HLA alleles and haplotypes in Sudanese population and their relationship with Mediterraneans

The contribution of migrated people from once green Sahara (about 10,000–6000 years bc) towards Mediterranean area had probably a double effect: both genetic and cultural connections have been described between Western Europe and North Africa. Sudanese populations from different ethnicities have been studied for HLA-A, -B, -DRB1 and -DQB1 antigens by a standard microlymphotoxicity method. Results found show that Nubians are genetically related with African Sub-Saharan populations and distant from other Sudanese tribes, who are closer to Mediterranean populations than to Sub-Saharan ones. This is concordant with other authors and meta-analysis data. Our present work is, to our knowledge, the first and only one HLA research that studies Sudanese people according to different Sudan ethnic groups: samples were collected before Sudan partition between North and South. A prehistoric genetic and peoples exchange between Africa and the Mediterranean basin may be observed and is supported with the results obtained in this Sudanese HLA study. However, demic diffusion model of agriculture and other anthropological traits from Middle East to West Europe/Maghreb do not exist: a more detailed Sahel and North African countries ancient and recent admixture studies are also being carried out which may clearer explain pastoralists/agriculture innovations origins in Eurafrican Mediterranean and Atlantic façade.

phylogenetic calculations.Other world populations were considered for genetic comparisons and are detailed in Table 1.The study was approved by Complutense University Ethical Committee (EC-0117.AK) and all subjects included in the study were adult unrelated blood donors who signed an informed consent to participate in it.

HLA typing, DNA sequencing and statistics
Generic HLA class I (A and B) and class II (DRB1 and DQB1) typing was done by two-step microlymphotoxicity technique on T or B lymphocytes 19 by using the 10th and 11 th International Histocompatibility Workshop and local reagents.This technique was used because it was necessary to study all Sudan available ethnicities and HLA low resolution was also available and it is sufficient to accurately distinguish or relate populations 4,11 .
Two types of analyses were done to compare HLA frequencies of Sudan with those of Mediterranean and other populations: (1) with low-resolution DRB1 and DQB1 data; and (2) with low-resolution DRB1 data.This latter analysis was performed in order to compare Nubians from Sudan and other Sudanese ethnicities in a proper way by using the same level of HLA allelism distinction, who have been typed/transformed for HLA-DRB1 generic but lack HLA-DQB1 typing.
Statistical analysis was performed with Arlequin v3.0 software 20 .In summary, this program calculates the linkage disequilibrium between two alleles at two different loci.Their level of significance (p) for 2 × 2 comparisons and also their relative linkage disequilibrium were calculated as previously described 21 .Computer program calculated the most frequent HLA extended haplotypes that were further analysed as follows; the most frequent complete extended haplotypes were tentatively deduced from: (1) the 4 HLA loci haplotype frequencies were obtained by software 21,22 ; (2) the previously described haplotypes in other populations were also taken into account 21 ; and (3) haplotypes if they appeared in two or more individuals and the alternative haplotype was well defined were also studied 21,22 .In order to compare phenotype and haplotype HLA frequencies with other populations, the reference tables of the 11th and 12th International HLA Workshop were used 6,12 (Table 1).Phylogenetic trees (dendrograms) were constructed with HLA-DRB1 frequencies by using the Neighbour-Joining (NJ) method 23 and genetic distances between populations (DA) 24 , by using DISPAN software which contains the programs GNKDST and TREEVIEW 25,26 .Also, correspondence analysis in three dimensions and

Ethical approval
This study was conducted according to the Declaration of Helsinki by the Ethics and has been approved by Complutense University Ethical Commitee (EC-0117.AK).

Consent to participate
All subjects included in the study were adult unrelated blood donors who signed an informed consent to participate in it.

Characteristic Sudanese HLA allele frequencies
The expected and observed genotype frequency values for HLA-A, -B, -DRB1 and -DQB1 loci do not significantly differ, and the population sample is in Hardy-Weinberg equilibrium.Table 2 shows the HLA allele frequencies found in our Sudanese analysed sample.Figure 2 depicts a low-resolution class II (DRB1) neighbour-joining (NJ) tree constructed form genetic distances between populations (DA, Table 3).Its topology shows how compared populations cluster in two main branches: in general, western (both North African and Europeans) and other Mediterraneans group together respectively and tend to converge in the same node; in the other branch Greeks and Sub-Saharans tend to cluster together with Nubians, Fulani, Rimaibe and Mossi (as described before in 5, 16 28 and with other genetic markers 29,30 .Our Sudanese studied group is placed especially close to most western Mediterraneans: North African (Algerians and Moroccans) and Iberians (Table 3).Genetic distance values (Table 3) give quantitative genetic distances and relatedness between populations and show that Turks are the closest to Sudanese population, followed by Egyptian-Bedouins and Iranians.Also, Mediterranean populations like Algerians, Moroccans, Spaniards and Italians are genetically close to Sudanese population studied in present work.Finally, Greeks and Sub-Saharans cluster together and behave as outgroups.In the same way, the correspondence analysis performed show again how Sudanese are placed within the eastern Mediterranean group, which is also related to the western one (Fig. 3).On the other hand, Nubians cluster together with Greeks and other black Sub-Saharan populations included in this work.Genetic distances between Nubians and our Sudanese mixed sample are similar to those among other Sub-Saharans (Oromo and Amhara) and Sudanese (Tables 3 and 4).These results in Greeks have been previously confirmed by different independent research groups and genes (see "Discussion" section).
To complete the genetic relationship study of our studied Sudanese from different ethnicities with other reported populations, a second set of analyses were carried out by using low-resolution HLA-DRB1 data.The constructed NJ tree (Fig. 2) and calculated genetic distances (Tables 3 and 4) show a very similar relatedness of Sudanese population in both types of analyses.Interestingly, Nubians, have a genetic distance value with respect to our Sudanese mixed sample as far as Greeks and Sub-Saharans (Table 3).In fact, when Correspondence analysis with generic HLA-DRB1 data is performed, Nubians are closer to sub-Saharans than any other compared population including our Sudanese mixed-ethnicities studied group (Fig. 3 and Table 4).Table 4 also shows that Nubians are more related to Sub-Saharans and also to Greeks, as it has been demonstrated by us and other researchers 4,5,16 .

HLA-A, -B, -DR and -DQ linkage disequilibria
The present study reports not only HLA class I (A and B) and class II (DRB1 and DQB1) allele frequencies but also two-loci and extended haplotypes for the first time in a mixed sample of different Sudanese ethnicities, allowing their comparison with previously reported results in other populations.
The most frequent HLA extended four-loci haplotypes (HLA-A-B-DRB1-DQB1) (Table 5) are marked by the high frequency of some of the two-loci haplotypes described above.A*30 and A*31 alleles (low resolution) are particularly frequent in Mediterraneans and also in our Sudanese sample (Table 2); the same occurs with B*49 allele.The most frequent HLA extended haplotype found in our Sudanese sample is A*02-B*51-DRB1*13-DQB1*06 (4.59%) and is identified as a Mediterranean/European haplotype; it is also present in Spaniards (1.0%) 12 .Also, this B-DRB1-DQB1 combination has an Iberian-Mediterranean background because is present with similar frequency in Portuguese, Italians, French, and African Americans from the US 12 .The second most frequent haplotype found in present work (A*02-B*49-DRB1*03-DQB1*02, 2.56%) has been found to be specific of the Sudanese sample studied, although its partial B-DRB1-DQB1 haplotype can be found in Italians 12 .Other Sudanese-specific HLA extended haplotypes with frequencies above 1% have been found in our  5).

Sudanese HLA Mediterranean background
Our present study shows that Sudanese population is related to Mediterraneans like Turks (who are genetically Mediterraneans except because of an Asian "elite" imposed language 18 ), Egyptians, Iranians, Algerians, Italians and Iberians.This is concordant with other North African, Middle East and Iberian HLA compilation study 31  to Sahel populations (Rimaibe, Mossi, Oromo, Amhara); it disconcerted us, and an allele-by-allele comparison was carried out between Greeks and Sub-Saharans to finally show that they both share some HLA alleles in common, confirming our phylogenetic results 5,16 .Typing of these populations has been done by many participant laboratories in the 12th Histocompatibility International Workshop under the direction of Julia Bodmer (Oxford) 5,6 and computer calculations were made by us.Also, another autosomal marker (3120 + 1 G → A cystic fibrosis marker) was found in Africans and also only in Greeks out of all other European populations tested 29,30 .It was written by Aeschylus that African peoples (Danaids) emigrated northwards firstly to Egypt, where they were rejected and then fled to Peloponnesus 32 .This may give us a sort of historical reference for these genetic results that have been independently confirmed by other groups and other genetic markers 5,16,[28][29][30] .

North African migrations and contacts with the Mediterranean Basin
Sahara Desert climate has undergone sharp variations over time, ranging from wet to dry over the past hundreds of thousands of years 33 .This variability is due to a 41,000-year cycle in which the axis of the Earth changes between 22° and 24.5°3 3 .In fact, the Columbia Shuttle detected a humid (traces of big rivers and lakes) and green  Genetic distances between populations were calculated with GNKDST software using low-resolution HLA-DRB1 frequencies.Bootstrap test showed values of 100 in all nodes after 1000 replicates.Note that Greek HLA relatedness with Sahel populations was detected by two independent different groups 4,5,16,28 and it is also supported by other autosomal genetic markers for African and Greeks specifically (3120 + 1 G → A cystic fibrosis marker) 29,30 .HLA-DRB1 allele frequencies is used because of the best discrimination between populations among HLA loci (more or less relatedness) and the fact that almost all the populations in data bases are HLA-DRB1 typed 31 .1).Three different clusters are obtained from the analysis: for simplicity, cluster A (light blue), cluster B (dark blue) and cluster C (yellow/orange).Mixed Sudanese population analysed in this study appears together with other Mediterraneans like Turks, Iranians, Egyptians, Italians or Jews in the eastern-Mediterraneans cluster (cluster A, light blue).However, Sub-Saharan Sahelians (Fulani, Rimaibe, Mossi) and Nubians cluster together with Greeks and other African Sub-Saharan populations (cluster C, yellow/orange), as previously obtained by two independent groups 4,5,16,28 .Correspondence analysis supports the results previously obtained in the genetic distances analysis (DA) (Table 3) and NJ dendrogram (see Fig. 2).Italians have been chosen from North Italy 6 , a population that usually cluster with northern Balkan Peninsula or Central European groups.
Vol:.( 1234567890 www.nature.com/scientificreports/Sahara Desert under the dunes 34 .Currently it is in a dry period, but the Sahara is expected to turn green again in about 15,000 years.Due to these desertification-greening cycles, it is possible that very different human populations inhabited in the last humid stage.The Tassili N' Ajjer National Park and Ahaggar Mountains in Algeria show a record of rock art that proves human settlement in the Sahara 33 .However, when the desertification began around 10,000 bc, the populations that inhabited the Sahara emigrated to other more habitable northern and other areas.It is well established that North Africans and southern Europeans are genetically related, and this may be due to a long lasting circum-Mediterranean cultural and genetic flow particularly during the last glacial peak 4,[35][36][37][38][39] , being a secluded population between European North Ices and Desert.On the basis of our present day genetic and linguistic studies, we have postulated that many people coming from what is nowadays the Sahara Desert started to move towards East, West, North, and also South, being an important part of the primitive people stock of Sumerians, Egyptians, Guanches (Canary Islands), Iberians, Etruscan, Minoans, Anatolians (currently named Turks on only linguistic bases because they show a clear Mediterranean HLA profile) 4 , Kurds, and other islanders or northern Mediterraneans 4,5,18,[36][37][38][39] .In the present work, HLA genetic background of Sudanese people from different ethnicities is studied, and results obtained inpart confirm these North African-Mediterranean peoples (genes) exchange.
Finally, there are common traits in Mediterranean, Iberian and North African cultures Circle, which may be conjointly named the Saharo-Canarian Circle 40 .Traces of incised lineal writing have recently been found in human sites in the central Sahara (Ti-m Missaou, Algeria) in what would be the first discovery of "megalithic lineal writing" in continental Africa 35 .This finding supports the theory that incised lineal writing was originated in fertile Sahara populations, which moved to the North when desertification began, carrying this type of lineal writing to Canary Islands (Guanche culture), the Mediterranean Basin and Europe.It is probable that these populations of the green Sahara moved towards the Mediterranean and the Canary Islands, originating common cultural and genetic traits among all the descendant populations (Guanches, Egyptians, Minoans, Etruscans, Iberians and Basques among others 38,39 ).The present study and other ones 4,5,30,31,34 , support this theory of movements from central Africa to other regions; the appearance of Mediterranean alleles and semi-haplotypes (like HLA-DRB1*03:05, *03:07, *04:11, *04:20, *11:10 or *13:10; results not shown) in different African populations endorse bi-directional population movements from the Sahara to the Mediterranean and South Europe in prehistoric and more recent times.In some cases, Lineal Megalithic Scripts and later classical letters (or syllables) are found together; it is the case of naviform scripts that are found mixed with Iberian-Tartessian semi-syllabary signs in the same context both in Iberian Peninsula and in Canary Islands.An evolutionary pathway of writing may be observed from primitive Lineal Megalithic Scripts to other more recent and complex alphabets or syllabaries like those found in Sitovo 41 , Gradeshnitsa 42 or Iberian scripts widespread in the Iberian Peninsula, Canary Islands and continental North Africa 43 .Also, other common archaeological elements have been found in Europe and Africa (Canary Islands), including Western Sahara territory 44,45 .The findings pointed out in this study in relation to Sudanese HLA profile support this emigrational theory from the Sahara (and Sub-Saharan areas) to the Mediterranean Basin.
However, other authors with different techniques and genetic markers (SNPs) put forward that admixture with Arabs in recent times has occurred in Sudan and that this area has been subjected to significant population movements in ancient African /Eurasian history 46 .Also, East Sahelians admixture with North Africans is confirmed with other studies 47 , and Fulani nomads, Tuaregs and other Arab speaking Sahelians show admixture with North Africans like most of Sahelian groups do 48 .

Figure 2 .
Figure 2. HLA-DRB1 dendrogram performed with Neighbour-Joining method showing genetic relatedness between Sudanese, Nubians, and other Mediterranean and Sub-Saharan populations.Genetic distances between populations were calculated with GNKDST software using low-resolution HLA-DRB1 frequencies.Bootstrap test showed values of 100 in all nodes after 1000 replicates.Note that Greek HLA relatedness with Sahel populations was detected by two independent different groups4,5,16,28 and it is also supported by other autosomal genetic markers for African and Greeks specifically (3120 + 1 G → A cystic fibrosis marker)29,30 .HLA-DRB1 allele frequencies is used because of the best discrimination between populations among HLA loci (more or less relatedness) and the fact that almost all the populations in data bases are HLA-DRB1 typed31 .

Figure 3 .
Figure 3. Correspondence analysis performed with HLA-DRB1 allele frequencies of Sudanese sample studied in present work together with those of other Mediterranean and African populations included (see Table1).Three different clusters are obtained from the analysis: for simplicity, cluster A (light blue), cluster B (dark blue) and cluster C (yellow/orange).Mixed Sudanese population analysed in this study appears together with other Mediterraneans like Turks, Iranians, Egyptians, Italians or Jews in the eastern-Mediterraneans cluster (cluster A, light blue).However, Sub-Saharan Sahelians (Fulani, Rimaibe, Mossi) and Nubians cluster together with Greeks and other African Sub-Saharan populations (cluster C, yellow/orange), as previously obtained by two independent groups4,5,16,28 .Correspondence analysis supports the results previously obtained in the genetic distances analysis (DA) (Table3) and NJ dendrogram (see Fig.2).Italians have been chosen from North Italy6 , a population that usually cluster with northern Balkan Peninsula or Central European groups.

Table 1 .
Populations used for the present work.N Number of individuals analysed for each population; 1 KZ Kafar Zubian, a Shia Muslim village.

Table 3 .
Genetic distances (DA) between studied Sudanese population and other world populations included in this work (HLA-DRB1).