Natural Selection On Human Height Doesn’t Measure Up To Much

Quantitative geneticists can sometimes seem to be obsessed with height. That’s because, frankly, it’s one of the easiest human traits to study using quantitative genetic analysis. Height is easy to measure, and variation in height has a clear genetic basis — if your parents had very similar childhood nutrition and health histories, and give you a similar childhood, your height can be very reliably predicted by taking the average of theirs.

Genomic data collected in well-controlled samples of people (who, again, have all grown up in very similar conditions) has identified hundreds of genetic variants in dozens of different genes that contribute to height. And genetic variants associated with differences in height occur at very different frequencies in populations with very different average heights — for example, tall northern Europeans versus shorter southern Europeans. That could indicate that natural selection at some time in the past is responsible for differences in height we see today, though why it would have been advantageous to be short in France, or tall in Norway, is unclear.

We recently had a hint, though, that this selection on height could be continuing in the Netherlands. The Dutch are known today as some of the tallest people in the world, but this is a relatively recent development: over the last century and a half, the height of the average Dutch man has increased from 5 feet 5 inches (1.65 meters) to over 6 feet (1.85m). Much of this is surely due to the fact that the Netherlands is a wealthy country with relatively low income inequality, and good access to healthcare for its citizens; but the fact that height is also strongly attributable to genetics had always left a possibility that this increased height could represent evolutionary change as well as social and economic progress.

Back in April, a paper in Proceedings of the Royal Society presented evidence of exactly this, showing that taller Dutch men have more children, the very definition of natural selection. This was based on analysis of data from more than 94,000 people, sampled across three generations, as part of an even larger public health study. The participants had all grown up in similar environments, and they reported their education levels and income — so it was possible to account for ways those factors might affect how many children a person had, independent of his or her height. In the final analysis, the paper’s authors estimated that taller men had more children, and this effect was greater than expected by chance, even after accounting for other possible causes.

However, there were two odd things about that paper. The first was that its analysis found women had more children if they were close to the average height — women who were shorter or taller than average had fewer children. It is somewhat surprising to think that the huge change in the height of the average Dutch person over the last 150 years is due only to selection on men. The second odd thing is that while the paper finds a significant relationship between men’s height and the number of children they had, it doesn’t explicitly estimate how that relationship would translate into evolutionary change over time.

That estimation is the focus of a brief new paper just released online ahead of print in the journal Evolution. Fairly standard evolutionary theory provides arithmetic to predict how much the average height of a population will change over time, using an estimated relationship between height and reproductive success (also called a “selection gradient”), the variation in height within a population, and the proportion of that variation attributable to genetic differences.

The actual differences in reproductive success associated with height are quite small: men who were taller than about 85% of the rest of the population had, on average, 1.4% more children. Given that this selection is observed only in men, that about 80% of variation in height is due to genetics, and that a 150-year span covers roughly six human generations, that accounts for an increase in average height of less than a tenth of an inch (2.3 millimeters) — about one percent of the total increase in the average height of Dutch people since the nineteenth century.

This is a prime example of the difference between significance in the statistical sense — whether or not an effect can be determined to be greater than zero — and significance in the sense of an effect that accounts for differences that mean something in human or biological terms. A very large, well-controlled sample can identify statistically significant effects that are nevertheless smaller than we can easily observe, or that really account for the things we’re interested in. In this case, a statistically significant estimate of natural selection suggests that, actually, the Dutch are tall for reasons that don’t have much to do with selection at all.

References

Allen, H. L., Estrada, K., Lettre, G., Berndt, S. I., Weedon, M. N., Rivadeneira, F., et al. 2010. Hundreds of variants clustered in genomic loci and biological pathways affect human height. Nature, 467(7317), 832-838. doi: 10.1038/nature09410.

Berg, J. J., & Coop, G. 2014. A population genetic signal of polygenic adaptation. PLOS Genetics, 10(8), e1004412. doi: 10.1371/journal.pgen.1004412.

Stulp, G., Barrett, L., Tropf, F. C., & Mills, M. 2015. Does natural selection favour taller stature among the tallest people on earth?. Proc. Royal Soc. B, 282(1806), 20150211. doi: 10.1098/rspb.2015.0211.

Tarka, M., Bolstad, G. H., Wacker, S., Räsänen, K., Hansen, T. F., & Pélabon, C. 2015. Did natural selection make the Dutch taller? A cautionary note on the importance of quantification in understanding evolution. Evolution. doi: 10.1111/evo.12803.