PDA

View Full Version : Effects of Migration and Assortative Mating on Admixture and IBD Analyses.



Angela
04-06-16, 02:46
The actual name of the paper is:
"The Effects of Migration and Assortative Mating on Admixture Linkage Disequilibrium, Noah Zaitlen et al"

It’s a pre-print that can be found at:
http://biorxiv.org/content/biorxiv/early/2016/05/31/056168.full.pdf

I was alerted to it by Razib Khan.
http://www.unz.com/gnxp/details-on-the-back-of-the-envelope/?utm_source=rss&utm_medium=rss&utm_campaign=details-on-the-back-of-the-envelope&utm_source=twitterfeed&utm_medium=twitter

As he says, "linkage disequilibrium based estimates of time since admixture often seem to give a relatively low figure in terms of generations. When it comes to non-human organisms in the field one doesn’t know always the demographic history, but with humans we have better records, and I too have noticed that the dates seem extremely skewed toward the recent. Part of this is accounted for by the dodge that these estimates are often sensitive only to the last pulse of admixture, but even then...

Exactly right…even then. This is the problem I have noticed in particular with both of the papers on recent admixture in Europe by the Hellenthal group. The dates they give don’t correlate very well with any documented large population movements.

The authors of the paper maintain that the problem is because " Statistical models in medical and population genetics typically assume that individuals assort randomly in a population. While this simplifies model complexity, it contradicts an increasing body of evidence of non-random mating in human populations. In this work we examine the effects of ancestry-assortative mating on the linkage disequilibrium between local ancestry tracks of individuals in an admixed population. To accomplish this, we develop an extension to the Wright-Fisher model that allows for ancestry based assortative mating. We show that ancestry-assortment perturbs the distribution of local ancestry linkage disequilibrium (LAD) and the variance of ancestry in a population as a function of the number of generations since admixture. This assortment effect can induce errors in demographic inference of admixed populations when methods assume random mating...We observe that LAD depends on the correlation of global ancestry of couples in each generation, the migration rate of each of the ancestral populations, the initial proportions of ancestral populations, and the number of generations since admixture. "

Just so we're clear, this is assortative mating:
http://anthro.palomar.edu/synthetic/synth_8.htm

The population the authors use to test their theories is the African American one, and about their results they go on to state that, " Our estimate of 15 generations since admixture in African Americans is larger than previous estimates[8, 9] and it fits considerably better the known history of African Americans[13]. This suggests that taking assortative mating into account may in some cases be critical in order to obtain the correct demographic history or other population parameters.”

A final comment from Razib Khan: “ The intuition isn’t that difficult. Assortative mating in this case often means that within population there are going to be correlations of segments of genomic ancestry which are the patterns you are tracking to infer backward to the initial admixture…If you see more dense local ancestry tracts in individuals because of positive assortative mating, you may confuse it in your model for very recent admixture…Assortative mating within a population may lead to higher heritable transmission of a trait across generations than you might expect. Many of the model based clustering algorithms which generate the bar plots now ubiquitous in admixture analyses assume Hardy-Weinberg Equilibrium in one’s populations.”

It also, as he and the authors make clear, affects IBD analysis.

If I understand this correctly, assortative mating is present where first cousin marriages are the norm, and also where more generally the population is more isolated because of religion, but also because of political divisions, terrain etc. Given those kinds of situations, in certain regions, certain countries even, dense local ancestry might confound admixture type analyses and even IBD analysis.

It’s something to keep in mind as we move forward.

LeBrok
04-06-16, 04:17
Did they discover that most people behave in conservative way? Wow! They should probably dive into mixed villages of india and see how farmers mix in context of arranged marriages, different religions and casts. This should pretty much set their algorithm right to cover last 10 thousand years.

Angela
04-06-16, 17:49
Did they discover that most people behave in conservative way? Wow! They should probably dive into mixed villages of india and see how farmers mix in context of arranged marriages, different religions and casts. This should pretty much set their algorithm right to cover last 10 thousand years.

It might change some of our understanding of the dating of admixture if these new mathematical methods are used.

As to assortative mating, researchers have known for years that in a modern context tall, thin people tend to marry tall, thin people, there tends to be a correlation in IQ between spouses, "attractiveness", as that is defined in the particular culture, class, educational level, all those things. It's not universally true, of course, as we know from observation. A fat, ugly man, if he's rich enough, can marry a beautiful, thin woman, sometimes the "Prince", marries Cinderella and raises her to his station, there have always been some unsanctioned relationships between upper class, elite males and lower class females, but the researchers are dealing in statistical probabilities. In terms of admixed populations in the U.S., for example, light skinned blacks or Puerto Ricans, absent some confounding circumstance, marry light skinned blacks. These are circumstances where some degree of choice is involved.

In the past, the choices weren't made directly by the people involved. Even in my mother's day, parents had a great deal of say in choices of mates, and so you really only married people of the same social station. Even among the peasantry, families sought out men with a good piece of land or tenancy, someone who would be a good provider. I don't know how it was in other places, but often it was the mother who took charge of the match making, looking for likely candidates often among her own family and acquaintance, although in northern Italy first cousin marriages were frowned upon. However, given how little people moved from, say, the 1400s to the 20th century in some rural areas, probably everybody in the vicinity was at least a fourth cousin of everybody else, even if they didn't know it. It was just a product of their circumstances, although there was an element of suspicion of outsiders involved. An old Italian rhyme is: Moglie e buoi dei paesi tuoi, or wives and oxen from your own town. Not very flattering, but I guess the point is, don't trust what you don't know. :)

bicicleur
04-06-16, 18:38
that is how it should be :

https://youtu.be/TzZTG2TnWiw

Angela
04-06-16, 19:27
that is how it should be :

https://youtu.be/TzZTG2TnWiw

I love that, Bicicleur! So funny.

I don't doubt that's how a lot of fathers feel, and the wilder they were in their youth, the more likely probably that they feel that way! My husband was an absolute horror and is to this day about our daughter and her boyfriends. They're all terrified of him, as well they should be...

My dad was a total anti-cleric, so when I routinely was the girl cast as the Virgin Mary in religious pageants in May or at Christmas, or as a nun in various school plays, he used to get very upset that the nuns were trying to brainwash me. (In a way I suppose they were...) On the other hand, he was worse than a horror when I was dating. Not only the boys were terrified...I was terrified too.

Anyway, as to assortative mating, until relatively recently, as I said, in Italy parents exerted a lot of pressure on mate choice, even if it wasn't actual matchmaking, and it involved regional identity as well as education, class etc. A second cousin once removed of mine wanted to marry a boy whose family had relocated from southern Italy. The wrangling went on for I don't know how many years. When it became clear that she absolutely wouldn't marry anyone else, and she started showing signs of what we'd now consider anorexia, they relented. Meanwhile, I like him better than her or her parents, so there you go. Heck, had my father's father been alive when I married, I would have had problems of my own.

So, I totally get what they mean by assortative mating.

LeBrok
04-06-16, 19:52
It might change some of our understanding of the dating of admixture if these new mathematical methods are used.

As to assortative mating, researchers have known for years that in a modern context tall, thin people tend to marry tall, thin people, there tends to be a correlation in IQ between spouses, "attractiveness", as that is defined in the particular culture, class, educational level, all those things. It's not universally true, of course, as we know from observation. A fat, ugly man, if he's rich enough, can marry a beautiful, thin woman, sometimes the "Prince", marries Cinderella and raises her to his station, there have always been some unsanctioned relationships between upper class, elite males and lower class females, but the researchers are dealing in statistical probabilities. In terms of admixed populations in the U.S., for example, light skinned blacks or Puerto Ricans, absent some confounding circumstance, marry light skinned blacks. These are circumstances where some degree of choice is involved.Yes, though these are rather modern phenomena of liberal society. It should have been taken under consideration analyzing pre modern population.


In the past, the choices weren't made directly by the people involved. Even in my mother's day, parents had a great deal of say in choices of mates, and so you really only married people of the same social station. Even among the peasantry, families sought out men with a good piece of land or tenancy, someone who would be a good provider. I don't know how it was in other places, but often it was the mother who took charge of the match making, looking for likely candidates often among her own family and acquaintance, although in northern Italy first cousin marriages were frowned upon. However, given how little people moved from, say, the 1400s to the 20th century in some rural areas, probably everybody in the vicinity was at least a fourth cousin of everybody else, even if they didn't know it. It was just a product of their circumstances, although there was an element of suspicion of outsiders involved. An old Italian rhyme is: Moglie e buoi dei paesi tuoi, or wives and oxen from your own town. Not very flattering, but I guess the point is, don't trust what you don't know. :) Got ya. More like this for Millennia. Knowing this we can assume that it took 1-2 thousand years for Bronze Age invaders to fully mix with locals. It is hard to start mixing process. You might have two populations, two villages close by, speaking different languages, having different religions, traditions and even looking differently. At this stage there is not much mixing, apart from occasional romantic love and rape. After few hundred of years it gets a bit easier. By common historical events, landlord and trade, they develop one common language, religion and tradition and they start looking alike too. At this point they consider themselves as one nation. There is no barrier for intermixing and marriages except geographical one. Though it still takes some time for complete mix.
Summarizing, population mixing is very slow at the beginning, but very fast at the end.


Moglie e buoi dei paesi tuoi, or wives and oxen from your own town. That's the natural conservatism I'm talking about. :)