Carmen Tomas1 , Juan J Sanchez1,2 , Anna Barbaro3 , Conxita Brandt-Casadevall4 , Alexis Hernandez2 , Mohamed Ben Dhiab5 , Misericordia Ramon6 and Niels Morling1
Published: 29 February 2008
Due to its history, with a high number of migration events, the Mediterranean basin represents a challenging area for population genetic studies. A large number of genetic studies have been carried out in the Mediterranean area using different markers but no consensus has been reached on the genetic landscape of the Mediterranean populations. In order to further investigate the genetics of the human Mediterranean populations, we typed 894 individuals from 11 Mediterranean populations with 25 single-nucleotide polymorphisms (SNPs) located on the X-chromosome.
Results
A high overall homogeneity was found among the Mediterranean populations except for the population from Morocco, which seemed to differ genetically from the rest of the populations in the Mediterranean area. A very low genetic distance was found between populations in the Middle East and most of the western part of the Mediterranean Sea.A higher migration rate in females versus males was observed by comparing data from X-chromosome, mt-DNA and Y-chromosome SNPs both in the Mediterranean and a wider geographic area.
Multilocus association was observed among the 25 SNPs on the X-chromosome in the populations from Ibiza and Cosenza.
Conclusion
Our results support both the hypothesis of (1) a reduced impact of the Neolithic Wave and more recent migration movements in NW-Africa, and (2) the importance of the Strait of Gibraltar as a geographic barrier. In contrast, the high genetic homogeneity observed in the Mediterranean area could be interpreted as the result of the Neolithic wave caused by a large demic diffusion and/or more recent migration events. A differentiated contribution of males and females to the genetic landscape of the Mediterranean area was observed with a higher migration rate in females than in males. A certain level of background linkage disequilibrium in populations in Ibiza and Cosenza could be attributed to their demographic background.
In Figure 1a, the first two axes accounted for 39% of the total variability (first axis: 25%, second axis: 14%). Thus, the Mediterranean populations are quite homogeneous. Equivalently to the results of the AMOVA analysis, the Moroccan population was considerably displaced in the first axis, and the Majorcans were placed in the opposite direction. In the second axis, the differentiation was less, but it is interesting that the Cosenza and Ibiza populations were located opposite to their neighbouring populations (i.e. Catanzaro and Valencia, respectively). Figure 1b shows the relative position of the Mediterranean populations in relation to a wider geographical area. The first two axes accounted for 51% of the global variability (first axis: 38%, second axis: 13%). The first axis displayed East Asian populations (Japanese and Chinese populations, which did not show a significant pairwise Fst value) opposite to the others (Middle East, European and African populations). In the second axis, the Mediterranean populations plus the CEU population (Utah residents with ancestry from northern and western Europe) formed a group between the North-European and African populations. Only the Moroccans showed a slight displacement from the others showing significant Fst values in most of the pairwise comparisons inside the Mediterranean group.
The analysis of the Mediterranean populations in comparison to the “worldwide” populations showed a clear intermediate position between African and North-European populations with Morocco somewhat closer to the African populations than the other Mediterranean populations. When only the Mediterranean populations were analyzed in a principal coordinate plot, opposite positions of Moroccans and Majorcans were observed (Figure 1). This supports the existence of a north-south gene flow “barrier” in the western part of the Mediterranean area [9,10]. It could be argued that the genetic differentiation of the Moroccan population was a result of the arab-berber sub-structure of the individuals sampled in this work, but this hypothesis does not seem likely. Several authors [26,28] reported a high genetic homogeneity between berbers and arabs in NW Africa, so they suggested that the Arabisation of this area was probably a cultural phenomenon, which did not imply a replacement of the ancestry population. Our results give support the hypothesis of an early settlement of NW Africa [26]. The original berber population seem to have received a low genetic influx from the surrounding areas.Different hypothesis have been suggested to explain the genetic differentiation of the Moroccan population. An initial genetic drift [26,30] could have caused differences in allele frequency distribution that have not been re-established due to a certain level of geographic isolation. The Strait of Gibraltar has been described by several authors [9,10] as an important genetic barrier. Even a certain level of genetic exchange probably occurred between NW Africa and the South of the Iberian Peninsula [10,27,32,33], sharp frequency changes have been described in this area [10,33]. Also the Sahara desert has been suggested as responsible of the genetic isolation of NW African populations from Sub-Saharan populations [30]. There is no consensus about the impact of the Neolithic demic diffusion in the Mediterranean area [29,30,34]. According to our results, a low impact of the Neolithic expansions and/or later migration events on NW African populations would have occurred.
An abbreviated version of this study into X chromosomes around the Med. There has been a fairly substantial movement of female sub Saharan DNA into North Africa due to the slave trade, which probably explains why Berbers group with the other African populations.
For anyone reading this not familiar with ancient North Africa; 12,000 year old bones from West Morocco have yielded Eurasian mt DNA haplotypes, and the Berbers are a lot more closely related to Europeans than sub Saharan Africans. The Berbers don’t appear to have changed much in the last 12,000 years.
Mathilda's Anthropology Blog. 
I read this one and I find it pretty inconclusive. They only show PC analysis, that tends to show bias for greater samples and bigger components in the overall sample.
What I see in the first graph is two geographical axis: an E-W one from Turkey to Valencia and a N-S one from Mallorca to Morocco (with the other samples neutral for this second dychotomy what rather suggests pecualirites of these two places). In the E-W axis though there is a geographically logical gradation with some exceptions: Ibiza appears extremely to the East, what may relate to a founder effect in this island: very possibly Phoenician-made) and much of the same can be said of Cosenza (in this case surely a Greek founder effect). Iraq appears strangely “western”, what to me says its major components are actually distinct from the ones shown in the PC graph and that what we are seeing here is just minor components of te Iraqi genetic pool probably.
The other graph shows also two axis, this time global ones: from East Asia to to West Eurasia (Morocco, Tunisia, Mallorca) and from Tropical Africa to West Eurasia again (Denmark as extreme here). Apart of seeing again the Iraqi sample strangely neutral (same reason as before probably) more surprisingly is the CEU (Utah Whites) one, that doesn’t even cluster with one of their major putative ancestors: the Danes.
To me all this what shows again is the limitations of PC analysis, that can suggest patterns that are not, specially when both PCs are very minoritary in any sampled population. I really prefer this kind of stuff with a good K-means analysis (with meaningful K depth).