I first met Nick at the very first ‘Twin Methodology’ workshop held in Leuven, Belgium, in 1987, and Since then I have had the pleasure of seeing him at almost every workshop (at 34 right now and counting). I was lucky enough to be able to attend the first workshop as it was held at my alma mater. Then I helped organize the next, and from then on, I was invited to help teach them. Nick and I have spent countless sessions teaching ‘the ACE model 101’ together to hundreds of workshop participants, with the same level of enthusiasm from Nick as at the first workshop. It was this enthusiasm for science and the desire to improve how investigators analyze their data that attracted me to pursue this line of research and led me to move to Richmond for a postdoc with Lindon Eaves.
Although Nick had already moved on from Richmond to Brisbane to start his own genetic epidemiology unit, his first major data collection project was clearly inspired by the work he had done with Lindon. They had conceived the extended twin (ET) model. Recognizing the limitations of the classical twin study, which typically partitions the variance in a trait in additive genetic (A), common (C) and unique (E) environmental sources, they sought to extend it to include other relatives such as the parents, siblings, spouses and children of twins (COTs). These extensions allow one to both evaluate the consistency of the estimates of the genetic and environmental contributions to the variance across a range of relationships and estimate additional sources of variance confounded with simpler study designs. The ET model (Eaves, Heath, Martin, Neale et al., Reference Eaves, Heath, Martin, Neale, Meyer, Silberg, Walters and Cloninger1999) provides a test for environmental or cultural transmission as well as genetic transmission, thus dividing the shared environment into sources shared with parents and those shared with siblings but not parents. In addition, dominance variance can now be simultaneously estimated with shared environmental variance. Further excess environmental sharing in twins compared to siblings can be quantified as ‘special twin environment’ or reflect potential age-specific effects of genes. Sex differences in all these sources of variance can equally be evaluated. Finally, relationship through marriage provide information about the extent of assortative mating.
Thus, from the theory of the causes of variation in human behavior, they developed a model system (Truett et al., Reference Truett, Eaves, Walters, Heath, Hewitt, Meyer and Kendler1994) for the analysis of family resemblance in extended kinships of twins, and collected data on health and lifestyle from a large sample of twins and their relatives and then fitted their model to the data and started the ‘stealth’ revolution. A path diagram of the model resembled a stealth bomber, a fitting name for a powerful model. Questionnaire data were collected on thousands of twins and their relatives both in Virginia (the Virginia 30,000) and in Australia (the Australia 25,000), allowing researchers to this day to explore the complexities of the causes of variation in complex traits ranging from social attitudes (Eaves, Heath, Martin, Maes et al., Reference Eaves, Heath, Martin, Maes, Neale, Kendler and Corey1999), depressive symptoms (Kendler et al., Reference Kendler, Walters, Truett, Heath, Neale, Martin and Eaves1994), panic and phobias (Kendler et al., Reference Kendler, Walters, Truett, Heath, Neale, Martin and Eaves1995), body mass index (Bergin et al., Reference Bergin, Neale, Eaves, Martin, Heath and Maes2012; Maes et al., Reference Maes, Neale and Eaves1997), church attendance (Kirk et al., Reference Kirk, Maes, Neale, Heath, Martin and Eaves1999), alcohol use (Maes et al., Reference Maes, Neale, Martin, Heath and Eaves1999; Verhulst et al., Reference Verhulst, Neale, Eaves, Medland, Heath, Martin and Maes2018), neuroticism (Boomsma et al., Reference Boomsma, Helmer, Nieuwboer, Hottenga, de Moor, van den Berg and de Geus2018; Lake et al., Reference Lake, Eaves, Maes, Heath and Martin2000), smoking initiation (Maes et al., Reference Maes, Morley, Neale, Kendler, Heath, Eaves and Martin2018; Maes et al., Reference Maes, Neale, Kendler, Martin, Heath and Eaves2006), political attitudes (Hatemi et al., Reference Hatemi, Funk, Medland, Maes, Silberg, Martin and Eaves2009), and so on. Many of these publications would not have happened had it not been for Nick’s generosity of data, time, encouragement, travel assistance and hospitality, discussing results over wine and good food, often accompanied by excellent classical music.
Through these interactions, Nick inspired graduate students and postdocs to further explore the ET model. While the first iterations and applications of the ET model were written in Fortran, we developed code in Mx and later in OpenMx (Maes et al., Reference Maes, Neale, Medland, Keller, Martin, Heath and Eaves2009) that allowed fitting it to raw data, to continuous and categorical data, incorporating covariates and extending it to the multivariate case (Maes et al., Reference Maes, Neale, Martin, Heath and Eaves1999). Alternative mechanisms of intergenerational transmission and assortment — phenotypic cultural transmission and social homogamy — were coded (Keller et al., Reference Keller, Medland, Duncan, Hatemi, Neale, Maes and Eaves2009). New programs were written to simulate data to evaluate bias, precision and accuracy of the parameter estimates (Coventry & Keller, Reference Coventry and Keller2005; Keller et al., Reference Keller, Medland and Duncan2010) as well as power associated with different family structures (Medland & Keller, Reference Medland and Keller2009). Additional relatives (Vinkhuyzen et al., Reference Vinkhuyzen, van der Sluis, Maes and Posthuma2012) and non-biologically related family members (Leve et al., Reference Leve, Neiderhiser, Harold, Natsuaki, Bohannan and Cresko2018; Maes et al., Reference Maes, Silberg, Neale and Eaves2007) were included. The utility of subsets of the data, such as COTs, to disentangle genetic from cultural transmission was explored (Docherty et al., Reference Docherty, Kremen, Panizzon, Prom-Wormley, Franz, Lyons and Neale2015; Eaves et al., Reference Eaves, Silberg and Maes2005) and expanded (Marceau et al., Reference Marceau, Narusyte, Lichtenstein, Ganiban, Spotts, Reiss and Neiderhiser2015; McAdams et al., Reference McAdams, Hannigan, Eilertsen, Gjerde, Ystrom and Rijsdijk2018). Cross-cultural comparisons were undertaken to test the reproducibility and consistency of findings (Lake et al., Reference Lake, Eaves, Maes, Heath and Martin2000; Maes et al., Reference Maes, Morley, Neale, Kendler, Heath, Eaves and Martin2018). Twin registries were expanded with data collected from other relatives (Boomsma et al., Reference Boomsma, Willemsen, Vink, Bartels, Groot, Hottenga and van der Kleij2008; Kaprio et al., Reference Kaprio, Rose, Sarna, Langinvainio, Koskenvuo, Rita and Heikkila1987; Ligthart et al., Reference Ligthart, van Beijsterveldt, Kevenaar, de Zeeuw, van Bergen, Bruins and Boomsma2019), recognizing the added value, not just in terms of power but in capturing more of the nuances of how genetic and environmental factors act and interact in creating individual differences. The list of phenotypes to which these models has been applied continues to grow, with publications on brain structure (Posthuma et al., Reference Posthuma, de Geus, Neale, Hulshoff Pol, Baare, Kahn and Boomsma2000), blood pressure (Kupper et al., Reference Kupper, Willemsen, Riese, Posthuma, Boomsma and de Geus2005), parturition timing (Kistka et al., Reference Kistka, DeFranco, Ligthart, Willemsen, Plunkett, Muglia and Boomsma2008), personality disorder (Distel et al., Reference Distel, Rebollo-Mesa, Willemsen, Derom, Trull, Martin and Boomsma2009), intelligence (Vinkhuyzen et al., Reference Vinkhuyzen, van der Sluis, Maes and Posthuma2012), political orientation (Kandler et al., Reference Kandler, Bleidorn and Riemann2012), proinflammatory state (Neijts et al., Reference Neijts, van Dongen, Kluft, Boomsma, Willemsen and de Geus2013), personality (Hahn et al., Reference Hahn, Spinath, Siedler, Wagner, Schupp and Kandler2012; Kandler et al., Reference Kandler, Penner, Richter and Zapko-Willmes2019) and political affiliation (Hufer et al., Reference Hufer, Kornadt, Kandler and Riemann2019; Kornadt et al., Reference Kornadt, Hufer, Kandler and Riemann2018). On a personal note, Nick has been extremely supportive in my career — and deserves every spot as coauthor and contributor. Furthermore, he genuinely cares about moving the field of (behavior) genetics forward and has clearly put his stamp on developing models for ET kinships, collecting relevant data and fitting ET models to them, and through it all mentored and encouraged his academic extended family, while enjoying their company during ‘just bring food’ dinners, good wine and, if possible, listening to some lovely music. Thanks, Nick!