New blogs

Leherensuge was replaced in October 2010 by two new blogs: For what they were... we are and For what we are... they will be. Check them out.

Friday, April 24, 2009

mtDNA tree (version 1.1)


After digging for some time in the sources at PhyloTree, I have come with a revised mtDNA tree that locates most (but not all) rare lineages.


Please check for any possible errors and/or further information I might have missed.


Click to enlarge

The tree is exactly the same as in the previous version. The only changes are in the location of some rare clades, reducing the number of "unknown" lineages to just 4 M subclades. In two cases (M22'25 and M47'50) I only know the location of one of the two subclades but I'm simplistically assuming that it applies for the whole lineage.

Most of the newly located rare lineages are South Asian, SE Asian (noted as East Asian) or Australian.

The revision still seems to keep M and R gravitating (or at least most highly diverse) at South Asia. N remains mysterious with its highest apparent diversity in Australia (but it is possible that some of the rare lineages are related) and then in South Asia (especially if we include R).


Visually:

· M forms basically a "T" with South Asia, East Asia and Sahul at the three extremes and a thin projection to the west.
· N(xR) follows the southern Asian coasts from West Asia to Sahul, with a rather reduced projection to the NE (i.e. middle-East Asia).
· R is more West-South-SE Asian axis, with only limited presence in Sahul.

There is a rather strong link between South Asia and West Asia at the R and N lineages, that in some cases extends to Australia (but always seems to exclude East Asia). This does not diffuminate the diferential personality of these regions though.


Further observations: comparison of N with M and R:

The high apparent diversity of N in Australia (take with a grain of salt) may suggest a SE Asian origin for this lineage or at least a high level of flow of this clade through SE Asia in the early stages of Eurasian expansion. I find a strictly Australian N urheimat rather unlikely anyhow.

If my reading is correct, the very first moment of Eurasian expansion was starred by the M explosion and that of many (9) of its derived lineages, mostly in South and SE Asia. N, which was surely lurking along M at that stage may then have been carried to SE Asia along with the other M lineages. It is a possibility.

Anyhow there is no single temporal instance (except its very original node) where you can talk of an "N explosion" clearly: its subclades expand locally, and even often ignoring regional divisions, without any particualr temporal concentration, scattered along time. This is quite different of what we can see in M (early stages especially) and, to a lesser extent maybe, R (second phase), whose nodes are succeeded by large scale expansion of their subclades.

Let me explain myself graphically (">" means the temporal equivalent of one SNP):
· M expands (35 branches) > 9 M subclades (26%) expand > 8 M subclades (23%) expand
· N expands (12 branches) > 2 N subclades (17%) expand > 2 N subclades (17%) expand (one is R)
· R expands (15 branches) > 6 R subclades (40%) expand > 2 R subclades (13%) expand

So not only the M node is much more vigorous than those of N and R but also its branches show more vigor as well. R seems also pretty vigorous at least in the first moment (at that point even much more than M) but, otherwise, N seems to trail behind the rest and scatter its expansion through time.

Only primary derived nodes were considered above.

--------------------------------------------------------------------------------------------

Check also my previous version and comments.

.

42 comments:

terryt said...

"I find a strictly Australian N urheimat rather unlikely anyhow".

I doubt that anyone would seriously claim that.

"N, which was surely lurking along M at that stage".

Not necessarily.

"Anyhow there is no single temporal instance (except its very original node) where you can talk of an 'N explosion' clearly".

Wouldn't an original 'explosion' and subsequent lesser diversification indicate intense selection? M, in the relatively benign climate of India, may not have been subject to so much selection.

terryt said...

I've had time to look at your diagram in more detail. I see the following:

Haplogroups M14, M15, and M42 in Australia. Haplogroups M28, M29 and Q in New Guinea. Now, tell me again Maju, which of these M haplogroups are shared?

Haplogroups N2a, N12, N13, N14 and S in Australia. Uh oh. I must have missed something. I can't find any N(xR) haplogroups in New Guinea.

Haplogroup R1 in Australia. Ah, and at last we have it: haplogroup P in both New Guinea and Australia, thus proving a vast, deep and prolonged connection between the two regions.

Maju said...

I doubt that anyone would seriously claim that.

The apparent diversity is highest there, so there are some grounds indeed. IMO, a scientifical mind should consider that possibility as a matter of fact.

Not necessarily.

Most likely. If N and M had different expansive frames, especially in the gerography, we should see it much more clear. We just do not: they overlap all the time.

Wouldn't an original 'explosion' and subsequent lesser diversification indicate intense selection? M, in the relatively benign climate of India, may not have been subject to so much selection.

I am not very much "selectionist" so I'd rather discard that hypothesis unless there's some evidence to back it.

I see N as the "young sister" of M. Or, if you wish, as a minority within a distinct majority. If I am right, N may have migrated to SE Asia (still undifferentiated) together with some M subclades like M10'42, M12'G, M31'32, M44'52 and surely others that expanded later on (the ones mentioned are those with a first flowering apparently simultaneous to that of N).

Then it spread in many directions including back to South Asia (N1'5, pre-R, pre-X, etc.) but also into East Asia (N9, pre-A) and southwards to Australia (S, pre-N12, etc.)

That route, that we can well call "the N road" (though maybe was more of naval route), remained "open" (i.e. actually transited, maybe just via the usual marriage exchange between neighbouring communities - unsure) until the W-N2a split, which seems the last epysode linking Australia with South/West Asia, at least on the mtDNA side.

Ah, and at last we have it: haplogroup P in both New Guinea and Australia, thus proving a vast, deep and prolonged connection between the two regions.

So you support the two routes hypothesis. Fine with me. But the continent was just one back then in any case.

And your conclusion anyhow says nothing in favor of your abuse of the concept of Wallacea anyhow: there would have been at least two Wallaceas: Sulawesi-Malukku and Sunda.

terryt said...

"there would have been at least two Wallaceas: Sulawesi-Malukku and Sunda".

Sunda is not Wallacea. Wallacea is the region between Sunda and Sahul.

Here's a couple of articles regarding Wallacean and Melanesian mtDNA haplogroups that you may find interesting.

http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1876738

From that article:

"All the ISEA populations studied have high levels of diversity, suggesting that they have maintained a comparatively large size over time and have not undergone substantial amounts of drift ... The Palu sample is unusual because of the much reduced level of haplogroup B types found; it is also one of the few eastern populations to contain haplogroups N9a6 and Y2, which could be due to recent arrivals from the west or north ... It is worth considering the possibility that this [mt-haps P and Q found to the west of Wallace's Line] may be a genetic trace of a conduit into ISEA for the root and tuber crops, perhaps of New Guinean origin, that arguably contributed far more to a change in subsistence in the Neolithic period of the region than did the introduction of rice farming from the north or west... In conclusion, the rather simple 'two-layer' settlement model of Australo-Melanesians ~50,000 years ago followed by 'Mongoloid' Austronesians ~4,000 years ago—even with allowance for considerable survival of indigenous lineages, as in more recent versions—clearly does not capture the complexity of demographic history in the region".

Food for thought there. And another one:

http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1201611

From this article:

"The implication is that isolation of these small island populations was an incomplete but persistent condition across the region for tens of thousands of years during the Pleistocene. By extension, movements between Near Oceania and Island Southeast Asia also would have been intermittent and small in scale... P is more widespread and heterogeneous than Q, and may therefore be the older of the two. With one clear exception, different branches of P occur either in Australia or New Guinea, but not both. This pattern suggests substantial isolation of Australia from New Guinea/Island Melanesia since around the time of first settlement... the three M haplogroups described here are centered in different locations within Northern Island Melanesia: M27 is most common and diverse in Bougainville, M28 in the interior of east New Britain, and M29 in southern New Ireland and east New Britain... A more complex scenario would have all of these haplotypes carried with Lapita groups as far as Fiji, and then have all been lost in separate founder events in every subsequent voyage into Remote Oceania, a most unlikely alternative... Their heterogeneous distributions [Mt-haps M] may also suggest they were introduced not only at different times but by different groups... Their highly structured distribution clearly contradicts some earlier notions of a loosely unified Melanesian, Old Melanesian, Australoid, or Australo-Melanesian population with a common ancestry, in favor of a far more complex population history".

Also from that article, "M27 was most frequent in central Bougainville, especially branch M27a, with the other M27 branches detected in Bougainville and sporadically in New Britain, New Ireland, the Solomon Islands, Santa Cruz, and Vanuatu".

Which may reduce the number of unkowns in your haplogroup diagram.

Maju said...

Sunda is not Wallacea. Wallacea is the region between Sunda and Sahul.

Lesser Sunda I mean. And yes all Lesser Sunda islands, excepting Bali, are Wallacea (and not Sundaland).

http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1876738.

Interesting the high ammounts of M* observed in several regions, notably Malay-Indonesians and NW Chinese.

All the ISEA populations studied have high levels of diversity, suggesting that they have maintained a comparatively large size over time and have not undergone substantial amounts of drift ...

Yah but does that mean "from the beginning"? Or just in recent times? The paper discusses origin of Austronesians in protohistorical times, not MP/UP founder events.

It is worth considering the possibility that this [mt-haps P and Q found to the west of Wallace's Line] may be a genetic trace of a conduit into ISEA for the root and tuber crops, perhaps of New Guinean origin, that arguably contributed far more to a change in subsistence in the Neolithic period of the region than did the introduction of rice farming from the north or west...

Or rather a remnant of when the colonization of Sahul happened, IMO.

It is very notable that:

1. MtDNA P and Q are found at significative and similar ammounts in many Malayo-Indonesian populations, notably in Sumba (4% each).

2. The main exception is Alor islands (north of Timor), which keep almost 30% of mtDNA Q (the main local lineage). Ambon (west of Papua) also has a notable apportion of this clade (11%) but neither of these two Q "refuges" show any P.

3. MtDNA N* (maybe related to Australian clades and/or the overall N spread) is concentrated in Java (14%). Hard to say more without proper resolution.

From the second paper:

is more widespread and heterogeneous than Q, and may therefore be the older of the two.

This is coincident with what I get in my tree: P seems to have spread before Q did (but M29'Q as a whole branched out before Q).

With one clear exception, different branches of P occur either in Australia or New Guinea, but not both. This pattern suggests substantial isolation of Australia from New Guinea/Island Melanesia since around the time of first settlement...

Ok, relevant.

the three M haplogroups described here are centered in different locations within Northern Island Melanesia: M27 is most common and diverse in Bougainville, M28 in the interior of east New Britain, and M29 in southern New Ireland and east New Britain...

Actually, in the map, both M28 and M29 appear strongest in East New Britain but whatever. What is most important is to notice that while these are Melanesian, they are not Papuan.

In my tree I get M27 expanding earlier, while M28 and M29 may even have arrived together. M29 is directly related to Q in any case and M29'Q branches out before M27 does, it seems.

Which may reduce the number of unkowns in your haplogroup diagram.

No because M27 is aready correctly placed in Melanesia. The unknown location clades are all high numbered (40s-50s).

terryt said...

"Interesting the high ammounts of M* observed in several regions, notably Malay-Indonesians and NW Chinese".

I don't doubt for a moment that M originated in India, as you can see from my diagram at remotecentral. It's mtDNA N I have problems with.

"Yah but does that mean 'from the beginning'? Or just in recent times?"

Hang on. One minute you're using high diversity to argue for an Indian origin of all haplogroups but now that it doesn't suit your belief you're arguing that high diversity is a product of a series of migrations. Make up your mind.

"Or rather a remnant of when the colonization of Sahul happened, IMO".

I disagree. And so do the authors of the paper obviously. Certainly the points you raise are easily explained by diverse back migrations.

"P seems to have spread before Q did".

That may be so, but certainly not before N had reached Australia. And P is derived from N.

"What is most important is to notice that while these are Melanesian, they are not Papuan".

In general you would be hard pressed to tell the difference in appearance between Melanesians and Papuans. Mind you, that could be a result of thousands of years of contact. Australian Aborigines on the other hand are fairly different from either. And to get to Melanesia yoy certainly have to go via NewGuinea, even if only along the coast. Anyway M29 is closely related to Q.

"No because M27 is aready correctly placed in Melanesia".

Sorry. I didn't think you had anything by M27.

Maju said...

I don't doubt for a moment that M originated in India...

I sometimes do. It'd be a lot easier if, after we could resolve that M/N hidden diversity in SE Asia, this region happened to be the most diverse. Geographically SE Asia is perfect to explain the geography of early human expansion in Eurasia and was not so badly affected by Toba as South Asia was (so if we would want to place the M/N migration before Toba, that would be the logical spot for them to survive).

But so far South Asia has highest diversity for M and R (and is second for N, after Australia - apparently).

In general, anyhow, I prefer to think of South and SE Asia as a continuum, almost a single region, at the early stages of human expansion in Eurasia.

Hang on. One minute you're using high diversity to argue for an Indian origin of all haplogroups but now that it doesn't suit your belief you're arguing that high diversity is a product of a series of migrations. Make up your mind.

As I said in the other discussion, you can question the diversity argument if you have good reasons to think it is product of immigration from different origins. Also, high diversity in general is not the same as high top-level diversity within a single haplogroup.

Anyhow I was just posting a caveat. I do think that ISEA (Indonesia, etc.) has been populated from old.

I disagree. And so do the authors of the paper obviously. Certainly the points you raise are easily explained by diverse back migrations.

I don't think so. The auithors' "explanation" (no explanation given in fact, just an unexplained opinion) sounds too much like the arguments on Iberian North African genetics being product of the Muslim period. Too simplistic and unrealistic.

Neither Papuans nor Australian Aboriginals are famed for their navigation skills, so I find it most unlikely that they migrated back to Wallacea. Instead I strongly suspect those remnants are real remnants of the original migration into Sahul.

That may be so, but certainly not before N had reached Australia. And P is derived from N.

S expanded long before P did or even before R existed as such (P is R-derived).

Again you're putting the cart (R, P) before the horses (N, S).

terryt said...

"Neither Papuans nor Australian Aboriginals are famed for their navigation skills, so I find it most unlikely that they migrated back to Wallacea".

But their ancestors must have gone at least one way across Wallace's line. And Melanesians were obviously reasonably good sailors, otherwise they couldn't have reached New Ireland and New Britain.

"S expanded long before P did or even before R existed as such (P is R-derived)".

That's what I keep telling you. MtDNA S was already in Australia by the time R developed, and therefore mt-hap P represents a later immigrant population into Australia, and probably New Guinea as well. Mt-hap Q (M derived) had probably also already reached New Guinea from mainland SE Asia by then.

"Again you're putting the cart (R, P) before the horses (N, S)".

I have consistently put mt-haps N and S before haplogroups R and P. It has consistently been you who has not been able to see the distinction between the two sets of haplogroups.

Maju said...

But their ancestors must have gone at least one way across Wallace's line. And Melanesians were obviously reasonably good sailors, otherwise they couldn't have reached New Ireland and New Britain.

Guess you can do wonders with a boat or a raft and determination.

Certainly there was a rather continuous flow for a time between SE Asia (and South Asia clearly) and the Sahulian landmass, so this asks for an explanation: either they were skilled boaters (not sailors probably, I really find too daring to suggest sails in that time) and/or the geography of the area has changed because of tectonics. I would seriously consider the possibility that there were more intermediate islands around in the MP. After all it is an active tectonic region and we do know well how islands like Krakatoa may appear or disappear in a matter of days in such areas.

That's what I keep telling you. MtDNA S was already in Australia by the time R developed, and therefore mt-hap P represents a later immigrant population into Australia, and probably New Guinea as well. Mt-hap Q (M derived) had probably also already reached New Guinea from mainland SE Asia by then.

We are in agreement here.

I have consistently put mt-haps N and S before haplogroups R and P. It has consistently been you who has not been able to see the distinction between the two sets of haplogroups.

I understand that you were suggesting that P evolved from N separately in Australia. It is this what I contend, as the origins of P must be the same as those of the rest of R (South Asia with all likelihood).

terryt said...

"Certainly there was a rather continuous flow for a time between SE Asia ... and the Sahulian landmass".

Intermittent I'd guess.

"either they were skilled boaters (not sailors probably, I really find too daring to suggest sails in that time)".

Agreed. No sails. Sails are a later invention.

"and/or the geography of the area has changed because of tectonics".

Changing sea level would be the most likely explanation. That would expose more islands as well as extending the mainland. You had a list of warm periods on one of your blogs. I wrote them down somewhere but can't find them. I presume crossing Wallacea was easiest at times of low sea level, although the Austronesian expansion took place after the latest rise of sea level. If you refer back to your original list you will be able to work out the most likely periods for crossing Wallacea. I believe it will pay to remember, though, that Homo erectus seems to have survived in Java until perhaps as recently as 30,000 years ago and Homo floresiensis on Flores until perhaps 12,000 years ago.

"I understand that you were suggesting that P evolved from N separately in Australia".

Actually in Wallacea, and then carried to both New Guinea and Australia at some time. Y-hap K (in the strict sense) is the most likely suspect as it's the only haplogroup common in both regions. I suggest mtDNA R had earlier evolved somewhere slightly to the west, perhaps around the South China Sea, from where it expanded to Wallacea (P) and north (F and B) and west (ancestors of HV, JT and KU).

Maju said...

Intermittent I'd guess.

Yah, possibly. Some have argued that the contact was kept for some time somehow and the remnants of some Sahulian clades (and "Papuan languages") in Wallacea may relate to that.

There may have been two main waves:

1. S to Australia and M29'Q in Melanesia (and M27 into island Melanesia especifically, a "moment" later).

2. N12 to Australia and P to Australia and New Guinea.

Or it may have been a more or less continous drip in which these episodes are just the most noteworthy ones. The sequence reads SNP by SNP:

1. S expands in Australia and M29'Q in Melanesia, soon after the N explosion.
2. M27 expands in Melanesia (specifically some islands) and M14 splits, sending one branch to Australia (the other to Arabia AFAWK)
3. N12 and P expand in Australia and P in New Guinea as well. R31 splits, sending one branch to Australia (the other to South Asia, where has some importance until today).
4. Nothing happens in relation with Sahul (that I can see at the resloution I'm working at).
5. N2 splits sending N2a to Australia (and W to West/South Asia).

End of the immigration process. Sahul becomes isolated until historical times (or at least Austronesian migrations of the Iron Age). This may have happened between 45 and 50 kya, I guess.

Changing sea level would be the most likely explanation. That would expose more islands as well as extending the mainland.

Of course but the usual maps still show big gaps, expecially between Timor and Australia. The temporary existence of volcanic islands may have helped this island hoping process, I guess.

You had a list of warm periods on one of your blogs. I wrote them down somewhere but can't find them.

I only have one relevant blog: this one.

Anyhow what you're looking for is: THIS.

In fact there is a warm peak (higher sea levels), more noticeable in the antarctic data, before 40kya and cold valley (lower sea levels) c. 45 kya. But the data from Greenland is somewhat different and it shows a marked warm peak c. 50 kya preceded by a rather long period of cooling. While this happened near the equator, the Greenland data seems more coincident with my previous guesses if anything.

I presume crossing Wallacea was easiest at times of low sea level, although the Austronesian expansion took place after the latest rise of sea level.

But that doesn't matter because Austronesians were truly oceanic sailors, able to reach as far as Madagascar, Easter island or your country, that is also extremely remote. Their geographic circumstances may be comparable (island enviroment) but the two phenomenons are not otherwise related.

I believe it will pay to remember, though, that Homo erectus seems to have survived in Java until perhaps as recently as 30,000 years ago and Homo floresiensis on Flores until perhaps 12,000 years ago.

I am not really aware of the survival of H. erectus in Java but, in any case, they survived in small pockets, segregated from H. sapiens. It doesn't seem they were ever any obstacle to AMH expansion.

"I understand that you were suggesting that P evolved from N separately in Australia".

Actually in Wallacea...
That is what I can't accept: P is derived from R and therefore must have split from R (not N) at (roughly) the same spot where other R sublineages did. We are talking of a starlike structure here, what implies a rapid expansion from a single origin, no intremdiate steps and certainly not separate evolution from N, through (I guess) a "parallel R", coevolved misterously against all odds in Wallacea for your convenience.

R has 15 top-level sublineages and only one reached Sahul (P), two if we consider the mainly South Asian R31. The rest went in other directons or remained near the R homeland (South Asia). It was a fast expansive drive from a single origin. We don't understand well the why or even the how but that's what happened.

Y-hap K (in the strict sense) is the most likely suspect as it's the only haplogroup common in both regions.

I can accept a parallel between Y-DNA K and mtDNA R as a whole. They may differ locally in many details though but its probable in my opinion that (Y)K and (mt)R respresent the same cultural and demic expansion - and therefore have the same South Asian origin.

I suggest mtDNA R had earlier evolved somewhere slightly to the west, perhaps around the South China Sea, from where it expanded to Wallacea (P) and north (F and B) and west (ancestors of HV, JT and KU).

I suggest that you review the many South Asian R subclades and stop focusing only on the ones with big capital letters, what can only cause you to get most confused. R2 or R30 are as important as U (not KU: K is just a subclade of a sublcade of a subclade of a subclade of U) or P when we try to understand the origins of R.

terryt said...

"S to Australia and M29'Q in Melanesia (and M27 into island Melanesia especifically, a "moment" later)".

With what Y-chromosomes? There must have been some or the mtDNA lines wouldn't have survived.

"Their geographic circumstances may be comparable (island enviroment) but the two phenomenons are not otherwise related".

I'm afraid they are related. The sequence of expansions across Wallacea can be read as being the consequence of contulally improving boating technology. Each expansion basically progressively gets to more remote islands.

"P is derived from R and therefore must have split from R (not N) at (roughly) the same spot where other R sublineages did".

OK. Where was that? Did P expand all the way from India to Australia and New Guinea without leaving a trace, or did some Rs expand from Sunda through India and as far as Europe leaving plenty of haplogroups on the journey?

Maju said...

With what Y-chromosomes? There must have been some or the mtDNA lines wouldn't have survived.

Parthenogenesis. We all know that rae island lizards manage that way. ;)

Seriously. I don't know. Y-DNA M is a good candidate for the first wave but is restricted to Melanesia. F* in Australia maybe?

Sincerely no idea because the two trees don't seem to fix well with each other. Also Y-DNA could perefectly have drifted much more than mtDNA.

I'm afraid they are related. The sequence of expansions across Wallacea can be read as being the consequence of contulally improving boating technology. Each expansion basically progressively gets to more remote islands.

Not for me. They are too widely separated in time. One thing is the Paleolithic colonization of Sahul and another almost totally different the Iron Age Austronesian expansion. The latest is more related to the other Eurasian sailors of that time than to Australians and Melanesians. They did touch Melanesia and even exchanged some DNA over there but that's about it.

It's like relating them with the OOA because they also sailed to Africa.

OK. Where was that? Did P expand all the way from India to Australia and New Guinea without leaving a trace, or did some Rs expand from Sunda through India and as far as Europe leaving plenty of haplogroups on the journey?.

I'm more for the first. But in fact they did leave traces. There's some P in SE Asia, as you showed me recently: 4% in Sumba, 1% in Ambon, 1% in peninsular Malaysia. Additionally there are many uniquely SE/East Asian R lineages (and some undefined R*) that could have diverged in the same expansive pulse that generated P.

As the difference between R and P is just one SNP, we can't pinpoint intermediary stages. It was a fast migration, that's for sure.

terryt said...

"Sincerely no idea because the two trees don't seem to fix well with each other".

But they should do if we're considering the first people into any particular region. I agree that drift may have eliminated original haplogroups, but if we keep using this an an excuse for missing haplogroups we have no way of determining anything about haplogroup history.

"F* in Australia maybe?"

And exactly how much F* is found in Australia? C is a much more likely candidate.

"another almost totally different the Iron Age Austronesian expansion".

First time I've ever heard it called an 'Iron Age expansion'.

Maju said...

But they should do if we're considering the first people into any particular region.

Maybe but it is a headache. Just finidng patterns within each type of lineages is confusing, so imagine when you try to overlap the trees.

They just don't ben easily to that excercise.

I agree that drift may have eliminated original haplogroups, but if we keep using this an an excuse for missing haplogroups we have no way of determining anything about haplogroup history.

Not eliminated in most cases surely but can have caused them to become minor, like that F* you can see in Australia. Whatever the case, Y can drift a lot.

And exactly how much F* is found in Australia? C is a much more likely candidate.

I got C expanding almost parallely to NO, so some time after K and NOP. Of course that can be an error induced by the low resolution of this macro-clade.

First time I've ever heard it called an 'Iron Age expansion'.

From my old African history manual: "the men who first stepped on Madagascar already knew of iron".

Later: "At what time can be located the arrival of these protomalagassy to the Great Island? After the beginning of Iron Age (that is soon before Christian age if we accept that they departed from Indonesia); but before the Hinduization of Indonesia, which happened between the 2nd and 8th centuries A.D."

So something like 2000 years ago for the Malagassy. Iron Age in any case. "Yesterday".

For the rest (from Wikipedia): arrival to Philippines c. 5000-2500 BCE but expansion beyond that core only since c. 1200 BCE. Maybe not Iron Age at the beginning yet but Bronze for sure, except for the earliest stages at Luzon.

Notice anyhow that the developers of Austronesian outriggers were not the aboriginal Negritos of Philippines but the newly arrived Austronesians from Taiwan. And that Wallacea proper stands as largely non-Austronesian (still many pockets of the catchall "Papuan languages" concept) originally.

terryt said...

"They just don't ben easily to that excercise".

I think I've shown that they do. It's just that my conclusions conflict with your beliefs.

"Of course that can be an error induced by the low resolution of this macro-clade".

And almost certainly is.

"From my old African history manual: 'the men who first stepped on Madagascar already knew of iron'".

I'm fairly certain that is incorrect. They arrived in Madagascar after iron was used in parts of the northern hemisphere but the Austronesians certainly did not know of iron. I agree with your dates but calling them (certainly the Austronesians in the eastern hemisphere) 'Iron Age' is incorrect. In fact they didn't even carry bronze east beyond island SE Asia. Bronze may even have been introduced to the Philippines later. The Austronesians were basically 'Stone Age', although in time 'Iron Age'.

"And that Wallacea proper stands as largely non-Austronesian".

That is so genetically. But interestingly most of the languages spoken in Wallacea are Austronesian. In other words the Austronesians were simply yet another migration through the region.

Maju said...

I think I've shown that they do. It's just that my conclusions conflict with your beliefs.

I have your graph well visible here an does not inspire me the least, sincerely. All I can think is of the flagrant errors like considering all K-number clades as single distinct subclade of K and stuff like that.

I don't know if you're onto something but I really don't see it anywhere.

In fact they didn't even carry bronze east beyond island SE Asia.

That might because of availabality of resources too. It's speculated that Iron Age began where copper or tin were scarce. Iron is aboundant almost everywhere but maybe not in coral reefs and volcanic atolls in the middle of the Pacific.

Bronze may even have been introduced to the Philippines later. The Austronesians were basically 'Stone Age', although in time 'Iron Age'.

Well, whatever. Neolithic is more different from Paleolithic than from Industrial Capitalism probably. Advanced gregarious stratified societies.

That is so genetically. But interestingly most of the languages spoken in Wallacea are Austronesian. In other words the Austronesians were simply yet another migration through the region.

In Wallacea strictu sensu? Psah. Now they are but a lot of large "Papuan" pockets remain. It is not the origin of Austronesians in any case.

And also their pattern of expansion is totally different from that of the early colonists of Australia-Melanesia. They never tried to colonize Sahul but the islands around it. They were (are) a genuinely mariner people. I don't think you can say that of Australo-Melanesians.

terryt said...

"all K-number clades as single distinct subclade of K".

They are.

"That might because of availabality of resources too".

I've checked Madagascar and the conclusion seems to be that the inhabitants of Madagascar probably picked up iron an their voyage to that island. Probably from the Swahili coast of Africa.

"In Wallacea strictu sensu?"

Yes. Just a few 'Papuan' pockets remain.

"They never tried to colonize Sahul but the islands around it".

Because those islands were already occupied. Similar motivation seems to have influenced earlier migrations through Wallacea.

Maju said...

No! K1, K2, K3 and K4 are distinct unique sublineages of K, exactly like NOP, M or L. I don't care if you are confused: I refuse to discuss anything else re. Y-DNA with you until you understand that.

I've checked Madagascar and the conclusion seems to be that the inhabitants of Madagascar probably picked up iron an their voyage to that island. Probably from the Swahili coast of Africa.

Would make sense if the Swahili existed back then but guess they could have picked steel from East Africans or Indians (they are also supposed to have stopped by at Ceylon or whatever).

I'm kind of surprised anyhow that steel (not really soft iron what has been made in the "iron age" but rather steel) was widespread in Africa back then and it had not yet arrived SE Asia.

"They never tried to colonize Sahul but the islands around it".

Because those islands were already occupied
.

It's generally tought of a continent and that it was occupied was no big deal when others came after them - and if there were several MP waves of colonization of Sahul, then they did not mind that either: they colonized over the first arrivals's layers.

Indonesia was also ocuppied and that was no real barrier for them.

terryt said...

"It's generally thought of a continent and that it was occupied was no big deal when others came after them - and if there were several MP waves of colonization of Sahul, then they did not mind that either: they colonized over the first arrivals's layers.

Indonesia was also ocuppied and that was no real barrier for them".

It's very true that human groups have mixed on continents. But even on many Wallacean, Pacific, Indonesian and Melanesian islands the various haplogroups were often obviously originally associated with some introduced technology, and habitat exploitation. Probably the same for new haplogroups on continents. It takes several generations for haplogroups and technology to mix. In fact usually technology mixes first. We see the same process in action today with European expansion into regions already occupied. In fact mixing has not yet been completed in most of those places.

Anyway evidence indicates that many islands the Austronesians were able to occupy had become uninhabited, possibly at least partly through inbreeding during times of reduced inter-island contact. The consequent fruitful bounty is what probably encouraged their rapid expansion. Improved inter-island communication introduced by the Austronesians then eventually allowed other haplogroups from the islands on either side of Wallacea to spread as well. So it wasn't that they "colonized over the first arrivals's layers". In many cases the first arrivals actually came back from either side. It's not a simple story.

Maju said...

I pretty much agree with the first paragraph and specially with technology being faster spreading in most cases than genes.

Just that:

We see the same process in action today with European expansion into regions already occupied. In fact mixing has not yet been completed in most of those places.

I have no reason to think that Paleolithic people were posessed by the same kind of modern racism as Europeans were in the last centuries, moreso when migrations back then always happened between short distances (the long distance "travels" we see in maps happened only through many many generations), what would allow admixture with rather akin groups to happen easily.

In historical terms race did not exist. There was people of different complexions sure, the word "race" may even have existed (meaning stock or lineage, often ethnicity) but we don't see the extreme pseudo-biological racism of modernity anywhere in the past. In ancient history the Cherokee would have never been expelled but rather been assimilated quite naturally, for example. In fact you see that still happening in the "less modern" Spanish and Portuguese empires.

terryt said...

"I have no reason to think that Paleolithic people were posessed by the same kind of modern racism as Europeans were in the last centuries".

I wouldn't be so confident. Perhaps so, but the people of Melanesia were certainly very unwelcoming of arriving strangers, whatever the colour of their skin. Most visitors failed to survive to tell the tale. Racism? or unwillingness to share resources? Fiji was famous for killing visitors until well into modern history. In contrast Polynesians were generally welcoming of Europeans, probably because they represented an advanced technology useful to the indigenous population.

"In ancient history the Cherokee would have never been expelled but rather been assimilated quite naturally".

Maori tribes many times did their best to annihilate enemy tribes. They were not usually sucessful until they aquired firearms.

I suspect that in Paleolithic times co-existence was possible if ecological specialisation allowed separation of resource use. But in cases of competition for resources the desire for annihilation, if not the ability, existed.

Maju said...

I just reject that such a modern concept as "race" and derived "racism" existed or had practical implications in the Paleolithic. Your neolithic insular examples are not really good for me either. I really need hunter-gatherer examples to illustrate a hunter-gatherer reality. Tell me about Australian Aborigines, Inuits, Pygmies or Bushmen... how xenophobic and bellicose are these or how open and pragmatical maybe instead.

terryt said...

"Tell me about Australian Aborigines".

Xenophobic and bellicose in the early days. To other tribes, not just (justifiably) to Europeans. Not sure about Inuits, Pygmies or Bushmen. But many tribes in SE Asia were certainly xenophobic and bellicose, headhunters in fact. Perhaps it would be unfair to call it 'racism' though. Although the tribes would have justified their attitude by myth. And I wouldn't really call Melanesians Neolithic.

By the way. I've just looked at your tree again and thought it would look really incredible if you could redesign it to place the haplogroups regionally. It would then be much easier to see patterns, although I realise it would be a lot of work.

Maju said...

... although I realise it would be a lot of work.

Yah.

Plus many lineages are split into several regions (probably some more than are noted there but that's as far as I can go with my references).

Not any practical endeavour, IMO. You can try to do it yourself, of course.

terryt said...

"I just reject that such a modern concept as 'race' and derived 'racism' existed or had practical implications in the Paleolithic".

Perhaps so. But something similar does seem to be the best explanation for the apparent lack of hybridism between Neanderthals and modern humans.

Maju said...

You may have a point. But Neanderthals were not just another "race" (i.e. variety, stock) within our species but another species. It is very possible that like lions and tigers, we did not mix because our socio-ecological behaviours and even our biology were way too different. Tigers and lions of different gender find each other's mating behaviour confusing and maybe even annoying, additionally the offspring may well have been rather inviable for one or another reason. We do not know of any Sapiens-Neanderthal mixed society, we do know of many "interracial" ones.

terryt said...

"Plus many lineages are split into several regions (probably some more than are noted there but that's as far as I can go with my references)".

There are not actually too many of them.

The research posted at Dienekes has G/M12 as an early branch. If we start there we get the following:

Japan (3 lines):
G/M12
C/Z/M8
M7

East Asia (5 lines):
M11
D
M13
M21
E/M9

East Asia/Australia (1 line):
M10/M42

Australia (1 line):
M15

Melanesia/New Guinea (3 lines):
Q/M29
M27
M28

Andaman Islands (1 line):
M31/M32

Borderland (6 lines):
M44/M52
M46
M51
M49
M48
M47/M50

India/Pakistan (12 lines):
M41
M40
M39
M36
M35
M34
M33
M22/M25
M6/M16/M17
M5
M4/M30/M37/M18/M43/M45
M3

India/SW Eurasia (1 line):
M2

SW Eurasia/Australia (1 line):
M14

SW Eurasia (1 line):
M1

It doesn't really look as though they emerged from a single region already differentiated and then became regionally drifted. Most seem to have differentiated in situ after they'd spread.

Maju said...

How many of those lineages is separated from the root by a single SNP? Mutations need time to accumulate, so it's only logical to think that a longer stem prior to expansion means a longer embrionic period of lurking as private lineages ("colesecence" in genetic terms). It's like seeds and trees: the seed may have travelled long before the tree roots and grows.

They must have emerged at a single spot or small area necessarily, because they all have a single common ancestor. You have butchered Eastern Asia in many somewhat arbitrary subregions but I cannot accept that division. Specially the division between Sundaland, "borderland" and other East Asia. Obviously most lineages are shared: M9 (and E) are found in aboundance in SE Asia and they probably coalesced there and that is also true for most other Eastern lineages. Btw M31 is shared between South Asia and Andaman (SE Asia) and this is not the only shared lineage across the Indo-Pacific arch certainly.

Also you'd need to butcher South Asia in a similar manner in order to be able to compare. But no worries: the smaller the regions the larger the number of haplogroups shared across their "borders".

And if "by the paper at Dienekes" (what a reference!) you mean Alexe 09, I 'm pretty sure we can ignore it unless further studies go coincident somehow: they did nt provide defning SNPs for the different "clades" and it conflicts with all other studies. Plus the range of their research is too small to compare even remotely with other much more perfected mtDNA trees. Discard until new solid data comes, if that ever happens.

terryt said...

"Obviously most lineages are shared".

Not true. Many are fairly localised, and those that are more widespread tend to occur just in neighbouring regions.

"M31 is shared between South Asia and Andaman".

So?

"How many of those lineages is separated from the root by a single SNP?"

I'd bet a fairly large sum of money that there are many divisions and connections within mtDNA M yet to be discovered. For example the Australian and Melanesian versions may share a common ancestor. I completely agree that mtDNA M itself 'must have emerged at a single spot or small area necessarily'. And all of her individual descendant lines must also 'have emerged at a single spot or small area necessarily' but they didn't all leave from the same region. So, although 'the smaller the regions the larger the number of haplogroups shared across their 'borders' we can still, to some extent, narrow down the separate regions of origin for most of these haplogroups.

Maju said...

So? .

That artificially locating it at Andaman alone causes confusion, the same that placing M9/E in East Asia (to the exclusion of SE Asia) is confusing. There are many other examples.

I'd bet a fairly large sum of money that there are many divisions and connections within mtDNA M yet to be discovered.

I would not bet against but a lot has been advanced anyhow. There have been major efforts to organize the phylogeny of haplogroup M, that was intially rather under-researched because of Eurocentrism.

Even if a few haplogroups happen to be related, the overall structure should not change that much, IMO.

For example the Australian and Melanesian versions may share a common ancestor.

The two M lineages I have marked as Australian are related to Eastern Asia and Western Eurasia respectively, while the Melanesian ones are on their own. Your idea would seem logic on first thought but that's not what the known data says.

Anyhow, you have already argued for a two routes migration into Sahul yourself, so this should not surprise you, right?

I completely agree that mtDNA M itself 'must have emerged at a single spot or small area necessarily'. And all of her individual descendant lines must also 'have emerged at a single spot or small area necessarily' but they didn't all leave from the same region.

This is confusing. The last sentence seems self-contradictory.

What I see is 8 lineages clearly expanding immediately after the M node: 4 in South Asia, 2 in East Asia, one (at least one, info may be incomplete) shared between East Asia (Andaman) and South Asia and yet another shared between East Asia and Australia.

That is what appears as the first Eurasian expansive moment. Additionally there would have been many private lineages, of which some expanded at later moments, others never really expanded but did survive and many others surely that were just drifted out.

At the second moment we see much of the same (including some N lineages now): same geography, just that Sahul is now clearly also affected directly. The pattern continues in the next phases, even if R becomes the main actor then.

So, although 'the smaller the regions the larger the number of haplogroups shared across their 'borders' we can still, to some extent, narrow down the separate regions of origin for most of these haplogroups.

Well, it can be an interesting excecise to do but is somehow daunting for me. After all, I'm just an amateur.

terryt said...

"That artificially locating it at Andaman alone causes confusion".

I don't see why. It's found on the neighbouring mainland. What's the problem? After all the mtDNA must have come to the Andamans from somewhere.

"I would not bet against but a lot has been advanced anyhow".

You have often pointed out that haplogroup diagrams become dated every few months as new data are collected. We can expect further alterations in the future.

"Even if a few haplogroups happen to be related, the overall structure should not change that much, IMO".

The link to you provided to your mtDNA expansions shows you actually do agree with my comment that 'there are many divisions and connections within mtDNA M yet to be discovered'. And that 'they didn't all leave from the same region'. Therefore your comment, 'The last sentence seems self-contradictory' should no longer be true.

You show just a few M haplogroups appearing first: M22/25, M6 etc., M4 etc., M33, M31/32, M44/52, M10/42 and G/M12. Basically all from India but moving slowly into the mountains of SE Asia. Your postulated early arrival of G/M12 in Japan to me seems unlikely, unless it moved north with a C Y-hap. The big expansion north of Y-chromosomes from the Y-hap F line occurrs only once Y-hap NO has separated from Y-hap P. Otherwise I have no argument with your diagram 2.

Your diagram 3 again shows support for my above quotes. Your mtDNA N, though, reveals what may be a problem. You have N1/5, N9 and S separating before the other members of what may also be an unresolved clade. What about mtDNA Y/N9, N2, N12, N13, N14, N21, N22, X, A and R?

Diagram 4 again shows you agree with the deeper connection idea. So, what could have been the 'decisive technological, sociological or ecological advantage' that carried mtDNA R? The boating technology and subsequent sociological and ecological advantage that you mention carried the Japanese clade 'following the East Asian coast in a northwards direction'? Of course to accept that you need to abandon a belief in the original out of Africa expansion having been fueled by a pre-existing advanced boating technology.

Finally, I'm pleased to note that you no longer take mtDNA N2/W as being evidence for a first coastal migration to Australia but concede it probably represents a later connection of some sort.

Maju said...

It's found on the neighbouring mainland.

The neighboring mainland of Andaman is SE Asia, notably Burma and the Kraa isthmus, not India.

We can expect further alterations in the future.

Surely but I do not expect anything too radical, rather improvements in the knowledge of the detail, notably in regard of those haplogroups that are less well known.

The link to you provided to your mtDNA expansions shows you actually do agree with my comment that 'there are many divisions and connections within mtDNA M yet to be discovered'. And that 'they didn't all leave from the same region'. Therefore your comment, 'The last sentence seems self-contradictory' should no longer be true.

What link?

I would expect some improvement but, as said before, nothing too radical.

Your postulated early arrival of G/M12 in Japan to me seems unlikely...

It's a possibility because both lineages are most iportant in Japan. Though guess a more inland origin for G may be reasonable too. Whatever the case G is found basically in NE Asia (notably in Japan) and M12 is an almost exclusively Japanese lineage.

...unless it moved north with a C Y-hap.

D IMO. Japanese are very high in Y-DNA D.

What about mtDNA Y/N9, N2, N12, N13, N14, N21, N22, X, A and R? .

Private lineages by the moment. Their expansion only happened at later moments.

Can you differentiate between the private lineage/stem stage and the haplogroup/flower one? This is important, as at the stem stage they were just like any other of the maybe thousands of lineages that have gone extinct - only later is when they became lucky and consolidated by means of expansions (or just lurking at minimal levels in some cases).

So, what could have been the 'decisive technological, sociological or ecological advantage' that carried mtDNA R?.

Honestly I don't know for sure but considering how this lineage was central in the "conquest" of Europe (and largely West Asia too), I'd say that the atlatl or otherwise the use of throwing weapons, both for hunting as for warfare. We know that a different between Cromagnoids and Neanderthals was that the former hurled weapons through all their life while the latter did not. This could be a critical advantage both vs. other H. sapiens bands and vs. Neanderthals.

Finally, I'm pleased to note that you no longer take mtDNA N2/W as being evidence for a first coastal migration to Australia but concede it probably represents a later connection of some sort.

Along with other lineages (and ther may be more hidden in the "asterisk" grey areas of many papers), it clearly indicates that there was almost constant flow for a period between S/W Asia and Sahul. Obviously this flow must have gone via SE Asia, even if we haven't been able so far to localize the remains.

terryt said...

"at the stem stage they were just like any other of the maybe thousands of lineages that have gone extinct - only later is when they became lucky and consolidated by means of expansions".

But many of the lineages I mentioned didn't expand at all. Presumably they appeared early in the regions in which they are now found. Therefore you can't validly use the excuse that they emerged later.

"What link?"

You have a very short memory. This one:

"What I see is 8 lineages clearly expanding immediately after the M node".

"The neighboring mainland of Andaman is SE Asia, notably Burma and the Kraa isthmus".

And Burma, not the Kraa isthmus, is where M31 is found. Is Burma part of India or part of SE Asia? That's where problems arise with your refusal to regard the East as being as finely divided regionally as is Europe. Calling Burmese SE Asians, or Indians, is like calling Germans Spanish, or Greeks. If anything we can be reasonably sure that a proportion of Burmese have come south from near Tibet. Perhaps they brought in Y-hap D although the arrival of the Tibeto-Burman language seems too late to be related to the movement to the Andamans. These immigrants mixed with pre-existing populations made up of people from both SE Asia and India. So again, is Burma part of SE Asia or part of India? Or best regarded as a separate region?

Maju said...

But many of the lineages I mentioned didn't expand at all. Presumably they appeared early in the regions in which they are now found. Therefore you can't validly use the excuse that they emerged later.

Such lineages have actually never emerged (probably) - just that they have not gone extinct either and they provide some valuable info because of their mere existence.

You have a very short memory.

Damn shit! "The link to you provided to your mtDNA expansions" could be anything. Most probably PhyloTree, as it's the reference I have used in my latest mtDNA posts. I don't even know what specific post you're talking about nor why does it "show" what you claim it does.

I don't know if I do have short memory or not but you certanly have some sloppy kind of discursive style. I ask you, please, to be more precise (a Virgoan mania: precission, but the devil is in the details).

This one:

"What I see is 8 lineages clearly expanding immediately after the M node"
.

Well, you just copied text here but now at least I can remember what link was associated to that text (yah, I must be short-memoried ^^). I pointed to my own tentative reconstruction maps.

But how does that: shows you actually do agree with my comment that 'there are many divisions and connections within mtDNA M yet to be discovered'. And that 'they didn't all leave from the same region'.

I don't know how the maps can show that. The letter M was carefully placed in a circle in red color roughly on Bengal. Admittedly I am absolutely NOT certain about this specific placement but I am quite certain that at some point there was one rather small population where undifferentiated M was dominant (and maybe N seconday).

A possible more specific location could be around Jwalapuram, where it's known that some Homo spp. survived the Toba catastrophe with a technology that seems nearly identical to South African MSA (what strongly suggests H. sapiens). (You need a free sibscription to access that article, btw).

Maju said...

And Burma, not the Kraa isthmus, is where M31 is found.

Would need to check but I'm pretty sure that:

1. Burma, like Afghanistan, have not really been DNA sampled ever - because of political reasons, I guess. We know no shit about Burma's genetics from direct data.

2. M31 is found in South Asia, in Southern South Asia specifically.

Will look for the specific link but stop making up "information" that is not there.

Maju said...

THIS PAPER mentions M31 in several Indian populations: Karnataka, Andrah Pradesh, Maharastra, tribal Lodha, Chenchu and Lambadi, a novel M31c (distinct from Andaman) among Austroasiatic tribals (making up as much as 17% among the Bhoi).

It is a very interesting review of South Asian mtDNA M sublineages that you should read anyhow.

Ebizur said...

"Recently, Reddy et al. (2007) reported a novel sub-haplogroup M31c (188-234-282-9269-14152-15300-15935-16136-16311), under the haplogroup M31, among the Austro–Asiatic Khasi populations (∼5%), with a maximum frequency among the Bhoi (∼17%)."

As I recall, a high frequency (23.5%) of haplogroup M7a in a sample of Waar Khasians has been reported in this same paper by Reddy et al. (2007), although the sample sizes leave something to be desired. I'm pretty sure that Reddy et al. have defined haplogroup M7a using the same mutations that have been used to define the M7a clade found in the Japanese Archipelago. Is the M31 found in Khasians (especially the Bhois) really the same M31 that has been found in Andamanese? The Khasians in M31 must at least belong to their own subclade distinct from the M31 subclade(s) of the Andamanese, though, judging from the novel "M31c" designation.

terryt said...

"just that they have not gone extinct either and they provide some valuable info because of their mere existence".

And are probably more valuable for unravelling human migration patterns than are more widespread haplogroups.

I have indeed discovered that M31 is perhaps more widely spread than just around the Andaman Sea. I'll check your link later but you might also be interested in this May 2008 abstract:

http://www.ncbi.nlm.nih.gov/pubmed/18186508

Quote, 'The enhanced resolution of M31 allows for the inference of a more recent colonization of the Andaman Islands than previously suggested'. I have understood for a long time that the Andamans were settled long after humans had reached Australia.

And this (from 2005):

http://www.andaman.org/BOOK/reprints/thangaraj/rep-thangaraj.htm

Quote, 'In our survey of -6500 mtDNA sequences from mainland India, none of the M lineages carried the coding region mutations specific to M31 and M32 (6). Furthermore, none of the haplogroup M complete sequences reported so far share any of the mutations that define M31 and M32, suggesting that these two haplogroups are likely to have evolved in situ on these islands'.

And while we're at it, another one:

https://www.sciencemag.org/cgi/content/full/311/5760/470a

Quote, 'The complete sequencing data illustrate that the Rajbanshi mtDNA differs markedly from the Andaman islanders–specific haplogroup M31 (here renamed M31a to distinguish the Rajbanshi branch, which we name M31b) by 11 specific coding-region mutations. This provides new phylogeographic evidence for an early divergence of these lineages on the Indian subcontinent and reduces the estimated time of isolation of the Andamanese'.

Again the authors suggest that the Andamans were relatively recently settled, agreed from India.

Maju said...

@Ebizur: my main pioint was anyhow to make clear that M31 (and hence also the larger haplogroup M31'32) is not exclusively Andamanese but also found in South Asia in several different populations. I fear I cannot clarify your doubts about the specific subclades beyond what that paper and the one mentioned by Terry say.

...

@Terry:

And are probably more valuable for unravelling human migration patterns than are more widespread haplogroups.

Depends. All info is valuable but it is extremely difficult, if not just plainly impossible, to trace the geographical history of a private lineage that has no relatives anywhere since the M/N expansion. In this sense, more widespread and branched out lineages are usually more informative in fact.

...you might also be interested in this May 2008 abstract...

Yah. Though I was already treating them as a single M31'32 basal haplogroup anyhow.

And this (from 2005)...

Obviously this is obsolete.

And while we're at it, another one:

https://www.sciencemag.org/cgi/content/full/311/5760/470a

Quote, 'The complete sequencing data illustrate that the Rajbanshi mtDNA differs markedly from the Andaman islanders–specific haplogroup M31 (here renamed M31a to distinguish the Rajbanshi branch, which we name M31b) by 11 specific coding-region mutations. This provides new phylogeographic evidence for an early divergence of these lineages on the Indian subcontinent and reduces the estimated time of isolation of the Andamanese'
.

This certainly adds more detail to the M31 phylogeny.

Again the authors suggest that the Andamans were relatively recently settled, agreed from India.

Well, very relatively. I get a simlar age for M31 and H expansions, what would still put the M31 branching at c. 40 kya possibly. Still the sister clade M32 may represent an older layer in the same islands, if we take the M31'32 split as reference, or may be instead a later related expansion from the same continental origin, whose mainland relatives have since been drifted out.

An alternative explanation could be that the lineage actually evolved in Andaman and expanded from there to the mainland.

Ebizur said...

Maju said,
"my main pioint was anyhow to make clear that M31 (and hence also the larger haplogroup M31'32) is not exclusively Andamanese but also found in South Asia in several different populations. I fear I cannot clarify your doubts about the specific subclades beyond what that paper and the one mentioned by Terry say."

My worry is that the authors of the paper to which you have linked (http://www.ias.ac.in/jgenet/Vol88No1/127.pdf), at least in the "Haplogroups M31 and M32" section of this paper, have referred to the presence of M31 both in the Andamans and in mainland India without explicitly indicating any shared mutations:

"Haplogroups M31 and M32Rajkumar et al. (2005) defined another new haplogroup, M31, characterized by a substitution at A5319G, in four samples, of Karnataka (Iyengar Brahmin and Lingayat), Andhra
Pradesh (Komati) and Maharashtra (Pawar). Thangaraj et
al. (2005a) designated two new clades M31 (4907-11176-15440-15530) and M32 (3817-9064-12189-15754) by the analysis of the complete mtDNA sequence of Onge and Great Andamanese and suggested an in situ origin of these haplogroups on the Andaman Island. The coalescence time of
M31 and M32 was estimated to be 3000 ± 2000 ybp and 12000 ± 4000 ybp, respectively (Thangaraj et al. 2005a)."

I'm just curious about the basis for connecting the continental Indian M31 of Rajkumar et al. (2005), which is described as being characterized by a substitution at A5319G, with the Andamanese M31 (4907-11176-15440-15530) of Thangaraj et al. (2005a). The fact that both these sources have been published within the same year (2005) makes me suspect a possibility of nomenclatural conflict.

Maju said...

Good hit, Ebizur!

At PhyloTree Rajkumar is not mentioned at all in regard to M31 but if you scroll around you see that his M31 is in fact M6'16'17.

Still the other paper, Palanichamy 2006, while not mentioned at PhyloTree seems fully in agreement with the main M31 sequence, suggesting a NE Indian (and Burmese???) connection for this Andamanese lineage. The NE Indian sublineage is exactly the "proposed" M31b at PhyloTree (but seems to have a good deal of support: 4 papers!).