the InMind blog

Where are the genes for psychological traits?

• written by Marcus Munafo

In this blog post, I will discuss the methodologies available for identifying genetic influences on psychological traits, why it has proved so difficult to reliably identify specific genes, despite 20 years of effort, and what this tells us about the nature of these genetic influences.

Why are people different – in their personality, cognitive ability, risk of experiencing mental health problems, and so on? For decades we have known, through twin, family and adoption studies, that some of the variability we observe is the result of inherited differences. These studies allowed us to calculate the heritability of a trait – broadly speaking, the proportion of variation in the trait that can be attributed to genetic influences. However, until relatively recently, we were unable to go much further than that. Knowing that some genes are involved in these traits is very different to knowing which genes… But knowing which genes is important – understanding the role of specific genes in variation in psychological traits should give us insight into the neurobiological mechanisms that underpin these.

There are around 20,000 genes in the human genome, and we each inherit two copies of these – one from each parent. Most genes can exist in subtly different forms (they are polymorphic), and these forms (alleles) give rise to specific combinations (our genotype). The correlation between our genotype and various observable characteristics (our phenotype) is the basis for genetic association studies. In the mid-1990s it became possible to directly measure an increasing number of genes, and therefore look for correlations between genetic variation and psychological phenotypes. One would simply measure the genotype and phenotype of a group of individuals, and look to see whether those who carry a particular allele score higher, on average, on some psychological trait. As genotyping costs reduced steadily, genetic association studies proliferated. This first wave of studies used candidate gene methods. These select individual genes for investigation on the basis of their known (or presumed) biology, and corresponding relevance to the trait of interest.

So what have candidate gene methods told us? Despite initial promise, a fairly consistent pattern quickly emerged – initial findings failed to replicate and a mixture of replications, partial replications and non-replications followed. The reason for these failures was, as it later turned out, that individual genetic effects on complex behavioural phenotypes such as personality and intelligence are very small. A study comprising a few hundred individuals is simply unable to reliably detect any true association. Therefore, by chance alone, some studies found evidence (i.e., a P-value less than 0.05) that a particular allele in a particular gene is associated with, say, the personality trait of neuroticism. But when other studies attempted to replicate this finding most could not, or even found an association in an opposite direction (that is, the allele associated with higher neuroticism in the first study might be found to be associated with lower neuroticism in a later study).

On the one hand, this failure is very informative. Specifically, a) what we thought we knew about the biological basis of many psychological traits may be incomplete, and b) genetic effects of common variants (i.e., those with a minor allele frequency of 5% or more) on these traits were probably very small. On the other hand, it’s worth considering what this approach could ever have really told us. Remember that candidate genes are selected on the basis of their known or presumed neurobiology. So scientists interested in anxiety-related traits, for example, might study genes in the serotonin pathway, because we believe that serotonin plays a role in these behaviours. But if an association is reliably observed between a gene in this pathway, this could only tell us what we already know. In principle a failure to detect an association could disprove the role of serotonin in anxiety, but in practice candidate gene studies are rarely or never used in this way.

Although candidate gene studies continue to be published, for the most part these methods have been supplanted by methods that interrogate the whole genome (such as genomewide association studies). The design of these studies has been informed by what was learned from candidate gene studies – namely, that the effects of individual genes on complex behavioural phenotypes will be very (very) small. They also build on parallel advances in human genetics, such as the Human Genome Project. In my next post, I will discuss how these whole genome methods have transformed our understanding of the genetic basis of many phenotypes.

References:

Chabris, C.F., Lee, J.J., Benjamin, D.J., Beuchamp, J.P., Gleaser, E.L., Borst, G., Pinker, S., & Liabson, D.I. (2013). Why it is hard to find genes associated with social science traits: theoretical and empirical considerations. American Journal of Public Health, 103, 152-166.

Munafò, M.R., & Flint, J. (2011). Dissecting the genetic architecture of human personality. Trends in Cognitive Science,15,395-400.