runs of homozygosity and inbreeding (and outbreeding)

here’s a really neat chart! (click on image for LARGER view. should open in new tab/window.):

what does it mean? well…

some very clever researchers/geneticists took a look for “runs of homozygosity” (roh) in the genomes of the individuals in the human genome diversity project (hgdp) — that’s 1043 individuals from 51 different populations. “runs of homozygosity” are stretches in the genome where identical dna was inherited from each parent. if you inbreed, you’re gonna have a greater number of longer runs of homozygosity in your genome than if you don’t.

apart from being just plain fun, sex shuffles up genomes from one generation to the next (presumably for some good reason or another). if you were to clone yourself, your descendants would have (pretty much) the same exact genome as you. if you were to mate with your mother or your sister (i know — ewwww!), your descendants would have different genomes from you, but they’d have lots of roh in their genomes ’cause their dna came from you and someone with whom you share a lot dna in common. the farther out you mate, the less homozygosity there’s likely to be.

you might also have lots of roh in your genome if you come from a population that has little genetic diversity — ’cause maybe your ancestors went through some sort of bottleneck or something.

inbreeding with close relatives — like marrying your first- or second-cousins (consanguineous matings) — leads to long roh since you share so much of your dna with your closest family members. endogamous mating — just mating within your population but not your close cousins — also leads to roh, but not ones as long as mating with your close relatives. you share dna with others in your population (say your clan or your ethnic group), but not so much of exactly the same dna or genes in certain stretches as with your closer relatives. a population will little genetic diversity, but that does not inbreed, will have lots of short roh — they share a lot of stretches of dna in common, but all of the outbreeding shuffles up the genomes within the population.

so that’s:

long roh = inbreeding, probably consanguineous (first-/second-cousin matings)
medium roh = endogmaous mating within a population
short roh = little genetic diversity in the population probably from an event like a population bottleneck

i’m oversimplifying, but that’s the gist of it.

so what did the researchers find when they looked at the 51 populations in the hgdp (see chart)?

– LOTS of short roh (1-2 Mb) within populations from oceania and central/south america, probably because those populations went through bottlenecks. the people from oceania have low amounts of long roh (>16 Mb), which means that they don’t inbreed closely much. however, the people from central/south america have the highest amount of long roh of all the groups, so that’s means they must inbreed closely a LOT.

– central/south asians, west asians, east asians, europeans and africans don’t have huge amounts of short roh — at least not compared to the folks in oceania and the americas. no big bottlenecks there. and africans, in fact, have the fewest short roh.

– central/south asians and west asians have pretty high amounts of roh in all of the middle ranges and the highest long roh after the native american populations. this indicates significant amounts of endogamy and close relative marriages (but we already knew that).

– the groups with the lowest amounts of long roh are the europeans, africans and east asians — in that order. in other words, it appears as though, of these three groups, africans and europeans inbreed more closely (first- or second-cousin marriage, say) than east asians.

if you’ve been following along, you know that’s not what hbd chick expected. i thought that east asians would’ve had more short roh than europeans ’cause they have a fairly recent history of close marriages. hmmmm….

i checked to see which populations of europeans are included in the hdgp (you can find a list in the article’s supplemental material here [opens pdf]) and they are: adygeis, basques, french folks, italians, orcadians, russians, sardinians and tuscans. apart from the french and the tuscans, all of these groups have recent (or current) histories of consanguineous or endogamous mating practices (see Inbreeding in Europe series below in left-hand column for more details), so they are not a fully representative sample of europeans. unfortunately, “core” europe, which contains the most outbred populations in europe, is not included in the hgdp and, therefore, not in this study.

*sigh*

still — this is interesting stuff! genetics. cool! i’m going to post more about this ’cause, for one thing, it should be possible to drill down further into these populations to compare them more specifically (there are some data available in the supplemental materials). so, more anon…!

thanks to prof. harpending for pointing out this article! (^_^)

*update 08/20: see also runs of homozygosity again

(note: comments do not require an email. lesson one.)

Advertisements

12 Comments

  1. Oh yeah, I remember this from Razib’s blog!

    http://blogs.discovermagazine.com/gnxp/2010/11/homozygosity-runs-in-the-family-or-not/

    with a lot of exposition…

    Note Razib’s post covers the difference between RoH in farmer and HG populations.

    (Think I assumed you knew all about it already for some reason! Probably your depth of knowledge everywhere else)

    I think there is other stuff relating to differences in RoH between urban and rural populations and between pastoral and farmer cultures.

    i checked to see which populations of europeans are included in the hdgp

    Good point and the patterns in the supplemental data do support that, though not sure I’d regard the East Asian populations included as likely to be the most outbreeding enthusiastic of East Asians either though.

    Reply

  2. With genome sequencing so cheap now (approaching $1000 soon) you would think this sample set could be expanded. What is the outlook?

    Reply

  3. @matt – “Think I assumed you knew all about it already for some reason!”

    i recalled seeing dienekes’ post on it after i was googling around about the paper up yesterday — but, no, otherwise i didn’t remember about this paper. (*facepalm* (^_^) ) there’s another one about orkney islanders, too — will have to take a look at that one of these days.

    @matt – “not sure I’d regard the East Asian populations included as likely to be the most outbreeding enthusiastic of East Asians either though.”

    no, absolutely. too bad the hgdp didn’t just have more straight-up han chinese, not to mention more central/nw europeans. i get the idea that they were trying to get their hands on some unique genomes, but still. oh well.

    Reply

  4. @luke – “What is the outlook?”

    i dunno! but as soon as the price does drop to $1000 (retail), you KNOW who’ll be asking for a full genome sequencing for christmas, right? (good stocking-stuffer!) (~_^)

    Reply

  5. This analysis is so interesting. Are there other similar studies going on as we speak? Or additional commentary on this particular one?

    The masses of data are so immense — a couple of billion base pairs on 24 chromosomes — that the mathematical challenge must be immense. What tools do they use?

    It occurs to me that Google search might come in handy here. After all the analytic challenge is similar: identify idential strings (words) on very long of documents (chromosomes). With all the different search options and operators Google supplies, well, it just seems a natural fit. And just think, you would be able to tap into all those gigantic server farms, which are so big (I hear) that they are often located close to nuclear power plants.

    I’d be surprised if this is an original idea.

    Reply

  6. @luke – “Are there other similar studies going on as we speak? Or additional commentary on this particular one?”

    i saw another reference to a similar article about orkney islanders, but i haven’t done a full search for other similar articles, so i dunno. definitely have to start rummaging around through google scholar/pubmed. (^_^)

    @luke – “What tools do they use?”

    they used some software called PLINK from hah-vahd — it’s free & open source. now, if i just had a few genomes on hand…. (~_^) (i wonder if you can use PLINK on 23andMe sort-of data, i.e. not full genomes?)

    Reply

  7. @luke – “Here’s a quote from a study: ‘Interestingly, if double-first cousin marriages are preferred to avuncular marriages in Central South Asia and the Middle East, the contrary is found in the Americas.'”

    you are AWESOME! (^_^)

    look at this! cool. and that’s just the first of the links i clicked on following the link from nature.

    i gotta go get some coffee so i can absorb all of this nummy data! (^_^)

    Reply

  8. http://www.nature.com/ejhg/journal/v19/n5/abs/ejhg2010205a.html

    “In these regions, first cousin (1C) marriages are the most frequent, but we also observed marriages between double first cousins (2 × 1C) and between avuncular (AV) pairs. Interestingly, if 2 × 1C marriages are preferred to AV marriages in Central South Asia and the Middle East, the contrary is found in the Americas.”

    I’m quite shocked at that. Even if the church couldn’t successfully impose the cousin ban you’d have thought they’d have the influence to stop uncle-niece – unless those groups are all hunter-gatherers?

    Reply

  9. @g.w. – “unless those groups are all hunter-gatherers…”

    a couple of them definitely are. here’s the list of the central/south american groups:

    Colombian
    Karitiana
    Maya
    Pima
    Surui

    the karitiana live somewhere in the rainforest in brazil. so do the surui. not sure who the colombians are, or if the pima are the ones in arizona or not (arizona’s not in central or south america, so that’s why i wonder). and the maya are the maya, i guess.

    Reply

  10. Hm, you wrote something about the study not including a diverse mix of Europeans, basically making the data a bit skewed, but also you mention being surprised about the data on East Asians because the Chinese have a history of close marriage or something. Is it just me or is that a bit of a double standard? Not really double, I mean it could apply to the other groups as well. But back to the point, remember, China may be huge but it’s not the only East Asian country. And anyway in China there are a bunch of different ethnic groups. Oh and one more thing, doesn’t roh length over time increase from consanguineous breeding (though cousin marriage is banned in China), but on a curve ’cause of recombination and tiny amounts of mutation? Sorry if I ramble.

    Reply

  11. @k – “Is it just me or is that a bit of a double standard? Not really double, I mean it could apply to the other groups as well.”

    yeah, you’re absolutely right. i’ve written a bunch more posts (which are not coherently put together, i regret to say — something else to put on the to-do list — but here’s a good one to start with) about some of the problems i can see with using the hgdp samples for these sorts of studies. i mean, i think they’re really awesome, of course, but there are problems with the samples — and they apply to all the populations (in different ways, though, perhaps).

    the samples from “china” of the han chinese were actually taken from chinese folks living in san francisco, so who knows where in china any of them came from and/or what their parents’ marriage patterns looked like (were they married to people from “back home” or had they married someone they’d met in the u.s. — that might affect how “inbred” or “outbred” the kids who were sampled appeared). there are prolly other potential problems with the chinese data that i haven’t thought of. there were, however, separate samples taken of minority groups living in china, so i don’t think there’s been any confusion there.

    @k – “doesn’t roh length over time increase from consanguineous breeding (though cousin marriage is banned in China)…”

    yes, that’s right. longer roh=close mating practices. btw, cousin marriage was only banned in china in 1980/81. it’s a very new thing there.

    @k – “Sorry if I ramble.”

    rambling works. (^_^)

    Reply

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s