I recently posted on Eurasian mtDNA. But I had to rethink my notes and conclusions on super-haplogroup N, largely on review of the somewhat complex (and badly named) strucure of this clade (see Ian Logan's mtDNA site) and the realization that haplogroup W is at least as South Asian as West Eurasian (see Metspalu 2004), and has its highest concentration in Northern Pakistan.
So, after due rethinking, I get the following haplogroup N map:
In other words, the true structure of haplogroup N has 6 subclades:
- macro-R: including huge and widespread haplogroup R (South Asian by origin without any serious doubt), some Indian N* and East Asian N9 (that includes Y)
- macro-W: including South & West Eurasian W, as well as Australian N* lineages. W as such is most frequent in Pakistan and hence I have depicted it there. In West Eurasia is most directly linked with West Asia.
- macro-X: including West Eurasian X and N1, as well as some of the Indian N*. X is also found occasionally in North Asia and notably, in the distinct X2a subclade, among the Natives of North America (apparently arrived there associated with Clovis culture). Nevertheless the highest diversity of X by far is in West Asia/North Africa, so it must be considered a West Eurasian clade. N1 includes several subclades, among which is haplogroup I of apparent ancient presence in Europe (always in small ammounts).
- S: Australian specific
- A: East Asian and American, with a long sequence of SNPs at the root
- Ket N*, with an even longer sequence of SNPs at the root
Origins and spread:
I would think that macro-R and macro-W appear to have originated in South Asia, maybe in NW and/or Western coastal South Asia, if we are to follow their modern distribution patterns. From there they would have migrated to the East, maybe with M and other N-derived clades (A, S), seeding N9 in East Asia, as well as W-related N* in Australia. The handful of R-derived clades in Eastern Asia and Sahul may have gone with them or soon after (broadly in the same process anyhow).
Macro-X instead appears to be most diverse in Western Eurasia, maybe indicating an older migration in this direction than the one starred by R subclades (U, R0 and JT) in the same direction. I am not sure in relation with which one would M1 (and possibly other M subclades such as M48) migrated in westward direction. It is possible then that X/N1-related Indian N* represents a backflow from West Eurasia (but hard to say with certainty).
Overall I'd say that N is most diverse in South Asia (if anywhere) and that, contrary to my previous post, it must have coalesced along with M in the subcontinent. Nevertheless it does seem to have a more western center of gravity than M, possibly indicating that the N-M split (from L3) happened in relation to events and geography of South Asia. It is possible that (broadly) M represents those that advetured further into the subcontinent, while N those that remained closer to the Indus river and Arabian sea area. Nevertheless the ability of N carriers to reach out to East Asia and Australia, among other remote places, along with a majority of M carriers, indicates that these geographic boundaries were by no means strict or rigid and that the different early Eurasian populations were admixing all the time within this process.
Important note/caveat: defining N as South or West Eurasian is largely a matter of where did macro-W coalesce. I have here suggested that W could be South Asian by origin but the philogeny of the haplogroup is complex and by no means clear (see Ian Logan's notes on W) . The presence of related macro-W "N*" in Australia would suggest a migration via South and SE Asia (no other explanation is possible) and the presence of W in Thailand also would suggest a South Asian center of gravity. But the matter is not fully clear and, if macro-W would be of West Asian origin after all, my previous model, with N (but not R) coalescing in West Asia, could stand.