Previous PageTable of ContentsNext Page

Genetic Relationships among five Ecotypes of
Sheep in the United Republic of Tanzania

John Stephen,[18] Clemens B.A. Wollny[19] and P.S. Gwakisa[20]


Indigenous sheep of the United Republic of Tanzania have evolved in diverse environments and represent unique combinations of genes that highlight productive and adaptive traits. These locally adapted sheep populations are endangered through breed substitution and lack of appropriate development. There is therefore an urgent need to act to prevent the rapid erosion of these valuable sheep genetic resources. This study was conducted to determine the genetic relationships among five Tanzanian sheep ecotypes kept by Maasai, Gogo, Sukuma and Konde communities in the country. These sheep ecotypes are mostly raised in pastoral, agropastoral and agricultural production systems. A total of 30 random amplified polymorphic DNA (RAPD) and 198 microsatellite polymorphic alleles were used to study genetic relationships among the sheep ecotypes. The results suggest closer relationships between ecotypes that are located close to one another than between ecotypes separated by greater geographical distance. Clear evidence of broad genetic diversity was observed among the studied sheep ecotypes. The diversity within the ecotypes makes genetic improvement through selection a possibility. Once they have observed that Tanzanian sheep possess different genetic attributes, farmers will have the benefit of choosing traits that will be included in future breeding programmes. The authors recommend the formulation of policies to promote the use of indigenous sheep genetic resources in breeding programmes rather than the usual practice of cross-breeding with exotic genotypes.


The United Republic of Tanzania is well endowed with a diverse range of livestock populations of various species, which contribute significantly to the food security and welfare of rural households. These livestock species are able to thrive in the harsh and varied environment found in the country.

In recent years, some livestock projects aimed at improving the productivity of indigenous animals have been undertaken. Most such programmes for small ruminants have either resorted to cross-breeding with exotic breeds or importing live animals (Das and Sendalo, 1991). The improvement of local breeds requires knowledge of the differences in genetic or phenotypic characteristics between breeds. Little is known about genetic or inherent valuable traits of the Tanzanian domestic animal breeds, many of which have doubtful classification. Animals are often characterized based on their phenotypic and/or genetic constitution (Madubi, Kifaro and Petersen, 2000).

The Tanzanian livestock industry is at present dominated by smallholder farmers in pastoral and agropastoral production systems. The systems account for about 99 percent of the total livestock production. The Tanzanian sheep population is primarily of local origin and not characterized as belonging to specific breeds. Yet, ecotypes or geographically separated populations do exist. Several such populations are spread over different regions of the country. Although quite different morphologically, African sheep are described as thin-tailed, fat-tailed or fat-rumped (Mason and Maule, 1960). This kind of classification is not enough for the rational utilization of the animal genetic resources available, or for conservation purposes.

One way to compare native breeds or ecotypes is to evaluate the level of divergence among them. In this way, highly divergent breeds can be chosen for evaluations. Moreover, if the level of divergence between local breeds is known, appropriate choices can be made for the conservation of genetic material.

In this study, five sheep ecotypes found in five Tanzanian regions were genetically characterized to establish the relationships between them. The regions were Arusha, Mwanza, Dodoma, Coast and Mtwara, and the sheep were kept by Maasai, Gogo, Sukuma and Konde communities. The ecotypes were compared using a variety of DNA markers and similarities between ecotypes were calculated. A hierarchy of relationships between the ecotypes was defined.

Materials and methods


A total of 94 blood samples were collected from unrelated Tanzanian sheep found in the Arusha, Mwanza, Dodoma, Coast and Mtwara regions. Two breeds, West African Dwarf sheep from West Africa and NorthRonaldsay sheep from England, were also included in the study as reference breeds.

Polymerase chain reaction (PCR) amplification

Two genotyping techniques, random amplified polymorphic DNA (RAPD) and microsatellite, were used in the study. RAPD loci included ILO525, ILO526, ILO1204 and ILO1205, whereas the microsatellite loci were TGLA53, OarJMP29, McM42, ILSTS005, OarFCB129 and HCS. The Oligonucleotide Unit of the International Livestock Research Institute (ILRI), Kenya supplied all loci used in the study. The RAPD loci cycling conditions consisted of 45 cycles, each for three seconds at 95 °C, 30 seconds at 94 °C, 45 seconds at 36 °C, two minutes at 72 °C and 15 minutes at 4 °C. The microsatellite loci cycling conditions consisted of 40 cycles each for four minutes at 94 °C, 45 seconds at 94 °C, one minute at X*, one minute at 72 °C, 20 minutes at 72 °C and 15 minutes at 4 °C (X* is the annealing temperature, which ranged from 50 °C to 60 °C, depending on the primer).

The RAPD PCR products were electrophoresed at 80V in 2 percent agarose gel immersed in Tris borate EDTA (Sigma Chemicals, St. Louis, United States) containing ethidium bromide, and the microsatellite PCR products were separated in 4.25 percent denaturing polyacrylamide gels using ABI PRISM 377 DNA sequencer.

Statistical analysis

The microsatellite data set was prepared using the microsatellite toolkit for MS Excel 97. RAPD allele frequency data were determined by direct counting. Genetic distance between breeds was calculated in two different ways. The neighbour-joining tree was constructed using allele frequency data, according to Takezaki (1998) for RAPD data. The DISPAN program (Tatsuya, 1993) was used in genetic distancing based on microsatellite data. Individual animal comparison was done using the neighbour-joining method (Saitou and Nei, 1987) and PHYLIP (version 3.5c) computer package (Felsenstein, 1993). Bootstraps resampling including 1 000 replicates was performed to test the robustness of the topology of the dendrograms.

Genetic distances were estimated using Nei’s (1972) standard genetic distance formula,

where D = Nei’s standard genetic distance, Xi and Yj are frequencies of the ith and jth allele respectively drawn in populations X and Y.


Genetic diversity analysis of local Tanzanian sheep as revealed by RAPD markers

Ten pair-wise comparisons were made for each primer. The average bandsharing indices were highest in Dodoma - 0.669 (0.148) - and lowest in Arusha - 0.521 (0.050) (see Table 1).

On average, the interecotype bandsharing values were lower than intra-ecotype bandsharing values. Arusha and Mwanza ecotypes showed the highest level of shared DNA fingerprints, 0.552 (0.091), whereas Mwanza and Mtwara ecotypes had the lowest bandsharing value, 0.302 (0.042), as shown in Table 2.

A total of 30 polymorphic alleles were amplified by the four loci in five ecotypes. The allele frequencies in the four loci were informative enough to make a comparison of the ecotypes. The average heterozygosity values varied from 0.137 (Dodoma) to 0.203 (Arusha).

Microsatellite DNA variation among the five Tanzanian sheep ecotypes

Six loci, namely McM42, ILSTS005, OarJMP29, TGLA53, HSC and OarFCB129, were typed in the five sheep ecotypes. The total number of polymorphic alleles was 198 and the mean number of alleles varied from 5.50 (Mtwara) to 7.17 (Mwanza). Observed heterozygosity values showed the highest value in the Arusha ecotype (0.719) and the lowest in the Mwanza ecotype (0.592).

The genetic distances, as determined using Nei’s distance formula between each pair of ecotypes, are shown in Table 3. The robustness of the resulting phylogenetic tree is represented in Figure 1. The most striking feature of the topology was the fact that the reference breeds were clearly outbred from Tanzanian sheep.

Relationships between individual animals

In order to authenticate the relationships between ecotypes displayed by genetic distance and dendrogram, an analysis of a comparison of individual animals, regardless of ecotype, was performed. Some reference breeds were included in the comparison to serve as out groups. The results show that the Tanzanian sheep individuals are clustered together away from the reference groups, as can be seen in Figure 1.


The study has demonstrated that by using random primers of arbitrary nucleotide sequences and microsatellite markers it is possible to show fingerprints of individual animals and ecotypes (or populations).

Higher intra-ecotype bandsharing indices suggest that animals sampled in Dodoma had the highest degree of homogeneity while the Arusha ecotype had the lowest (and the lowest indices). As expected, the interecotype bandsharing indices were lower than the intra-ecotype values. The interecotype bandsharing indices showed that there were more similarities between Arusha and Mwanza - 0.552 (0.091) - than there were between Mtwara and Mwanza ecotypes - 0.302 (0.042). Such a trend can be explained by the geographical separation of these animals: animals from regions close to one another show more similarities than those from geographically distant regions. It was not possible to find any ecotype-specific RAPD fingerprints from the ecotypes studied. This may be because the range of RAPD loci screened was narrow and the study sample size small.

The high level of average heterozygosity within the Arusha ecotype (0.203) and the low level of this parameter in the Dodoma ecotype (0.137) are the reverse of bandsharing indices. This finding suggests that there are higher proportions of heterozygote genotypes in the Arusha ecotype than in the Dodoma ecotype. High heterozygosity values might be a result of the introduction of outbred populations and random mating among the animals of the Arusha ecotype. This supports the notion described above that ecotypes with a high level of bandsharing values have a low level of genetic differences while the reverse is true of ecotypes with low heterozygosity values.

Table 1. Bandsharing within the five ecotypes


n = 20

n = 20

n = 14

n = 20

n = 20

























Average (SEM)

0.521 (0.050)*

0.651 (0.125)

0.616 (0.174)

0.669 (0.148)

0.548 (0.148)

*Indicates the standard error of the mean (SEM).

Table 2. Bandsharing between ecotypes







Average bandsharing for four primers (SEM)

Arusha - Mwanza





0.552 (0.091)*

Arusha - Mtwara





0.428 (0.057)

Arusha - Dodoma





0.487 (0.061)

Arusha - Coast





0.357 (0.054)

Mwanza - Mtwara





0.302 (0.042)

Mwanza - Dodoma





0.376 (0.062)

Mwanza - Coast





0.367 (0.062)

Mtwara - Dodoma





0.383 (0.116)

Mtwara - Coast





0.339 (0.126)

Dodoma - Coast





0.448 (0.171)

*Indicates the standard error of the mean.

Table 3. Standard genetic distances between the five sheep ecotypes






W. African Dwarf















W. Afric. Dwarf













According to Nei (1975), heterozygosity is a good measure of genetic diversity of polymorphic loci. The higher the heterozygosity values, the broader the genetic diversity. In this study, higher genetic diversity has been displayed within the Arusha ecotype than within the other four ecotypes. This suggests that animals sampled from the Arusha ecotype showed a high degree of unrelatedness. Higher heterozygosity values suggest that individual sheep were totally unrelated, but another possible explanation is the occurrence of recent mutation after the divergence of the ecotypes, which could also have been influenced by the presence of fitness traits such as disease resistance. However, this issue was not addressed in the study.

The standard genetic distance matrix showed that the Mwanza and Mtwara ecotypes were the furthest apart - 0.2214 (0.0589) - whereas Arusha and Mwanza were the closest - 0.0190 (0.0218). This observation was not surprising, considering that the former two ecotypes were sampled from two opposite ends of the country. The Mwanza and Mtwara regions are over 2 000 km apart. Accordingly, a "genetic barrier" created by the lack of contact between some regions must have had an influence on genetic distance among these ecotypes.

The phylogenetic tree (Figure 1) showed that Arusha was grouped together with Mwanza; Dodoma was in the same group as Mtwara. The West African Dwarf sheep, which was used to replace the Tanzanian ecotypes as an outbred population, is seen to be closer to Tanzanian ecotypes than NorthRonaldsay (from England), which was the furthest away. This again can be explained in terms of history and the geographical separation of these ecotypes. Genetic distances between these ecotypes conform to the expected relationships among the ecotypes, given that the Mwanza and Mtwara regions are distant from one another whereas the Mtwara and Coast regions are not - both have an Indian Ocean coastline.

Figure 1. A neighbour-joining tree showing the genetic relationships among the local Tanzanian sheep ecotypes and two reference breeds using genetic distance. The numbers indicate bootstrap values (1 000 resampling)

Individual animal relationships (as seen in Figure 1) showed that Tanzanian sheep were clustered together. This implies that there is a closer relationship between the Tanzanian ecotypes than there is between these ecotypes and the outbred populations. The West African Dwarf sheep and NorthRonaldsay sheep were clustered away from the Tanzanian ecotypes. This indicates that there is some similarity between the two breeds and they are genetically different from Tanzanian ecotypes.

In summary, the findings suggest a closer relationship between ecotypes in close geographical proximity than those separated by greater physical distance. Clear evidence of broad genetic diversity was observed among the sheep ecotypes studied. Because of the existence of broad genetic diversity in Tanzanian sheep, genetic improvement through selection, with the main emphasis on community-based breeding programmes, is possible.

Having observed that Tanzanian sheep possess different genetic attributes, farmers will have the benefit of choosing traits that will be included in future breeding programmes. This will help to dispel the misconception that Tanzanian sheep are of low genetic potential and are all of the same type.

The authors recommend the formulation of policies that will promote the use of indigenous sheep genetic resources in breeding programmes rather than the usual practice of cross-breeding with exotic genotypes.


The authors are grateful to the SACCAR/GTZ programme for financial support to the first author in the form of an M.Sc. scholarship, which made this study possible. Support to enable the first author to attend the workshop was provided by the Centre for Tropical Agriculture (CTA) and channelled through GTZ. The authors wish to thank Dr Olivier Hanotte of ILRI, Kenya, for supervising the experimental design, microsatellite genotyping and statistical analysis of the study.


Das, S.M. & Sendalo, D.S.C. 1991. Small ruminants highlights in Tanzania, pp. 16-18. Dar-es-Salaam, Ministry of Agriculture.

Felsenstein, J. 1993. PHYLIP - Phylogeny Inference Package, Version 3.5c. Department of Genetics, Washington University, Seattle, WA.

Madubi, M.A., Kifaro, G.C. & Petersen, P.H. 2000. Phenotypic characteristics of three strains of indigenous goats in Tanzania. AGRI, 28: 43-51.

Mason, I.L. & Maule, J.P. 1960. Indigenous livestock of eastern and southern Africa. Technical Communication No. 14. Commonwealth Bureau of Animal Breeding and Genetics, Commonwealth Agricultural Bureau, Farnham Royal, UK. 240 pp.

Nei, M. 1972. Genetic distance between populations. American Naturalist, 106: 283-292.

Nei, M. 1975. Molecular genetic population and evolution, pp. 128-170. Amsterdam, North-Holland Publishing.

Saitou, N. & Nei, M. 1987. The neighbour-joining method: A new method for reconstruction of phylogenetic trees. Molecular Biology and Evolution, 4: 406-425.

Takezaki, N. 1998. Neighbour-joining tree construction from allele frequency data (NJBAFD). Shizuoka, Japan, National Institute of Genetics.

Tatsuya, O. 1993. Genetic distance and phylogenetic analysis, "DISPAN". Institute of molecular evolutionary genetics. Pennsylvania State University, Pennsylvania, USA.

[18] Mogabiri Farm Extension Centre, Diocese of Mara, PO Box 134, Tarime, United Republic of Tanzania
[19] Bunda College of Agriculture, University of Malawi, PO Box 219, Lilongwe, Malawi
[20] Department of Veterinary Microbiology and Parasitology, Sokoine University of Agriculture, PO Box 3019, Morogoro, United Republic of Tanzania (E-mail:

Previous PageTop of PageNext Page