Friday, February 1, 2008

NRY Haplogroup E3a: A Multidisciplinary Proposal of Its Origins

Having touched on this matter in an earlier posting but in a broader topic involving more lineages, this post will focus on one lineage—haplogroup E3a.

Geological and geographical evidence

~23 ky ago: Ogolian aridity takes hold, and renders much of north Africa desertified.

The advent of Hg E3a's TMRCA essentially coincides with rise of the Ogolian aridity.
  • PN2 clade (E3) bearers in the vicinity of the general expanse straddling Sudanese-Central African Republic -Ugandan-Kenyan region [get a map aid, if necessary] give rise to E3a ~ between 21 and 18 ky ago [see Semino et al. 2004 for TMRCA dates, pending additional or new info]; E3b-M35* would have likely arose relatively earlier than E3a* [as evidenced by its near absence in some the populations that carry this], sometime prior to the Ogolian and the LGM period.
Hence, indications are that the initial territory of an ancestral Hg E3a bearing group would have likely started south of the Sahara, at a time when:
  • The Ogolian aridity rendered much of North Africa with adverse weather conditions, turning much of that general region into desert. The Sahara at this time, extends south beyond its current boundaries to a certain point, likely as far as a little beyond the Niger bend.
  • The above mentioned geological conditions in the aforementioned geography indicate that the population density in those arid regions would have been quite sparse—meaning, that most populations would have then been concentrated south of the broad desert region.
Reference to the geographical extent of this desertification is laid out here in maps: AFRICA DURING THE LAST 150,000 YEARS [clickable]

Archaeological evidence
Archaeology indicates that much of west Africa during the height of the Ogolian aridity was abandoned, with many of the populations living therein in the pre-Ogolian periods seeking refuge in the last vestiges of vegetation beyond the desert, which at the time - as noted above, extended to as far as beyond the Niger River bend, and possibly few moving to the far northwestern coast of Africa, if not crossing over to the Iberian or Southwestern European region [perhaps to this end, possible case can made from findings in Goncalves et al. 2005?], which was then a major refuge center for European groups trying to cope with the side-effects of the LGM in their neck of the woods. Vivid archaeological indicator of this comes from:
  • "After a favourable climatic period, characterised by relatively dense and diversified Palaeolithic occupations, the arid Ogolian begins locally around 23000 years BP and is represented at Ounjougou by a significant depositional and archaeological hiatus." — Aziz Ballouche [see: Link ]
Incidentally, when this was going on in west Africa, arid conditions extended all the way to the horn-shaped coast of the African Horn, possibly encouraging populations to reside relatively more inwards—that is, away from that aridified horn-shaped coastal region; rather, populations of that region were likely encouraged to seek refuge in the general region straddling southern Sudan, Ethiopia, Kenya and Uganda or even further—i.e. region straddling Uganda, Kenya, and Tanzania. [See AFRICA DURING THE LAST 150,000 YEARS, for visual aid—geology conditions via geography in the Ogolian period]

It is again, from archaeological evidence, that we get a picture of when repopulation events make their appearance, and in what pattern they occur, i.e. things which are all relevant to understanding the circumstances and nature of how the Hg E3a bearing group makes it appearance in west Africa:
  • Consequently, it has to be seen in the context of heavy rainfalls and a resettlement of the vegetation cover, during the 10th millennium BC, that a **new population** arrives on the Plateau of Bandiagara." — Human population and paleoenvironment in West Africa [see: Link ]
...and we see how these repopulation events occur sequentially from archaeological record, showing a movement from an east-to-westward direction, and thereafter a south-to-northward movement in west Africa itself:
  • From 30,600 to 10,000 BC: "A cultural flow, from the southeast of Subsaharan Africa and to the Sahara, could explain the diffusion of the microlithic industries all the way through West Africa. We observe them initially in Cameroon at Shum Laka (30.600-29.000 BC), then at the Ivory Coast in Bingerville (14.100-13.400 BC), in Nigeria in Iwo Eleru (11.460-11.050 BC), and finally in Ounjougou (phase 1, 10th millennium BC)." — Human population and paleoenvironment in West Africa [see: Link ]
This east-to-west movement shows that a new group of people moved into west Africa, because the "microlithic traditions" that appear at Shum Laka had been in existence for as long as ~ 30 ky ago, and yet didn't show up in west Africa proper until ~ 14 ky ago, a late date which works well within the confines of the TMRCA age of Hg E3a. So the people originally responsible for that "microlithic tradition" at Shum Laka were not the people who likely brought that tradition to west Africa, but by a new group of people who passed through Shum Laka. These new group of people first appeared in vegetation holdout regions beyond the Niger bend, and gradually moved northward, eventually reaching to regions as far as that of Ounjougou, as evidenced by the above carbon-dated archaeological finds. It should be remembered that the Ounjougou region was abandoned in the intense desertification period.

*Details of the "microlithic traditions" had been provided in a related earlier post: P2 Clades: The Arrival of E3a and E3b Haplogroups [clickable]

Genetic evidence
The east-to-west movement of new groups [the candidates being E3a bearing group] into west Africa observed above, seems to be lent further support from genealogical finds.

Observations of genetic differentiation within contemporary E3a bearing populations, suggests that the earliest Hg E3a bearing groups, carrying mainly M2, P1 and M180 mutations devoid of the M191 mutation, had moved to the far west region of Africa...from an eastern-oriented origin:

Consider for one,
  • Although haplotypes 22, 24, and 41 were probably all involved in the Bantu expansion, the processes that determined the current distribution of these haplotypes in the Sudanese belt (a region south of the Sahara extending from western to central Africa) seem to have been more complex and perhaps involved a separate expansion.
  • In particular, haplotype 24 and its derivative, haplotype 22, harbor opposite clinal distributions in the region, a finding that is at odds with the hypothesis of a parallel dispersion of these two lineages in the area.
  • Haplotype 22 has a frequency of 23% in Cameroon (where it represents 42% of haplotypes carrying the DYS271 mutation), 13% in Burkina Faso (16% of haplotypes carrying the DYS271 mutation) and only 1% in Senegal (Semino et al. 2002), whereas haplotype 24 reaches its highest frequency (81%) in Senegal (Semino et al. 2002).
  • A possible explanation might be that haplotype 24 chromosomes were already present across the Sudanese belt when the M191 mutation, which defines haplotype 22, arose in central western Africa. Only then would a later demic expansion have brought haplotype 22 chromosomes from central western to western Africa, giving rise to the opposite clinal distributions of haplotypes 22 and 24.
Source: Cruciani et al. 2002, A Back Migration from Asia to Sub-Saharan Africa

This piece shows that the oldest Hg E3a bearing populations, characterized by high frequencies of haplotype 24 and relatively modest frequencies of haplotype 22, remain localized in West Africa; the contrasting distribution and frequency patterns of these haplotypes suggest that the higher the population with haplotype 22, the less older it is in its genetic composition relative to those bearing less haplotype 22 in a gene pool dominated by the relatively older haplotypes, like haplotype 24. Senegal appears to have the highest concentration of haplotype 24, and interestingly, the lowest concentration of haplotype 22...thus, making it likely the oldest Hg E3a bearing group.

The Senegalese samples also have other peculiarities that further support the aforementioned idea of their being the oldest Hg E3a bearing group:
  • Hg E3 (P2 or PN2 clade):

    Bantu (South Africa) - E3* = 1.9%, Senegalese - E3* = 2.9%, Ethiopian (Amhara) - E3* = 10.4%, Ethiopian (Oromo) - E3* = 12.8% in the ascending order.
...the 'older' groups appear to have relatively wider distribution of E3, as the Senegalese sample in particular seems to indicate vis-a-vis other Niger-Congo speaking groups elsewhere; a lineage which has by far the highest frequencies and distribution in Ethiopia [which is in East Africa fyi]

The Senegalese sample shows considerably more E-M35* than that from other Niger-Congo speaking groups of sub-Saharan West Africa, which are not immediately neighbouring major "ancient" African groups in East Africa or like Khoisans.

Hg E-M35* lineages:
  • In descending order…Ethiopian (Oromo) - E-M35* = 19.2%, KhoiSan (South Africa) - E-M35* = 16.7%, Ethiopian (Amhara) - E-M35* = 10.4%, Berber (North-Central Morocco) - E-M35* = 7.9%, Berber (Southern Morocco) - E-M35* = 7.5%, Senegalese - E-M35* = 5%, Tunisian - E-M35* = 3.4%, Algerian - E-M35* = 3.1%, Arab (Morocco) - E-M35* = 2.3% , Burkina Faso -E-M35* = .9%
The Senegalese group is also the only sub-Saharan 'predominantly-Niger-Congo' speaking group that shares the rare microsatellite motif with Ethiopians on the E-M78 bearing chromosome:

It is interesting that both E-P2* and E-M35* and their derivatives, E-M78 and E-M123, exhibit in Ethiopians the 12-repeat allele at the DYS392 microsatellite locus, an allele scarcely seen (Y-Chromosome STR Database), especially in other haplogroups and other populations (A.S.S.-B., unpublished data).

In addition, the Ethiopian DYS392-12 allele is usually associated with the unusually short DYS19-11 allele, which is typical of this area. These findings are not easily explained. One possible scenario is that an ancient differentiation of the E-P2 haplogroup occurred in loco (East Africa). However, this also implies a low mutability of the associated microsatellite motif (DYS392-12/DYS19-11). Alternatively, the microsatellite motif may be due to homoplasy.The first scenario is more likely, since this unique microsatellite haplotype occurs in E-P2*, E-M35*, and E-M78 but is almost *absent in all other haplogroups and populations*.

In addition, the high stability of the DYS392 locus (Brinkmann et al. 1998; Nebel et al. 2001) and of the shorter alleles of DYS19 (Carvalho-Silva et al. 1999) has been reported elsewhere.

Moreover, the observation that the derivative E-M78 displays the DYS392-12/DYS19-11 haplotype suggests that it also arose in East Africa. This is illustrated by the microsatellite network (fig. 3, shaded area), which reveals that the Ethiopian branch harboring DYS392-12 is not shared with either Near Eastern or European populations. — Semino et al. Origin, Diffusion, and Differentiation of Y‐Chromosome Haplogroups E and J, 2004.

The Ethiopian sample may not share the said allele with those populations mentioned, including the northwest African samples as far as the present author can tell, but it does share the said allele with the Senegalese sample [see fig. below], which would suggest that the Senegalese M78 derivative didn’t come from interaction with its northwest African neighbors; rather, they may well be relics of ancient migrations from east to west.

 -
For larger image, click on this link: Source

So what do all these genetic indicators say?

Senegalese sample has the highest concentration of haplotype 24 and the lowest concentration of haplotype 22 - making it the most likely ancient Hg E3a bearing group, it also has higher Hg E3 vis-a-vis other Niger-Congo speaking group of west Africa, has the higher Hg E-M35* vis-a-vis other Niger-Congo speaking groups not neighbouring major "ancient" populations of east Africa or south Africa [aka KhoiSans], and has the rare Ethiopian/East African 12-repeat allele at the DYS392 locus.

The advent of Hg E3a itself coincides with the rise of the Ogolian aridity, which rendered much of North Africa to as far as beyond the Niger River bend largely inhabitable, inducing populations to be densely populated south of the Sahara, which is therefore the likely region of the rise of TMRCA of Hg E3a bearers.

Bearing "rare" lineages predominantly found in east Africa - i.e. the likely point of origin, along with sequential archaeological evidence for [east-to-west and thereafter, in situ west African south-to-north] repopulation events in west Africa, much of which was abandoned in the Ogolian desertification, show that the earliest Hg E3a bearers— which finds expression in Senegalese samples - could not have arose in situ west Africa, but originated in an eastward oriented geography and migrated to west Africa, as the Ogolian aridity relaxed, bringing along with them new microlithic traditions picked up from the Shum Laka region, settled therein and thereafter underwent demic expansion, resulting in the "high diversity and frequency" of the Hg E3a distribution in west Africa.

All this becomes even more apparent, when one realizes that Hg E3 (PN2 clade) is quite rare even in sub-Saharan west Africa, but even rarer in the Sahara. It has the highest frequency in sub-Saharan east Africa [Ethiopia], where it is deemed to have originated— and so, the likelihood of the emergence of Hg E3a in a region with significantly higher frequencies of Hg E3 makes more sense than vice versa.

The above are all the reason the present author found it particularly simplistic in a discussion that he was a part of in the past, a strange question asked by a discussant, as to how one could possibly reconcile 'high diversity and frequency' of Hg E3a in west Africa with the prospect of its origins outside of west Africa; that question was no doubt a mentality born out of lack of familiarity with genetics and how multidisciplinary science works.

Update!

Getting back to the point raised earlier about the rare 12-repeat allele at DYS392 by Semino et al...

It is interesting that both E-P2* and E-M35* and their derivatives, E-M78 and E-M123, exhibit in Ethiopians the 12-repeat allele at the DYS392 microsatellite locus, an allele scarcely seen (Y-Chromosome STR Database), especially in other haplogroups and other populations (A.S.S.-B., unpublished data). 

In addition, the Ethiopian DYS392-12 allele is usually associated with the unusually short DYS19-11 allele, which is typical of this area. These findings are not easily explained. One possible scenario is that an ancient differentiation of the E-P2 haplogroup occurred in loco (East Africa). However, this also implies a low mutability of the associated microsatellite motif (DYS392-12/DYS19-11). Alternatively, the microsatellite motif may be due to homoplasy.The first scenario is more likely, since this unique microsatellite haplotype occurs in E-P2*, E-M35*, and E-M78 but is almost *absent in all other haplogroups and populations*.

In addition, the high stability of the DYS392 locus (Brinkmann et al. 1998; Nebel et al. 2001) and of the shorter alleles of DYS19 (Carvalho-Silva et al. 1999) has been reported elsewhere.
Moreover, the observation that the derivative E-M78 displays the DYS392-12/DYS19-11 haplotype suggests that it also arose in East Africa. This is illustrated by the microsatellite network (fig. 3, shaded area), which reveals that the Ethiopian branch harboring DYS392-12 is not shared with either Near Eastern or European populations. — Semino et al. Origin, Diffusion, and Differentiation of Y‐Chromosome Haplogroups E and J, 2004.

To which the present author responded...

The Ethiopian sample may not share the said allele with those populations mentioned, including the northwest African samples as far as I can tell, but it does share the said allele with the Senegalese sample [see fig. below], which would suggest that the Senegalese M78 derivative didn’t come from interaction with its northwest African neighbors; rather, they may well be relics of ancient migrations from east to west.

Cruciani et al. in their 2007 publication, Tracing human male movements in northern/eastern Africa and western Eurasia: new clues from Y-chromosomal haplogroups E-M78 and J-M12, had this to say about the 12-repeat allele in question:

“An eastern African origin for this haplogroup was hypothesized on the basis of the basis of the exclusive presence in that area of a putative ancestral 12-repeat at the DYS392 micro satellite , found in association with E-M78 chromosomes (Semino et al. 2004). North-eastern African populations were not represented in that study.

In order to test this hypothesis, we analyzed for DYS392 a geographically widespread subset of the E_M78 chromosomes here identified. We observed that the DYS392 12-repeat allele is associated with the majority of the chromosomes belonging to the north-eastern African E-V12* (15 out of 18) and to the eastern African E-V32 (21 out of 23), with about half (9 out of 21) of the E-V22 chromosomes (both in eastern and north-eastern Africa), with a few of the European E-V13 (2 out of 23) and with some north-African E-V65 (3 out of 16) chromosomes.

These findings show that the DYS392 12-repeat allele is common in different regions characterized by high frequencies of E-M78, and suggest that it was most likely generated by multiple mutation events occurring in different UEP-defined sub-haplogroups. Thus, the DYS392 allele distribution is not informative to infer the place of origin of E-M78 chromosomes.”

Cruciani et al. point out that Northeast Africa wasn't represented in the 2004 Semino et al. study, presumably to emphasize the point that its detection therein would have shown that the 12-repeat allele in question wasn't confined to sub-Saharan east Africa, particularly Ethiopia. It is certainly true that northeastern Africa was not represented, but in fact, a close examination of Semino et al.'s 2004 study, shows that the 12-repeat allele wasn't confined to sub-Saharan east Africa; it makes single or very low appearances in the "Near Eastern" and "European" samples, and interestingly none in north African [sans northeast Africa] samples.

Cruciani et al. on the other hand, did detect the 12-repeat allele in north African [sans northeast Africa], but still in relative low frequency (3 out of 16). The same applies to their findings in European samples, where it appears to be yet rarer than that of the north African samples. Considerable frequency is however noted in their northeast African sample, and consistently, in sub-Saharan east African samples. So, even going by Cruciani et al.'s findings, the greater distribution of the 12-repeat allele in eastern Africa in general, would argue for its introduction from eastern Africa.

Cruciani et al. reckon that simply because the 12-repeat allele is not confined to east Africa, it therefore doesn't allow one to infer the place of origin of E-M78 bearing chromosomes. This ignores the point just made about its relatively rare occurrence outside of east Africa. Cruciani et al. reach this conclusion on the understanding that the 12-repeat allele's appearances in distinct sub-clades invokes homoplasy or parallel mutational events, and indeed, Semino et al. 2004 did not rule out that possibility, but unlike Cruciani et al. they took additional material into consideration:

1) "It is interesting that both E-P2* and E-M35* and their derivatives, E-M78 and E-M123, exhibit in Ethiopians the 12-repeat allele at the DYS392 microsatellite locus, an allele scarcely seen..."

So while there is notable presence of 12-repeat allele bearing E-M78 chromosomes in northeast Africa as well, only in sub-Saharan east Africa does one come across not only considerable frequencies of E3-P2 [PN2 clade], but also the more immediate precursor of Hg E-M78, i.e.— Hg E-M35*, which bear this 12-repeat allele. As far as the present author can discern at this point, pending introduction to studies that suggest otherwise, the ensemble of at least three different PN2 macro-cladesincluding the ancestral onesbearing the said "12-repeat allele" only occurs in sub-Saharan east Africa. Case in point:

"...first scenario is more likely, since this unique microsatellite haplotype occurs in E-P2*, E-M35*, and E-M78 but is almost *absent in all other haplogroups and populations*."

2) "One possible scenario is that an ancient differentiation of the E-P2 haplogroup occurred in loco (East Africa). However, this also implies a low mutability of the associated microsatellite motif (DYS392-12/DYS19-11)...

In addition, the high stability of the DYS392 locus (Brinkmann et al. 1998; Nebel et al. 2001) and of the shorter alleles of DYS19 (Carvalho-Silva et al. 1999) has been reported elsewhere."

So again, while not ruling out independent parallel microsatellite mutational events across different PN2 clades, the repeat occurrence of the 12-repeat allele across different PN2 clades, including the ancestral ones, coupled with the possibility of "low mutability", which translates into "high stability" of the locus in question, indicates that the 12 repeat could well have also been inherited from a precursor PN2 clade(s). If the latter is the case, certainly sub-Saharan East Africa as the likely place of origin for the allele, can be put forward as a strong argument. And in getting back to the appearance of this allele in sub-Saharan or Sahelian west Africa, in the Senegalese sample for instance, would tend to favor an east African origin at some point in history, rather than as a remnant of interactions with coastal northwest Africans, where the allele is rare, even going by Cruciani et al.'s findings.