How fast do gene frequencies change? : Gene Expression

Search

Q & A

Books

How fast do gene frequencies change? permlink

Share this: Facebook Twitter Stumbleupon Reddit Email + More

Posted on: August 13, 2007 10:25 PM, by Razib Khan

Ezra Klein asks "How Quickly Do Genes Change?" in response to Andrew Sullivan gushing over Greg Clark's new book, A Farewell to Alms. Clark offers the hypothesis that the industrial revolution in England was catalyzed in part by changes in behavior which might have been reinforced by selection for particular alleles. In terms of the specific hypothesis, I'm skeptical and would probably bet that Clark has overplayed his hand and put too many eggs in one basket. But, in response to Ezra's question I threw down a flurry of comments (with a lot of grammatical errors due to my haste) which you can see over at the post. The short answer of course is: the rate of change of allele frequencies depends on the parameters. Molecular genetics is often about positing deterministic biophysical pathways, but evolutionary genetics is just as interested in the fluctuations and variations of alleles over time, and those variations are contingent on a host of factors. There isn't one specific answer, just a range depending on your assumptions or the empirical realities (e.g., selection coefficient, heritability of the character, demographic history, meta-population dynamics, etc.). So whether Clark's hypothesis is right or wrong as an empirical question, not a theoretical one. One can't rule it out a priori.

Share this: Facebook Twitter Stumbleupon Reddit Email + More

Comments

But the parameters necessary would be unlikely? I'd bet there would have to be pretty heavy selection pressure (i.e. pretty big differentials in getting grandchildren).

A lot would hang on his assertion that "the modern population of the English is largely descended from the economic upper classes of the Middle Ages."

One obvious place things may have gone wrong is his reliance on wills to track numbers of children. If the truly poor weren't likely to make wills, it may be that what Clark has found is that the upper middle reproduced more than the lower middle, which I find easy to believe, rather than the upper middle reproducing more than everyone else.

Posted by: Oran Kelley | August 13, 2007 11:45 PM

But the parameters necessary would be unlikely? I'd bet there would have to be pretty heavy selection pressure (i.e. pretty big differentials in getting grandchildren).

i don't think he's offered a formal model to really bite into. but i haven't read the book. the descriptions leave me skeptical, though i don't dismiss it out of hand.

Posted by: razib | August 14, 2007 1:43 AM

I suppose one can generalize that a *new* allele will always take a long time (in generations) to get going, because it starts rare, and even a high percentage advantage in fitness compared to other alleles will only gradually raise its frequency in the population.

Posted by: David B | August 14, 2007 10:25 AM

It shouldn't be hard to ballpark this stuff. The wealth differential has been found everywhere anyone has looked in pre-demographic transition groups. What he describes in England is the same as what happened in Portugal or Sweden.

For a ballpark, assume time preference has an additive heritability of 25%. Assume that everone with time preference more that 1 sd above the mean of the distribution has double the fitness of everyone else. About 16% of the population then has twice the number of offspring as everyone else on average.

After a generation of reproduction the new mean time preference will be increased by (0.2 * .25) or 5% of a standard deviation. In 20 generations, 500 years, time preference should go up by a full standard deviation.

The best talking sociologist I have ever read is Edward Banfield. He thought that time preference was at the core of class differences. Time preference does also have a very high IQ correlation. (Clark's token arm waving about IQ in his manuscript does him no credit).

Henry

Posted by: henry harpending | August 14, 2007 10:51 AM

I see here and elsewhere people saying they haven't read the paper yet. The core chapter is an engaging easy read and it is all over the web. Search for "Genetically Capitalist?"

As Peyton Manning says, it's not that hard.

Henry

Posted by: henry harpending | August 14, 2007 10:55 AM

Harpending is probably right to emphasize time preference (although I have a different interpretation of its significance). Time preference is an important part of life history strategy. Namely, low time preference in predictable, low risk environments and high time preference in unpredictable, high risk environments.

Human life history strategy is largely determined by prenatal and preadult environments (Chisholm, Daly and Wilson et al.) a legacy of adaptation to social and ecological environments that are fluctuating over various time scales. Changes in the technological-economic infrastructure constructed an environment for certain socioeconomic groups that favored a life history strategy that included, among a suite of traits, low time preference. Economic growth brought more socioeconomic strata into that life history shift. Later, the trade-off between the high embodied capital investment (Kaplan et al.) required for a technological economy and number of offspring, combined with memetic emulation of reference groups (not followed by certain groups e.g. Mormons), resulted in the demographic transition.

It could be argued that this constructed niche may have shifted the genetically based reaction norm for life history strategy through genetic assimilation, but I doubt it since economic mobility favored the genetic maintenance of plasticity.

Posted by: Colugo | August 14, 2007 8:01 PM