Are heritable individual differences explained by balancing selection or mutation-selection-drift balance?
– by Brendan Zietsch
A key question for evolutionary psychologists is: what selection pressures have shaped human traits and how do they vary and covary across individuals? Recent genomics studies have revealed a wealth of evidence that sheds light on these questions. In my paper, “Genomic findings and their implications for the evolutionary social sciences”, I aimed to bring together these findings while explaining the conceptual and technical background that is often assumed knowledge for reading the primary reports. I also outlined what I see as the implications of these findings for psychological life history theory and for our interpretation of individual differences more generally.
The key question that genomic studies can answer is, which form of selection has shaped genetic variation in human traits: negative selection or balancing selection? Negative (or purifying) selection removes harmful variants and depletes genetic variation. Genetic variation is maintained due to a balance between this depletion against the constant influx of new genetic variation from mutations. Balancing selection, on the other hand, refers to forms of selection that actively maintain genetic variation. It can occur when the relationship between trait value and individual fitness varies over time or place (fluctuating selection) or sex (sexually antagonistic selection), or when it depends on the rarity of the trait in the population (negative frequency-dependent selection), or when an allele’s effect on fitness depends on the other allele at the same locus (heterozygote advantage). So, the question: is the genetic variation in traits today shaped by a history of negative selection or balancing selection?
Genomics studies can be evolutionarily informative because they reveal the genetic architecture of human traits. Roughly, genetic architecture refers to the character of the genetic variation that underlies trait variation, especially the number of genetic variants that contribute to heritable variation and the how the frequencies of those variants relate to their effect sizes. Negative selection and balancing selection produce different genetic architectures (see below),. Therefore, from the genetic architecture of traits, we can make inferences about which form of selection has shaped each trait.
Certain features of genetic architecture are consistent across many traits that have been subject to genomic analysis, including traits that are of interest to evolutionary psychologists: life history traits such as age at puberty; morphological traits like waist-to-hip ratio, BMI, and height; cognitive-based traits like educational attainment; personality traits like neuroticism; and mental disorders like schizophrenia.
One common feature is that the heritability of such traits is spread evenly across thousands of genetic variants. Under no selection, or under balancing selection, we would expect that many variants might influence a trait but that a small number of these account for most of the trait variance. That’s because we know that traits are influenced by rare variants with large effect sizes, and there is no reason, other than negative selection, for this not to be true of common variants as well. As a mathematical necessity, in that case, a relatively small number of common variants with large effects would account for most of the trait variance. Instead, we see that any one variant only accounts for a tiny percentage of trait heritability, which is exactly what we would expect under negative selection.
Another feature shared among traits is that both common and rare variants contribute substantially to trait heritability. Several lines of evidence suggest that rare variants contribute disproportionately to trait variance, relative to what is expected under neutral selection or balancing selection, where virtually all the variation is expected to be accounted for by common variants. This pattern is expected under negative selection, because selection is less effective against rare deleterious alleles than common deleterious alleles. Modelling shows that balancing selection can only maintain variation at intermediate frequencies.
A third feature of genetic architecture observed across traits is that variants’ effect sizes are negatively associated with their minor allele frequency. Rarer variants tend to have larger effects than common variants, which invariably have tiny effects. The only known explanation for this pattern is that selection against harmful variants (i.e. negative selection) eliminates any common variants with large (or even moderate sized) effects, whereas rarer variants, being less visible to selection, are able to remain at low frequency in the population even with larger effects.
A fourth feature is that younger alleles (i.e. arose by mutation more recently) explain more heritability per-locus. This is expected under negative selection: deleterious alleles that have not been around as long have had less time to be eliminated by natural selection. Under balancing selection we would expect the opposite, since it would maintain variation at loci that affect the trait under selection for longer than under evolutionary drift.
These observations constitute pervasive evidence that the genetic variation in complex traits has been shaped by negative selection, and provide no evidence that it has been shaped by balancing selection. This conclusion is backed by formal tests for negative and balancing selection. These tests aggregate evidence across significantly trait-associated variants identified in genomewide association studies. They reveal that traits of interest to evolutionary psychologists show significant evidence of having been shaped by negative selection and significant evidence in the opposite direction of the criteria regarding balancing selection.
Overall, these findings mean we should not reach for balancing selection as an explanation of individual differences, as has been very common in the evolutionary social sciences. Balancing selection has been argued to have maintained a plethora of individual differences including promiscuous and monogamous individuals, cheaters and cooperators, progressives and conservatives, risk takers and hesitators, long-term planners and short-term opportunists, and aggressive hawks and peaceful doves. Indeed, various authors have argued that variation in personality traits in general is maintained by balancing selection. Genomics findings suggest that such explanations are highly unlikely.
The findings also have implications for psychological life history theory, insofar as proponents have argued that genetic covariation among traits is aligned along a fast-slow life history dimension due to balancing selection. If balancing selection has not shaped genetic (co)variation in traits, as the evidence suggests, then this claim does not get off the ground. In the paper I also discuss implications for dimensional theories of personality variation. In short, I argue that if personality variation is the result of a mess of countless genetic variants across the whole genome, many of which are rare in the population or even private to the individual, variation in personality probably does not have a simple dimensional structure (e.g. the Big Five). Rather individuals probably vary in every way possible. The Big Five may just reflect the dimensions of variation that matter to perceivers. ] We are most interested in a relatively narrow segment of all the ways people vary – we have words for (and make personality questionnaire items primarily about) the Big Five personality factors because these are relevant to our social and self-perceptions. But we don’t have words relating to blink rate, for example, even though it’s a behaviour that is socially visible (though usually unnoticed) and varies widely between individuals. The same would apply to countless other ways individuals vary that are not socially relevant or important. So all these other ways individuals vary do not make it into our personality models.
In all, the wealth of recent genomic findings gives strong insights into the history of selection on the traits we are interested in as evolutionary psychologists, as well as pointing to surprising new ways of interpreting individual differences.
Read the original article here: Zietsch, B.P. (2024). Genomic findings and their implications for the evolutionary social sciences. Evolution & Human Behavior, 45(4), 106596.
