Brisbane Adolescent Twin Study: Outline of study methods and research projects

An ongoing program at the Queensland Institute of Medical Research (QIMR) uses a resource of adolescent twins (2720 individuals) and their singleton siblings (1179), constituting 1324 families. A large number (4 500 families) have participated in two main studies (the mole and the cognition studies) with the aim of identifying genetic polymorphisms associated with moliness and cognitive function respectively, as well as several other projects that run in parallel with these studies. Most recently 950 families were invited to participate in a mail and phone study assessing reading ability, taste and smell sensitivity, and health and wellbeing. Although the program measures a broad range of phenotypes, the common focus of all studies is to reveal the genetic architecture underlying the trait, and thereby facilitate the search for quantitative trait loci. The present paper provides an overview of the studies, and highlights the significance of the resource, as well as some of the unique aspects of twin research. Twin studies have been used for centuries to investigate the heritability of a particular trait, and provide a method for investigating and partitioning the relative contributions of genes and the environment to phenotypic variability in all human traits. The classical ‘‘twin design’’ capitalises on the fact that identical, or monozygotic (MZ), twin pairs are similar because they share 100% of their genes and, when reared together, 100% of their shared (common) environment, with differences between them attributed to their nonshared (unique) environment. Nonidentical twins, or dizygotic (DZ) twins, on the other hand, are similar because they share on average only 50% of their genes and, when reared together, 100% of their shared (common) environment. Assuming that common environmental factors promote resemblance equally in MZ and DZ twin pairs (Plomin, De Fries, McClearn, & Rutter, 1997), greater similarity of MZ than DZ pairs is attributed to genetic factors. However, if the degree of similarity is the same for MZ and DZ twins this suggests the trait is influenced by common environmental factors. The key statistical tool in twin research is structural equation modelling (SEM), also referred to as path analysis, in which genetic and environmental effects are modelled as latent variables on the phenotypic difference of a measured trait between individuals. The contributions of the latent variables are estimated as regression coefficients in the linear regression of the measured trait on the latent variable. Modern twin analyses extend beyond the classical twin study for estimating the heritability of a trait, and allow the simultaneous modelling of two or more (dependent) variables in one analysis to investigate the contribution of genetic and environmental factors to the covariation among traits. This allows one to examine to what extent common genes pleiotropically influence multiple traits, and what extent there are specific genetic effects. For a detailed explanation of the various methods applicable to twin samples, the reader is referred to the following: Boomsma, Busjahn, and Peltonen, 2002; Evans, Gillespie, and Martin, 2002; Martin, Boomsma, and Machin, 1997; Neale and Cardon, 1992. More recently a twin pair-based approach (cf. the more usual extended pedigree or sibpair design) to linkage and association studies to map quantitative trait loci (QTL: a chromosomal region that contributes to variability in a complex quantitative trait) Correspondence: M. J. Wright, Genetic Epidemiology Laboratory, Queensland Institute of Medical Research, Post Office, Royal Brisbane Hospital, Qld 4029, Australia. Tel.: + 61 7 3362 0225. Fax: + 61 7 3362 0101. E-mail: [email protected] Australian Journal of Psychology, Vol. 56, No. 2, September 2004, pp. 65 – 78. ISSN 0004-9530 print/ISSN 1742-9536 online # The Australian Psychological Society Ltd Published by Taylor & Francis Ltd DOI: 10.1080/00049530410001734865 has been developed with notable success (Knoblauch et al., 2000). The purpose of this paper is to provide an overview of the adolescent twin studies (Table I) that are being undertaken in the Genetic Epidemiology Laboratory, at The Queensland Institute of Medical Research (QIMR) in Brisbane (information accessible at http:// www.genepi.studies.au). Although the studies are somewhat diverse, a common focus is to elucidate the genetic and environmental architecture underlying the phenotype, thereby facilitating the search for QTLs that influence these traits. The discovery of novel genes or novel variants of known genes that directly influence a trait may potentially lead to the isolation of a specific protein and/or possible regulating mechanism at the molecular level. The research program is based upon an adolescent twin sample initially recruited in 1992, with new twins being added at various intervals, and referred to as the Brisbane Adolescent Twin Sample. The sample includes both MZ and DZ twin pairs, including opposite-sex twin pairs, and although the focus has been on the recruitment of twins, other multiples (triplets and quadruplets) and singleton siblings of twins/multiples are also included. The use of an extended twin design by including nontwin siblings greatly increases statistical power to detect genetic and shared environmental influences on a measured variable (Posthuma & Boomsma, 2000). In addition, by adding a singleton sibling, MZ pairs become informative for mapping genes using linkage and within-pair association analysis (by themselves they are not), and DZ twins become more than twice as informative. We have also included the parents of the twins, and although only minor demographic information is routinely collected from parents, a large majority of the parents have donated a blood sample and therefore are useful in error detection in linkage studies as well as haplotype (a region on a chromosome that is bounded by recombination sites such that genetic loci within a haplotypic region are usually inherited as a unit) determination. At present there are three ongoing research studies that involve in-person assessment: the mole study, which began in 1992; the cognition (or memory, attention, and problem solving [MAPS]) study, which began in 1996; and the pelvic floor study (which is restricted to women) that began last year (2002). For all three studies data collection is labour intensive and time consuming, and thus expensive, and necessarily involves a smaller number of pairs than is usually available with studies in which phenotypes can be collected via questionnaire and at significantly lower cost. However, given the nature of the studies, the sample sizes thus far collected are considerable. In parallel with the mole, cognition, and pelvic floor studies, are several additional studies for which the phenotypes are collected at the same time, and most recently, a large mail and phone study was initiated. The objectives of these studies are briefly summarised below.