The good result of this congress (the fifth of the series), in terms of quality of papers and number of participants, proves we were on the right track when in 1974 we founded the International Society for Twin Studies, and even more proves that the interest in gemellology is increasing.

I wish to thank our many colleagues who contributed to the promotion, construction and success of this field of biomedical research, and particularly Gordon Allen, Ian MacGillivray, Walter Nance, Paolo Parisi, and of course Aldur Eriksson.

For my part, I would like to emphasize that going from the title “Study of Twins” given to my book in 1951 to the word gemellology placed in the title of our quarterly Acta Geneticae Medicae et Gemellologiae in 1952, I wanted to affirm: first, that the study of twins is a branch of modern genetics; second, that the study of twins represents a “logos,” that is, a subject composed of acquired concepts, of constant phenomena and specific methods of research on twins in normality, in pathology and in population. With this word, now also adopted by Professor Nance in his Presidential Address, “Introduction to Gemellology,” twin research has become an identified field of science.

In order that the Amsterdam Congress may be as fruitful as the previous ones, I believe it useful to look out of the window of the future with you to see some roads which could be new frontiers of gemellology.

We consider a study of MZ and DZ twin pairs ascertained because one or both twins have a disease. Genotypes at a major locus are known and putative environmental risk factors have been measured for all individuals. The power of the study to estimate the effect on liability of the measured and residual genetic and environmental effects (Gm, Gr, Em, Er) and all two-way interactions between them (except Gr × Er) is estimated by simulation. If liabilities can be indexed on a continuous scale (eg, blood pressure as an index of liability to hypertension), then a study of 600 MZ and 600 DZ pairs would have sufficient power to detect quite subtle interaction effects, even if ascertainment is greatly biased toward MZ twins. If liabilities cannot be measured and only affection status is known, then the power of the study would be much lower, although not impracticably so. There appears to be no advantage in augmenting the twins with a sample of control individuals who have been drawn at random from the population regardless of disease status, at least for the case we have considered in which the disease threshold on the liability scale is assumed to be known without error. The argument is developed in terms of the utility of the design for research into breast cancer.

Advances in computer technology have made possible a greater sophistication in the statistical analysis of pedigree data, however this is not necessarily manifest by fitting more comprehensive causative models. Planned twin and family studies measure numerous explanatory variables, including perhaps genetic and DNA marker information status on all pedigree members, and the cohabitation of all pairs of individuals. A statistical analysis should examine the contribution of these measured factors on individual means, and in explaining the variation and covariation between individuals, concurrently with the postulated effect of unmeasured factors such as polygenes. We present two models that meet this requirement: the Multivariate Normal Model for Pedigree Analysis for quantitative traits, and a Log-Linear Model for Binary Pedigree Data. For both models, important issues are examination of fit, detection of outlier pedigrees and outlier individuals, and critical examination of the model assumptions. Procedures for fulfilling these needs and examples of modelling are discussed.

When the univariate twin design is extended by including parents of twins, it is possible to assess additive genetic effects in the presence of assortative mating and genotype-environment correlation, the effects of parental influence, as well as the extent of residual shared environmental influences. The analysis of data obtained in such an extended twin design can be carried out by means of constrained maximum likelihood confirmatory factor analysis. Specifically, the structural model underlying this design can be represented as a LISREL model with nonlinear constraints. This representation offers the possibility to consider extended multivariate twin designs involving common genetic and environmental factors. The proposed method will be illustrated with applications to simulated and real data.

The effects on twin data of social interaction between spouses is examined. When social interaction leads to an increase in marital resemblance (eg through reciprocal imitation), the variance of married individuals is increased, compared to the variance of unmarried individuals. Furthermore, the expected correlations between concordant married twin pairs will be lower than the expected correlations between concordant unmarried twin pairs, with the discordant twin correlations being intermediate in value. It is therefore possible, in principle, to detect the effects of marital interaction without using either longitudinal data or data on spouse pairs. However, to be detectable in twin data, marital interaction must be strong, or must exhibit marked asymmetry of effects between males and females. Genotype × environment interaction can also produce heterogeneity of correlation between concordant married, discordant, and concordant unmarried twin pairs, when genetic and environmental effects interact with marital status. However, this will usually produce increased estimates of the genetic component of variance in unmarried twins, whereas marital interaction produces increased genetic variance in married twins.

The genetic analysis of physiological time series has to accommodate the presence of autocorrelation. This can be accomplished by means of orthogonal transformation of the series, thus enabling the use of standard genetic analysis techniques for the sequence of uncorrelated transforms. In view of the oscillatory character which typifies various physiological time series, it is customary to invoke spectral techniques for the analysis of these series. It can be shown that spectral analysis is an orthogonal transformation that asymptotically resembles principal component analysis. Consequently, standard genetic analysis methods for the uncorrelated spectral transforms may be used. This approach will be illustrated with simulated and real (heart rate) data for univariate twin time series. Furthermore, it will be indicated that the proposed analysis can be readily generalized to multivariate time series.

A Vietnam Era (1964-1975) Twin Registry of American male-male veterans born between 1939 and 1955 has been developed to provide a study sample for research evaluating the impact of Vietnam service on the medical and psychosocial aspects of health. In preparation for developing the Registry, several alternative sources of twins and methods for identifying twins were investigated. A computerized database of veterans discharged from the military after 1967 was selected as the source because it contains about 50% of the total Vietnam era veteran population, is reasonably unbiased, and provides a feasible method for identifying twins. Twins were identified using an algorithm which involved matching entries on the database for same last name, different first name, same date of birth, and similar social security number. Twin status was confirmed by review of military records. The registry, now complete, is composed of 7,400 twin pairs. It will be an important resource for future research projects.

An examination of ascertainment bias in identification of twin pairs in the Vietnam Era Twin Registry has been conducted. A complete listing of all male-male Vietnam era veteran twin pairs born in Connecticut between 1939 and 1955 was obtained (N = 150). An attempt was made to match these pairs with a listing of Vietnam era veteran twin pairs derived from the United States Department of Defense's Defense Manpower Data Center (DMDC) computer files. The results indicate that the DMDC files identified only 46.7% of the 150 Connecticut born Vietnam era veteran pairs. Statistically significant differences (P < 0.05) between pairs found on the DMDC files and Connecticut veteran pairs missing from the DMDC files are observed for the following variables: a) year of discharge from military service, b) total length of active military service, c) branch of service, and d) foreign service. No consistent pattern of bias is observed for factors related to the physical and psychosocial healt of veteran pairs. The implications of the ascertainment biases in the Vietnam Era Twin Registry are discussed.

Kinships composed of twin parents, their spouses and children, offer a robust and flexible sampling design for research in genetic epidemiology. Families-of-twins designs circumvent some of the sampling problems that arise when independent data sets are combined, and these designs provide unique evaluations of maternal influences, assortative mating and X-linkage. Unfortunately, empirical studies of families of twin parents have been limited by relatively small samples and by the self-selection biases intrinsic in ascertainment of families from volunteer twin registries.

A large and representative cohort of monozygotic and dizygotic twin parents, drawn from a population-based twin registry, provides the optimal sampling frame for twin-family research. This paper reviews the sampling considerations underlying the initial family study based on the Finnish Twin Cohort and evaluates the representativeness of the sampled twins. Spouses and adult children (over 18 years) of 236 pairs of twins, about equally divided by gender and zygosity, were evaluated by a postal questionnaire. Individual response rates exceeded 86% and in 464 of the 472 nuclear families (98.3% ), at last one member of the twin's family completed the questionnaire. The sampled twins, selected for fecundity to maximize statistical power of the obtained data, were broadly representative of non-selected twins drawn from the Cohort, with whom they were matched on age, gender, and zygosity. Such results suggest that the Finnish Cohort has excellent potential for extended twin-family research designs.