Thursday, November 17, 2005

Heterozygosity and vigor   posted by Razib @ 11/17/2005 08:45:00 AM

An interesting paper is out in PNAS titled Genetic diversity and reproductive success in Mandrills. The money shot:

We found that heterozygous individuals showed greater reproductive success, with both females and males producing more offspring. However, heterozygosity influenced reproductive success only in dominant males, not in subordinates. Neither the acquisition of alpha status in males, nor social rank in females, was significantly correlated with heterozygosity, although more heterozygous alpha males showed longer tenure than homozygous ones. We also tested whether the benefits of greater genetic diversity were due mainly to a genome-wide effect of inbreeding depression or to heterosis at one or a few loci. Multilocus effects best explained the correlation between heterozygosity and reproductive success and tenure, indicating the occurrence of inbreeding depression in this mandrill colony.

Inbreeding depression often occurs when recessive and or deleterious alleles are unmasked in the genome. Well all carry a mutational load, and I have read/heard that each human harbors 3 de novo unique lethal mutations. The importance of outbreeding is that it perpetuates complementation so that a good copy is present to compensate for the faulty copy of a given gene. The logic is crystal clear as to why brother-sister matings are extremely deleterious: they are likely to share bad copies from their parents, so complementation will not occur and the lethal or deleterious alleles will be unmasked. For example, if a lethal copy of a gene is present at a frequency of 0.1% in the population, it would be unmasked due to random mating at a rate of 1 in 1,000,000 (0.1%*0.1% = 0.001*0.001, va p2 + 2pq + q2 = 1). If parents are heterozygous for that same lethal copy (that is, both carry a good copy as well as the bad copy), simple Punnet Square logic tells us that 25% of the offspring are dead on arrival. If you constrain the breeding population to descendents of this pair, the initial generations will have a high mortality rate because of the reexpression of the lethal copies. Obviously selection will work against it, but in a small population random genetic drift is strong enough to fix even deleterious alleles! If it was a lethal, you get population extinction, but that is not the only scenario. One could have mildly deleterious alleles scattered across the genome, for which fixation and reproduction of this state during subsequent generations are both possible.1 These loci would be a non-trivial fitness hit, but they would not be the kiss of reproductive death. The implication above is heterozygosity, the tendency to possess two versions of the same gene on a locus, as opposed to the same version (homozygosity), tends to correlate well with a low expression of negative mutational load (due to cumulative masking effects across many loci). On the other hand, the researchers point to no evidence of heterosis, which would imply that the heterozygote phenotype was more fit than either homozygote alternatives. This is not surprising, it is relatively hard to find this in nature (over the long term this might be true of the MHC loci, though the line between that and frequency dependent selection is fine in my opinion).

A mandrill colony is interesting because they are primates, our cousins. Biology is a science of generalizations with many exceptions because the species sample space is so large, and looking toward complex mammals, and primates no less, can give us insights into humans (though they are not determinative or definitive! Monkeys we not be). Steve points me to this piece which offers that:

...British Pakistanis account for 30% of all British children with recessive disorders, which include cystic fibrosis.

Why? See here:

The research, conducted by the BBC and broadcast to a shocked nation on Tuesday, found that at least 55% of the community was married to a first cousin.

I have elucidated the prediction equations in regards to recessive diseases and cousin matins before. Though the individual risk may be low in absolute terms (i.e., 2% vs. 0.1%), the social impact of cousin matings can be great as the frequency of many diseases can jump by orders of magnitude. Additionally, as I noted previously, many Westerners neglect that the coefficient of inbreeding of one's ancestors must also be a consideration, so highly inbreed clans who have engaged in successive generations of cousin marriage may exhibit more inbreeding depression than one would expect based on paper registrations of "first cousin" relationship because of random genetic drift within the extended family.2 I am skeptical than in the British Pakistani community reproductive skew is strong enough to purge the genetic load from the population...perhaps there is no reason to fear the demographic impact of Muslims?

1 - See Muller's Ratchet.

2 - If two first cousins mate, the coefficient of relatedness is 1/8. But, that assumes that the grandparents are unrelated. What if both of grandparents are first cousins? And the parents are first cousins? I leave it to you to draw the pedigree.