Monday, December 26, 2011

Genetic Profile: Maghrebi populations are Not "Satellite-Europeans" in Africa


Many observers will note that at some point in time or another, they witnessed (if not perpetrators themselves) internet inquiries around the origins of Maghrebi populations, whom as indigenous Africans (let's face it; they are not endemic to any other region outside of Africa), are notable not only for having elements among them who are relative "outliers" in terms of epidermal pigmentation, but also for their own original distinctive language family, properly called Tamazightotherwise inconsiderately known in English lingo as "Berber". These populations have wittingly or unwittingly been subjects of racism, in the sense that their "African-ness" have been questioned primarily because of appearance of those populations among them with "light skin", "depigmented skin" or "white skin'the terms change depending on who's speaking. If for example, they were unanimously dark-skinned like the remainder of African populations, it is highly doubtful that "western" researchers would even consider tacitly passing them off as some sort of lost-Europeans (or extensions of Europeans) in Africa, notwithstanding any genetic similarity to the patterns seen today of Maghrebi populations.

They have as such, become the subject of all sorts of speculation and dubious science, from claims that they are descendants of the so-called "Mechtoids" of the Epi-Paleolithic, descendants of European "Cro-Magnons", descendants of Vandals, "hybridized" descendants of some African element that met up with European females on the north coast of Africa, underwent some "jungle fever situation" and created the "Berbers" out of that, descendants of Arabian peninsula "Arabs",  to unconfirmed "connections" such as that of descent from the people that the ancient Egyptians referred to as "Tamahou". To cut through all that noise, we will undertake the most objective scrutiny of Maghrebi gene pool, with the aid of a study that just came out, titled "Human Alu insertion polymorphisms in North African populations", by Cherni et al. (2011), and the recall of another study, which had already been discussed on this site, under the heading of "Population Relationships in the Mediterranean Revealed by Autosomal Genetic Data (Alu and Alu/STR Compound Systems)", by Gonzalez-Perez et al. (2010).


The discussion will open with a quote from Fadhlaoui-Zid et al., in a study that just came out, because it expresses a concern that has come up in previous studies, about the astonishing lack of insight into a complete picture of Maghrebi gene pool, considering that the Maghreb is one of the most frequently sampled regions of Africa by 'western' research teams:
"However, the lack of genetic information of certain geographical areas and the focus of some studies in parts of the North African landscape have limited the global view of the genetic pool of North African populations." - Fadhlaoui-Zid et al. (2011), Mitochondrial DNA Structure in North Africa Reveals a Genetic Discontinuity in the Nile Valley.

On cue here, in terms of traditionally restricted and highly selective sampling choices of "western" researchers in coastal Northern African landscape, which has been pointed out on this site on occasions. Limitations of such choices have found expression in such studies focusing on western African Imazighen and/or Arabized populations, like say, the Cherni et al. 2005 finding in the work "Female gene pools of Berber and Arab neighboring communities in central Tunisia: microstructure of mtDNA variation in North Africa", wherein a coastal north-based Tamazight-speaking ethnic group "unexpectedly" displayed a "sub-Saharan" component that was substantially more extensive than its neighboring coastal Arabized counterpart, or that of Pereira et al. (2010) in the work "Linking the sub-Saharan and West Eurasian gene pools: maternal and paternal heritage of the Tuareg nomads from the African Sahel", wherein the "sub-Saharan" Kels Tamasheq of the Malian area departed from previous DNA reports (see Rando et al (1998), and/or Claudio Ottoni et al. (2009) for example) when samples turned up with extensive so-called "Eurasian" component, respectively.

Aside from selective geographic sampling choices, "western" DNA research teams are on record for making wide-reaching claims about Maghrebi populations being "more related to Europeans than Africans [mainly "sub-Saharan" Africans]", despite repeated demonstrations of "asymmetrical" uniparental gene pool patterns of said populations. Recurring DNA reports about lopsided "Eurasian" constitution of Maghrebi mtDNA or maternal gene pool, have undoubtedly supplied a platform for said researchers to make such claims. On the other hand, consistent DNA reports of a lopsidedly predominant "African" component, on the paternal side, have generally been overlooked. It is any wonder how any population characterized by such asymmetrical genetic pattern can be simplistically adjudged as being more related to peoples of a different continent (Europe) than those of its actual geographic origin (Africa).

It is a common practice of "western" researchers to analyze Maghrebior coastal Northern African in generalpopulations against fairly geographically-representative or comprehensive sampling of Europe and/or from the so-called "Near East", but against only either fairly scattered [i.e. geographically] small samples or single-locality sampling of the so-called "sub-Saharan" region. Western Africa has generally been afflicted by such sampling choices than any other location. As such, DNA reports portray abrupt changes in patterns across populations in the Western hemisphere of the African continent.

To give an example of the aforementioned skewed sampling choices, consider this from Cherni et al. (2011):

"The objective of this study is to compare the genetic structure of some North African populations with respect to the other populations of North Africa, Western, Eastern and Central Europe, by using MDS andAMOVA methods."

Notice how Europe is extensively covered, for a relatively small landmass, comprising "Western, Eastern, and Central Europe". Whereas, Maghrebi populations are often compared with one or just a few scattered and highly geographically-fragmented sampling of so-called "sub-Saharan" Africathis, for a far larger landmass and a continent reputed to sport the greatest human diversity anywhere. Even on a regional level, i.e. Western Africato repeat, one population is often picked from an area below the coastal-north Maghrebi territories and taken for granted as a "representative" population of the remainder of entire region, while "Afro-Asiatic"-speaking populations generally situated in between coastal north Maghrebi populations and those on the southern fringes of Western Africa, are passed off as "North Africans", rather than be counted along with their "Niger-Congo"-speaking andon occasions"Nilo-Saharan"-speaking neighbors, under the "sub-Saharan" banner.

"48 from Thala which is a town in western Tunisia located in the mountainous backbone of Tunisia. It is a municipality of 13,968 inhabitants. The traces of human presence in the region date back to the Paleolithic. Several cemeteries have been saying Capsian discoveries. Thala had a rich and eventful history, the human presence dating back at least 50,000 BC marked the Berber identity throughout the region. Its growth was partly caused by its industry based on pottery."

While Tamazight ("Berber") gene pool ultimately trace their origins in the Paleolithic, which pretty much goes for any human alive, it has been determined that the bulk of their gene pool is of Neolithic or post-Neolithic provenance. Some Paleolithic era recent common ancestry does surface in their gene pool, but these constitute a fairly modest component of said gene pool overall [see Arredi et al. (2004), for example], generally on the mitochondrial DNA or maternal side. This appears to be supported by osteological data on the paleontological record of the Maghreb, which does not link the well-known Paleolithic, Epi-Paleolithic or Pre-Neolithic era specimens to contemporary Amazigh ("Berber") populations [See, Brace et al. (2005) for example]. Thus, the authors' claim that human presence dating back "at least 50,000 BC", denoting "Berber identity" in coastal Northern Africa, has very little substance to stand on.

With one exception, consistent with patterns observed elsewhere, the authors note that they observed population structuring characterized by an African/Non-African divide, with the Maghrebi populationsrepresenting the Africansgenerally assuming their own position in the MDS plot, away from the European bunch:

"In agreement with previous studies on human diversity using polymorphic Alu insertions (Stoneking et al. 1997; Watkins et al. 2001; Garcia-Obregon et al. 2006) African and non-African populations clearly segregated along dimension I in the MDS representation. Thus, North African populations (North Morocco; West Morocco; Southeast Morocco; Algeria; Tunisia; North, Centre-South Tunisia; Sahara; Sejnane; Takrouna; Libyan; Smar, Bousalem and Zarzis ) were plotted in the positive segment of the first axis (dimension I). The exception is for the Tunisian population of Thala which is isolated and seems to be the most genetically dissimilar. This North African population is characterized by having the highest insertion frequencies for B56, ACE, PV92, APO-A1, D1 and F13B (table 1)...

Study populations (Libya, Smar, Zarzis and Bou Salem) are closest to North African populations whereas Thala is isolated and is closest to Western European populations."


"On the other hand, the European populations are spread along dimension II: Western and Central European populations (Andalusia; Basque Country; Catalonia; Valencia;Canaries; France ; Swiss; Germany and Genova) are concentrated in the negative segment of dimension II, whereas Eastern European populations (the Aromuns of Albania; Albania; Greek Cypriots; Turk Cypriots) are situated in the positive segment of dimension II. For Tunisian populations, the two Berber communities Sejnane and Takrouna are separated from North African populations and are closest to the Syrians and some Central European or Mediterranean populations (Germany and Genova) as already described by Frigi et al. 2010a."

These observations attest to a definite pattern that can be assigned to coastal North Maghreb in general, which lends the Maghrebi gene pool a genetic peculiarity in its own right, that separates it from those of Europe. Such structuring belies notions which simply lump coastal North Africans with Europe and portray them as "mere extensions" of Europeans, rather than reflect the noticeably complex genesis of said Northern Africans; the authors themselves acknowledge complexity...

"The different positions of North African populations suggest that the North African gene pool is hybrid in nature, most probably because of a relatively high rate of mixture with other populations that have occupied the Mediterranean area over different periods of history. It indicates also the complex history of human settlement in NW Africa."

...which is a far cry from the often-heard one-sided assigning of genetic-relationship with Europeans, while ignoring other important genealogical sources that make part of the complex mosaic of Maghrebi genealogy.

"The important finding of this study is the proximity of Libya to North African populations. Regarding Tunisian populations, we remark that Bou Salem and Smar have affinities to North-African populations. Zarzis which is relatively isolated is also closer to these populations. The closeness between Libya and some Tunisian populations indicates the gene flow between Tunisia and Libya across the Sahara."

But that:

"The position of the population of Thala is worthy of attention. It’s isolated and separated from North African populations but is closer to Western European populations. This finding reflects a high Eurasian genetic component for some North African populations and confirms the heterogeneity of North-African populations."

Indeed, recurring reports of high "Eurasian genetic component", as noted above, are reflective of just snapshots of the complex coastal Northern Maghreb gene pool; other sections of this gene pool tell a different story. However, from this point on, the authors tread into the problem that has been the ongoing subject of this entry. They are guilty of the "North Africans more related to Europeans [than "sub-Saharan" Africans just below them]" line, despite having noted the African vs. non-African structuring through the MDS plots, involving the Maghrebi and European samples, having acknowledged the lopsided African paternal component, of "sub-Saharan African provenance", with essentially negligible to no European component, and having just acknowledged that reports of "Eurasian component" in Maghrebi gene pool provide just snapshots of a rather complex gene pool...

"The sub-Saharan origin of North African populations have been proposed on the basis of results indicating local evolution of Y chromosome and mtDNA African haplogroups (Ennafaa et al. 2009; Frigi et al. 2010). The much later transaharan trade in enslaved persons no doubt played a role in genetic contributions, but the egress from a dessicating Sahara with subsequent population formations would explain some of the “sub-Saharan” variation be it from western or eastern Africa. However, from our results, the North African populations appear more related to European than to sub-Saharan populations. The “Eurasian” component seems to have come in over a longer period of time (Keita, 2010). A small amount of gene flow per generation into a population/geographical region can drastically change its original gene frequencies in only a few thousand years as noted by Cavalli Sforza (1991)."

This again speaks to, at least in part, what these researchers choose to assign to "sub-Saharan" banner and what they choose not to. Moreover, what many of these researchers are reluctant to point out, is that a good part of the "relationship" they see between Maghrebi samples and European counterparts, is the legacy of gene flow from the Maghrebi gene pool into the European counterpart; whereas, many of said researchers prefer to talk about gene flow scenarios from the European end into the Maghrebi counterpart. Again, this can be picked up in earlier reports obtained from using autosomal Alu indel markers, although involving different sets of Alu markers from those utilized by Cherni et al. (2011); the results are still consistent:

Multidimensional scaling plot (stress 0.036) applied to the Reynolds’ genetic distance matrix based on 18 autosomal Alu markers.
Click on the image for hi res.

The same phenomenon plays itself out in the plotting of three Alu/STR combinations, in terms of the Maghrebi samples forming their own cluster away from not only the other implicated African samples, but importantly, from the European samples that are most geographically closest to the Maghreb, outside of Africa:
Fig. 3. Multidimensional scaling plot (stress 0.049) applied to the Reynolds’ genetic distance matrix based on three Alu/STR compound systems.
Click on the image for hi res.

These data come from Gonzalez-Perez et al. (2009) [Population Relationships in the Mediterranean Revealed by Autosomal Genetic Data (Alu and Alu/STR Compound Systems)], and an in-depth discussion of that study can be accessed here: link

Gonzalez-Perez et al. (2009) point to persistent genetic exchange from across both sides of the Mediterranean, from times of antiquity to recent, culminating in the distribution of what they call elements of "sub-Saharan gene flow" on the northern Mediterranean end (aka southern Europe). However, in terms of signs of Maghrebi gene flow into the "northern Mediterranean" (southern Europe), specific examples of Alu markers had been named, along with their flanking tandem repeats.

The derived alleles of CD4 110 and DM 107 for instance, serve as imprints of Maghrebi ancestry among southern European samples, by way of not only their heavy concentration in the "southern Mediterranean", even more prominently so in the Maghreb, but also in the pattern of diversity elicited by Gonzalez-Perez et al, involving the derived CD4 and DM loci in general. The authors noted that there was an observable trend in the reduction of variation associated with the CD4 locus for example, with the least variation in southern Europe and the highest in the lone "sub-Saharan" sample of Ivory Coast:

 "The most obvious pattern of haplotype variation is observed in the CD4 system. The ancestral CD4(+) chromosomes show a decreasing pattern of copy number variation from sub-Saharans to Southern and Northern Mediterraneans. Among these latter populations, the 85(+) and 110(+) haplotypes are the most frequent (Supporting Information Table 2). The derived CD4 Alu(-) chromosomes present a lower variation than the ancestral Alu(+) chromosomes, which is statistically significant for Northern Mediterraneans (P < 0.01) and Southern Mediterraneans (P < 0.05), but non-significant for the sub-Saharan sample. This reduction trend is considerable in Northern Mediterranean samples (gene diversity: 0.174 for derived chromosomes vs. 0.554 for ancestral ones), moderate in Southern Mediterraneans (0.458 vs. 0.705), and less marked in sub-Saharans (0.721 vs. 0.778)."

On the DM front:

"In the DM system, ancestral Alu(+) chromosomes show a predominance of 77-bp allele in Mediterranean groups (frequencies > 28%). A majority (>24%) of the derived Alu(-) chromosomes are carriers of the 98-bp allele in sub-Saharan and Southern Mediterraneans, suggesting that the DM Alu deletion probably occurred on a 98-bp variant chromosome (Brook et al., 1992). In Northern Mediterraneans, the most frequent DM(-) variants corresponded to 98- and 101-bp alleles (15.3 and 15.5%, respectively) (see Supporting Information Table 4)."

Taking a quick look at the distribution map involving the aforementioned alleles of the two aforementioned loci, we have as follows:
Fig. 4. Frequency distribution maps of haplotypes CD4 110(-) (A) and DM 107(-) (B).
Click on the image for hi res.

Careful inspection shows Maghrebi genetic penetration into Europe, and particularly more extensive in southern European territories around the Mediterranean sea. Gonzalez-Perez et al. allude to the real possibility of this gene flow:

"the distributions of frequencies for the Mediterranean haplotypes CD4 110(-) and DM 107(-) (Fig. 4A,B) are suggestive of gene flow processes across this geographical region."


"The highest frequencies of CD4 110(-) and DM 107(-) have been found in the High Atlas region (7 and 5.5%, respectively) of Morocco, reaching polymorphic frequencies in all the North African samples [barring the Mozabites for the CD4 110(-) combination]. They have also been found in the Iberian Peninsula, scattered along the northern Mediterranean shore to Greece and Turkey, and on the main islands of the western Mediterranean (Majorca, Corsica, Sardinia, and Sicily; Gonza´lez-Pe´rez et al., 2007). The CD4 110(-) haplotype has also been reported in West Saharans and Mauritanians (Flores et al., 2000) and on five of the seven Canary Islands (Flores et al., 2001), as well as in Adygei from the Northern Caucasus (Tishkoff et al., 1996).

Assuming from their frequency distribution that the place of origin of these particular haplotypes is located in the westernmost extreme of North Africa (Fig. 4A,B), their current ample distribution along both shores of the Mediterranean most likely reflects the effect of gene flow across the region since ancient times, even though specific ages cannot be accurately estimated with our data."

As a result of genetic exchanges between the "Southern Mediterranean" (African) and the "Northern Mediterranean" (European), both the Maghrebi AND southern European populations positioned between the more northward European and the lone sample from sub-Saharan Africa, the Ivory Coast sample. That notwithstanding, the Maghrebi samples maintained their distance even from the southern European populations, nearest to them, of all of Europe.

It seems that these observations had not been lost on Cherni et al. (2011) either, as they proceeded to acknowledge as follows, with respect to the alleged European genetic contribution:

"This genetic flow from Europe seems have happened since Neolithic period. Despite the fact that Neolithic expansion had the same effect in Northern Africa as in Europe, the Straits of Gibraltar acted as a barrier between the two continents, limiting gene flow between North-western Africa and Western Europe through the Iberian Peninsula (Comas et al. 2000; Garcia-Obregon et al. 2006; Varela et al. 2008; Frigi et al. 2010). This justifies the fact that the majority of North African populations appear in the MDS analysis as a separate group from European populations."

So-called "Mediterranean" samples from both sides of the Mediterranean sea, i.e Africa and Europe, in the aforementioned maps both assume intermediate positions, but especially the Maghrebi samples (Gonzalez-Perez et al.'s "Southern Mediterranean")i.e. separate from the "Mediterranean" European samples (Gonzalez-Perez et al.'s "Northern Mediterranean"). Part of the reason the Maghrebi samples form their own clusters in an intermediate position between the African samples on the extreme positions of the plots and the European samples on the other extremes, is the noticeable higher prevalence of markers shared with the "sub-Saharan" sample in them than in the "Northern Mediterranean" (European) samples. This is reflected in the 18 Alu marker data of Gonzalez-Perez et al., and the FXIIIB locus, one of the three loci considered for flanking STR analysis.

Take the FXIIIB locus, for instance...

"FXIIIB system exhibits general high diversity both in (-) and (+) chromosomes (0.565–0.697 vs. 0.575–0.692, as shown in Supporting Information Table 3). Ancestral Alu(-) chromosomes show relatively high frequencies of low copy number alleles (172 and 176) in sub-Saharan population, values higher than 15% of the 172, 184, and 188 variants in Southern Mediterraneans, and a clear predominance of the 188-bp allele (0.319) in Northern Mediterraneans (Supporting Information Table 3). In the derived FXIIIB Alu(+) chromosomes, the most frequent allele is the 180-bp variant (>20%) in Mediterraneans, whereas in sub-Saharans the most frequent allele is the 184-bp variant."

The Maghrebi ("Southern Mediterranean") samples apparently shared visible frequencies of otherwise a "low" copy number (and hence, relative distribution) allele such as that of the 172 allele, and another marker fairly distributed in the African samples, the 184 allele; these markers were not implicated in the European samples on the other hand.

At the same time, the 188 allele was only implicated for the Maghrebi and southern European samples, and not the lone "sub-Saharan" sample from Ivory Coast. With this kind of sharing of alleles with both southern European and "sub-Saharan" samples, it is easy to see why the Maghrebi sample almost always assume an intermediate position between the European samples and the African ones. The higher concentration of "sub-Saharan" prevalent markers in the Maghrebi samples characteristically sets them apart from the European counterparts...

"As for individual populations, the sub-Saharan gene flow in North Africa based on the Alu data collection ranges between 6 and 17% (Table 3), except the Siwa Berbers where that influence was negligible. Admixture values based on Alu/STR combinations indicate that sub-Saharan flow in North Africa ranged from 16% (North East Moroccan Berbers) to 35% (remaining samples) with the exception of Siwa Berbers who showed the highest admixture value (51%)."

As a matter of precision, the so-called signal of "sub-Saharan" African gene flow actually ranged between 17% and 37% for the "remaining coastal northern African samples" (outside the Siwa) in the authors' Alu/STR data rather than the "16% and 35%" mentioned above. Gonzalez-Perez et al. characterized the seeming dis-concordance between "sub-Saharan contribution" estimates from the 18 Alu markers and those of the 3 Alu/STR combinations as "a wide range probably related with the different mutational nature of the markers analyzed and with the effect of repeated homoplasic mutation in STRs".

While any of these things can certainly a driving factor, particularly the "different mutational nature of the markers", the relatively more visible "sub-Saharan" connection in the Alu/STR combinations than in the 18 Alu collection is not all that surprising, because the Alu/STR compounds were more likely to preserve the inherited alleles due to the effect of linkage disequilibrium, as another driving factor for the uneven information about possible "sub-Saharan" contributions across the 18 Alu markers and the Alu/STR combinations. Gonzalez-Perez et al. eventually go onto acknowledge the role of linkage disequilibrium on the polymorphic nature of the loci they studied, and indeed, it is more than likely an important contributing factor behind the better preservation of "sub-Saharan" contribution in the Alu/STR loci than the individual 18 Alu loci.

Whereas, the so-called signal of "sub-Saharan" African gene flow in the southern European samples,...

"The sub-Saharan contribution in Northern Mediterraneans was imperceptible for the Alu data set, but reaches values around 6% for the STRs set and Alu/STR haplotypes."

Again, as a matter of precision, the reported "values around 6%" is in reference to the estimated "overall" sub-Saharan contribution in the southern European samples, which must not be reflective of an average, as they don't quite match up with the individual Alu/STR data; but for the individual southern European Alu/STR data, the "sub-Saharan" contribution actually ranged from 7% to 15%, still not as extensive as that of the Maghrebi counterpart.

In the Maghrebi sample, the estimated "overall" contribution for the Alu, STR and Alu/STR combinations, i.e. all samples combined, was reportedly:

"In the case of Southern Mediterraneans, the overall sub-Saharan contribution ranges from 13% for the Alu markers to 46 and 40% for the STRs and Alu/STR data, respectively."

...apparently higher than those estimated for the southern European counterpart. The authors do admit that they omitted the STR data, as follows:

"Gene flow in Mediterraneans (North and South shores) was analyzed with LEADMIX simulations checking for different parental groups. The only consistent results had been summarized in Table 3 (admixture based on STRs are almost identical than those obtained from Alu/STR haplotypes and, hence, they are not included in Table 3)."

There must have been a typing error in there, as it does not make sense for them to omit STR data, if it were a matter of "admixture based on STRs" being "almost identical than those obtained from Alu/STR haplotypes", but it would make sense, if it was omitted because "admixture based on STRs are almost identical to those obtained from Alu/STR haplotypes".

Getting back to that matter of very real (with statistical significance, ruling out chance occurrence) genetic identity differences between the Maghreb and southern European samples, in that the former cannot be dismissed as being an extension of the latter, consider this revelation from Gonzalez-Perez et al.:

"The genetic variance of Alu/STR haphaplotype frequencies attributable to the differentiation between Northern and Southern Mediterraneans (between groups FCT of 2.2%, P < 0.001, and total FST of 2.9%, P <0.001) indicates significant differences between these two groups."

Whereas signals of "sub-Saharan" gene flow were estimated in the Maghrebi samples, the authors claim that they were not able to do the same in terms central or northwest European gene flow into southern European samples:

"When gene flow in Northern Mediterraneans was tested, taking Central Europe and Southern Mediterraneans as parental populations, the results were statistically inconsistent, indicating the limited power of our markers to discriminate gene flow within Caucasoid populations."

The possible reasons for this may be two-fold; it could speak to the fact that Europeans come from a constricted subset of OOA migrantswhom in turn are generally considered a subset of an ancestral 'eastern African" source, not to mention that they occupy a relatively smaller landmass, and hence, the inter-population differentiation across these markers have manifested modestly and not statistically clearcut; whereas differentiation in these markers between the Maghrebi samples and the "sub-Saharan" counterpart are more marked, as the case might be between any other two African populations, because African gene pool is far more diverse than the European counterpart. We've already seen examples of this trend in at least two out of three Alu/STR combination markers discussed by Gonzalez-Perez et al.

On the other hand, it could very well be, as the authors say, an artifact of the limited discriminatory power of their selected loci; or it could yet be the interplay of both just-mentioned reasons.

Though they used different Alu loci, the results of Cherni et al.'s (2011) analysis has the same effect, wherein the Maghrebi samples mainly keep to themselves, and clearly differentiate themselves from European samples. Cherni et al. do note at least one "outlier" in their Maghrebi bunch, in the form of the Thala sample, which they attribute to possible higher "Eurasian component" than in other samples...

"The position of the population of Thala is worthy of attention. It’s isolated and separated from North African populations but is closer to Western European populations. This finding reflects a high Eurasian genetic component for some North African populations and confirms the heterogeneity of North-African populations."

This revelation again goes to show that the Maghrebi populations have a definite genetic profile of their own, such that any excesses from the outside alters the preexisting set up and causes the receiving segment of the Maghrebi gene pool to stand out from the rest. Of course, one can make the case that the Thala sample had received relatively higher external gene flow than the other Maghrebi samples, but Cherni et al. (2011), later on open up the possibility that:

"The method of Harpending and Ward (1982), which uses the theoretic regression of population heterozygosity and distance to the centroid as a valid indicator to analyze the geographic model of genetic structure in populations, has been applied to the populations of North Africa in order to evaluate the relationship of these populations with other Mediterranean populations in this area. In this case, the Smar population is very close to the theoretical prediction line, as is the case with most of the populations in its geographical area included in the analysis, even Northern Morocco, Western Morocco, North-Central-South of Tunisia and other Tunisian populations. The samples from Algeria, Sejnane, Takrouna and Zarzis are low outliers, which is interpreted as being indicative of these populations being receivers of a limited gene flow from the area. This result suggests that Zarzis, which is characterized by its low heterozygosity and these populations have either had a smaller effective population size, or they have been more isolated. The samples from Thala and Libya are high outliers, a position that does not necessarily imply that they recently received a greater external contribution, but which is indicative of a series of genetic differences with respect to the populations included in the analysis that may be due to the effect of the genetic contributions inherent to their origin."

It is as if Cherni et al. would like to think of "outliers" like the Thala sample, as a population that started out as some distinctive "hybrid" product of parental populations from two different landmasses from the very onset, in obvious contrast to the rest of Maghrebi populations which largely adhere to the general Maghrebi genetic profile, rather than the more obvious scenario, which would entail that the "oultiers" received more external genetic contribution at some point in time. It's not like the Maghrebi populations speak some pidgin language family that is a composite of African and "Eurasian" language, or some autochthonous tongue that also typically spoken outside of their African homeland; Arabic is only spoken due to acculturation, and even that has not replaced the original Tamazight dialects of the region.

The samples from Algeria, Sejnane, Takrouna and Zarzis, especially the Zarzis sample with its relatively lower heterozygosity, seemed to have received the least external gene flow of the Maghrebi bunch, including from surrounding regional populations; yet, these populations still managed to maintain close genetic distances with the remaining Maghrebi collection. Whereas, as plainly noted, the Thala and Libyan sample distinguished themselves as "high outliers", which Cherni et al. interestingly attribute to "genetic contributions inherent to their origin", as a more probable alternative reason than subsequent external gene flow, but it is worth recalling how Cherni et al. claim that the Thala sample, for example, assumes this extreme outlier position, which to recall from above, was attributed to a "high Eurasian" genetic contribution:

"The position of the population of Thala is worthy of attention. It’s isolated and separated from North African populations but is closer to Western European populations."

And then coupling the above, with this earlier mentioned bit:

"the Straits of Gibraltar acted as a barrier between the two continents, limiting gene flow between North-western Africa and Western Europe through the Iberian Peninsula (Comas et al. 2000; Garcia-Obregon et al. 2006; Varela et al. 2008; Frigi et al. 2010). This justifies the fact that the majority of North African populations appear in the MDS analysis as a separate group from European populations."

None of these speak to a general North African social genesis that could have started out as a composite African-European "hybrid" population from the onset. An outlier population in this regard, should therefore be seen first and foremost as a very possible recipient of subsequent gene flow, before entertaining any other consideration. Interestingly enough, Cherni et al.go back to entertaining that very scenario:

"Taken together, results on Y chromosome, mtDNA and Alu Insertions in North Africa allow to propose a scenario for this region. The ancient sub-Saharan settlement would have been followed by admixture with Iberian populations."

They go onto explain how this genetic exchange very likely occurred, made apparent from Maghrebi diversity pattern:

"But, as the North African Y chromosome remained dominant in the region, we could argue that this admixture have been realized in one direction: North African men and Eurasian women, explaining the gene flow from Europe and high frequency of European types of mtDNA in North Africa as compared with Y chromosome. This situation would not be the result of drift toward Eurasian mtDNA. Our results on Alu insertions interestingly confirm that this gene flow happened several times probably always on the same direction. These matrimonial exchanges between North Africa and Europe should be considered in a context of patriarchal societies with men attached to territory and women from different regions including Europe. Hence, genetic diversity on one hand and relationship with Europe should have been due to women."

The general coastal Maghrebi genetic pattern cannot primarily be attributed to subsequent regional homogenization through gene flow between coastal Maghrebi populations themselves, since the authors themselves claim that their results negate such a scenario [also see preceding notes about "low outlier" samples like that Zarzis], in that "different genetic structures of North African populations" indicate "weak gene flow between them due to high endogamy rates that also should be considered in patriarchal socioeconomic context".

Furthermore, Cherni et al. (2005) has shown that the African-European ancestry pattern, repeatedly mentioned across several studies published over time, does not neatly spread across all coastal Maghrebi populations; they found that their Kesra Tunisian sample produced a profile that was expected of the "Arabized" counterpart, sporting fairly substantial African maternal component, while their "Arabized" sample (the Zriba), produced the profile a priori expected of the Tamazight-speaking counterpart, characterized by a relatively modest "sub-Saharan" African component. The authors then, attributed this finding to insufficient sampling choices of previous research, and thereby resulting in the reconstruction of Maghrebi gene pool from just a snapshot of Maghrebi overall genetic diversity.

It is necessary to point out that when Cherni et al. (2011) elicit the "relationship of the Maghrebi samples with "sub-Saharan" Africans vs. the "relationship of the Maghrebi samples with Europeans", that they did not actually sample any so-called "sub-Saharan" population. On top of that, their study produced the same finding as that of earlier studies like that of Gonzalez-Perez, which was extensively looked at here as well; accordingly, the Maghrebi populations form their very own group, in between other African samples and European counterparts. The logical conclusion of this, is that the Maghrebi populations have a genetic profile of their own in its own right, and that they cannot be reduced to some "satellite Europeans" in Africa.


The coastal Maghrebi samples assume an intermediate position, as they should, because they have received gene flow from both directions, internally from Africa and across the Mediterranean sea from Iberia. This is a far cry from saying that they are lopsidedly related to one population (Europe) than another (Africa). Part of the reason 'western' research teams see the relationship between Maghrebi samples and those from right below them, in "sub-Saharan" Africa, versus that of "relationship between Maghrebi samples and Europeans" is due to their sampling procedures and clouded preconceived "balkanization" of Africa.

The "relationship between Maghrebi samples and "sub-Saharan" Africans" is almost always going to be misleading, because most 'western' research teams compare coastal Maghrebi samples with limited samples from "sub-Saharan" western Africa, while expunging certain groups from that very region and counting them as "North Africans" instead of counting them as part of the said "sub-Saharan" region. These populations usually involve nomadic Tamasheq groups from Mali and Niger, and "Afro-Asiatic"-speaking sedentary populations of Mauritania, while neighboring populations of those very same locations are included in the "sub-Saharan" block.

Other times, geographically-intermediate populations in between coastal Maghrebi populations and those well-below the Saharan desert in the western African general region are altogether excluded, giving an abrupt change within the region, instead of a clinal relationship. Indeed, Gonzalez-Perez et al. for example, made a passing note of incidences of CD4 110(-), which they touted as an example of "Mediterranean haplotypes", in geographically-intermdiate populations such as West Saharans and Mauritanians, who are positioned between coastal northwest populations and those well below the Saharan desert but were excluded in the Gonzalez-Perez et al. (2009) study. Such incidences go to show that continuity of genetic connection can be achieved, if comprehensive sampling of the Western African region were taken, and that said gene pools were all equally recognized as that of western Africa, rather than the region being "balkanized" into "North African"/"sub-Saharan" African dichotomy, wherein certain populations of territories otherwise traditionally called "sub-Saharan" by "westerners", are conveniently assigned to the "North African" block.

Furthermore, given the oft acknowledgment by these researchers that Maghrebi paternal gene pool is lopsidedly of "sub-Saharan" origin, it is bizarre that they do not generally consider their relationship with "sub-Saharan" eastern Africans, whereby the latter represents part of the "sub-Saharan" continuum. Even in rare occasions when Maghrebi samples are compared against "sub-Saharan" eastern African counterparts, the latter are often treated as part of "northeast" African block rather than "sub-Saharan" Africans. After all, it would make sense that some of the genetic components that coastal Maghrebi populations don't seemingly share with "sub-Saharan" western Africans, could well be part of their original "sub-Saharan" eastern African gene pool.

Also take note, that when authors speak of "Maghrebi relationship with Europeans", they don't specify that any reputed "close relationship" is essentially with only southern European samples; 'western' researchers tend speak of Maghrebi "close relationship" with Europeans, without specifying which Europeans this situation actually applies to, yet they generally take care to point out that only certain Africans, namely "sub-Saharan" Africans (usually limited selective western African samples), are being implicated in their estimation of "Maghrebi relationship with sub-Saharan Africans". As the plots (above) show, northwest and/or central European samples showed a relationship no more considerably closer than the lone "sub-Saharan" from Ivory Coast did to the Maghrebi samples. In fact, in some angles, the Ivory Coast sample was even more closer to Maghrebi cluster than certain European samples.

In the earlier posted 18 Alu marker MDS plot for example, along the vertical dimension, the lone "sub-Saharan" sample from Ivory Coast assumed much closer distances from the Maghrebi cluster than the German sample did, while just as close as the general southern European cluster did, if not even closer when compared to some southern European samples, like say the Pas Valley, from the Iberian peninsula. Along the horizontal dimension, both all the European samples and the lone "sub-Saharan" sample from the Ivory Coast assumed almost equidistant positions from the Maghrebi cluster.

Other segments seemingly absent in "sub-Saharan" western African sample(s), could well be relics of "sub-Saharan" eastern African ancestry that underwent micro-evolutionary polymorphic events within coastal Maghrebi populations themselves in isolation from the so-called "sub-Saharan" populations. Then factor that in with subsequent gene flow between Maghrebi populations and those from across the Mediterranean sea. This would allow Maghrebi genes to enter southern European gene pool and vice versa, giving a veneer of similarities to the populations exchanging genes. Other occasional similarities between the Maghrebi gene pool and their southern European counterparts could very well be rooted in their common "sub-Saharan" eastern African ancestry, which are generally tested against limited samples from "sub-Saharan" western Africa.

Such oversight of information will give misleading impressions of actual "sub-Saharan" African relationship with coastal Maghrebi populations, outside of uniparental markers. Then comes into equation, that observation just mentioned in the Cherni et al. piece above, about the African-European asymmetrical lineage among Maghrebi populations. Much of that asymmetrical ancestry very likely comes from institution of historic slavery enterprise involving mostly European females in the Maghreb. Of course, a lot of devout Eurocentrists balk at such a prospect, but it is a fairly reasonable explanation for the aforementioned asymmetric parental pattern, and importantly, it is backed by history.

The skeptic-wisdom is, after all, that European male contribution seems all but negligible in the Maghrebi gene pool, and yet, that the slave enterprise in the Maghreb would have included European males. For one, slave-male genetic exchange with any preexisting local maternal gene pool would have been strongly discouraged by slave-holders, as was the case with many locations that have put in place institutions of slavery, while the Maghrebi male population would have had a freer hand in exchanging genes with enslaved European females. Additionally, Maghrebi enthusiasm for European male slave market would have been relatively modest compared to that for the European female slave counterpart, since the Maghrebi would have then had firsthand access to preexisting slave trade with "western African" polities just geographically beneath them, where a labor pool of more physically-robust males could have been extracted than that from European counterparts.

In an ironic twist, devout Eurocentrists like to portray any so-called "sub-Saharan" African contribution in Maghrebi gene pool as "slave"-mediated, when in fact, it appears to have been the other way around: the asymmetric African male-to-European female ancestry speaks more to a European contribution that was mediated by and large through slavery than the case is for the "sub-Saharan" contribution in coastal Maghrebi gene pool, which Frigi et al. (2010) for example, determined to have been around since prehistoric times.

*As with other entries on this site, this entry is subject to modifications in response to new revelations in scientific research and/or additional information brought to the attention of the author of this entry.

Fadhlaoui-Zid et al. (2011), Mitochondrial DNA Structure in North Africa Reveals a Genetic Discontinuity in the Nile Valley. 

Cherni et al. (2011), Human Alu insertion polymorphisms in North African populations. 

Pereira et al. (2010), Linking the sub-Saharan and West Eurasian gene pools: maternal and paternal heritage of the Tuareg nomads from the African Sahel. 

Frigi et al. (2010), Ancient Local Evolution of African mtDNA Haplogroups in Tunisian Berber Populations.

Gonzalez-Perez et al. (2009), Population Relationships in the Mediterranean Revealed by Autosomal Genetic Data (Alu and Alu/STR Compound Systems). 

Claudio et al. (2009), First Genetic Insight into Libyan Tuaregs: A Maternal Perspective. 

Cherni et al.(2005), Female gene pools of Berber and Arab neighboring communities in central Tunisia: microstructure of mtDNA variation in North Africa. 

Brace et al. (2005), The questionable contribution of the neolithic and the Bronze age to European Craniofacial form. 

Arredi et al. (2004), A Predominantly Neolithic Origin for Y-Chromosomal DNA Variation in North Africa. 

Rando et al. (1998), mtDNA analysis of Northwest African populations reveals genetic exchanges with European, Near Eastern and sub-Saharan populations.

*Personal notes, 2011.


khufu said...

Hey mistery solver This is not a comment.I only want to say that i missed you two months already. What's going on? everything ok.Don't stop this blog cause many of us are following you even if we do no place comments all the time.

D. W. Reynolds said...

All I can say is "WOW", and finally! This is greatly balanced writing and objective inclusion of the latest dna evidence with the known history of the peopling of North Africa.

And please keep up the good work because I'm spreading the news about your blog 'round the net. :)

Mystery Solver said...


I had no intention of ending my blogging. I haven't been publishing as of late, only because I've been held up by other activities; I've even cut back on posting on ES for the same reason. I'll resume publication on new material soon. So, be on the lookout. Cheers :)