DS in Guinea-Bissau can be classified into three “formats”: “large format” in the Bafatá region, “intermediate format” sheep in the Gabù and Oio regions and sheep of “small format” in the Cacheu region. Indeed, the average values of the quantitative characters (CG, CD, HW, BL, EL and TL) of the Bafatá DS subpopulation were significantly higher than those obtained in the Gabú, Oio and Cacheu regions. This format gradient could be explained by the differences in the agro-ecological conditions and the farming practices. In fact, the agro-ecological area of the North-East, which includes the Bafatá, Gabú and Oio regions, is characterized by savannah trees and clear forests, which offer rich natural pastures to pastoralists who are Fulani and Mandingos. Moreover, the livestock is dominated by ruminant species. Contrariwise, in the North-West agro-ecological zone including the Cacheu region, ruminant species (sheep, goat and cattle) are mainly raised for ritual ceremonies by breeders who are rather animistic . In addition, this zone is covered with wooded savannahs and dense forests hardly accessible by animals, hence the predominance of the sedentary system in the Cacheu region. At the cultural level, Bafatá region is mainly populated by Fula-speaking people, practicing the Muslim religion and traditionally attached to animal husbandry compared to the other regions (Cacheu and Oio) where the populations are strongly Christianized and more attached to pig farming. The Bafatá region is also a large area of ruminant species concentration during the transhumance period and hosts the most important livestock market in the country. This region generally receives animals from Gabú and both Gabú and Bafatá regions have more than 70% of the country's ruminant livestock . During the dry season (November to May), ruminants from the Gabú region migrate to the Bafatá and Oio regions .
Sheep from the Cacheu region had the smallest size in the study area. In fact, Cacheu is one of the regions of the North-West agro-ecological zone with high humidity favorable to parasitism and vectors of pathogens such as tsetse flies which transmit the trypanosomes causing African animal trypanosomosis.
DS subpopulations of the Gabú and Oio regions were highly heterogeneous with an “intermediate format”, probably due to the introduction of improving rams in these regions in the past . This heterogeneity is observed not only between regions but also within region (Figure 1). The effect of the agro-ecological zone on the format of ruminants, especially sheep, has been previously reported in Côte d'Ivoire in DS , in Senegal with Peul-peul (Fulani) sheep  and in Togo in Vogan Sheep and DS . A recent morphobiometric characterization of DS in the sudano-guinean zone of Cameroon revealed three genetic types  as observed in the present study in Guinea-Bissau. In Burkina Faso, Traoré et al.  described a sheep population named "Mossi sheep" which is a savannah DS found in an agro-ecological zone between the sudano-sahelian zone and the sudano-guinean zone with an "intermediate format" between DS and sahelian sheep.
The average values of HW obtained (55.67 ± 4.16 cm for the Bafatá region, 54.67 ± 3.44 cm for the Oio region, 53.44 ± 2.85 cm for the Cacheu region and 53.23 ± 3.77 cm for Gabú region) are closed to those reported by Dayo et al.  in DS in Togo (HW = 54.63 ± 8.23 cm; BL = 58.47 ± 6.30 cm and CG = 74.72 ± 8.28 cm) and Sangaré  in DS in West Africa and Gueye  in Senegal. Similar results have also been reported in other populations of DS in Ghana (HW = 57.06 ± 0.28 cm; BL = 54.87 ± 0.35 cm and CG = 69.19 ± 0.41 cm) by Birteeb et al.  and Asamoah-Boaheng and Sam  and in Côte d'Ivoire (HW = 59.60 ± 5.40 cm; BL = 57.80 ± 5.40 cm and CG = 70.80 ± 6.50 cm) by N'Goran et al. . However, the values of the present study were higher than those previously reported by Hadzi  in DS in Togo and in Guinea-Bissau . These results could be explained by the differences of climatic conditions of the agro-ecological zones in which these studied populations are bred, the study periods of the year (season effect), the farming systems or the genetic variability that could be observed between DS populations across the countries. It has been reported the existence of two sub-categories of DS [22, 23] and DS of savannah are larger than those of forest zones , demonstrating once more the effect of the agro-ecological zone on the format of this sheep breed.
The tail of the DS is thin and relatively long. The average TL (28.06 ± 4.36 cm) is similar to those reported by N’Goran et al.  in DS in Côte d'Ivoire (24.70 ± 3.40 cm) and in Togo (27.47 ± 8.05 cm) . This TL is longer than those reported in the DS (West African Dwarf) by Gbangboche et al.  in Benin (17 cm), in Nigeria (19.42 ± 0.63 cm)  but shorter than those of the Sahelian sheep (48.20 ± 5.37 cm) and Vogan sheep from Togo (45.24 ± 6.23 cm) .
Concerning the ear length, the value obtained (9.85 ± 1.12 cm) is similar to value reported by Gbangboche et al.  in West-Africa, who found that DS has small ears, about 10 cm. However, the value in the present study is lower than those reported in DS in West Africa: 13.03 ± 0.39 cm in Nigeria , 11.61 ± 2.61 cm in Togo ) and in the Peul-peul (Fulani) sheep (13.30 ± 1.20 cm) in Senegal ; and significantly shorter than those recorded in Vogan sheep (18.45 ± 2.08 cm) and Sahelian sheep (21.63 ± 2.48 cm) . No sexual dimorphism was observed for this trait contrary to Gueye  who showed that male sheep and goats had slightly longer ears than females in Senegal.
The coat color pattern in DS in Guinea-Bissau is dominated by the uniform white pattern and the spotted white and brown / fawn pattern in all regions. In the Bafatá region, the frequency of the spotted pattern is higher than in the other three regions. Indeed, for the Muslim populations in Bafatá and Gabú regions, the rams are preferentially slaughtered while the uniform white or spotted ewes are kept for the reproduction in order to have the offspring with white coat color. This explained the presence of only few rams in most of herds. The higher proportion of animals with uniform white color pattern could also be due to a strong selection of animals expressing the white coat color to meet the livestock market demands (higher price than other coat colors) and the cultural preference in the country (religious sacrifices or gifts during baptism celebrations and the “Eid El-Kebir” (Tabaski) celebration or for the dowry). The preferences for the coat color of animals differ from one society to another. For example, in southern Ethiopia, red coat color for ewes is the most suitable for market demands . In Côte d'Ivoire, the DS had at 55.00% patchy white-black coat color compared to 24.00% uniform white coat , and only 5.88% of the DS were white in southern Togo . This diversity for coat color in DS in West Africa is linked to the choices made by the societies in which these animals are raised. In Ferlo zone in Senegal, the dominant coat color of the Peul-peul sheep has evolved from patchy (white-black or white-red/fawn)  to spotted of white and black / red/fawn .
The ears of DS in Guinea-Bissau are mostly erected horizontally and only 2.67% of animals in the Bafatá region had slightly drooping ears. These results agree with those of Dayo et al.  in DS from Togo (86.27%) and N’Goran et al.  in Côte d'Ivoire (87.00%). Drooping ears in DS are considered to be the result of Sahelian sheep genes introgression [12, 14]. Thus, the presence of animals with slightly drooping ears in the Bafatá region (2.67%) could be explained by crosses occurred with Sahelian sheep from neighboring countries, especially from Senegal.
Sexual dimorphism has been observed for the presence of horns with only 6.75% females horned in our study. This proportion is higher than the 2.30% often reported for ewes wearing horns (most are stumps); but lower than the 14.60% of Mossi ewes carrying horns in Burkina Faso . The horns are developed for rams and absent or in stumps in ewes. In the current study, the most of horned ewes were from the regions of Bafatá and Oio where small ruminants and cattle move during the transhumance in the dry season . Horned ewes are thought to have come from crossing with transhumant animals. It is important to highlight that in half of these ewes, the horns are in stumps.
The horn shapes were significantly different according to the zone: horns laterally straight were the most observed in Bafatá, Gabú and Oio regions while spiral horns facing forward predominated in the forest and humid Cacheu region similarly to the one reported by Dayo et al.  in the south of Togo.
Molecular genetic diversity
The current study provides the first information on molecular genetic characterization of DS in Guinea-Bissau and is complementary to the morphological characterization of this breed. This study presents a comprehensive genetic analysis of DS, the assumed only sheep breed of Guinea-Bissau, from four administrative regions covering two agro-ecological zones. The genetic diversity of subpopulations was influenced by the agro-ecological zones. Similar observations were reported by prior studies in West African DS . Indeed, these authors had reported that Malian, Gambian and eastern Guinean DS populations had higher genetic diversity than those from Senegal and southern and western Guinean using expected heterozygosity (He) and the mean number of alleles (Na). Based on the He, Cacheu and Oio DS subpopulations would be closer to Senegalese, Gambian southern and western Guinean populations while Bafatá and Gabú DS presented similar expected heterozygosities to Malian and eastern Guinean DS. The Na in the current study (7.42 ± 2.19) was similar to those obtained by Wafula et al.  in Guinean and Malian DS and Agaviezor et al.  in West African Dwarf sheep in Nigeria. However, the allelic richness (adjusted mean number of alleles) values were lower than those reported by Wafula et al.  and Agaviezor et al.  and probably due to the small sample size used for genotyping in our study.
Genetic structure of the population
Using different population differentiation parameters (FST, GST, genetic distance, genetic identity) and representation (NJ Tree and FCA), our results showed that the population differentiation over the 4 subpopulations is very low since the multi-locus FST and GST values indicated that only 2.9% and 4.3% respectively of the total genetic variation were due to the subpopulation differences. The remaining 97.1 for FST and 95.7 for GST corresponded to differences between individuals within the subpopulations. These values were lower than those (8.8% for FST and 12% for GST) reported by Agaviezor et al.  in four sheep populations in Nigeria (Udah, Balami, Yankasa and West African Dwarf sheep also known as DS). Even though the genetic differentiation observed between the four DS subpopulations in Guinea-Bissau was low, the current study pointed that the subpopulation from Cacheu region slightly differs from those in Gabú, Bafatá et Oio regions. Indeed, these three subpopulations are genetically close even though they come from geographically different locations. This similarity is shown by: i) the high genetic identity (from 0.9603 to 0.9017) of the three subpopulations while this value decreased to 0.8683 between Bafatá and Cacheu subpopulations, ii) the low genetic distances between the three subpopulations. The closest Nei's  unbiased measures between Bafatá and Gabú, and the farthest between Bafatá and Cacheu may be due not only to their geographical locations but also to the breeding systems, the presence of the livestock market in Bafatá and the cultural behavior of the breeders in the different regions. Ira et al.  reported that Bafatá, Gabú and Oio regions had 95.88% of the sheep population of Guinea-Bissau and breeders practice transhumance breeding system, mixing cattle and sheep while in Cacheu region the breeding system is rather sedentary in association with agriculture (production of mangrove rice, sorghum, millet, beans, peanuts and cashew).
The molecular study on DS population of Guinea-Bissau confirmed the results obtained from phenotypic study and three genetic groups were distinguished as shown by FCA and NJ tree.
Genetic improvement strategies of Sheep breed in Guinea-Bissau
From the findings of the results of this study, different guidelines may be suggested for the development of breeding strategies of the local sheep breed of Guinea-Bissau. These include:
- the improvement of livestock management through the strengthening of technical capacities of livestock farmers (herd management, animal health monitoring, techniques for the production and storage of livestock feed) and other actors in the local sheep value chain (public services, processing, marketing);
- the rehabilitation of Governmental sheep stations for the improvement of the DS and the establishment of a breeding and dissemination schemes adapted to the local context and ex-situ conservation (cryopreservation of genetic materials);
- the establishment of breeder associations for community-based genetic improvement, in-situ conservation and promotion of local breeds.
The objectives of these strategies should be to improve growth and meat production traits, while maintaining reproductive and adaptive traits by involving local communities. The question of the choice of strategies must be dealt with taking into account the cultural consideration of each region of Guinea-Bissau. The use of crossbreeding schemes requires rigorous monitoring of Governmental services to avoid un-controlled crossbreeding and therefore a heterogeneity of Djallonké breed in the country.
The different options to consider must be based on the complementarity of traits (production, adaptation to the local environment and cultural activities of the breeders).