Everything's a Little Bit Heritable
It's the first law of behavior genetics, but how do we know it?
Posted Jun 17, 2019
Although "genetics" is in the name, behavior genetics is not just about genes. Behavior genetics is the study of anything and everything that makes us different from one another. We start by sorting all potential causes of individual differences into two bins: genes and environments. Into the gene bin goes the DNA that is present in every cell in your body. Into the environment bin goes literally everything else: family, culture, hormones, gene expression, social experiences, toxin exposures, whether you've had a stressful day, or a stressful life, and measurement error. It is an exceptionally broad approach to sorting through the multitudes of causes of the complexity of the human condition, but it's a starting point. Using a handful of basic statistical methods, over several decades and millions of participants, behavior geneticists have identified four laws that apply to everything we've studied that might be called a complex phenotype (or outcome).
- Everything's (at least) a little bit heritable.
- Shared environments have little effect.
- Unique environments always play a role.
- Heritability comes not from one or two genes, but from hundreds or thousands of genes, each of which has a very, very, very small individual impact.
How do we know that individual differences are heritable?
Let's start with the first law: everything's (at least) a little bit heritable. We infer that something is at least partly heritable when people who are more genetically similar are also more similar in their outcomes. There are many ways it can be estimated, but a straightforward (and common) approach is to compare twins. Twins come in two types: identical and fraternal. Identical (or monozygotic) twins result from a single sperm fertilizing a single egg, to create a single zygote, which then splits in two before continuing to grow. In the end, identical twins have the same set of DNA. Fraternal (or dizygotic) twins result from two different eggs being fertilized by two different sperm, resulting in two zygotes that grow in parallel. Fraternal twins are no more closely genetically related than any pair of full siblings (that is, sharing on average 50 percent of the DNA that can differ between any two people), although they (conveniently for researchers) happen to be born at the same time, so we don’t need to worry about differences between twins within a pair being due to age differences. When twins are raised together, we now have a very straightforward way to estimate the influence of genes. The only thing that is more similar between identical twins than fraternal twins is their DNA. So, when identical twins end up more similar to one another than fraternal twins, we infer that the trait is partly heritable, or partly influenced by genes.
How similar are twins?
If a trait was 100 percent heritable, then identical twins would be exactly the same as one another. If a trait was zero percent heritable, then identical twins would be no more similar to one another than fraternal twins. In practice, we don't see either of these two extremes occurring. In what's commonly referred to by behavior geneticists as "The Heritability of Everything" paper (a meta-analysis, or systematic combination of available data, of twin studies across all outcomes) showed that the average correlation between identical twins was 0.6, and the average correlation between fraternal twins was 0.3. So in neither group are the twins exactly alike (that would be a correlation of 1.0), nor are they entirely dissimilar (a correlation of 0.0). And, consistently across outcomes, identical twins are on average more similar to one another than are fraternal twins. All of this together means that, across everything we've studied, heritability isn't 100 percent, but it isn't zero percent either.
A neat trick with twin correlations
Heritability is usually given as a percentage, to make clear that it can range from zero to one, or zero percent to 100 percent. Most behavior genetics research uses statistical approaches that are better suited to computer estimation than to calculation by hand (structural equation modeling, maximum-likelihood estimation). But we can get what's referred to as Falconer's estimates for heritability just by comparing the correlations between identical and fraternal twins. Heritability is approximately equal to twice the difference between the identical and fraternal twin correlations, or (using the correlations from the Heritability of Everything paper):
- heritability = 2 x (identical twin correlation - fraternal twin correlation)
- heritability = 2 x (0.6 - 0.3) = 0.6 or 60%
Falconer's estimates hold up quite well compared to more computationally intensive approaches. In the Heritability of Everything paper, for example, more complex statistical methods gave an average heritability estimate of 49 percent—not far off from our Falconer's estimate of 60 percent. Raw twin correlations are typically reported in behavior genetics research papers so that the reader can form expectations before the presentation of more complex statistical results.
It's not just twins
Although much behavior genetics research has focused on twins, heritability can also be estimated from non-twins, whether inferred from family trees (such as with extended family data, or comparing how similar adopted children are to their adoptive and biological parents) or from measured genotypes from unrelated individuals. It does happen that some things are more heritable than others (for example, schizophrenia and autism are strongly heritable, but depression is weakly heritable). But across different methods and traits, the conclusions are remarkably similar: everything is at least a little bit heritable.
It's now well established that we're never talking about nature versus nurture. To say that something is heritable is neither groundbreaking nor informative. Nor does something being heritable mean that a person's life is pre-determined by their genes. Even more frustrating, heritability doesn't tell us anything about any one person. In fact, the influence of genes or environments on any given person may differ substantially from the group-average statistic of heritability. In order to understand the causes of anything that differs between people requires understanding a complicated mix of genes and environments.
Turkheimer, E. (2000). Three laws of behavior genetics and what they mean. Current Directions in Psychological Science, 9(5), 160-164. https://journals.sagepub.com/doi/pdf/10.1111/1467-8721.00084
Chabris, C. F., Lee, J. J., Cesarini, D., Benjamin, D. J., & Laibson, D. I. (2015). The fourth law of behavior genetics. Current Directions in Psychological Science, 24(4), 304-312. https://journals.sagepub.com/doi/pdf/10.1177/0963721415580430
Polderman, T. J., Benyamin, B., De Leeuw, C. A., Sullivan, P. F., Van Bochoven, A., Visscher, P. M., & Posthuma, D. (2015). Meta-analysis of the heritability of human traits based on fifty years of twin studies. Nature Genetics, 47(7), 702-709. https://www.nature.com/articles/ng.3285