diagnostic and Statistical Manual of Mental Disorders (DSM) criteria; (3) twin studies of OCD using a dimensional approach, comparing resemblances in monozygotic and dizygotic twins; and (4) twin studies of OCD using a dimensional approach, analyzing the data with Structural Equation Modeling. It is concluded that only the studies using the last method have convincingly shown that, in children, obsessive–compulsive (OC) symptoms are heritable, with genetic influences in the range of 45% to 65%. In adults, studies are suggestive for a genetic influence on OC symptoms, ranging from 27% to 47%, but a large twin study using a bio- metrical approach with continuous data is still needed to provide conclusive evidence. Strategies for future twin studies of OCD are discussed.

The classical twin design (CTD) circumvents parameter indeterminacy by assuming (1) negligible higher-order epistasis; and (2) either nonadditive genetic or common environmental effects are nonexistent, creating two potential sources of bias (Eaves et al., 1978; Grayson, 1989). Because the extended twin-family design (ETFD) uses many more unique covariance observations to estimate parameters, common environmental and nonadditive genetic parameters can be simultaneously estimated. The ETFD thereby corrects for what is likely to be the largest of the two sources of bias in CTD parameter estimates (Keller & Coventry, 2005). In the current paper, we assess the extent of this and other potential sources of bias in the CTD by comparing all published ETFD parameter estimates to CTD parameter estimates derived from the same data. CTD estimates of the common environment were lower than ETFD estimates of the common environment for some phenotypes, but for other phenotypes (e.g., stature in females and certain social attitudes), what appeared as the common environment was resolved to be assortative mating in the ETFD. On average, CTD estimates of nonadditive genetic factors were 43% lower, and additive genetic factors 63% higher, than ETFD estimates. However, broad-sense heritability estimates from the CTD were only 18% higher than ETFD estimates, highlighting that the CTD is useful for estimating broad-sense but not narrow-sense heritability. These results suggest that CTD estimates can be misleading when interpreted literally, but useful, albeit coarse, when interpreted properly.