The genetics of marriage

Marriage is about procreation, as the Catholic Church carefully explains. We all have a eugenic interest in our children so it would not be too surprising that married couples might show some interesting genetic correlations. We already know about assortative mating for intelligence, but could there be more?

A recent study, 'Genomic Assortative Mating in Marriages in the United States' discusses

"genome-wide genotype data from the Framingham Heart Study (FHS; number of married couples = 989) and Health Retirement Survey (HRS; number of married couples = 3,474), this study investigates genomic assortative mating in human marriages."

The paper was published six months ago; their main result:

"Overall, our data suggest a degree of genomic assortative mating at the allelic level in married couples who were born in the first half of the 20th century in the United States. Apparently, this degree of genetic assortment averaged over the human genome is much smaller than the 0.20 Pearson had conjectured based on the observed correlations in height and arm span between husband and wife. As alluded earlier, certain genetic variants such as those underlying height are likely to be heavily assorted; however, the level of overall assortment in the genome seems much less.

However, a genomic correlation of 0.015–0.02 with married couples, estimated for the “positive” assorting SNPs in HRS, can represent an important genomic assortment for at least two reasons. A married-couple correlation may be compared with genetic relatedness among biological relatives. A genomic correlation of 0.015–0.02 is close to the average genomic correlation (0.0312) among second cousins (or the genomic correlation [0.0312] of an individual with his grandfather’s grandfather). While an individual passively and unselectively inherits half of his or her genes from each of the two parents, married individuals consciously or unconsciously assort on genes that play a strategic role in their reproductive marriages."

A genomic correlation  of 0.03125 = 1/32. According to Richard Dawkins,

"For relationships as distant as third cousin, 2 x (1/2)⁸ = 1/128, we are getting down near the baseline probability that a particular gene possessed by [an individual] will be shared by any random individual taken from the population."

Dawkins is referring to an ethnically-homogeneous population, as in the phrase 'old English stock'. The GWAS research above proposes that, on average, husbands and wives are more closely related by a factor of four than random members of the population. Assortative mating indeed.

Expect much more on this down the line.

Frank Salter and the genetics of immigration

I've been thinking about Frank Salter, who wrote "On Genetic Interests: Family, Ethnicity, and Humanity in an Age of Mass Migration" back in 2003 (I reviewed this last year). From being a taboo subject a while ago, immigration policy is now plastered across the front pages and camps out in op-ed pieces. When people immigrate to our country we gain their phenotypes in the short run => whatever skills they have to offer. If they stay and reproduce, we also get their alleles. Should we care?

Salter wants to argue that co-ethnics (people of the same race as yourself) are to be considered kin - part of your extended family. The figure I heard was that a random pair of co-ethnics are related at about 1/128 (apparently this is third cousin relatedness). Based on Hamilton's theory of kin selection, Salter suggests that this should lead us to favour co-ethnics over those more distantly related ('other races').*

There is more.

In a review of Salter's book (p.10)  this is what Peter Gray (Department of Psychology, Boston College) had to say:

"Salter uses evolutionary theory not to explain behavior but to prescribe it. He clearly equates genetic interest with human good. In summing up his argument, in the “Afterword,” Salter writes, “My primary aim has not been to explain human behavior, but rather to offer social and political theory about what individuals should do if they want to behave adaptively” (p. 325, Salter’s italics). "

"What we should do, according to Salter, is discriminate by race. We should do this because it is in our genetic interest to do so. Races differ genetically, and we share more genes with people of our own race than with those of different races, so it is in our genetic interest to favor our own race. To Salter, unlike to the rest of us who use evolutionary theory, genetic interest is not just the metaphorical “interest” of the gene, it is the real interest of the person. Salter writes, “Genetic interest residing in a population is a public good that belongs, as it were, to its individual members” (p. 43)."
...
"I’m sure that Salter is not in favor of rape. But his logic states clearly that it would be immoral to pass laws against rape if that behavior is in someone’s genetic interest. When you think of it, most of our laws — laws against rape, murder, stealing, exploitation, slavery, and the like — are interfering with someone’s ability to pursue their genetic interest."

Yes, it's usually pro-social doves banding together to protect themselves against hawks. Insofar as doves tend to achieve more civilization-wise, this has to be judged a good thing. And here I think is the crux of the argument.  A civilised nation is better for all its members in terms of future survival (provided people actually get on and reproduce in the first place!). Looking around the world, and contra-Salter, a rational policy of genetic interests should be trying to boost alleles for intelligence, conscientiousness and general pro-sociality within our own societies. Strangely, that's exactly what immigration policies based on encouraging talent actually do.

---

* There is very little evidence of any built-in instinct which identifies somewhat genetically close kin. It seems very unlikely that we have spontaneous emotional connections to a random co-ethnic living in a town miles away. Even in a tight family environment, emotional bonding is apparently as much about growing up in close proximity as about cues from close genetic relatedness.

Your illustrious medieval ancestors?

I caught a "Who Do You Think You Are" programme featuring Celia Imre the other day. She sought the origins of her son's radical politics and her own feminist feistiness in her ancestors. Here's what the program found for her:

"Imrie’s particular wish to uncover an ancestor to inspire her politically inclined son, Angus, unearthed a corker in her “eight-times grandfather” William, Lord Russell, son of the Earl of Bedford in the time of King Charles II. A leading Whig politician who truly had the courage of his convictions, Russell was such an intransigent defender of Protestantism and lover of constitutional liberty that he was accused of plotting to kill the King, and promptly beheaded.

"At Woburn Abbey, the Russell family seat, Imrie set out on the still more dramatic trail of Frances Howard, grandmother of the aforementioned William. A victim of one appalling dynastic marriage, and caught up in vicious courtly intrigues while trying to secure happiness in a second, she was packed off to the Tower of London with her new husband, accused of murder. Frances was eventually pardoned but history was not so forgiving."

So she had found her 'good genes' then? Not so fast, here's Richard Dawkins:

"For relationships as distant as third cousin, 2 x (1/2)⁸ = 1/128, we are getting down near the baseline probability that a particular gene possessed by [an individual] will be shared by any random individual taken from the population."

Celia Imre's “eight-times grandfather” William, Lord Russell, is nine generations separate from her and shares a relatedness of  1/512 = (2^-9). That distant relatedness could be greater if her lineage includes a degree of inbreeding - but it's unlikely to be more than 1/128 - Dawkins' rough figure for the genetic relationships of any two random ethnic English people.

So Celia Imre's traits for general lefty feistiness are certainly in her genes, but not through the good offices of those specific medieval ancestors.

More generally, you get a dilution down to 1/128 in seven generations. At four generations per century, you may assume that any specific traits of a specific ancestor more than 175 years ago (i.e. before c. 1839) have since been diluted out by Mendelian segregation and recombination. (Also more from 23andMe's reseaches here).

Genetic immortality? fuhgeddaboudit..