Twin Studies: A Unique Epidemiological Tool

506 views
Lin Li
About the Author

Lin Li is a PhD candidate at the School of Medical Science, Γ–rebro University in Grythyttan (Sweden) and works on the long-term influences of ADHD on nutrition and life-style factors.

Prof. Henrik Larsson
About the Author

Prof. Henrik Larsson is a member of the School of Medical Science, Γ–rebro University and Department of Medical Epidemiology and Biostatistics, Karolinska Institute, Sweden.


Twin studies have been used for decades to estimate the relative importance of genes and environments for traits, behaviors and disorders. A very large meta-analysis of all twin studies conducted during the past 50 years (almost 3000 publications) revealed that across all studied traits the average reported heritability was 49%, meaning that about 50% of the variation in traits is due to genetic factors [1].

1. Methods and theory of classical twin design

By comparing the differences and similarities between twins, researchers use them as a natural experiment to study whether a trait, phenotype or disease is due to nature (genetic predisposition) or nurture (environmental factors).

In order to get a better understanding of twin studies, one must first understand the two types of twins:

  • Monozygotic (MZ) or identical twins were conceived in a single egg, which split and forms two embryos. Therefore, MZ twins share all their genes (100%), and are definitely the same sex.
  • Dizygotic (DZ) or fraternal twins were developing from a separate egg and each egg is fertilized by its own sperm cell, and therefore sharing on average 50% of their genes. DZ twins could be of the same sex or different sex.

Based on the different degree of genetic and the similar extent of prenatal and later environmental factors sharing between MZ and DZ twins, MZ twin pairs may show a higher similarity on a given trait, as compared with DZ twins, if genes significantly influence that trait. On the other hand, if MZ and DZ twin pairs share a trait to an equal extent, it is likely that the environment influences the trait more than genetic factors.

The similarity for a given trait is estimates via intra-class correlations (ICC), and similarity across different traits by the cross-twin cross-trait correlations (CTCT). Comparison of correlations across MZ and DZ pairs allows for the variance (V) of a given trait to be decomposed into three factors:

  • Genetic factors, including additive genetic factors (A), and dominant genetic factors (D)
  • Shared environmental factors (C), that is events that happen to both twins, affecting them in the same way. For example, the socio-economic status of the family, the general personality and general parenting styles and beliefs of the parents.
  • Non-shared or unique environmental factors (E), that is events happen to one twin but not the other one, or the events affect either twin in a different way. For example, school and classroom environment, also including measurement error.

Under then assumptions of no interaction and no covariance between A, C, D, and E, the total variance of a phenotype (P) can be expressed as:

π‘‰π‘Žπ‘Ÿ,𝑃.=𝐴+𝐷+𝐢+𝐸

Narrow sense heritability is defined as the proportion of variance in a trait due to additive genetic effects (A):

,β„Ž-2.=,π‘‰π‘Žπ‘Ÿ(𝐴)-π‘‰π‘Žπ‘Ÿ(𝑃).

Broad sense heritability as the proportion of variance due to additive and dominance genetic effects (A+D):

,β„Ž-2.=,π‘‰π‘Žπ‘Ÿ(𝐴+𝐷)-π‘‰π‘Žπ‘Ÿ(𝑃).

The classical twin model can be extended to explore bivariate and multivariate traits association and test for differences between males and females by using sex-limitation models. More information on how to conduct classical and advanced twin model-fitting analyses, please refer to [2] and [3].

2. Important advantages of twin studies

  • Estimate the relative importance of genetic factors (i.e., heritability) of one or more traits
  • Help identify shared genetic factors that influence different traits, behaviors and disorders.
  • Explore the causal status of environmental risk factors by controlling for genetic and shared environmental confounding.
  • Offers unique opportunities to study the gene-environmental interplay, including both gene-environmental correlations and gene-environmental interactions.

In summary, the twin study design is considered an important behavioral genetic approach that has been used in many fields, including biology, psychology and sociology. Using a substantial amount of the published twin research (and other genetic informative studies, e.g. sibling comparison, adoption studies), Plomin et al. summarized the top 10 replicated and important findings [4]. These findings included:

  • All psychological traits show significant and substantial genetic influence;
  • No traits are 100% heritable, highlighting the importance of environmental factors, and
  • The heritability is caused by many genes of small effect.

Most of these findings or discoveries that could only have been found using genetically sensitive research designs.

In the Eat2BeNice project, we are currently using data from Swedish Twin Register (https://ki.se/en/research/the-swedish-twin-registry) to estimate the heritability of unhealthy eating habits and ADHD symptoms in adults, and also to investigate the relative importance of genetic, shared environmental and non-shared environmental factors for the overlap between adult ADHD symptoms and different dietary habits diets. We will also test a specific hypothesis regarding gene-environmental interactions.