[Posted on 19 August, 2020 by Lisa Dinkler]
The issue of nature versus nurture has been debated for centuries. From the beginning of the 20th century, the twin study design has helped us to shed light on this debate since it offers the possibility to disentangle the effects of genes (nature) and environment (nurture). The first formal twin registry was established in Denmark in 1954. As of 2019, there are more than 30 twin registries all over the world (1).
Twin studies investigate whether differences between us (called individual differences or individual variation) are influenced by differences in our genes or environments. More precisely, twin studies estimate the proportion of genetic and environmental influences on human characteristics (often called traits), such as personality traits, cognitive traits and psychiatric disorders. Twin studies do not tell us which and how many genes that are involved in, for example, a psychiatric disorder. This is traditionally the subject of the field of molecular genetics and genomics.
Over the past five decades, twin studies have shown that all psychological traits are under substantial genetic influence (even things you never considered to be genetically influenced such as the act of voting (2)). Twin studies have also shown that genetics on average explain half of our individual differences in having psychiatric disorders (3, 4), which has dramatically changed our understanding of psychiatric disorders and guided genomic research within psychiatry.
The classical twin design
All humans share circa 99.9% of their DNA. The 0.1% that is not shared causes individual differences and are called segregating alleles **.Identical twin twins share 100% of their segregating alleles, because they develop from the same maternal egg, fertilised by a single sperm, which splits after fertilisation. These twins are therefore called monozygotic. Non-identical twins share on average 50% of their segregating alleles (just like any other couple of non-twin siblings), because two different eggs were fertilised by two different sperms (dizygotic twins).
Another special feature of twins is that both identical and non-identical twins grow up in the same home environment at the same age. Therefore, both types of twins share the home environment (making the twins more similar to each other) to the same extent. This is called the equal environments assumption and enables us to “control” for the environmental influences of the home environment (nurture) when estimating heritability (nature). In other words, identical and non-identical twins differ only in the genetic influences they share, but not in the environmental influences they share.
We estimate heritability by comparing the degree of similarity between identical and non-identical twins. If identical twins are more similar to each other than non-identical twins, we can conclude—due to the equal environments assumption—that this is due to genetic influence on the trait (i.e., the trait is heritable). The more similar identical twins are to each other and the less similar non-identical twins are, the higher the heritability. In the simplest twin study design, we estimate the proportions of heritability, shared environmental factors and non-shared environmental factors, based on the correlations between identical and non-identical twins. Environmental factors include all non-genetic influences and therefore comprise a very broad range of influences such as drug use during pregnancy, birth complications, parenting behaviours, media consumption, and exposure to toxins, just to name a few. Some environmental factors are usually shared between the twins (e.g., neighbourhood, parental education, parenting behaviours, or the amount of conflict in the household) and therefore make the twins more similar to each other. These factors are called shared environmental factors. Other environmental factors are not shared between the twins—especially as they become older—and therefore make the twins more different from each other. Examples are having different peers, teachers, and hobbies. These non-shared environmental factors are the reason that even identical twins are not completely similar in everything. Whether specific environmental factors are shared or non-shared between twins is very individual. While the neighbourhood is usually shared because the twins live in the same place, parenting behaviours can indeed be non-shared if the parents act very differently towards the twins.
As you can see, the twin study design is quite intriguing, but also has its limitations. For instance, genetic and environmental factors might interact and correlate with each other. Another problem is the assumption that identical twins always share 100% of their segregating alleles, as it is not always completely true. Certain mutations (called de novo mutation ***) can happen after the fertilised maternal egg has split and therefore lead to small genetic differences between identical twins (5). In this case, heritability would be overestimated using the twin model described above.
Now that we have covered the basics of the twin study design, it is time to get to the ESSENCE part.
|What is heritability?
Heritability is a bit of a tricky term. It measures how well differences in people’s genes account for differences in their traits. In other words, heritability is the proportion of differences between people explained by genetic factors. Since heritability is a proportion, it can take values between 0 and 1. However, heritability is often expressed as a percentage instead. So, when we say that, for example, autism is 74% heritable, it does not mean that 74% of an individual’s autism is caused by genes and 26% are caused by environmental factors; nor does it mean that the chance of a parent with autism to have a child with autism is 74%. Instead, a heritability of 74% means that the individual variation in autism (i.e., the different number of autism traits different people show) is 74% due to genetic differences between people and 26% due to different environmental factors people are exposed to.
Let’s take an example to make this clearer. I think we all agree that the fact that humans walk on two legs is genetically determined. However, the heritability of walking on two legs will be very low. Why is that? Because there is little variation in humans regarding their ability to walk on two legs and if there is variation, it’s very often due to accidents (which are environmental *** factors), where people lose the function of their leg(s). Therefore, the proportion of genetic influence on the variability of walking on two legs (i.e., its heritability) will be close to 0, while walking on two legs still is genetically determined in humans. This and more examples helping to understand heritability can be found here: https://theconversation.com/explainer-what-is-heritability-21334.
Twin studies and ESSENCE
1). Autism and ADHD are highly heritable conditions
The certainly most-studied conditions within the ESSENCE spectrum are autism and ADHD. Both also seem to have the highest heritabilities within the ESSENCE spectrum (ca. 74%), but the estimates differ considerably between studies and range between 60-90% (7, 8). When trying to interpret these estimates it can be helpful to bear in mind (a) that no psychological trait or psychiatric disorder is 100% heritable (which is why they are called complex traits and complex disorders) and (b) that on average, psychological traits have a heritability of 50% (3).
That autism and ADHD could be (largely) due to genetics has long been a taboo subject, until twin studies have shown over and over again that indeed most of the individual variation in autism and ADHD is due to genetic factors. This has certainly helped to remove the blame that was put on parents and their parenting styles, as in the case of the long-held belief that “refrigerator” mothers cause autism in their children (9).
The heritability of other ESSENCE conditions seems to be slightly lower than the heritability of autism and ADHD. Here are a few examples: developmental coordination disorder 70%, tic disorders 56%, conduct disorder 55%, oppositional defiant disorder 50-62%, and dyslexia 52-64% (4, 10-13).
2). Everyone is “on the spectrum”- some of us lower and some of us higher
Basically every one of us has traits of autism or ADHD, very few people have no traits whatsoever, many people have some traits and few people have many traits (14). This is what we call the normal distribution of traits. Disorders such as autism and ADHD are more and more viewed as representing the extreme end of this distribution of traits, at least in a large portion of affected individuals. Twin studies have helped us to understand that this view also seems to hold on a genetic level (15, 16).
Let’s take the example of autism. With twin studies we can investigate if it is the same genes that influence (a) the whole distribution of autistic traits (if a person has no, some or many autistic traits; can also be thought of as the degree of autistic traits) and (b) if a person has an autism diagnosis or not. Twin studies have shown that genetic influences on the degree of autistic traits and having an autism diagnosis overlap to up to 50% (15, 17). Furthermore, genomic studies have found thousands of genetic variants associated with risk for autism (18). These genetic risk variants all have only very tiny effects on the risk for autism, but the more of them an individual has, the higher their risk for autism (this is called a high genetic vulnerability). Together, these studies support the idea that the genetic vulnerability for autism is normally distributed in the whole population, which also leads to a normal distribution of autistic traits in the population (14).
There are, however, a few caveats to the hypothesis that genetic vulnerability for autism is normally distributed in the whole population. You might have noticed that the overlap of genetic influence on the degree of autistic traits and having an autism diagnosis was not 100% (but 50% at the maximum). This means that there are also genetic influences specific to having a diagnosis of autism (meaning that they do not influence the degree of autistic traits in the rest of the population). Such specific genetic influences might play a role in individuals where autism is a by-product of genetic syndromes, such as Fragile X syndrome. Fragile X syndrome is caused by genetic mutations with large effects that are very uncommon in the population, leading to a stronger genetic distinction between the autism in these individuals and the autistic traits in the rest of the population.
3). Autism and ADHD are relatively stable over time due to genetic influence
The stability of autistic traits from childhood to adulthood is mainly (85%) due to genetic influences, however, there are indications that a part of these genetic influences on the stability might be qualitatively different between childhood and adulthood, that is, some of the genetic variants that influence autistic traits in childhood are not the same as the genetic variants influencing autistic traits in adulthood (19, 20).
4). The overlap of genetic influence on different ESSENCE conditions is huge
Twin studies have found that there is something called a general genetic factor, which suggests that many ESSENCE conditions may share a considerable part of their genetic influences (21, 22). Interestingly, these findings provide more evidence for the ESSENCE concept itself. The genetic relationship of autism and ADHD has been studied the most in this context. Twin studies have found that autism and ADHD share between 29-77% of the genetic factors that influence them, which admittedly is quite a broad range of estimates (10, 23-25). That not all genetic influences on autism and ADHD are shared between them means that there also are genetic factors specific to autism and genetic factors specific to ADHD. Furthermore, some studies suggest that the overlap of genetic influences on autism and ADHD is, at least in part, dimension-specific. Of the two autism dimensions, restrictive and repetitive behaviours and interests seem to be more strongly associated genetically with ADHD than social interaction and communication problems (26, 27). Such knowledge can guide further research by suggesting that, for example, molecular genetic (“gene-finding”) studies might benefit from focusing on the specific dimensions of autism and ADHD.
5). If you are a parent with autism, it does not mean your child definitely will have autism, too
As mentioned before, a heritability of 70% does not mean that a child of parents with autism has a likelihood of 70% to also have autism. Of course, it is likely that parents pass on at least a part of their autism risk genes to their children. This is why autistic traits in parents and children are moderately correlated (28). How high the autism risk for the offspring is will depend on if one or both parents have autism, how many risk genes the child inherits and which environmental factors play a role (remember: the heritability of autism is around 74%, and not 100%, therefore environmental factors are important as well). To explain the interplay of genetic and environmental factors in increasing the risk for a psychiatric condition the metaphor of mental illness jars or mental health jars is often used. You can read more about mental illness jars here: https://blogs.kcl.ac.uk/editlab/2019/11/18/mental-health-problems-in-your-family-do-not-mean-that-you-will-definitely-develop-a-problem-too/
The future of twin studies
By now, the heritability of most known psychological traits and psychiatric disorders has been estimated repeatedly and for many of them heritability estimates have even been summarised in meta-analyses (4). The classical (simple) twin model as described above might therefore be a bit outdated today, however, twin studies with more complex designs are still extremely valuable. Such extended designs comprise, for example: 1) studying whether the same genetic factors influence a trait at different stages of life (see point 3 above); 2) studying the genetic influence shared between traits/disorders (see point 4 above); 3) including data from other family members (parents, siblings, spouses or children) allowing for a much larger range of hypotheses to be tested; 4) studying identical twins who are discordant (i.e., not similar) for a certain trait or disorder, making it possible to investigate causal effects of environmental factors as well as epigenetics (changes in gene activity and gene expression), and 5) combining information on traits and disorders with DNA from blood or saliva and other biological materials (stool, hair, skin etc.) (29). Apart from enabling studies on a genetic level, twin studies are usually designed as long-term follow-up studies and therefore provide rich data on the development of traits and disorders over time for epidemiological studies.
* ESSENCE is an acronym for Early Symptomatic Syndromes Eliciting Neurodevelopmental Clinical Examinations that was coined by Christopher Gillberg in 2010. It is an umbrella term describing early symptoms of conditions such as autism and ADHD (but also many others such as intellectual disability, specific learning disorder, developmental coordination disorder, Tourette syndrome, etc.) that lead parents to seek clinical help for their children. You can read more about ESSENCE.
** Genes exists in several variant forms. An allele is one of the forms a gene can take.
*** de novo mutation is a non-inherited genetic variant. It arises either in the parental germ cells or in the fertilised egg during early embryogenesis .
**** To be exact here, accidents/injuries are not “pure” environmental factors, but they are under some genetic influence, meaning that genes explain a small part of why people become involved in accidents/injuries. However, since it is only small part that is explained by genes, they can still be called environmental factors (6).