The Genetics of Height

Physical height is one of those human traits which nearly everyone seems to care about, with most people either wishing they had a little more of it or wishing they could find a partner who did. In addition, height - unlike a nebulous concept like "intelligence" - is one particular trait which is quite easy to objectively measure, and it is beyond dispute that how tall one will become is heavily determined by one's genetic inheritance. As such, one would think there'd already be a huge body of work out there which nails down the genetic variations which primarily contribute to this trait, but the surprising thing is that very little was actually known about its genetic determinants until very recently (subs reqd).

In separate efforts, each team [of three in total] pooled genetic and height data collected from other studies to amass information on 13,000 to 31,000 people. The teams then scanned their respective databases for single-nucleotide polymorphisms--regions where the DNA differs by one base--that were associated with being either taller or shorter than average. Researchers led by Kári Stefánsson, a geneticist with deCODE Genetics in Reykjavik, Iceland, identified 27 such regions, and the other two groups, led by geneticist Timothy Frayling at Peninsula Medical School in Exeter in the U.K., and Joel Hirschhorn, a geneticist at the Broad Institute in Cambridge, Massachusetts, spotted 20 and 10 regions, respectively. (The teams have not yet had a chance to compare their findings, so it's not yet clear how much overlap there is.) People who carried mostly "tall" versions of these genetic variations were 3.5 to 5 cm taller on average than people carrying mostly short versions, the teams report online this week in three papers in Nature Genetics.

Now, 3.5-5 isn't really very much in terms of human variation in height, and indeed, the article discloses as much:

The researchers don't yet know how these genetic variations influence height. Some occur in genes that play a role in skeletal growth and development, which makes sense, but others lie in genes that trigger cancer or in undeciphered regions of DNA. What's more, the genetic variations account for less than 4% of the height variation found in the general population, so the researchers think numerous more regions remain to be identified.

There are two things to note here, I believe. The first is that even what seems to be a "simple" trait heavily under genetic control is turning out to be controlled by perhaps several hundred genes rather than a few all-powerful ones, and as such, knowing that two populations differ in the frequency of any particular one of these genes isn't going to tell us anything much about the limitations on height imposed by their genes; this is especially true when one considers that the genes with the biggest effects are precisely the ones which are going to be easiest to find, so if these 30-50 contribute only 4% of variation in total, we'd be talking about an effect of under 0.1% per gene in the undiscovered loci. If height is going to be this difficult to nail down, it shows how utterly bereft of justification any hopes of finding a few "IQ" genes to "explain" the supposed intellectual superiority or inferiority of human groups will be.

The second point this article brings up, albeit indirectly, is that the whole idea of genes "determining" height is itself a misnomer. It is true enough that the genes of parents will determine the ultimate height of their children when all else is held equal, but that crucial "all else being equal" condition is rarely met, so when one hears, for instance, that height is a "90% heritable" trait (as has indeed been indicated by some studies), this really means a lot less than it might seem to, for the simple reason that "heritability" as a technical term in population genetics means something very different from what a lay audience might assume it to. It is possible, for instance, that height is "90% heritable" in a group of Netherlanders as well as a group of Taiwanese, but that tells us nothing about the causes of the height variations between the two groups, only that - ethical issues aside - if it were possible to selectively breed members of both groups for height as if they were livestock, we would expect to get excellent results in raising group mean height by allowing only the tallest members of each to breed: to simply things somewhat, the high heritability figures only tell us that within each group, environmental variation is low enough that genetic differences can come to the fore.

It is a widely held bit of folk wisdom that East Asians are shorter than people from other parts of the world, and that this is so because of their genes, but when one takes the ideas above into account, this bit of received wisdom seems much less clearly based on solid foundations. The fact of the matter is that as heavily genetically determined as height may seem to be, changes in diet can account for massive jumps in mean height in populations over the course of a single generation, and until recently East Asia was much poorer than Europe and North America. What is more, diet continues to matter even when differences in wealth have disappeared, and for raising extremely tall adults, a diet heavy in milk and meat is a much better one than a diet leaning a great deal on tofu and rice - even if the end result is an epidemic of obesity and premature deaths from cancer and diabetes ... Looking further into the issue, however, it isn't even clear that East Asians really are any shorter than people from elsewhere, at least nowadays: if this Wikipedia entry is to be believed, the average young Japanese or South Korean is already taller than the average Ghanaian or Ivorean of similar age, despite likely consuming less milk and meat than a typical member of these other populations, while even young Argentinean and Spanish men are statistically indistinguishable from their East Asian counterparts, dietary differences notwithstanding; such numbers certainly do not support the idea of "innate" racial differences in height, discounting small groups like the pygmies of Central Africa.

The long and short of it: not even something as straightforward to measure as height is going to prove amenable to the simplistic fantasies of armchair genetic reductionists, and what is truly known about the genetics underlying the subject remains close to nothing, even with these new results.