Ars technica covered this in a well written, well explained article [1] yesterday. Worth a read for folks who don't necessarily have the right background to understand the significance and / or importance.
To quote one of the many good excerpts:
"They looked at siblings-in-law, and first-cousins-in-law, and then further afield, at relationships like “the sibling of a sibling’s spouse" (your brother’s wife’s sister) and "the spouse of a spouse’s sibling” (your husband’s sister’s husband). Even at these distant relationships, lifespans were correlated—if your spouse’s sibling’s spouse lived to a ripe old age, that means you’re a bit more likely to do the same."
Thanks; we changed the link to that from http://www.genetics.org/content/210/3/1109. We try to stick with the best popular article when a technical paper is outside most HN readers' specialty. In that case it's standard to link to the paper from the comments, for anyone who wants to go further.
Never heard the term assortative mating. From Wikipedia [1]:
Assortative mating is a mating pattern and a form of sexual selection in which individuals with similar phenotypes mate with one another more frequently than would be expected under a random mating pattern...
Is there any evidence people prefer assortative mating given enough racial heterogeneity (i.e. people who look like you aren't the only choice)? I'm personally mostly sexually attracted to people who look nothing like me, and anecdotally I think most people I know are like this too? In my experience people are appealed to assortative mating since they don't live with people from other racial/ethnic backgrounds so there are cultural barriers (language, prejudice, differences in social class, education, slang etc). But in multiracial areas (e.g. the Bay Area), it seems to me that people do prefer mating with individuals from other races. Am I wrong? Any papers investigating this question?
In modern usage, assortative mating is less about race and more about socioeconomic group. Would you marry a woman (or man) significantly poorer than you, from a poor family, etc. I mean, my wife and I are by no means the same race, but we're absolutely the same in socioeconomic strata, which means our children will be resultingly advantaged.
It doesn't even necessarily have to be a matter of preference- in general, throughout life you'll most likely meet more people of your own socioeconomic group, and less people of other groups. From there, simple probability predicts assortative pairing.
I don't think we're there yet. It will be interesting to see how this changes overtime. I'm curious how to works in largely homogeneous countries like Sweden.
Sweden is no longer ethnically homogeneous; about 17% of the population was born outside of the country. While some of those are from nearby EU countries (Finland and Poland), most are refugees from the Middle East (esp. Syria and Iraq), Africa, and the Balkans resettled under Sweden's generous asylum policies.
> But in multiracial areas (e.g. the Bay Area), it seems to me that people do prefer mating with individuals from other races. Am I wrong?
In my experience it has nothing to do with race and everything to do with culture and socioeconomic status. Odds are there's more mixed marriages in SF because it attracts those with a certain socioeconomic status (given how expensive it is) and people assimilated into the culture.
It's much easier to make a life with someone who has similar desires, beliefs, food preferences, holiday celebrations, etc..., versus having to constantly compromise and bicker over every tiny, normally insignificant life choice.
The interracial relationships of the Bay Area are purely a result of the numbers. If you’re in the Google employee socioeconomic class, your social environment is about 30% Indian, 20% Asian, 20% American born white, and 20% other. The chance that you’ll marry someone of a different race, with no preference, is >70%.
However relationships are very homogenous on social class, education, values, politics, and physical attractiveness.
Can’t comment on the Bay Area, but I’m familiar with Los Angeles and Portland. In my experience, there isn’t actually much intermarriage between races. And the intermarriage that does take place tends to follow very specific patterns: Asian woman + white man, white woman + black man, etc.
Sexual attraction is complicated. There are cultural influences, especially when it comes to races, at specific time periods some races are seen as cool to date, some are fetishisized, while others are avoided:
Sorry but how is this comment helpful? I was clearly talking about phenotypical resemblance e.g. hair, eye, skin color, height, face shape etc... Gender or sex is irrelevant. Obviously heterosexual men are attracted to women but the question is given enough racial diversity are white heterosexual men mostly attracted to white women or not.
I think you're over estimating how much those sorts of individual physical things impact people's choice of long term spouse. Nobody (or at least very few people) is putting off marrying someone they otherwise would marry because they don't have their preferred hair/eye/skin color. I also think you're underestimating how much those physical features are mapped to personality/cultural features. People want to date and marry people who they have a lot in common with. Once people have found someone they're very compatible with on a personality level I think they tend to be less choosy about the physical stuff.
Yes, I was being snarky initially. I was trying to imply that being the opposite sex is enough difference to satisfy most people's minimum requirement.
I would guess you're correct (I'd still be interested to see data) for long-term relationships but I wouldn't agree with you on short-term attraction i.e. seeing people on the street and finding them attractive. This is purely physical (even non-physical aspects e.g. associating someone's appearance with a certain type of personality have a physical source) so it is possible hair/eye/skin color can matter, so whether people prefer people looking similar/dissimilar to them is still an open question, I think.
EDIT: But of course, since we're talking about reproduction, you have a good point I think.
Clothing style is also physical, but heavily influenced by income and lifestyle.
Proximity is another factor. When rock climbing you meet people healthy enough to be rock climbing, college etc. In a less positive context your likely to meet someone at an AA meeting with a similar genetic propensity for alcoholism.
From a naïve perspective if people choose to mate along longevity phenotypes, does that same mechanism affect people choosing mates whose DNA expresses genetic defects as well? Or does this go contra the understanding?
The article suggests that these longevity phenotypes are not even inherited. For example, one trait that's assortatively selected for is geographic location. Another is socioeconomic status. Neither of these are encoded in genes, but they correlate with traits that may be encoded in genes. Silicon Valley draws a certain type of person, Flint Michigan draws a certain type of person, and a tech worker in Silicon Valley is far more likely to marry another tech worker in Silicon Valley than an unemployed factory worker in Flint Michigan. However, that double-tech-worker couple will have the money to tear down the old house with lead paint and replace it with new construction, and they will live in an area with the tax base to pipe in fresh water from Hetch Hetchy rather than drink polluted well water from the ground. As a result, their kids may have decades longer longevity than kids in Flint.
With this model, it's not so much that people are choosing to mate with people with genetic defects. It's more that that's who they end up with, because people tend to mate with others who have roughly equal resources to give their children a shot at better longevity. Some of that can be because of circumstance, and some of it could be that their particular genetic make-up is no longer adapted to the modern world. The article is suggesting that the portion due to genetic makeup is likely less than has previously been estimated, because people choose mates who are in the same general health & fitness bucket, and so any differences in genetic fitness gets magnified by your partner's (and other community members') fitness as well.
Not interested at all in derailing good conversation here but let's give a shoutout to the authors for bucking up the $1500 to Genetics [1] to make this article Open Access. It's fortunate that this work was funded by Calico and Ancestry, because otherwise that likely wouldn't have happened.
First, all the authors work for Calico or Ancestry. One could view this as a ‘published whitepaper’, in which case they obviously want to have their paper available broadly.
> because otherwise that likely wouldn't have happened
This is not true - there are open access fees at many publishers, and grant-funded labs regularly choose to pay these fees.
They can afford it easily: "J. Graham Ruby, the lead author on the paper, works for Calico Life Sciences, a research and development company funded by Alphabet."
It was interesting to see them mention shared environments between closely related persons, which was something I hadn't thought about before as another sort of "heritability".
(Where "shared environments" could be things like multi-generational poverty or a dangerous family profession or living somewhere with a long history of excess carcinogenic pollution. You get it from your parent and share it with your siblings, but it's not genetic.)
This is pretty much the central problem in distinguishing genetic and environmental factors: those that are most closely genetically related also tend to share much of their environment (hence the utility of things like twin and adoption studies).
Well, being precise here (sorry, biologist talking)––the whole point of this article is that the "shared environments between closely related persons" is precisely not heritability.
The issue at stake here is that genetics play a far lesser role in determining longevity than other non-genetic inheritable factors (like the "shared environment" examples you give); e.g. the heritability of longevity is considerably lower than previously thought.
Firstly I will admit that the article was a bit much for me to process as a non-biologist, and that it didn't register that that was the entire point of the article until you said so :-)
But to be a little defensive, yeah, I know shared environmental factors aren't really "heritable", I'm just saying that it honestly hadn't occurred to me before reading this article that they would be confused for heritable genetic factors, even though it does kind of seem obvious once it's pointed out.
The heritability of traits increases as the environment improves. For example historically height wasn't very heritable, it depended mostly on environment, which is to say nutrition. Now in countries where adequate nutrition is widely available, height is much more heritable.
Could the same effect apply to longevity? I don't really know, because I don't even know what environmental factors improve longevity, except for the grossly obvious ones.
>Where "shared environments" could be things like multi-generational poverty or a dangerous family profession or living somewhere with a long history of excess carcinogenic pollution. You get it from your parent and share it with your siblings, but it's not genetic.)
This is the key here. If you grew up poor with parents that weren't kind to their bodies and died young you'll probably do the same, marry someone who does the same and have kids who do the same.
> If you grew up poor with parents that weren't kind to their bodies and died young you'll probably do the same, marry someone who does the same and have kids who do the same.
Considering that 90% of the population used to engage in manual labor on farms to provide for themselves, if your lineage never made the transition to non-manual labor like most of the developed world, genetics are probably a more likely explanation.
Is it likely that the same factors have led to inflated numbers for the heritability of other attributes as well? Or is this something specific to longevity?
Yes. Particularly because we "assort" by phenotype (appearance.) We are more attracted to those who look like us, even if illness has affected our appearance. (For example a genetic collagen disorder.) What may be one of two similar illnesses being inherited (most of the time) is mistaken for one illness being dominantly inherited. for example.
In case, like myself, you come looking for the punchline: "...concluded that the true heritability of human longevity for birth cohorts across the 1800s and early 1900s was well below 10%"
I think you are thoroughly confused. They are not saying assortative mating changes heritability, they are saying assortative mating is messing with our techniques to estimate heritability, and so the true number is less than we thought, while still stable. We need better statistical techniques to infer heritability in the face of this problem.
Of course it changes the genetics. In fact it's expected with traits like this. Observing it doesn't make estimating the narrow-sense heritability any harder (or easier).
To quote one of the many good excerpts:
"They looked at siblings-in-law, and first-cousins-in-law, and then further afield, at relationships like “the sibling of a sibling’s spouse" (your brother’s wife’s sister) and "the spouse of a spouse’s sibling” (your husband’s sister’s husband). Even at these distant relationships, lifespans were correlated—if your spouse’s sibling’s spouse lived to a ripe old age, that means you’re a bit more likely to do the same."
Worth reading, IMHO.
[1] https://arstechnica.com/science/2018/11/genetics-play-less-o...