Improved phylogenetic resolution and rapid diversification of Y-chromosome haplogroup K-M526 in Southeast Asia
Tatiana M Karafet, Fernando L Mendez, Herawati Sudoyo, J Stephen Lansing and Michael F Hammer
Abstract
The highly structured distribution of Y-chromosome haplogroups suggests that current patterns of variation may be informative of past population processes. However, limited phylogenetic resolution, particularly of subclades within haplogroup K, has obscured the relationships of lineages that are common across Eurasia. Here we genotype 13 new highly informative single-nucleotide polymorphisms in a worldwide sample of 4413 males that carry the derived allele at M526, and reconstruct an NRY haplogroup tree with significantly higher resolution for the major clade within haplogroup K, K-M526. Although K-M526 was previously characterized by a single polytomy of eight major branches, the phylogenetic structure of haplogroup K-M526 is now resolved into four major subclades (K2a–d). The largest of these subclades, K2b, is divided into two clusters: K2b1 and K2b2. K2b1 combines the previously known haplogroups M, S, K-P60 and K-P79, whereas K2b2 comprises haplogroups P and its subhaplogroups Q and R. Interestingly, the monophyletic group formed by haplogroups R and Q, which make up the majority of paternal lineages in Europe, Central Asia and the Americas, represents the only subclade with K2b that is not geographically restricted to Southeast Asia and Oceania. Estimates of the interval times for the branching events between M9 and P295 point to an initial rapid diversification process of K-M526 that likely occurred in Southeast Asia, with subsequent westward expansions of the ancestors of haplogroups R and Q.
Link
I change thread as I find Tobus comment on the post "More ancient Scandinavians” interesting with respect to MSMC paper and this new post. According to Tobus ”Consider 30-40kya ago, a Tianyuan-like population splits into two groups, one which (eventually) gives rise to modern Europeans, South Asians and Central Asians (ie a "West Eurasian" branch), the other to modern East Asians and Amerindians (ie an "East Eurasian" branch). ANE is an ancient example of the former branch, thus it's *both* ANE (a specific sub-branch) and "West Eurasian" (the parent lineage).
ReplyDeleteWhen I look at the picture on the post MSMC, I see that there are three Eurasian branches and I suppose that their main thrust was from south to north and not the other way round. I would guess that the Tianyuan like northern population originated in South Asia and spread to East and West. In the west they encountered people that had originated in West Asia and in the east they encountered people who had originated in Southeast Asia. However, people mixed also via the southern route, i.e. South Asia with West Asia and South Asia with Southeast Asia.
The picture gets more exciting when we try to fit the refined structure in haplogroup K-M526 with MSMC paper. It seems that there has been an ancient movement of people from Southeast Asia to the west and it has also affected Africa. However, it seems that we cannot see that movement in Figure D of the MSMC paper. As the relationship between mtDNA M, N and L3 is not completely clear, we may come across surprises in the future in this respect. I have even seen trees in which N is above L3, and several L3 branches and M stem from the N root. Anyway, y-DNA seems to be highly mobile. It looks like the typical European and Native American y-DNA originated in Sunda, and perhaps the typical Australian y-DNA line belongs to the same line with the first settlers of Northern Eurasia.
"Estimates of the interval times for the branching events between M9 and P295 point to an initial rapid diversification process of K-M526 that likely occurred in Southeast Asia, with subsequent westward expansions of the ancestors of haplogroups R and Q".
ReplyDeleteI have long assumed that to be the case. The idea is further supported by:
"Interestingly, the monophyletic group formed by haplogroups R and Q, which make up the majority of paternal lineages in Europe, Central Asia and the Americas, represents the only subclade with K2b that is not geographically restricted to Southeast Asia and Oceania".
Sounds suspiciously as though the whole haplogroup developed in SE Asia and P moved off from there. The idea is further supported when we consider that K2a and K2c are also 'eastern'. I can find no older reference to K2d-P402 unfortunately. As a result I have no idea where it was found. It must be one of the new clades the authors discovered.
"Estimates of the interval times for the branching events between M9 and P295 point to an initial rapid diversification process of K-M526 that likely occurred in Southeast Asia, with subsequent westward expansions of the ancestors of haplogroups R and Q."
ReplyDeleteThis is what many of us have been saying for years. Note that this work does not change the fact that P (and thus Q and R) branch off about the same time as what later become M and S. What does change is that {M, P, S} share a common ancestor over NO - of which N and at least some of O migrated north (some migrated back south recently with rice farmers and island travelers).
I still maintain that is it likely that P split off during the eastward migration in the vicinity of the Brahmaputra river, but given the early NO split and its current distribution, I now agree with Terry that a split further SE (towards Sundaland) is also possible - although I still find it less convincing, unless we postulate a just-post-Toba process.
I don't really like calling this haplogroup and many of its subgroups "K." Like "NO," it should be called "KT." Also, either get rid of L, T, N, O, S, M, P, Q, R - or add more letters for the many equivalent K sub-branches.
This is a bad blow to Eurocentrists. However it is not the worst yet, but it may be in the future as the status of K2* K2c K2d becomes known.
ReplyDeleteBut even as of now, the notion that M and S represent Caucasoid intrusions into SE Asia is quite laughable to anyone with even a grain of intellectual and academic integrity.
Maybe I am being naughty but I cannot help but laugh big time. Hahaha.
More on this later.(It may be that K2a and M526 as a whole may be of East South Asian origin rather than SE Asian origin. Though either way it would not bother me one bit.)
@Kristiina
ReplyDelete"change thread as I find Tobus comment on the post "More ancient Scandinavians” interesting with respect to MSMC paper and this new post. According to Tobus ”Consider 30-40kya ago, a Tianyuan-like population splits into two groups, one which (eventually) gives rise to modern Europeans, South Asians and Central Asians (ie a "West Eurasian" branch), the other to modern East Asians and Amerindians (ie an "East Eurasian" branch). ANE is an ancient example of the former branch, thus it's *both* ANE (a specific sub-branch) and "West Eurasian" (the parent lineage)."
I responded to Tobus in the post "More ancient Scandinavians” post. He is wrong.
"it looks like the typical European and Native American y-DNA originated in Sunda,"
Amerindians bookend the whole Y-DNA tree with having both most upstream C and most downstream Q lineages. But they don't have any of the lineages postulated as connecting the two, including K lineages. And they don't have the DE node either. What this likely means is that the phylogenetic pathway between C and P was likely different from the one postulated in the current phylogenies. Those non-IJK lineages were simply lost in America as hg Q expanded across the whole continent. So the way to think about the Y-DNA tree from CT down is that it's split between DE (Asia-Africa), F (East Asia-West Eurasia) and CP (America-Asia-West Eurasia). Amerindians likely split off first followed by West Eurasians, East Asians and Africans.
I always enjoy reading posts by Eurologist. That is because his posts satisfy my (probably) racist fantasy of how Eurocentrists would try to mask their hurt feelings against overwhelming evidences.
ReplyDeleteBut actually I partially agree with him this time. But the early split off that occurred in the Bengal region may have been between NOX and the rest or... K2a and K2b.
That still means that P ultimately originates from a place deep in the Sundaland.
Another interesting fact from this paper is the existence of "pre-P". P295+ P27-. The distribution of this group should indicate the early history of P haplogroup. I suspect it is mostly South Asian but cannot help but hope that it is SE Asian, in which case guys like Eurologists will be literally crying.
How on earth is it a blow to 'eurocentrists'? If anything it's a blow to out of africa. Timing is right for an out of india expansion right when 'modern' humans show up in europe.
ReplyDeleteThis is a bad blow to Eurocentrists and to those who adhere to the northern Central Asian origin hypothesis.
ReplyDeleteK2c-P261 is found in Bali, and K2d-P402 is found in Java. K2b2*-P295 is found among the Aeta people of the Philippines, in Timor, in Sumba, and in Sulawesi. Most of the K2 branches are found in insular Southeast Asia.
Does any one have the paper or any quotes from it
ReplyDeleteI really really am dissapointed the the paper is closed access
ReplyDeleteDo the Australians have K2b2?
The Aeta is all of their K K2b2??
"K2b2*-P295 is found among the Aeta people of the Philippines, in Timor, in Sumba, and in Sulawesi."
ReplyDeleteIf true, shocking and delightful. Any reference?
Aeta are a typical Oceanic pygmy race along with batak and others.
It seems that P haplogroup not only originated in SE Asia but spent some time there in their early history.
Eurocentrists like Maju should be crying by now. If they are not, they should.
If they are not shocked by all this they do not understand uni-parental phylogenetics.
Elliot . . . I mean Hector,
ReplyDeleteSorry dude I'm afraid reality continues to favor "Eurocentrists".
The branch lengths connecting M and P to the exclusion of LT are very short relative to the spans of time we're dealing with here. The authors of this paper find only one SNP -- M526 itself -- uniting NO and MP to the exclusion of LT, and five SNPs uniting M and P to the exclusion of NO. This represents maybe five hundred years. A SE Asian origin for K-M526 has never been credible with the broader phylogenetic context in mind. Deal with it (or don't -- and enjoy having your hopes slowly crushed as ancient DNA results eventually pile up).
@ TheXanian:
ReplyDelete"and K2d-P402 is found in Java".
Thanks. That answers my earlier question.
"K2b2*-P295 is found among the Aeta people of the Philippines, in Timor, in Sumba, and in Sulawesi".
The phylogeny shown has two basal K2b haplogroups, neither of which is 'K2b2-P295'. In fact the P295 mutation is at the root of the P'Q'r branch. Would you mind clarifying the situation? It makes sense that the two K2b branches are present in 'the Philippines, in Timor, in Sumba, and in Sulawesi'.
@ Grognard:
"Timing is right for an out of india expansion right when 'modern' humans show up in Europe".
Everything indicates that any 'out of India' expansion was preceded by an 'out of island SE Asia expansion'.
@ Hector:
"But the early split off that occurred in the Bengal region may have been between NOX and the rest or... K2a and K2b".
Possibly in Bengal, but unlikely. Of the four basal K-M528 branches two (K2c-p261 and K2d-P402) are definitely island SE Asian (see TheXanian's comment), a third is the NO branch and the fourth is the K-P331 branch most of whose branches are from island SE Asia, Australia New Guinea or Melanesian. That places virtually all the branches a considerable distance from Bengal.
"That still means that P ultimately originates from a place deep in the Sundaland".
In fact quite possibly even further east, in the islands between Sundaland and Sahul.
@ Kristiina:
"I would guess that the Tianyuan like northern population originated in South Asia and spread to East and West".
This paper strongly suggests not South Asia but SE Asia. Mind you, K has to have reached SE Asia from somewhere and my guess is via South Asia.
"As the relationship between mtDNA M, N and L3 is not completely clear, we may come across surprises in the future in this respect".
Yes. In fact Ebizur has nearly convinced me that mt-DNA M moved east through Central Asia rather than through South Asia.
@ Eurologist:
" this work does not change the fact that P (and thus Q and R) branch off about the same time as what later become M and S"
But after NO, K2c and K2d had branched off, all very much 'eastern' haplogroups. That makes it unlikely that 'P split off during the eastward migration in the vicinity of the Brahmaputra river'. It places the split further east.
@ German Dziebel:
"He is wrong".
It must be very lonely being the only person who is correct.
"But they don't have any of the lineages postulated as connecting the two, including K lineages".
Surprising, don't you think? In fact impossible if human Y-DNA originated in America.
"What this likely means is that the phylogenetic pathway between C and P was likely different from the one postulated in the current phylogenies".
You are really clutching at straws now.
@Kristiina and German -- it is incredibly hard to follow in depth discussions taking place in parallel on two separate blog posts. It would have been better if Kristiina had updated her response there rather than plunging those of us unfamiliar with the prior discussion right into the thick of it at this blog spot.
ReplyDelete@Eurologist: It is important to retain continuity of references so that one can easily compare old literature to new literature. By analogy, in the classification of composite particles made out of quarks in physics, the non-systemically named particles still retain the historic names even though this confounds what could otherwise e a very simple nomenclature system. I couldn't follow this paper without at least dual designations to track the old names.
FWIW, I also dislike the less informative allele based notation like K-M526, even though this is far less prone to change as opposed to designations like K2b2 which convey model dependent meaning in a compact and unobtrusive way.
@terryt and eurologist Is there anything in the paper about the relative timing of (new) K1 (which is old L and T which are both mostly West Eurasian) v. P? Presumably, the phylogeny assumes a K1 v. K2 split before a split of P from the rest of K2. Does K1 v. K2 happen ca. Iran with K2 migrating to SE Asia and then P back migrating?
Is there any indication re whether Q makes it to Siberia via NW India v. East Asia?
Any idea what would have caused the rapid expansion of K-M526 in SE Asia? Presumably this is post-Toba and pre-LGM. Could it be dog domestication? On the other hand doesn't it have to be pre-Australian/Papuan migration (which probably happened pre-dog domestication)? Could it be Upper Paleolithic technology innovation or is that too late too? Could the wave of volcanic eruptions in Europe ca. 40kya have been good for SE Asia while bad for Europe?
It appears that C splits off before K in a separate and earlier wave that is also pre-Australian settlement.
terryt,
ReplyDeleteEverything indicates that any 'out of India' expansion was preceded by an 'out of island SE Asia expansion'. [. . .] In fact quite possibly even further east, in the islands between Sundaland and Sahul.
This theory would also brilliantly explain where the all that (sadly, for Hector, non-existent) Denisovan ancestry in Europeans came from.
@ n/a
ReplyDeleteIt's you and the other eurocentrists who should deal with the fact that new evidences favor the southern origin hypothesis.
At the moment I don't have access to university libraries as I am between jobs. This article will be available free once it goes past the advanced pub stage so please don't bother the authors, begging for the article.
ReplyDeleteIn the mean time I obtained the link to this picture.
http://snag.gy/LqLuf.jpg
This says pretty much all. P295+ P27-(K2b2*) is found only in SE Asia and Indonesia(at a lower frequency).
The idea of P's origin somewhere else is just laughable at this point.
@n/a
I have used various aliases but Elliot is not one of them. My last DNA-forums ID was Manthus, I think.
You don't appear to understand any of this so I suggest you just cry. very hard. That is because crying is good for your health and in this case, by crying, you make yourself appear to understand far more than you actually do, which is... nothing.
@terryt
My speculation was before I saw Xanian's post. However it is still viable if K2c and K2d coalesce with K2b instead of K2a.
NO(or K2a)'s odd man out status seems to be reminiscent of C3(or C2 in the new tree) that is securely East/Northeast Asian.
Also notice that P295+ P27- is absent in South Asia, Australia, Oceania. I don't think P's origin is too far out to the Ocean side of Sundaland.
@ Andrew:
ReplyDelete"I couldn't follow this paper without at least dual designations to track the old names".
The diagram Dienekes provided has the old nomenclature on the left with the new on the right.
"designations like K2b2 which convey model dependent meaning in a compact and unobtrusive way".
Agreed.
"Is there anything in the paper about the relative timing of (new) K1 (which is old L and T which are both mostly West Eurasian) v. P?"
I certainly wouldn't go as far as to claim Ki is SE Asian. It is almost certainly South Asian, and formed at some time during a movement east from somewhere near Africa. I agree the timing would be very interesting.
"Presumably, the phylogeny assumes a K1 v. K2 split before a split of P from the rest of K2".
P actually splits from K2b according to the phylogeny. K2a, K2c and K2d are all 'eastern' haplogroups which tends to place K2b somewhere nearby.
"Does K1 v. K2 happen ca. Iran with K2 migrating to SE Asia and then P back migrating?"
That I what I have long maintained the evidence indicates. This paper merely reinforces what I have long accepted to be the case. And IJ split from K1'K2 even earlier.
"Any idea what would have caused the rapid expansion of K-M526 in SE Asia?"
Again, I have long believed the boating technology necessary to cross Wallace's Line to Australia allowed increased mobility and spread back through the already inhabited regions. That is somewhat near your conclusion: ' doesn't it have to be pre-Australian/Papuan migration'.
@ Daniel Szelkey:
"I really really am dissapointed the the paper is closed access"
I guess we all are. I'm thinking of buying it.
"Do the Australians have K2b2?"
Evidently not. Hector says it is found only in SE Asia and Indonesia. We'll have to leave it to Hector to answer the Aeta question. I'm sure K is not their only Y-DNA though.
@ Hector:
ReplyDelete"Eurocentrists like Maju should be crying by now. If they are not, they should".
Maju banned me from his blog mainly because I insisted he was wrong on this matter. I doubt he will say anything now.
"My speculation was before I saw Xanian's post".
I realised that.
"However it is still viable if K2c and K2d coalesce with K2b instead of K2a".
But that is not what the data as we have it at present allows. I'm afraid you might become compared with German if you're not careful!
"NO(or K2a)'s odd man out status seems to be reminiscent of C3(or C2 in the new tree) that is securely East/Northeast Asian".
I don't think NO is particularly 'odd man out'. It is still eastern and could easily have evolved first in SE Asia and moved north, as NO. I'm sure that both N and O developed in northern China, or somewhere near there. However I believe it unlikely that C2 moved north into China. To me it is far more likely that southern C1 moved south from China. In other words C and F had separate migration routes east. As Andrew said, 'It appears that C splits off before K in a separate and earlier wave that is also pre-Australian settlement'.
"Also notice that P295+ P27- is absent in South Asia, Australia, Oceania. I don't think P's origin is too far out to the Ocean side of Sundaland".
But you did say:
"This says pretty much all. P295+ P27-(K2b2*) is found only in SE Asia and Indonesia(at a lower frequency). The idea of P's origin somewhere else is just laughable at this point".
Much of Indonesia is not actually part of Sundaland. Eastern Indonesia was never connected to either mainland region even at times of lowest sea level.
@ n/a:
"This theory would also brilliantly explain where the all that (sadly, for Hector, non-existent) Denisovan ancestry in Europeans came from".
Do you really believe that the K-derived Y-DNA haplogroups spread through previously uninhabited regions? However an origin in SE Asia would explain the mysterious affinity with Papuans that pops up in admixture studies in populations around the world.
"I always enjoy reading posts by Eurologist. That is because his posts satisfy my (probably) racist fantasy of how Eurocentrists would try to mask their hurt feelings against overwhelming evidences.
ReplyDeleteBut actually I partially agree with him this time."
Hector, you must be confusing me with someone else. While I am mostly interested in Europe and where its population originated - this does not make me "Eurocentric" in any way, shape, or form.
Extant European populations' origin is clearly mostly African but that via S and SE Asia, a bit of NE Asia, in recent times also from W Asia, and a bit of ancient populations from all over the place.
http://www.ranhaer.com/attachments/forumid_97/140606182790b0cedf4d0872a7.jpg.thumb.jpg
ReplyDeleteP-P295* actually found in Aeta,Sulawesi,Sumba and Timor.Guys,it's yrue
This comment has been removed by the author.
ReplyDeleteHowever, whatever one theorizes from haploid evidence (mtDNA, Y-DNA), the autosomal evidence gives a far broader, and deeper, account. Hammer et al (I believe) had first pointed out that the autosomal DNA evidence nevertheless sees South Asia as a distinct staging point of migrations throughout Eurasia, SE Asia and the New World (outside Africa).
ReplyDeleteThis comment has been removed by the author.
ReplyDelete@Andrew
ReplyDelete"@Kristiina and German -- it is incredibly hard to follow in depth discussions taking place in parallel on two separate blog posts. It would have been better if Kristiina had updated her response there rather than plunging those of us unfamiliar with the prior discussion right into the thick of it at this blog spot."
Your opinion has absolutely no value and you are a persona non grata for me (http://anthropogenesis.kinshipstudies.org/2012/08/on-dziebel-out-of-america-and-anthropology-response-to-razib-khan/). I personally don't care what's convenient for you. Kristiina did the right thing bringing what's relevant from other strings into the new string. She just didn't know that Tobus was wrong.
I doubt they came from island. If anything more like sundaland connected them both together.
ReplyDelete"Also notice that P295+ P27- is absent in South Asia, Australia, Oceania. I don't think P's origin is too far out to the Ocean side of Sundaland".
ReplyDeleteHowever TheXanian wrote this:
"K2b2*-P295 is found among the Aeta people of the Philippines, in Timor, in Sumba, and in Sulawesi".
That places P's origin securely east of Wallace's Line. Every one of those islands lies between Sunda and Sahul. In other words: in Wallacea.
"A SE Asian origin for K-M526 has never been credible with the broader phylogenetic context in mind".
On the contrary it is the only possible solution given the phylogeny. The only K-M526 haplogroups that can possibly have originated west of Wallace's Line are K2a (NO) and K2d. And then K2d is only just west of the line, in Java. K2c is admittedly only just east of the line, in Bali, but as far as I'm aware K2b1 is entirely east of the line (certainly 'restricted to Southeast Asia and Oceania') and it seems from TheXanian's comment that K2b2 (P) originated east of the line.
The only remaining problems is: where are these K2b1 haplogroups found:
K2b1b-P336
K2b1c-378
K2b1a-P315
K2b1a-P401
Does anyone know? We know they are from somewhere in Southeast Asia and/or Oceania.
Glad to hear the geographic mystery of YDNA MP indeed implied that SE Asian and also Australian K* and Papuan YDNA S are all closely related. And naturally, it wasn't just one biblical ancestor but an ancestral population having closely related YDNA that moved in. Like already expected - and expressed in previous comments - this result downgrades most of the different K branches to further downstream K*: below M526 and even dowstream P331. The confidence in "ancestral" subclades in the Southeast Asian "Terra Incognita" was never justified. Instead, with L/T (K1) around an ultimate origin of YDNA K in the wider neighbourhood of Indonesia remains as unlikely as anywhere else.
ReplyDeleteSo, having accepted that M526 moved into SE Asia, apparently including mainland SE Asian YDNA NO, we are left with a huge dating problem. Such a diverse group that includes YDNA P can't have moved into SE Asia as pioneer moderns, since these are presumed much older here than the ~35 ky indicated by YDNA genetic dating for YDNA P. Does genetic dating now necessarily imply that YDNA P indeed originate in Island SE Asia? The time between the common ancestor P-P295 and that of (non-SE Asian) P-P27 is now estimated at another 12.3ky, while MA-1 already carried an extinct YDNA R1 at 24 ky. Moreover, MA-1 seems to contradict a more recent expansion of P-27 as insinuated by the authors. I rather suspect dating issues still exist, and M526 moving in together with early moderns, probably through northern Eurasia. But what if M526 indeed entered SE Asia much earlier and started a subsequent back migration that could have affected MA-1: then why it should not have affected eg. Ust-Ishim at ~45 ky as a potential ancestor of MA-1? In the SE Asian scenario Ust-Ishim can't have been ancestral to YDNA P and should have been from a completely different lineage. This has been discussed earlier on Popperian grounds. But none of this, nor any result for Ust-Ishim, resolves the origin of all remaining - ancestral, non SE-Asian - YDNA F+ haplogroups. I gather Karafet et al. should have taken this in consideration, especially in case they have some inside information about Ust-Ishim, but somehow I think they haven't. They are perfectly happy with YDNA dating of one branch out of context.
Another thing, also mentioned by Karafet et al., is the possible massive extinction of haplogroups in northern Eurasia. This is possible in case northern Eurasia was the scene of exceptional selective forces, what I have always deemed very likely. Also MA-1 is a testimony of genetic purge, including its extinct mutations at mtDNA U.
Last but not least, albeit not mentioned by Karafet et al., is the possibility that some mutations in the YDNA tree are reversed to the effect that Hg R is actually older than Hg P. This would explain MA-1 probably better. It is still remarkable that Hg R1a has the ape-like SRY10831.2 mutation, to date only shared by the African YDNA A haplogroup.
Contrary to the Asiacentrists that think the origin question is like a football game that can be won by cheering and jeering, I think more possibilities than ever are now on the table. I wonder how this new result could be conciliated with my idea the Cromagnoid type could have been an Asian hybrid, originally. At least the result challenges both dating and the phylogenetic tree.
All very fascinating!
ReplyDeleteI still stick to my hypothesis that P radiated Northwards from Western India, just south of Gujurat.
It's origin much further East is very likely based on this data.
Interestingly, my 100% Native Irish father, has a small segment of Oceanian, per 23andMe... maybe that's the explanation!
First off a correction. Bali is west of Wallace's Line and therefore is part of Sunda, not Wallacea.
ReplyDelete"Eurocentrists like Maju should be crying by now. If they are not, they should".
To my surprise Maju has posted regarding the paper, and supports the findings:
http://forwhattheywereweare.blogspot.co.nz/2014/06/y-dna-macro-haplogroup-k-m526.html
What's more he has managed to find where my mystery haplogroups were found. I do disagree with him on his placement of K2b's origin. Every single haplogroup in K2b has its major presence in Wallacea (counting the Philippines as Wallacean as they are certainly not Sunda), not in Sunda as he shows. K2b1b's presence in Borneo and M's presence in Mentawai almost certainly represent back movement from Wallacea rather than being the place of origin for each haplogroup. His suggestion that the odd Q and R in SE Asia is a remnant of the original development of these haplogroups is also likely incorrect. They are most likely much more recent arrivals. I did find this statement from him amusing:
"Overall this study underlies and vindicates my repeated claim of SE Asia playing also an important role in the formation of the Asian+ branch of Humankind, together with South Asia".
Wonder where he repeatedly claimed it?
@ Andrew:
"Is there anything in the paper about the relative timing of (new) K1 (which is old L and T which are both mostly West Eurasian) v. P?"
Maju has done some work on this. He places the 'P' node 5000 years after K, but no word on when K formed or when K1 formed.
I now see that the usual definition of Sundaland does not include most of Java and nearby islands. But they are very closely located and if the sea level is lower much of it is accessible from Sundaland by land bridges.
ReplyDeleteMy inference was based on the larger percentage of P295* among SE Asians vs Indonesians in the map-diagram from the paper I provided. Comparing with Xanian's post I now see that most or all of P295* West of Wallace line, come from Aeta. But that is fairly substantial.
I still claim that P295's origin was on the land-side of Sundaland.
When there are many small isolated populations there is a greater chance of rare lineages surviving. You have a similar situation in N China vs S China.
N China is metropolitan whereas S China is divided into many river valleys.
I am a bit tired of explaining why this is so.
Rokus man, you crack me up. R being older than P... Do you think geneticists are fools?
Do I have to explain why this is hilarious?
Rokus cracked everyone up back in the days of DNA-forums. He did it everyday but there was one particular one that was priceless but I forget. I only remember my reaction. "man, why do you invest so much of your time in things you know nothing about?"
"I doubt they came from island. If anything more like sundaland connected them both together".
ReplyDeleteSunda and Sahul have never been connected. Even at times of lowest sea level they were separated by a bunch of eastern Indonesian islands that are most conveniently called 'Wallacea'. The data presented in this paper makes it almost impossible to believe that K2b evolved anywhere other than in those islands. As Hector said, 'The idea of P's origin somewhere else is just laughable at this point'.
"Interestingly, my 100% Native Irish father, has a small segment of Oceanian, per 23andMe... maybe that's the explanation!"
Very likely to be the explanation.
"Extant European populations' origin is clearly mostly African but that via S and SE Asia, a bit of NE Asia, in recent times also from W Asia, and a bit of ancient populations from all over the place".
Yes, we are all a little bit of everything. Even those 'Papuans' who crossed Wallace's Line were already a mix.
"Instead, with L/T (K1) around an ultimate origin of YDNA K in the wider neighbourhood of Indonesia remains as unlikely as anywhere else".
I don't think anyone has claimed an eastern origin for K-M9 (KLT). But what this paper makes obvious is that the K-M526 branch originated in Sunda and the K-P331 branch originated in Wallacea.
"having accepted that M526 moved into SE Asia, apparently including mainland SE Asian YDNA NO, we are left with a huge dating problem".
I see no problem.
"Such a diverse group that includes YDNA P can't have moved into SE Asia as pioneer moderns"
It is extremely unlikely they were 'pioneer moderns' in SE Asia. Y-DNA C and mt-DNA N had beaten them to it.
"while MA-1 already carried an extinct YDNA R1 at 24 ky."
Fits the data perfectly. Those SE Asian Q and R haplogroups almost certainly did not originate in SE Asia. As I suggested above, they are later arrivals. That explains the apparent 'more recent expansion of P-27 as insinuated by the authors'.
"then why it should not have affected eg. Ust-Ishim at ~45 ky as a potential ancestor of MA-1?"
I doubt Ust-Ishim is Y-DNA R (the dating is too early) although Ust-Ishim is quite possibly one (amoung several) ancestors of MA-1.
"In the SE Asian scenario Ust-Ishim can't have been ancestral to YDNA P and should have been from a completely different lineage".
We may be about to find out exactly which lineage, it seems.
"I now see that the usual definition of Sundaland does not include most of Java and nearby islands. But they are very closely located and if the sea level is lower much of it is accessible from Sundaland by land bridges".
ReplyDeleteWhere did you see that? They are certainly east of Wallace's Line which passes between Bali and Lombok. Placental mammals (which, of course, includes humans) have generally been unable to cross that line and so Java and Bali are usually considered part of Sunda. Wiki certainly considers those islands as part of 'Sundaland':
http://en.wikipedia.org/wiki/Sundaland
I see they exclude the Philippines but really those islands, apart from Palawan, are not part of Sunda.
"I now see that most or all of P295* West of Wallace line, come from Aeta. But that is fairly substantial".
But, as I said, we can hardly regard the Philippines as being part of Sunda. And with other members of P295* being found in Timor, Sumba and Sulawesi we can make a pretty good guess as to what islands provided the route to the Philippines.
"I still claim that P295's origin was on the land-side of Sundaland".
Unlikely. If you exclude the Philippines from Sunda then every K2b haplogroup has a major presence in the islands east of Sunda, many exclusively so.
"When there are many small isolated populations there is a greater chance of rare lineages surviving".
I agree, but in the case of the SE Asian islands don't forget that Y-DNA O's expansion has had a huge influence. The survival of the minor lineages in this case prove to be most revealing. Look at K2 as a whole. We have K2a in East Asia but all the others are SE Asian: K2* in Sumatra and Sulawesi (the latter island almost certainly somewhat recent), K2d in Java, K2c in Bali, leading inexorably to K2b beyond Bali across Wallace's Line.
The paper states that the P326 group, which is L and T never ever reached south-East Asia
ReplyDeleteas someone who summarized it ( and I agree with it )
chronological notes underline is that the expansion (split) of K, K2b and K2b1 (P) happened in "rapid" sequence with few millennia between each node.
Or in other words:
1. There was a split between K1 (LT) and K2 as this one migrated eastwards.
2. Then K2 split in Sundaland
2. Then K2b expanded in what is now Borneo.
3. Then P expanded from ~ Northern Borneo, with a branch (P1) back-migrating to South Asia
@ Rokus:
ReplyDelete"So, having accepted that M526 moved into SE Asia, apparently including mainland SE Asian YDNA NO"
I now see why you see a difficulty. When K-M526 'moved into SE Asia' NO was yet to form. The same goes for P. There was no mass migration of a variety of haplogroups. Well, there may have been haplogroups other than K-M526 but just a single K-M526 in SE Asia gave rise to the descendant K2 haplogroups.
" I rather suspect dating issues still exist"
In fact the authors don't give a date for K-M9's appearance. They just give dates from its formation. A dating for early K branches (and F) would be very useful.
@ Hector:
"I now see that the usual definition of Sundaland does not include most of Java and nearby islands".
Concerning the Philippines and Sunda: I am on Huxley's side rather than on Mayr's on this matter. Have a look at the map in the biogeography section:
http://en.wikipedia.org/wiki/Wallace_Line
"It is extremely unlikely they were 'pioneer moderns' in SE Asia. Y-DNA C and mt-DNA N had beaten them to it. "
ReplyDeleteNo in southeast asia+oceania mtdna macrohaplogroup M is associated with y-dna macro haplogroup C. While, mtdna N+(R) are associated with y-dna haplogroup K2b. For example 90+% of aeta have N and 90+% of them have K2b. Mtdna haplogroup P is found in popualtions that have K2b1-P405 (basically S Y-DNA), while mtdna haplogroup S is directly associated with y-dna haplogroup M.
Change that to
"It is extremely unlikely they were 'pioneer moderns' in SE Asia. Y-DNA C and mt-DNA M had beaten them to it. "
"The paper states that the P326 group, which is L and T never ever reached south-East Asia"
ReplyDeleteThat's what I have always claimed. It dropped off along the way.
"in southeast asia+oceania mtdna macrohaplogroup M is associated with y-dna macro haplogroup C".
That is certainly not correct for far Oceania where we have Y-DNA C closely associated with R-Derived B. And in New Guinea Y-DNA C is alost absent and so is mt-DNA N. Modern haplogroup association is not necessarily an indication of common origin.
"Mtdna haplogroup P is found in popualtions that have K2b1-P405 (basically S Y-DNA)"
P is especially common in Australia where K2b is virtually absent apart from derived K2b1. And S is certainly absent.
"mtdna haplogroup S is directly associated with y-dna haplogroup M".
I don't follow that one. Mt-DNA S is exclusively Australian and Y-DNA M is unknown there.
"For example 90+% of aeta have N and 90+% of them have K2b".
By the time the Aeta arrived in the Philippines they had had ample opportunity to become a real smorgasbord of haplogroups.
One can still make the argument that 4 of the 8 subgroups under K are not SE Asian (L, T, NO, K2e), and a 5th one is split into a SE Asian one (K2b1) and a more western-oriented one (K2b2 = P -> Q, R).
ReplyDeleteSo, unless very basal subgroups of P and NO are only found in SE Asia, I think the jury is still out where and over what region the incredible expansion of K occurred. Note that the expansion of K2b1 and K2b2 lag considerably behind that of K{xLT}, which supports a secondary event after the people entered and expanded at the footsteps to SE Asia (the Ganges/ Brahmaputra region). For now, I stick with the Brahmaputra river region as the first area of split after LT, firstly because due to its geography it provides a good explanation of why "P" would move N/ NW and "NO" N/ NE, and secondly, it is a good "holding position" before the Toba super-volcano explosion, which may have enabled the spread of a multitude of evolving K-subgroups into and in Sundaland, as well as the spread of "NO" N/NE.
The distribution of P-P295* is what population geneticists call "relic distribution".
ReplyDeleteUnless you fancy extensive and free "island hopping", it is safe to assume that most successful lineages are not born on an island. Also the center of mass is more suggestive of Sundaland proper rather than islands to the east of Sundaland as the birthplace of P.
I read the paper and the percentage of P-P295* is really low and sampling was quite uneven favoring Indonesian islands.
I won't surprised even if some South Asian P295*s are found in the future. Then I shudder at the thought of Maju being too happy about it(and drawing a wrong self-serving conclusion when that happens).
@TerryT,
ReplyDelete"I rather suspect dating issues still exist"
In fact the authors don't give a date for K-M9's appearance.
They do, their article strictly conforms to their own standards in Karafet's 2008 paper, only to narrow down on the dating of YDNA P - previously: 34,000 (26,600–41,400).
Thus, in Karafet's 2014 paper this would result to the following dates:
CT-M168: 70,000 ka
K-M9: 47,500 ka
K-P331: 44,500 ka
P-P295: 39,500 ka
P-P27: 27,200 ka
Note that Karafet's date for R1 was: 18,500 (12,500–25,700) ka, while MA-1's R1 was dated 24,000 ka. This date may be close to Karafet's own upper limit, though this may also indicate Karafet actually underestimated the dates by ~30%. Applied to Karafet's dates above this would lead to the following correction:
CT-M168: 90,000 ka
K-M9: 62,000 ka
K-P331: 58,000 ka
P-P295: 51,200 ka
P-P27: 35,300 ka
Such dates would make a lot more sense in view of an early modern Cromagnoid expansion that originated in SE Asia. Ust-Ishim (45,000 ka) thus could turn out to be P-P27- without any problem. However, then we still don't know whether northern Eurasia was the source of an early mass migration to SE Asia or on the receiving end.
Sorry, a few more thoughts.
ReplyDelete"Then K2b expanded in what is now Borneo".
Borneo certainly does not feature prominently in Maju's list of K2b's distribution. It is only mentioned in K2b2b-P336, also listed as being present in eastern Indonesia apparently. In most papers I have seen
'eastern Indonesia' is a synonym for the triangle of islands that Sumba, Timor and Flores/Alor form. Even as early as K2b-P331 (or as Hector called it P295+ P27-) we find two of these islands mentioned. The triangle is the first islands any voyagers would reach on crossing Wallace's Line from Bali. Where we find K2c-P261. And next to it K2d-P402 in Java. And K2* next to Java in Sumatra.
The newly revealed K-M9 phylogeny and the various clades' distributions is fairly unequivocal as to the haplogroup's history. K2-M526 moved down what became the chain of SE Asian islands leaving the string of haplogroups. K2b-P331 developed in the eastern Indonesian island triangle beyond Wallace's Line.
From that triangle K2b2 moved back west but K2b1-P397 spread through more islands. Members of K2b1a-P404 reached New Guinea, where it diversified. Members of one subclade, K2b1a, moved south from New Guinea into Australia. Members of K2b1b-P336 moved back across the line to Borneo. K2b1c-P378 moved north to the Philippines. Members of M-P256 also reached New Guinea, especially Melanesia.
"90+% of aeta have N"
ReplyDeleteWhat N haplogroups do they have? As far as I'm aware the only mt-DNA haplogroups specifically Philippine are M derived (M17 and M80) although R24 is specifically Filipino.
"It is extremely unlikely they were 'pioneer moderns' in SE Asia. Y-DNA C and mt-DNA M had beaten them to it. "
The biggest problem you have with that connection is New Guinea. No mt-DNA N haplogroup is present there apart from R-derived P (both branches shared with Australia where the haplogroup almost certainly originated) and B4a1, almost all B4a1a1a at that. The other New Guinea mt-DNA haplogroups are M27, M28 and M29'Q. It is impossible to connect their presence with Y-DNA C of any sort.
On the other hand New Guinea Y-DNA is dominated by K2b1-P397 haplogroups, and predominately K2b1a-P405. C1c-M38 is present but mostly as the derived C1c1-M208. Interestingly C1c-M38 almost certainly coalesced in the same island triangle as K2b-M526 did. Which was first? Presumably east Indonesian C1c is reasonably closely related to Australian C1d-M347. And we have quite a selection of basal mt-DNA Ns in Australia. It is therefore more than likely that any 'original' Y- and mt-DNA connections were thoroughly disrupted in eastern Indonesia.
And don't forget that Australian K belongs to just a single branch of New Guinea K2b1a-P405, and most likely entered Australia from there. Why was K2b unable to colonise Australia before reaching New Guinea? Because Australia was already occupied?
P295+ P27- means K2b2*, not K2b*.
ReplyDeleteOne can imagine what would have happened if NO and P1's place were exchanged. It that were the case, P1 would have been declared "West Eurasian" and Hamar Fox and his friends would have been claiming moral victory of some sort. Also they would have had absolutely no qualm about designating NO as "Negritto" or whatever the most derisive description they could find and in particular they would not have questioned the SE Asian origin of NO.
Hamar, Maju etc. these guys exactly fit the description of Hikkomori except that they are not Japanese. They think that the White race has been immensely wronged by revisionist left leaning hippy scientists and they fancy themselves as the "tough no nonsense" guy who stands firm and corrects this "misinformation" to restore the honor and greatness of the White race.
Of course the orginal Japanese Hikkomori have gripes with the notion that Japanese originated from Korea. They spend enormous amount of time and effort to correct this "misunderstanding" conspired and propagated by Korean CIA or something.
In an odd way this may show that the human race is a one big family after all. Dysfunctional and unhappy one but a family nonetheless.
Hector,
ReplyDeleteOne can imagine what would have happened if NO and P1's place were exchanged. It that were the case, P1 would have been declared "West Eurasian" and Hamar Fox and his friends would have been claiming moral victory of some sort.
I chuckled quite a bit, because for all the talk above about people crying this and people crying that, I could definitely picture tears streaming down your face as you typed this.
For the record, once you peel away the many layers of your own projection -- which have nothing to do with me -- all I actually did was to correlate haplogroup Q with the autosomal composition of its likely population of origin at the likely time of its mutation. Feel free to quote evidence of my saying anything about P or NO being 'West Eurasian'.
Also they would have had absolutely no qualm about designating NO as "Negritto" or whatever the most derisive description they could find and in particular they would not have questioned the SE Asian origin of NO.
Again, projection. Nowhere on this blog will you find one single derisive remark about any human (or non-human) population from me. I can, of course, easily find numerous examples of you making such remarks.
Hamar, Maju etc. these guys exactly fit the description of Hikkomori except that they are not Japanese.
They think that the White race has been immensely wronged by revisionist left leaning hippy scientists and they fancy themselves as the "tough no nonsense" guy who stands firm and corrects this "misinformation" to restore the honor and greatness of the White race.
Feel free to quote any piece of political rhetoric I've used anywhere on this blog. In fact, I'm laughing at how incorrect your assessment of my character is. I just don't get it. Don't get me wrong, I couldn't care less what you think of me. It's not as though you're even remotely relevant to me. You are, however, attempting to smear me, so I feel it's necessary to correct you on this, just as you've forced me to correct you on basic logic in other threads.
As for Maju, I don't know much about him, but his politics seem to be left, so I'm not sure what you're talking about with him either.
In an odd way this may show that the human race is a one big family after all. Dysfunctional and unhappy one but a family nonetheless.
I agree. Nothing appeals to me less than the fact I belong to the same species as you. While you're obviously more repellant than the average person, I still can't deny that I'd happily replace every other human being on earth with a cat.
@ Rokus:
ReplyDelete"MA-1's R1 was dated 24,000 ka."
Is MA-1 actually R1, or just R*? If the latter it would fit the first list of dates you provided. Three thousand years for a haplogroup rapidly expanding from SE Asia to Central Asia is not impossible.
@ Eurologist:
One can still make the argument that 4 of the 8 subgroups under K are not SE Asian (L, T, NO, K2e), and a 5th one is split into a SE Asian one (K2b1) and a more western-oriented one (K2b2 = P -> Q, R)".
Not really. You have to exclude L and T for a start. They belong on a separate branch. And K2e's placement in the tree is far from resolved at present. So that leaves just NO as the only non-SE Asian haplogroup within K2-M526.
"Note that the expansion of K2b1 and K2b2 lag considerably behind that of K{xLT}, which supports a secondary event after the people entered and expanded at the footsteps to SE Asia (the Ganges/ Brahmaputra region)".
The expansion of K(xLT), (K-M526), obviously predates that of K2b-P331 as the latter is a branch within the former. But take a look at those K2-M526(xP331) branches. As I pointed out in an earlier comment they string out along the SE Asian mainland with NO in the north, K2b1* in Sumatra, K2d in Java and K2c in Bali. That pretty much eliminates South Asia as the region of origin. It does leave open the likelihood of an origin near the NE India/SW China border region however.
@ Hector:
"The distribution of P-P295* is what population geneticists call 'relic distribution'".
Yes. But interestingly all members of its closest relation, K2b1-P397, are found east of Wallace's line, and the fact that the individual K-P295(xP27)s are also found east of Wallace's line has to be an important consideration as to the origin of the K2-P295 haplogroup.
"Also the center of mass is more suggestive of Sundaland proper rather than islands to the east of Sundaland as the birthplace of P".
What are you using as 'the center of mass'? If a particular haplogroup expanded unidirectionally the centre of mass would give no indication of centre of origin at all.
"Unless you fancy extensive and free 'island hopping', it is safe to assume that most successful lineages are not born on an island".
It would surely depend on what factor provided an advantage for the particular haplogroup.
"I won't surprised even if some South Asian P295*s are found in the future".
It has to be there really. I'm not actually claiming P-P27 actually first appeared in Wallacea, just that its immediate ancestor in the form of P-P295(xP27) must have originated there.
So ignoring the arguments about the details of the sequence "Out of Africa" has been replaced by a sequence of "Out of X" events?
ReplyDeleteI don't understand the West vs East Eurasian argument here. This pre-dates the split of both groups.
ReplyDeleteEither way, the North Eurasian branch became West Eurasian genetically, with no East Eurasian. Also, the great majority of European autosomal DNA came from the F branch (G,I,J, and such). Only a 20% max came from North Eurasia (R,Q), which is West Eurasian anyway.
Let's leave the ad hominem and ethnocentric arguments out of the conversation.
@ Daniel Szelkey:
ReplyDelete"90+% of aeta have N"
I've just followed that commnet up and see the Aeta have 40% mt-DNA P, actually an R-derived haplogroup rather than N. But anyway, that supports the notion that humans arrived in the Philippines from the south and after they'd reached Australia. P is almost certainly an Australian haplogroup. Four of the seven basal haplogroups are exclusively Australian (P5, P6, P7 and P9), two are shared between Australia and New Guinea with the earlier branches in Austraklia (P3 and P4) leaving just one not particularly Australian although a small level of P1 has been found there.
I read somewhere (sorry, I forget where) that both Q1 and P1'2'8'10 moved back to 'Eastern Indonesia' from New Guinea/Australia before both expanded from that region once more. P8 and P10 both reached the Philippines and presumably that is the 'N' haplogroup you refer to. But its association with Y-DNA K2b need date no earlier than its voyaging to the Philippines. And certainly not as early as its original arrival in Australia.
I'm presuming the remaining 50% Aeta mt-DNA 'N' is actually B, most of which is probably no earlier than the Austronesian expansion, even in the Philippines.
"I still can't deny that I'd happily replace every other human being on earth with a cat".
ReplyDeleteWho was it said, 'the more I see of humans the more I admire dogs'? But, living in New Zealand, I definitely wouldn't want to replace humans with cats. Introduced mammals, and not just predators, have devastated the native wildlife.
"So ignoring the arguments about the details of the sequence 'Out of Africa' has been replaced by a sequence of 'Out of X' events?"
What has become more and more obvious is that we have backwards and forwards, to and fro, etc. Surely it has become impossible to believe that 'modern humans' originated in a single region and spread throughout the world from that region.
This comment has been removed by the author.
ReplyDeleteChad,
ReplyDeleteGood points, and I agree.
"MA-1's R1 was dated 24,000 ka."
ReplyDeleteIs MA-1 actually R1, or just R*? If the latter it would fit the first list of dates you provided. Three thousand years for a haplogroup rapidly expanding from SE Asia to Central Asia is not impossible.
How careless of me. Raghavan (2014)'s supplement Figure SI 5a reveals MA-1 as an extinct branch - otherwise already on its way on the branch leading to R1+R2 sharing 19 derived base paires and just having 5 ancestral positions. Moreover, having 35 "private" mutations against 18 listed R1* mutations MA-1 is confirmed ancestral for, and even counting the 17 R1* mutations that resulted in no-calls, MA-1 is quite comparable to R1*.
Such a genetic distance makes three thousand years for a haplogroup rapidly expanding from SE Asia to Central Asia (P-P27: 27,200 ka => MA-1, 24,000 ka) quite impossible. Moreover, this timing is not supported by archeological evidence. Dispersal of P-P295 P27- between 51,200 ka and 45,000 ka, thus applying a 30% correction to the dates of Karafet (2008), would solve this problem.
"Raghavan (2014)'s supplement Figure SI 5a reveals MA-1 as an extinct branch - otherwise already on its way on the branch leading to R1+R2 sharing 19 derived base paires and just having 5 ancestral positions".
ReplyDeleteI was fairly sure it was something like that. Thanks for the clarification.
"Such a genetic distance makes three thousand years for a haplogroup rapidly expanding from SE Asia to Central Asia (P-P27: 27,200 ka => MA-1, 24,000 ka) quite impossible".
Do youi really think so? Don't forget it would have been expanding through a region already occupied to a large extent. It wouldn't necessarily have had to carry a particular mt-DNA along with it.
"Dispersal of P-P295 P27- between 51,200 ka and 45,000 ka, thus applying a 30% correction to the dates of Karafet (2008), would solve this problem".
Possibly. I'm more than a little inclined to agree with Maju on 'molecular-clockology' anyway.
This comment has been removed by the author.
ReplyDelete"Such a genetic distance makes three thousand years for a haplogroup rapidly expanding from SE Asia to Central Asia (P-P27: 27,200 ka => MA-1, 24,000 ka) quite impossible. Moreover, this timing is not supported by archeological evidence. Dispersal of P-P295 P27- between 51,200 ka and 45,000 ka, thus applying a 30% correction to the dates of Karafet (2008), would solve this problem."
ReplyDeleteRokus,
I largely agree.
"Such a genetic distance makes three thousand years for a haplogroup rapidly expanding from SE Asia to Central Asia (P-P27: 27,200 ka => MA-1, 24,000 ka) quite impossible".
ReplyDeleteI don't see a problem. Especially when we turn our minds back to the old idea of a great southern coastal migration. It was supposed to have carried a whole major human population from Africa all the way to Australia in an instant. So what has changed minds now?
"Moreover, the structured distribution of derived subgroups suggests an initial rapid dispersal into .virgin territory"
I don't think the authors at all propose an expansion into virgin territory, although its quite possible the expansion was aided by exploitation of a so-far unexploited habitat. It is surely obvious the ancestors of the K expansion must have arrived in SE Asia from somewhere. The phylogeny suggests strongly that the route was via South Asia.
"with little subsequent gene flow between 'demes' (ie genetically defined 'regions')".
On the contrary it looks as though there was (perhaps later) substantial gene flow between populations. At present the haplogroups do not define phenotype.
"Such a genetic distance makes three thousand years for a haplogroup rapidly expanding from SE Asia to Central Asia (P-P27: 27,200 ka => MA-1, 24,000 ka) quite impossible".
ReplyDeleteI don't see a problem. Especially when we turn our minds back to the old idea of a great southern coastal migration. It was supposed to have carried a whole major human population from Africa all the way to Australia in an instant. So what has changed minds now?
Only the context changed: Out of Africa could always be cited to quell any serious discussion with a lot of nonsense, and nobody cares - or dares - to disagree. The comfort of such a mindset can't be taken as a valid argument against a serious attempt to deal with real data.
Further thoughts on:
ReplyDelete"Such a genetic distance makes three thousand years for a haplogroup rapidly expanding from SE Asia to Central Asia (P-P27: 27,200 ka => MA-1, 24,000 ka) quite impossible".
The Romani, or Gypsies, made it all the way from northern India to the far west of Europe in much less than 3000 years. Something like 400 years:
http://en.wikipedia.org/wiki/Romani_people
It might be possible to argue that these people had a superior method of transport than did the Y-DNA K people, but to me that argument doesn't really stand up to any sort of scrutiny.
"Moreover, the structured distribution of derived subgroups suggests an initial rapid dispersal into .virgin territory' with little subsequent gene flow between 'demes' (ie genetically defined 'regions')".
Surely it is undeniable that the Y-DNA K people mixed with others they met along the way. Most admixture analyses show a widespread 'Papuan' element which can best be explained by the Y-DNA K expansion from a Papuan SE Asian region. As to what happened to the Denisovan element: Papuans have no more than 6% Denisova. Most Eurasians have no more than 2-3% Papuan. Six per cent of two or three per cent is an extremely small amount.
"the puerile debate about the supposed SE Asian origin of ancestral Hg K"
'Puerile'? Are you not interested in our remote ancestry? Creationist perhaps?
@ Terry
ReplyDelete"with little subsequent gene flow between 'demes' (ie genetically defined 'regions')".
On the contrary it looks as though there was (perhaps later) substantial gene flow between populations. At present the haplogroups do not define phenotype."
That was the author's statement; not mine. Of course, there was probable *later* gene flow
This comment has been removed by the author.
ReplyDelete@ Dr Rob:
ReplyDeleteI see what you're getting at now.
"the structured distribution of derived subgroups suggests an initial rapid dispersal into .virgin territory' with little subsequent gene flow between 'demes'"
There was certainly 'subsequent gene flow' and the expansion certainly was not strictly into 'virgin territory' (people already lived in the region) but the expanding group must have formed a coherent population to be able to expand so rapidly. The comparison to Roma may be very apt. The apparent likelihood of a coherent population points strongly to one (or more) accompanying mt-DNA lines. Although some M haplogroups look to have expanded west from SE Asia to some extent, and Y-DNA C1 also looks to have been involved, the only real possibility is mt-DNA R.
I have long noticed there is a significant gap between South Chinese (N21 and N22) and Australian (N13, N14, O and S) mt-DNA. That gap is filled with mt-DNA R haplogroups(B4'5, R9, R14, R21, R22 and R23). As a result I have concluded that mt-DNA originated, and expanded from, that region. The distribution of South Asian mt-DNA R haplogroups can easily be interpreted as indicating an entry along India's east coast and subsequent expansion up the Godavari and Ganges River systems. Perhaps we can regard as supporting evidence the MA-1 haplogroups Y-DNA R and mt-DNA U.
Can a Y haplogroup defining mutation happen only once? Is there only one branch per mutation?
ReplyDeleteBecause if a defining mutation can happen twice, then the first time could have arisen a thousand miles from and thousands of years before the second time. And there would be two of the same haplogroup out there, easily confused as being the same descent.
You would have two groups showing the same haplogroup but being only indirectly related to one another. There would be two different lineages, for example, for R1a.
Any detection of R1a would not tell you which lineage you were looking at.
The phylogenetic tree we see in the diagram in this paper is based on nodes that begin new branches that represent defining mutations.
What if any of those mutation happened twice and survived?
Only as an example of how this would work, the branch P is defined by mutation P295. P originates at that point/node in the phylogenetic tree, branching from a parent branch K (P331).
At some point in time, in parent branch K (P331), mutation P295 occurs, beginning a new branch.
At some later point in time, in parent branch K (P331)mutation P295 occurs again, at some other geographic location.
These two P haplogroups would show the same Y haplotype, but they would not be the same branch. They would only be indirectly related to one another.
You would have two populations showing Y haplotype P. They could easily be confused as a single branch. But in fact they would represent two completely different genetic histories.
BTW, this does happen in microbe populations. Some mutations, being highly probable, occur again and again. When cultures contain these mutations, there is sometimes no way to tell whether the mutation spread to or originated in a particular culture.
"Because if a defining mutation can happen twice, then the first time could have arisen a thousand miles from and thousands of years before the second time".
ReplyDeleteYes, but by then other mutations will have occurred and the recurring ones become reasonably obvious. There are some disagreements concerning the various trees but actually a fairly wide consensus has been reached. I have observed that in general those who question haploid phylogenies are those who have a pre-existing belief that cannot be made to fit the phylogeny. Consequently rather than studying the phylogeny as to what it can reveal they dismiss it outright. And I have every confidence in these authors' work after all those years.
I wrote:
ReplyDelete"Because if a defining mutation can happen twice, then the first time could have arisen a thousand miles from and a thousand years before the second time".
terryt replied “Yes, but by then other mutations will have occurred and the recurring ones become reasonably obvious.”
Terry, if you look at this again, I think you’ll see it doesn’t work.
The two lines of descent would be in homogeny with this mutation -- they would be identical. As a matter of how the tree works, relevant mutations on this tree would create new branches. Mutations outside the branches of the tree would be on a different tree of decent.
Telling the difference between the two duplicated branches would need to be a matter of direct observation. But given the ages of these mutations, that would be difficult or impossible. Remember that a good part of this Y-haplotype tree is reconstructed, not observed. I’m not saying I have any problem with that, but it does mean this tree as is would not “see” parallel evolution.
And based on the average rate of mutation we see on the tree, there’s no reason to think recurring mutations would occur often enough to be detectible, much less obvious.
terryt also replied:
“I have observed that in general those who question haploid phylogenies are those who have a pre-existing belief that cannot be made to fit the phylogeny. Consequently rather than studying the phylogeny as to what it can reveal they dismiss it outright. And I have every confidence in these authors' work after all those years.”
Terry, this is something different. I actually expect that this matter of parallel evolution in specifically human Y-haplotype trees has already been addressed somewhere in the literature. It’s real. It’s not a quibble or diversion. It’s a scientific question.
And the fact of parallel evolution could affect whether a phylogenetic tree reflects the historical events or distorts them.
BTW: interesting paper with a twist on parallel evolution in clover just out: Olsen, Kooyers and Small. Adaptive gains through repeated gene loss: parallel evolution of cyanogenesis polymorphisms in the genus Trifolium (Fabaceae). Philosophical Transactions of the Royal Society B, 2014 DOI: 10.1098/rstb.2013.0347
"The two lines of descent would be in homogeny with this mutation -- they would be identical".
ReplyDeleteThe Y-chromosome is subject to a high mutation rate therefore the chance of identical Y-chromosomes being present in two geographically separated populations is pretty low. Identical Y-chromosomes are extremely unlikely except in father/son relationships or perhaps as far as grandson. At a more distant relationship more mutations would appear. Besides which any difficulties you see with Y-DNA K are surely just as valid for any other Y-DNA phylogeny.
"As a matter of how the tree works, relevant mutations on this tree would create new branches".
But the mutations would only bee seen as a 'tree' if a reasonable number of descendants survive.
"I actually expect that this matter of parallel evolution in specifically human Y-haplotype trees has already been addressed somewhere in the literature. It’s real. It’s not a quibble or diversion. It’s a scientific question".
Any reference?
"Relationships of DNA sequence haplotypes at the cyanogenesis loci and flanking genomic regions suggest independent evolution of gene deletions within species. This study thus provides evidence for the parallel evolution of adaptive biochemical polymorphisms through recurrent gene deletions in multiple species".
The genes involved in cyanogenesis involve selection and I don't think any evidence indicates particular mutations on the Y-chromosome are considered a product of selection. And don't forget that the remainder of the genome in the various clover lines would be quite different from each other. It is just the 'adaptive biochemical polymorphisms through recurrent gene deletions' that are the same in the various lines.
terry t wrote: “The Y-chromosome is subject to a high mutation rate therefore the chance of identical Y-chromosomes being present in two geographically separated populations is pretty low. Identical Y-chromosomes are extremely unlikely except in father/son relationships...”
ReplyDeletemy reply:
Terry, please understand that this phylogenetic tree is not based on the whole Y-chromosome. The nodes you see are a SINGLE MUTATION on the Y chromosome. The parent branch of the tree represents a single shared mutation. The split at the node represents differentiating mutations.
In this kind of phylogenetic tree, lineage mainly follows a series of single mutations from ancestor to descendants.
Parallel evolution (or convergent evolution) would mean that single mutation that creates the new node happened more than once, but this kind of tree would not show it.
terry t wrote:
“Besides which any difficulties you see with Y-DNA K are surely just as valid for any other Y-DNA phylogeny.”
my reply:
Yes. It means that, for example, the haplogroup labelled R1a may have arisen more than once in a parent group R(1). Two different populations that arose at different times and places, but with a common descent as far as the parent haplogroup.
terry t wrote: “But the mutations would only be seen as a 'tree' if a reasonable number of descendants survive.” my reply:
Technically, the tree needs only one descendant to survive and reproduce. But remember that some nodes are reconstructed but unattested. The tree assumes they had to have occurred for the sequence of descent to have happened. terry t wrote: “The genes involved in cyanogenesis involve selection and I don't think any evidence indicates particular mutations on the Y-chromosome are considered a product of selection.”
It’s not selection but persistence. The clover study is mainly of note because it points to different measures of likelihood that a particular mutation will recur because of a matter of differential biochemistry in the DNA (or RNA). In other words, some single mutations may be mechanically more likely than others, and therefore more predictable.
See, eg, http://www.nature.com/nrg/journal/v14/n11/full/nrg3483.html Whether this applies to the mutations that are used in this phylogenetic tree is a much more complicated matter than whether or not parallel evolution might have occurred.
A different question is the historical meaning of the tree. The moment these kinds of tree go beyond sequence of descent and are applied to chronology and geography, the problems begin.
"The clover study is mainly of note because it points to different measures of likelihood that a particular mutation will recur because of a matter of differential biochemistry in the DNA (or RNA). In other words, some single mutations may be mechanically more likely than others, and therefore more predictable".
ReplyDeleteWell no. In the case of the clover mutations they have only lasted because they have been selected for in independent lines. Yes, it shows that the same mutation can occur multiple times but in this case it has not occurred within the same genetic landscape.
"Terry, please understand that this phylogenetic tree is not based on the whole Y-chromosome".
That doesn't really matter. Under your scheme it would be impossible to construct any sort of Y-DNA phylogeny. As it stands the phylogenies are based on mutations within a defined series of previous mutations. Perhaps I should have been more specific. and said, 'The Y-chromosome is subject to a high mutation rate therefore the chance of Y-chromosomes with a particular set of identified mutations being present in two geographically separated populations is pretty low. Y-chromosomes with an identical set of identified mutations are extremely unlikely except in father/son relationships...' The chance of the same mutation occurring in two branches of a haplogroup with the same identified mutations is very remote considering the number of mutations possible.
"It means that, for example, the haplogroup labelled R1a may have arisen more than once in a parent group R(1)"
So with regard to one of my earlier comments where would you like R1a (and K) to have originated?
I wrote: "The clover study is mainly of note because it points to different measures of likelihood that a particular mutation will recur because of a matter of differential biochemistry in the DNA (or RNA). In other words, some single mutations may be mechanically more likely than others, and therefore more predictable".
ReplyDeleteterryt replied:
“Well no. In the case of the clover mutations they have only lasted because they have been selected for in independent lines. Yes, it shows that the same mutation can occur multiple times but in this case it has not occurred within the same genetic landscape.” My reply: Terry, the research focuses on two closely related variants. The defined observational difference between them are two mutations. The researchers believe they are seeing recurring mutations, not recessives. BOTH mutations arise in both environments. One is selected in one environment. The other is selected in the second environment. The mutations would of course precede any selection.
I wrote:
"Terry, please understand that this phylogenetic tree is not based on the whole Y-chromosome".
terryt replied:
“That doesn't really matter. Under your scheme it would be impossible to construct any sort of Y-DNA phylogeny.”
My reply: As with any kind of inference from historical evidence, there’s always uncertainty. The best we can do is admit the uncertainties and figure where they may be and how they might offer alternative explanations. The one thing not to do is to pretend they are not there.
The tree in the article is pretty impressive. But this kind of phylogenetic tree tells you only the order of descent, NOT where and when. That kind of information is not contained in the data used to make the tree. It’s an extra conjecture away from what the tree is able to say.
terryt wrote:
“The chance of the same mutation occurring in two branches of a haplogroup with the same identified mutations is very remote considering the number of mutations possible.”
My reply: Yes, that’s the crux of it. Whether EACH of these mutations on the tree is a once in 70,000 years occurrence. Or could some have recurred. I don’t think we know how to measure that. But we do know that some mutations recur. It appears to be a matter of bio-chemical structure. This IS a problem with phylogenetic tree on a grander than scale than human Y-chromosome, as the other paper I mentioned points out. So it’s not totally outside the realm of possibility regarding this tree.
I wrote:
"It means that, for example, the haplogroup labelled R1a may have arisen more than once in a parent group R(1)"
terryt wrote:
“So with regard to one of my earlier comments where would you like R1a (and K) to have originated?” My reply: No opinions and no preferences.
But simply as a matter of its historical implications, remember that in theory each of these nodes represents a single individual -- one founder in a pre-existing parent haplogroup of men.
If the parent group ends up in a far-away place from that single individual’s offspring, we might picture how that could happen. How did they separate so cleanly from the parent group? Especially since none of them would be aware of the mutation.
And if the offspring end up in two very different far away places from each other, we might wonder if there was more than one founder.
"If the parent group ends up in a far-away place from that single individual’s offspring, we might picture how that could happen. How did they separate so cleanly from the parent group?"
ReplyDeleteIn the Paleolithic it is extremely unlikely a single individual moved far from its origin without being accompanied by a number of relations. The only time we would get a huge movement would be into an as yet unoccupied region.
"But this kind of phylogenetic tree tells you only the order of descent, NOT where and when. That kind of information is not contained in the data used to make the tree. It’s an extra conjecture away from what the tree is able to say".
The phylogenetic tree tells us where a particular haplogroup's closest relations are found. From that we can get a very good idea of where it originated. Of course a haplogroup could expand greatly but be later replaced by the later expansion of another haplogroup. We can often see where that has happened.
I wrote:
ReplyDelete"If the parent group ends up in a far-away place from that single individual’s offspring, we might picture how that could happen. How did they separate so cleanly from the parent group?"
terryt wrote: “In the Paleolithic it is extremely unlikely a single individual moved far from its origin without being accompanied by a number of relations. The only time we would get a huge movement would be into an as yet unoccupied region.”
my reply:
The point is that the male carrying the new mutation -- the founder -- would already be part of a group -- the parent haplogroup. So that population would now consist of two nodes on the tree -- the parent and the new mutation. But the mutation is neutral, so none of them know that.
Therefore there is no reason for them to separate. Unless there is strong drift, the two nodes should stay together. And travel together. So why do we find clean breaks between these nodes later on in their history?
I wrote: "But this kind of phylogenetic tree tells you only the order of descent, NOT where and when. That kind of information is not contained in the data used to make the tree. It’s an extra conjecture away from what the tree is able to say".
terryt wrote: “The phylogenetic tree tells us where a particular haplogroup's closest relations are found. From that we can get a very good idea of where it originated. Of course a haplogroup could expand greatly but be later replaced by the later expansion of another haplogroup. We can often see where that has happened.”
The tree would look exactly the same if all those node mutations happened in Las Vegas or Antarctica 200 years ago. All you are seeing in the tree is lines of descent. There is no actual time or place represented in this tree. Only the order of appearance in relation to one another.
With parent and offspring nodes, we can assume some kind of proximity, but the tree does not tell you where on earth either were located.
The tree itself and the process that forms it contains NO information about the actual time or location of any haplogroup. That has to be inferred from external info -- current location, conjectured rate of mutation and of course ancient DNA.
And obviously if there was any parallel occurrence of the same mutation, the tree would diverge from what actually happened, unless it showed two branches for the same mutation.
"The point is that the male carrying the new mutation -- the founder -- would already be part of a group -- the parent haplogroup"
ReplyDeleteYes. A member of a tribe.
"So that population would now consist of two nodes on the tree -- the parent and the new mutation. But the mutation is neutral, so none of them know that"
They don't need to know anything about haplogroups. Most living today know nothing of the subject. But members of the tribe would know the father/son relationships.
"Therefore there is no reason for them to separate".
It's extremely doubtful that they would separate.
"Unless there is strong drift, the two nodes should stay together. And travel together. So why do we find clean breaks between these nodes later on in their history?"
We see the explanation in the recent mt-DNA tree:
http://dienekes.blogspot.co.nz/2014/08/new-estimates-of-human-mtdna-node-dates.html
What is obvious from that paper is that Polynesian B4a1a1a left Taiwan accompanied by a number of related haplogroups including its parent haplogroup. Polynesian-specific B4a1a1a3 and Madagascar-specific B4a1a1a2 had already formed and were already quite diverse. Along the route to those two extremes other haplogroups dropped out leaving the two apparently isolated in their particular regions.
"The tree would look exactly the same if all those node mutations happened in Las Vegas or Antarctica 200 years ago".
True but the geographic distribution in either case would look completely different.
"With parent and offspring nodes, we can assume some kind of proximity, but the tree does not tell you where on earth either were located".
But the geographic distributions of the related haplogroups go a considerable way to telling us where the origins of the various branches lie.
"The tree itself and the process that forms it contains NO information about the actual time or location of any haplogroup. That has to be inferred from external info -- current location, conjectured rate of mutation and of course ancient DNA".
Exactly We actually have far more information on current distribution and so that makes the greatest contribution to our understanding.
"And obviously if there was any parallel occurrence of the same mutation, the tree would diverge from what actually happened, unless it showed two branches for the same mutation".
Only if the two mutations happened within a line that had already been defined by exactly the same pre-existing mutations. The chance of that occurring are very remote.
Terry - sorry for the late reply.
ReplyDeleteI've seen researchers acknowledge the possibility of recurring mutation (and therefore parallel evolution) only in cases where the geographical separation just absolutely demands it and mainly with regard to human female lineage. But that's unfortunate because extreme cases are not the only ones where a recurring mutations is possible and the probabilities are unknown and so the history has to be unsure.
Compare your case of the Polynesian and Madagascar split to a case like mtDNA haplogroup X, where "recurrence appears most plausible:
"In a parsimony tree, this Iranian mtDNA would share a common ancestor with the Native American clade (fig. 2). Yet, the nonsynonymous substitution at np 12397 converting threonine to alanine cannot be regarded a conservative marker, as it has also been observed in two different phylogenetic contexts—in haplogroups J1 and L3e—among 794 complete mtDNA sequences (Finnilä et al. 2001; Maca-Meyer et al. 2001; Herrnstadt et al. 2002). Therefore, the scenario that the threonine to alanine change in the haplogroup X background is indeed due to recurrence appears most plausible."
Maere Reidla et al, Origin and Diffusion of mtDNA Haplogroup X (2004)
In the same paper, "distinct geographical distribution" was actually used against reversion as a possible alternative explanation of temporal relationship, despite good evidence in the other direction:
"“Moreover, haplogroup X is subdivided into two major subhaplogroups, designated “X1” and “X2.” Subhaplogroup X1, represented by a single Druze mtDNA in figure 1, differs from the root of haplogroup X by eight coding and three control region transitions and lacks the two transitions (195 and 1719) that characterize X2. These two nucleotides are rather mutable (Finnilä et al. 2001; Herrnstadt et al. 2002); thus, it cannot be completely ruled out that X1 is indeed a subset of X2 that reverted at both nucleotide positions. However, this possibility appears very unlikely, especially when one considers the time depth and the distinct geographic distribution of X1"