Analysis of genetic diversity and differentiation of sheep populations in Jordan

Background: Genetic diversity of sheep in Jordan was investigated using microsatellite markers (MS). Six ovine and bovine MS located on chromosomes 2 and 6 of sheep genome were genotyped on 294 individual from ten geographical regions. Results: The number of alleles per locus ( A ), the expected heterozygosity ( H e ) and observed heterozygosity ( H o ) weremeasured.Overall A , H e and H o were12.67,0.820and0.684,respectively.Ontheotherhand,geneticdistances undoubtedly revealed the expected degree of differentiation among the studied populations. The ﬁ nding showed closenessofthreepopulationsfromsouth(Maan,ShowbakandTafeilah)toeachother.Populationsfromthemiddle regions of Jordan (Karak, Madaba, Amman, AzZarqa and Mafraq) were found to be in one cluster. Only two populations of the middle region were an exception: AlSalt and Dead Sea. Finally, sheep populations from Irbid were located in separated cluster. It was clear that the studied prede ﬁ ned populations were subdivided from four populations and would be most probably accounted as ancestral populations. These results indicate that number of population is less than the prede ﬁ ned population as ten based on geographical sampling areas. Conclusions: The possible inference might be that geographical location, genetic migration, similar selection forces, and common ancestor account for population admixture and subdivision of Awassi sheep breed in Jordan. Finally, thepresentstudy sheds newlight on themolecular andpopulation genetics ofAwassisheepfromdifferentregions of Jordan and to utilize the possible ﬁ ndings for future management of genetic conservation under conditions of climate changes and crossbreeding policy.


Introduction
Sheep is common and popular small ruminant livestock in Jordan. Sheep in Jordan is a fat-tailed breed known as the Awassi sheep [1] which possesses great adaptability to tropical environmental conditions. Awassi sheep is often used as a triple purpose sheep, better for high milk production, in most of the countries of the Middle East [1]. They have little variation in morphological characteristics and production and reproduction traits. As a consequence to geographical rearing area and rearing system, there are different Arabic names given such as Baladi (local), Naiemi, and Sahrawi (desert). The differentiation between these sheep strains or breeds has not yet been established on ground. A molecular genetic differentiation of sheep in Jordan was limited to using few populations or using arbitrary and/or limited DNA markers in genetic differentiation studies [2,3]. Advance markers such as microsatellite (MS) and single nucleotide polymorphism (SNP) have not been applied for sheep genetic diversity and differentiation. In particular, MS markers have provided wide opportunities to analyze genetic variability at DNA level in universe sheep breeds. Microsatellite DNA markers are widely used since they are polymorphic and are randomly distributed in the organism's genome [4]. These markers have also been successfully used to study the biodiversity and genetic relationship and differentiation between and within breeds [5,6].
On the other hand, Jordan has experienced sharp reduction in sheep numbers as a result of persistent drought since 2007 [7]. Most of the sheep population is found in the north region of Jordan where drought has mostly been hit. In addition, another major threat to their genetic diversity is from unplanned crossbreeding with exotic improved Awassi strain and other exotic breeds. Based on Galal et al.'s [1] recommendation that limited information is available on the molecular biodiversity of Electronic Journal of Biotechnology 17 (2014) [168][169][170][171][172][173] Awassi sheep in Arabian countries like Jordan, the present study aims to analyze their genetic biodiversity and differentiation.

Sheep populations and sampling
Thirty one populations of Awassi sheep were studied in ten different regions of Jordan as described in Fig. 1. The ten targeted regions are Irbid, Mafaraq, AzZarqa, AlSalt, Madaba, Dead Sea, Kark, Petra, Tafeilah (Dana), and Maan. The populations were of small, medium, large size and few in numbers that were grazed on road sides or reared in backyards. A total of 294 samples of mature unrelated ewes and rams were samples. Sample of 0.5 cm tissue was taken from each animal's ear using an ear puncher. The collected samples were stored at -18°C until extracted for DNA.

Sampling and DNA extraction
DNA extraction was performed using a commercially available protocol of E.Z.N.A.® MicroElute Genomic DNA Extraction Kit [8]. Subsequently, DNA concentrations were estimated by a Nano-DNA spectrophotometer in which the quality of DNA was evaluated using the ratio of A260/A280.

DNA genotyping
Six ovine and bovine MS (Table 1), located on chromosomes 2 and 6, were employed for genotyping experiments using Silver Sequence™ DNA System of Promega® [9]. Selection of the markers was based upon their close linkage to each other on chromosomes 2 and 6. On the other hand, their primers were selected for ease of use in PCR reaction with special regard to the annealing temperature and MgCl 2 concentration in particular. Primer sequences were taken from the Australian Sheep Gene Mapping website [10] and synthesized by BioEngland® (Table 1). PCR reaction utilized a 10 μl volume of DNA and reagents for genotyping. DNA samples were liquated into a 48 well PCR plate. Thermal cycling was performed on an MJ Research PTC-100 thermal-cycler. The Amplified PCR products were resolved on a 5% polyacrylamide gel electrophoresis using a Sequi-Gen GT gel rig for Silver staining [9]. Sequencing ladders were prepared using a fmol® DNA Cycle Sequencing system [9] and 3 μl of each of the four reactions loaded onto the gel, so that the size of the MS alleles was determined. When the electrophoresis run was completed, the gel was recovered and developed. Then, the gel was dried and viewed by the APC Film Development method [9]. The film was developed as a photo picture to be ready for scoring the genotypes. Allele sizes were scored by visual comparison with the sequencing ladder; pGEM®-3Zf(-) Vector.

Genetic analysis
A total of 294 samples were fully informative for the six studied MS loci without any missing genotypes (Table 1). Population genetics of the studied populations was then investigated. Allele frequency and polymorphism under Hardy-Weinberg equilibrium were measured using Genetic Data Analysis (GDA) software [11] and by CERVUS software [12]. The parameters were observed heterozygosity (H o ) and expected heterozygosity (H e ) at each locus [13], and polymorphic information content (PIC) [14].
On the other hand genetic distance matrix between populations was measured using GDA software which utilizes the most widely used measure of genetic distances [15]. The phylogeny tree was drawn using MEGA software [16]. The population structure was analyzed using STRUCTURE software [17] considering an admixture model and correlated allele frequencies between studied populations. The lengths of the burn-in Monte Carlo Markov chain (MCMC) were 1000 and 10000 in 100 runs for possible number of clusters (breeds/population) (K) from 2 to 5. For each K value, logarithmic likelihood probability of data (L[K] = ln Pr(X|K)) and F st values for each cluster were estimated. In addition, GENETIX (ver 4.05) software was used to predefine and identify the studied goat breeds [18].

Genetic variation within populations
AMOVA analysis indicated that the genetic variation percentage among individuals within population was 5.55%, whereas it was 12.26% among populations. Estimates of allelic number at every locus were calculated to describe the genetic variation within and between populations ( Table 2). The average number of alleles per loci for all population and at studied loci was 12.67. At a single locus, the number of allele was lowest (9) at INRA40 and highest (18) at ILSTS30. This result indicates a high number of MS alleles in studied sheep populations and thus could be used to predict the level of the genetic variation and genetic differentiation within and between populations. Similar observation was reported by Arranz et al. [19,20] who estimated MS variation of Spanish Merino sheep. Furthermore, Iovenko [21] and Peter et al. [22] reported that high sheep population genetic variations were due to high allele variation profiles. On the other hand, Meadows et al. [23] reported that the analysis of genetic diversity within five sheep breeds showed that Merino contained the highest genetic diversity as average number of alleles observed per locus was 8.13, whereas Macarthur Merinos contained the lowest amount of diversity, with an average number of alleles of 3.03.
The H e and H o as measures of genetic diversity at a single locus, are shown in Table 2 for all populations and in Table 3 for each locus.
Overall H e and H o values were 0.678 and 0.734 for all populations. In most of the cases, H o was lower than H e , except that for Tafeilah sheep population and resulted overall in a slightly higher average H e of 0.734 for all populations (Table 2). In particular, sheep population of AlSalt has the highest variation value of 0.841, whereas sheep of AzZarqa has the lowest value of 0.696 (Fig. 2). Fig. 2 shows the heterogeneity level of each population at each studied locus. The average H e and H o for all loci were 0.820 and 0.684, respectively  Table 3). At all loci, H o was lower than H e . It is good to mention that, in this study, all loci had significant deviation from HWE, showing heterozygote advantage. These results mean that Jordan sheep are genetically subdivided. In general, deviation from HWE is probably a result of inbreeding, selection, and/or migration, then populations can be considered structured. These results were similar to a study used the same MS markers on three flocks of Awassi sheep in south Jordan, and H o and H e were 0.67 and 0.70, respectively [3]. The heterozygoties of Awassi, Kivircik, and Akkaraman breeds and two of their crossbreeds of Turkey were high and ranged from 0.667 to 0.782 [24]. Arora et al. [25] reported that both H o and H e averaged 0.665 and 0.786, respectively and ranged from 6.40 to 7.92 in six Indian sheep breeds. In addition, Meadows et al. [23] reported that the analysis of genetic diversity within five sheep breeds showed that Merino contained the highest genetic diversity as average H e was 0.70, even though it was as low as 0.4 for Macarthur Merinos. The studied populations revealed a high level of genetic variation seen from a high number of alleles per locus, H e , and H o . These results were similar to those reported in other studies of MS markers in sheep populations such as those of Bancroft et al. [26] and Coltman et al. [27]. The most genetically variable population was found to be AlSalt, which had more alleles per locus, higher H e and H o . This might be a result of gene flow into this population as well as due to sampling variation. This observation was explained as when genetic material, in the form of male gametes mainly, migrates from one population to another [28] as cases of introducing breeding rams from other flocks. However, the high level of genetic variation in populations could be due to several possible factors such as management and breeding practices under which selection took place [29,30,31]. Table 2 and Table 3 show inbreeding coefficients (F is ) for each population and at each locus respectively. F is was positive in all  population except Tafeilah, indicating that more inbreeding was found than it was expected in all sheep populations (  (Fig. 3). Overall results indicate an evolutionary distinction of sheep of the south regions from those of the north regions of Jordan, whereas sheep of the middle regions was less differentiated. For example, Madaba sheep had least genetic differentiation from Karak sheep (Fig. 3). Similar results were found in literature indicating that convergence and divergence between studied populations. Overall F st value was low for the North Spain sheep populations (0.061) [28]. In addition, F st estimates in Spanish sheep reached similar (0.073) and an estimation was also obtained including were slightly greater (0.092) [20].

Genetic distances and phylogenetic tree
The pairwise genetic distance between populations was relatively short for population such as Maan, Tafeilah and Showbak but long for population of Kark with Mafraq, AzZarqa, Irbid, Dead Sea and AlSalt. The distances between these populations were described by drawn the UPGMA evolutionary phylogenetic tree which is shown in Fig. 4. In details, the cluster analysis based genetic distances revealed relatively short distances between (Node 11 = 0.057) Karak sheep and Madaba sheep. Fig. 4 revealed a very clear degree of differentiation of three clusters. Northern populations (Irbid, Mafraq) and some middle region populations (Madaba, Kark; Node 12 = 0.103115) and AzZarqa (Node 13 = 0.110) were grouped in one cluster. Other sheep populations of the middle region (AlSalt and Dead Sea; Node 17 = 0.348) were in a separated second cluster. The third cluster was comprised of sheep populations of the south region (Petra, Tafeilah and Maan; Node 18 = 0.526). In general, this result was similar to previous finding of closeness three native Awassi sheep populations of southern Jordan [3]. In addition, it is indicating a long evolutionary distance separation between southern sheep populations and populations of other geographical areas. In addition, the notable result was that populations of the middle region were located in the middle of the phylogenic tree. In general, this finding firstly supports the closeness of populations geographically close to each other. These results are in agreement with the known history of the populations in regard to their location and thus possible gene flow and their common ancestors. Overall, the clustering of populations in consensus UPGMA trees followed previously mentioned genetic differentiations between studied sheep populations (Fig. 3). Another scope for benefiting the results is that it might shed light on the traits of interest of closely related populations to understand the mechanisms of evolution considering the interaction with their own habitat on one hand. On the other hand, considering closer phylogenic populations for introgression and crossbreeding programs is hope of benefiting.

Genetic structure and admixture
Genetic structure analysis of studied sheep population was performed using STRUCTURE software with population number (K) ranging from K = 2 to K = 5 assuming admixture and correlated allele frequencies models. The likelihood values of the bootstrap samples were for choosing the optimal K value leading to the most reliable results membership coefficients to four clusters (K = 4; average likelihood = -7638.84 and alpha value = 0.062) (Fig. 5). It was clear that four populations would be most probably accounted as ancestral populations for current studied populations. This resulted case indicates that the number of population groups is less than predefined population as ten based on geographical sampling areas. As a consequence, individuals of all studied populations are subdivided into a group of four populations as presented in Fig. 6. Fig. 6 represents that all individuals of Irbid sheep population were assigned by more than 80% to the first cluster, except few individual of less than 80% assignment probability. On the other hand, most individuals of southern sheep populations (Petra, Tafeilah, Maan) were solely assigned to the fourth population. The second cluster has individuals of Mafaraq sheep with different assignment probabilities (~60%) (Fig. 6). On the other hand, the third cluster was formed from AlSalt and Dead Sea sheep individuals, mostly by~90% admixture proportion (Fig. 5). The notable result was that admixture evidence was found in majority of Mafaraq, AzZarqa, Madaba and Kark sheep individuals. This result is probably shared ancestry between those studied populations. On the other hand, it might be due to the migration of individuals that usually occurred in the regions. In fact, Mafaraq and AzZarqa regions were considered as livestock rearing and trading regions where sheep from other regions are gathered for meat production and marketing. Structure and admixture analyses were performed on different sheep populations. For example, Álvarez et al. [32] reported that admixture analysis performed on the parental role of Burkina-Sahel and Djallonké sheep breeds from Africa and Mossi sheep breeds was a hybrid population nearer to the Djallonké breed. Furthermore, Ligda et al. [33] provided the genetic structure of ancestral populations of Greek sheep breeds. Turkish Awassi sheep as a fat-tail sheep was separated from other Turkish sheep breeds based on correspondence analysis [34].
In conclusion, the studied populations of Jordan Awassi sheep revealed a high level of genetic variation expressed by the number of alleles, H e and H o measurements at population and loci levels. The latter showed a high allele number which reflected in high estimates of H e and H o at the six studied loci. At population level, these values were high and close to each other and reflected in explaining the level of genetic differentiation between the studied sheep populations. Genetic distances and phylogeny also undoubtedly revealed the degree of differentiation in the populations. It was clear that the studied predefined populations were subdivided from four populations and would be most probably accounted as ancestral populations. This resulted case indicates that the number of population groups is less than predefined population as ten based on geographical sampling areas. The possible inference might be that geographical location, genetic migration, similar selection forces, and common ancestor account for population admixture and subdivision of Awassi sheep breed in Jordan. Finally, the present study sheds new light on the molecular and population genetics of Awassi sheep from different regions of Jordan and to utilize the possible findings for future management of genetic conservation under conditions of climate changes and crossbreeding policy.