search this blog

Friday, January 9, 2015

Very rare SNPs reveal distant genealogical ties between the UK and China


As far as I can see, the paper doesn't say anything about the direction of gene flow. It might have been either way, I suppose, considering how active Great Britain was in the Far East during colonial times. Two of the British individuals mentioned in the abstract below are from Cornwall and one is from Kent.

Abstract: Nucleotide sequence differences on the whole-genome scale have been computed for 1092 people from 14 populations publicly available by the 1000 Genomes Project. Total number of differences in genetic variants between 96,464 human pairs has been calculated. The distributions of these differences for individuals within European, Asian or African origin were characterized by narrow unimodal peaks with mean values of 3.8, 3.5, and 5.1 million respectively and standard deviations of 0.1-0.03 million. The total numbers of genomic differences between pairs of all known relatives were found to be significantly lower than their respective population means and in reverse proportion to the distance of their consanguinity. By counting the total number of genomic differences it is possible to infer familial relations for people that share down to 6% of common loci identical-by-descent. Detection of familial relations can be radically improved when only very rare genetic variants are taken into account. Counting of total number of shared very rare SNPs from whole-genome sequences allows establishing distant familial relations for persons with 8th and 9th degree of relationship. Using this analysis we predicted 271 distant familial pair-wise relations among 1092 individuals that have not been declared by 1000 Genomes Project. Particularly, among 89 British and 97 Chinese individuals we found three British-Chinese pairs with distant genetic relationships. Individuals from these pairs share identical by descent DNA fragments that represent 0.001%, 0.004%, and 0.01% of their genomes. With affordable whole-genome sequencing techniques, very rare SNPs should become important genetic markers for familial relationships and population stratification.

Citation...

Al-Khudhair et al., Inference Of Distant Genetic Relations In Humans Using “1000 Genomes”, Genome Biol Evol (2015), doi: 10.1093/gbe/evv003

41 comments:

Chad Rohlfsen said...

Ahh... Looks like the same technique for showing Brits are much closer to Iron Age Celts, rather than Anglo-Saxons.

Davidski said...

Were there any Han Chinese living in Cornwall and Kent during the 19th century?

Chad Rohlfsen said...

British brought home Asian and Indian wives.

Chad Rohlfsen said...

There was a prime minister with Indian ancestry... The name escapes me...

Chad Rohlfsen said...

Lord Liverpool

Chad Rohlfsen said...

Just Wiki, but something...

http://en.m.wikipedia.org/wiki/History_of_Chinese_immigration_to_the_United_Kingdom

Srkz said...

8th or 9th degree of relationship cannot be predicted precisely. It can be 8th degree, or 18th degree, or even 88th degree. My opinion is this is not XIX century ancestry, but something (probably) much more ancient.

Chris Davies said...

Chinese seamen living in British port cities during the 19th century married English women, especially in Liverpool and East London.

gerschwint said...

There's a lot of barking up the wrong tree. The size of the segments are the clue: the connection is ancient. Unless I misread, the paper looked especially at the British-Chinese relationship, so if they'd looked at any two populations, they may well have found a similar phenomenon. The paper is not unveiling any special relationship, as far as I can tell, but rather tentatively making the point that even populations toward the extremes of variation can have some tiny degree of ancestry in common.

Davidski said...

They looked at all of the 1000 Genomes samples, which also include Finns and Tuscans.

gerschwint said...

Correct, and they did find a similar phenomenon, with LWK sharing with at least one Japanese and LWK sharing with at least one Finn matching or exceeding two instances of the GBR-CHB sharing. I'm working my way through the paper. Hopefully, there'll be some comprehensive coverage of the tyes and degree of sharing between all members of all populations covered.

Chris Davies said...

Plenty of ex-Londoners have settled in both Kent and Cornwall over time. Kent is in close proximity to London and Cornwall is popular as a holiday destination or for retirement, second homes, etc. Perhaps the people from Kent and Cornwall have ancestors from East End of London.

gerschwint said...

Davidski,

They looked at all of the 1000 Genomes samples

IBS is missing.

Chris Davies,

Plenty of ex-Londoners have settled in both Kent and Cornwall over time. Kent is in close proximity to London and Cornwall is popular as a holiday destination or for retirement, second homes, etc.

All Kent and Cornwall samples have four grandparents born in Kent and Cornwall respectively. Holiday destinations and second homes are modern phenomena.

Perhaps the people from Kent and Cornwall have ancestors from East End of London

Percentages of IBD are calculated at 0.0011%, 0.0043% and 0.012%. To get to even the highest percentage, you have to go back around 13 generations. If I go back 13 generations, I find my ancestors living in the 1500s, and I'm only 29. British-Chinese relations didn't begin in earnest until the early 19th century.

gerschwint said...

Although, of course, I should add that the common ancestor may be further back than the actual admixture event.

gerschwint said...

Apologies for triple post, but I need to correct myself: IBS was included, although there is little said about them.

Chris Davies said...

Maybe it is somehow through Portuguese admixture. HLA haplotype A2-B46-DR9 is a common marker in Chinese populations [generally 2.5% - 5.0%] but found at <0.01% in nearly all of Europe, except for Portugal [0.09% in 1176 Portuguese].

gerschwint said...

Chris Davies,

Maybe it is somehow through Portuguese admixture. HLA haplotype A2-B46-DR9 is a common marker in Chinese populations [generally 2.5% - 5.0%] but found at <0.01% in nearly all of Europe, except for Portugal [0.09% in 1176 Portuguese].

Genes can randomly pop up anywhere in the (connected) world. Asian EDAR can be found in West Africans (probably by way of East Africa, where EDAR occurs in reasonable frequencies). To draw any conclusions from this paper, we'd have to see global populations rigorously tested against one another in order to create a reasonable context for the interpretation of this data. That was not the case. 44,278 international pairs were studied, even though all of the GBR samples alone, tested against all members of other world populations (even excluding New World populations) comes to around 31,056. The testing was not exhaustive. However, the authors were exhaustive in testing one particular relation: All GBR against all CHB individuals.

Other interesting findings, such as LWK sharng with FIN & JPT, are alluded to briefly and then left unexplored, because the paper is essentially about their ability to detect relations, not to study type or degree of admixture in any population. As such, only minimal data pertaining to the latter is available.

Davidski said...

From the paper...

"Due to the differences in population structures, we observed significant variations in the number of shared vrGVs between the first and the second order relatives in different populations (see Figure 4 and Supplementary Table S3). First order relatives (shared 50% common genetic materials) have 28,000-46,000 shared vrGVs in Africa and only about 14,000-20,000 in Asia. This number is proportionally decreased for the second order relatives and further on."

It looks like sharing rare SNPs with east Africans isn't as much of a big deal as sharing rare SNPs with East Asians, because the latter have fewer of them.

gerschwint said...

Both FIN and JPT have an excess of sharing with LWK relative to similar populations:

http://dienekes.blogspot.co.uk/2012/10/a-thousand-and-ninety-two-genomes.html

Interestingly, CHB shares more with all Europeans than does CHS. Maybe the admixture is from Europe to CHB?

As for the present paper, only 44,278 of 191,384 possible global pairings were studied, and the info from this isn't reported (AFAIK), at least not extensively, so not much can really be drawn from it about population movements. GBR-CHB relations are the only ones exhaustively studied, from what I can make of the paper, and there's really not much to compare this with (except that one of the relationships had the highest inter-continental IBD sharing).

Davidski said...

Some of the CHB have minor ANE admixture. See the North Han Chinese listed here.

https://docs.google.com/spreadsheets/d/1x8pm8sVcHqceiNFJMO082kxaBF5ePr4__bAK05VQRFw/edit?usp=sharing

But the Chinese individuals who share the rare SNPs with Brits aren't part of this group.

NA18614
NA18627
NA18541

Graham Little said...

It's well known that British soldiers did have Indian wives in colonial India. So I guess this shouldn't be a surprise in East Asia too.

Ebizur said...

This is most likely a coincidence, but when I looked up the information for each of those samples mentioned by Davidski, every one of them turned out to be a Han Chinese female from Beijing who belongs to one or another subclade of mtDNA haplogroup B4:

Sample=NA18541
Population=CHB
Population Description=Han Chinese in Bejing, China
Gender=female
Haplogroup=B4i1
Exome Sequencing Platform=Agilent
Coverage=99.70
Per-base depth=52.84
"Han Chinese from Beijing, China; at least three out of four grandparents are Han"

NA18614 Haplogroup B4h
"gender: Female
cell line: lymphoblast cell line NA18614
ethnicity: Han Chinese from Beijing (HapMap code: CHB)"
"Han Chinese from Beijing, China; at least three out of four grandparents are Han"

NA18627 Haplogroup B4b1a2a
"gender: Female
cell line: lymphoblast cell line NA18627
ethnicity: Han Chinese from Beijing (HapMap code: CHB)"
"Han Chinese from Beijing, China; at least three out of four grandparents are Han"

They apparently do not share a common direct matrilineal ancestor until the level of B4, which is surely too ancient to be relevant to their sharing of rare SNPs with some British individuals. However, the fact that every one of them is from Beijing may be relevant.

Krefter said...

Check out this realistic reconstruction of Loschbour. That's probably what alot of our Mesolithic ancestors looked like in person 8,000 years ago.

http://s10.postimg.org/sbetmupnd/Loschbour.png

Krefter said...

http://blobsvc.wort.lu/picture/669df97ed00d98457db11dc05f9d92f3/519/291/wortv3/d676c960a6340a291eb05742f6dd2c1782fb53bc

Krefter said...

The new way to look at ancient mtDNA.

http://www.anthrogenica.com/showthread.php?3701-The-new-way-to-look-at-Ancient-mtDNA&p=64340#post64340

I made a bunch of add ons to the spreadsheet of ancient mtDNA I already had.

Davidski said...

Who is that in your avatar?

Krefter said...

It's a picture of me with a Torc and sword added on. I got the idea from a Roman coin of Gaulish prisoners.

Grey said...

So if it's ancient it's very ancient and if no ANE then more likely coastal?

Krefter said...

Off topic: I just had a light-bulb moment.

Because U5a was the main U5 clade of Early Neolithic Central Europe, it's probably a WHG lineage not ANE.

U4 and U2e should be most associated with ANE. U2 happens to be very popular and diverse in south Asia, along with Y DNA P, where ANE ancestry peaks in Eurasia.

mtDNA U itself might be THE lineage of the ancestral clade of ANE-WHG. It can't be random all 5 samples from upper Palaeolithic west Eurasians have U.

There are only 3 non-U samples from Euro hunter gatherers, and 2 are a C1 founder effect. The other is pre-H2a2b, which could be from near eastern admixture.

Anders PĂ„lsen said...

Maybe there is something in the PCA variation connection I found between Chinese populations all the way to Europe.

http://fennoscandia.blogspot.no/2014/01/is-there-east-asian-influence-in_29.html

andrew said...

Lots of ways to get a connection that faint. Silk Road trader has a child who migrates to Persia who has a child who migrated to the Levant who is brought home by a sailor delivering Crusaders to Italy who has a child who becomes a sailor and ends up settling in England, over the course of many generations, with dilution of a single event over time that leaving different fragments of ancestry in two or three people who cryptically and remotely descend from that person.

If a Martian had that degree of genetic connection it would be interesting. One Eurasian to another, not so much.

Ryan said...

David - Presumably your models wouldn't detect someone who has 0.001% ANE admixture though, so I'm not sure that these samples not showing an ANE component really says anything relevant.

Anders - even the British Royal Family has documented Magyar and Cuman heritage, so I wouldn't discount a Uralic route out of hand.

Chad Rohlfsen said...

A lot of us have Magyar and Cuman ancestry, but that is questionable as a source.

Chad Rohlfsen said...

That's 800 years ago and about impossible to locate in modern folks.

Mike Thomas said...

I read that genetically, in any individual , ancestry doesn't go back beyond the high Middle Ages

Davidski said...

Is that so? Then why is it possible to detect ancestry from Denisovans among Papuans?

I can give you plenty of other examples that contradict what you just said.

ryukendo kendow said...

@ Mike Thomas
Where did you read that?

Mike Thomas said...

Yeah I meant with any actual genealogical relevance.
Try "The Malthusian parameter of ascents: What prevents the
exponential increase of one’s ancestors?"
and "On the Common Ancestors of All Living Humans"

ryukendo kendow said...

@ Mike Thomas
Those are very fascinating papers, thanks for raising them.

However, genealogical relevance and population genetic relevance are very different things. The first paper focuses on the first but not the second.

From the second paper, where ACA refers to that generation where, due to the phenomenon of 2 parents, 4 grandparents, 8 great-grand etc., one's ancestors expands to the entire population living at the time:
"It is likely that the notion of a relatively recent ACA
point may lead to some confusion. If we consider only ancestors
who lived prior to the ACA point, a Japanese and a
Norwegian today share the exact same set of ancestors. At
first glance this seems patently ridiculous. Certainly the
Japanese and Norwegian have quite different genotypes
due to very different ancestry. The confusing fact is that both of these statements are true. Although the Japanese and Norwegian have the same set of ancient ancestors,
they did not receive an equal hereditary contribution from
each of those ancestors. The Japanese owes a small proportion
of his genetic makeup to people living in northern Europe several thousand years ago, and a large proportion
to people living in and around Japan, while the opposite is true of the Norwegian."

If you consider that, at the ACA point population structure akin to that between present-day Norwegians and Japanese already existed for the differential inheritance to take place from, it becomes clear that this in no way invalidates population structure as a concept.

In fact, we can even make an ACA estimate for all primates alive today, despite the fact that we are different species now.

Also, the second paper assumes a small but constant level of exchange takes place in all localities at all times, preventing any tree-like divergences from forming, when I think it is becoming quite clear from what we've seen so far that sudden exchanges taking place in pulses from separated geographical locales with isolated populations with little intermixing before the admixture event seem to be the case in reality, making a criss-crossing tree structure a more accurate representation of what's going on in many places at least.

So both papers support that genealogical relevance decays very rapidly, but neither support the assertion that discernible differences in population genetic terms declines that fast.

Mike Thomas said...

Nice commentary, Ryu

Ryan said...

Chad - That's my point. You inherit DNA scholastically, so the 0.001% doesn't give a reliable date. The shared sequences don't necessarily mean British admixture into China or visa versa - but possibly a third group mixing with both. If there is this one documented route there are certainly other undocumented routes.