Turkish Journal of Biology
Turk J Biol (2013) 37: 176-183 © TÜBİTAK doi:10.3906/biy-1203-64
http://journals.tubitak.gov.tr/biology/
Research Article
Genetic variability among native dog breeds in Turkey 1,
2
3
1
Metin ERDOĞAN *, Cafer TEPELİ , Bertram BRENIG , Mine DOSAY AKBULUT , 1 4 5 Cevdet UĞUZ , Peter SAVOLAINEN , Ceyhan ÖZBEYAZ 1 Department of Medical Biology and Genetics, Faculty of Veterinary Medicine, Afyon Kocatepe University, Afyonkarahisar, Turkey 2 Department of Animal Science, Faculty of Veterinary Medicine, Selçuk University, Konya, Turkey 3 Institute of Veterinary Medicine, Göttingen University, Göttingen, Germany 4 Division of Gene Technology, KTH-Royal Institute of Technology, Science for Life Laboratory, Solna, Sweden 5 Department of Animal Science, Faculty of Veterinary Medicine, Ankara University, Ankara, Turkey Received: 30.03.2012
Accepted: 07.09.2012
Published Online: 25.03.2013
Printed: 25.04.2013
Abstract: In this study, the genetic structures and relationships of native Turkish dog breeds were investigated using 20 polymorphic loci (17 microsatellites and 3 proteins). For this aim, a total of 141 blood samples were taken from Turkish shepherd dogs and Turkish Greyhounds located in several geographical regions of Turkey. Multilocus FST values indicated that around 1.92% of the total genetic variation could be explained by breed differences and the remaining 98.08% by differences among individuals. The gene flow between populations within each generation varied between 8.4 (Akbash–White Kars Shepherd dog pairs) and 62.3 (Black–Grey Kars Shepherd dog pairs). Four different groups appeared in the 3-dimensional factorial correspondence analysis, and among these, dogs from the Akbash, Kangal, Kars Shepherd, and Turkish Greyhound breeds grouped in clearly separated clusters in distant parts of the 3-dimensional graph. These results clearly show that Akbash and Kangal Shepherd dogs are different populations with different genetic structures. Therefore, the generalised grouping of Turkish shepherd dogs into a single breed called Anatolian or Turkish shepherd dogs is incorrect. Key words: Turkish dog breeds, microsatellite, F-statistics, genetic variability, factorial correspondence analysis, genetic distance
1. Introduction Archaeological findings show the presence of domestic dogs in Germany, Israel, and North Iraq at 14,000–16,000 years before the present (BP) (1,2). Still earlier dates for the presence of domestic dog have been suggested from Europe (3), but the evidence does not seem conclusive (4). Morphological, behavioural, and genetic data clearly indicate that domestic dogs (Canis familiaris) originated from domesticated wolves (5–7). It is not clear from archaeological findings whether dogs originated from one or more wolf population(s). Studies of mitochondrial DNA and Y-chromosomal DNA diversity world-wide indicate that dogs originated in the southern part of East Asia (8–10). A study of genome-wide single-nucleotide polymorphism variation among domestic dogs and wolves showed dogs to share more unique multilocus haplotypes with wolves from the Middle East than with wolves from North China, Europe, and America, but did not include samples from southern East Asia (7). According to archaeological findings, dogs similar to mastiff breeds lived in Anatolia at 9000 years BP (11). It has been suggested that these dogs served as * Correspondence:
[email protected]
176
the origin of today’s dog breeds in Turkey. Dogs brought by different cultures or nations conquering Anatolia also contributed to forming the modern dog breeds in Turkey (11,12), e.g., the Turkish immigrants from Central Asia who settled in Anatolia and brought dogs. However, possible traces of the origins of Turkish dog breeds can be found from across the Anatolian Plateau to Central Asia as well as to the plateau of Afghanistan (13), and the history of Turkish dog breeds is still obscure. Carvings of dog pictures in caves (approximately 6000 BP) discovered in Tibet resemble the modern shepherd dogs in Turkey, as well as mastiff breeds, Greyhound, and Saluki (11,12). Kırmızı (11) claimed that the origins of Turkish shepherd dogs were in Central Asia and that these breeds spread to Anatolia, the Middle East, and Europe through the emigration and immigration of Turks from Central Asia to Anatolia and Europe. These dogs were brought to Europe by Turks during the Ottoman era or earlier, possibly serving as the origin of shepherd dog breeds such as the Great Pyrenees, Chuvatch, Greek Shepherd dog, Kuvasz, Sharplaninatz, Komondor, and Maremma Shepherd dog in Europe (12,14).
ERDOĞAN et al. / Turk J Biol
There is no consensus on the origin of Akbash and Kangal Shepherd dogs and their subvarieties (12). In a study carried out by Erdoğan and Özbeyaz (15), Kangal and Akbash Shepherd dogs were found to be located in different clusters and to have different genetic structures according to data from polymorphic loci. These findings changed the perception about the idea that these 2 breeds have a close relationship. The aim of the present study was to determine the microsatellite polymorphism in Turkish Greyhound, Kangal, Akbash, and Kars Shepherd dog breeds, to estimate the genetic relatedness among these dog breeds, and to address the following questions: What are the current levels of gene flow between breeds? Has gene flow been the main factor in the current genetic similarity between these populations from different regions of Turkey? 2. Materials and methods 2.1. Animal samples and microsatellite markers analysed In this study, a total of 141 dogs from 6 breeds were used. Kangal Shepherd dogs (n = 30), Akbash Shepherd dogs (n = 33), White Kars Shepherd dogs (KW, n = 15), Black Kars Shepherd dogs (KB, n = 23), Grey Kars Shepherd dogs (KG, n = 9), and Turkish Greyhounds (TG, n = 31) were chosen as nonrelative to each other and as best representing their breed characteristics (Figure 1). In the phenotyping of the transferrin, postalbumin-1 (Poa-1) and postalbumin-3 (Poa-3), with the methods described by Erdoğan and Özbeyaz (15), were used. DNA from blood samples was extracted according to standard phenol–chloroform methods (16). DNA concentrations were determined using a NanoDrop ND-1000 spectrophotometer (Peqlab, Erlangen, Germany) and adjusted to 25 ng/µL. One microlitre of each DNA was mixed with primer mix 1 (PEZ1, FHC2054, FHC2010 labelled with FAM; PEZ5,
PEZ12 labelled with JOE; PEZ6, PEZ8, FHC2079 labelled with NED), primer mix 2 (FH2247, FH2164 labelled with FAM; FH2001, FH2326 labelled with JOE; PEZ22, FH2289 labelled with NED), or primer mix 3 (PEZ11, FH2324 labelled with FAM; FH2161 labelled with NED). The PCR buffer supplied by QIAGEN (Hilden, Germany) contained 6 mM MgCl2. In a total reaction volume of 14 µL with HotStar HiFidelity Taq polymerase (QIAGEN), 32 cycles were performed on an MJ Research Dyad Tetra 4D thermocycler (MJ Research Inc., Waltham, MA, USA). The DNA samples were denatured for 15 min at 95 °C and then subjected to cycles of 30 s at 94 °C, 1.5 min at 59 °C, and 1 min at 72 °C. Next, 1.5 µL of the PCR products was loaded on an ABI Prism 3100 Genetic Analyser (Applied Biosystems, Darmstadt, Germany). Profiles were analysed using Software 3100 GeneScan Analysis Module 2 and Genotyper V3.5 NT. 2.2. Statistical analysis The average heterozygosity in a population was calculated according to the method of Nei (17) by using the values of heterozygosity calculated in each locus. The significant differences on an estimated average heterozygosity index between populations were calculated with the t-test. F-statistics were estimated in the form of F, θ, and ƒ as described by Weir and Cockerham (18) according to FIT, FST, and FIS, respectively. The gene flow (Nem) between populations was calculated with Nem = (1 ‒ FST) / 4 FST based on the FST value of one locus and all loci (17,19). These were calculated using the GENETIX 4.05 computer programme (20). Classification of dogs according to their neighbourhood in the factorial space was drawn using the GENETIX 4.05 computer packet programmes (20). Further assessments of genetic structures of populations were performed with STRUCTURE version 2.3.2 (21).
Figure 1. Geographical locations and illustrations of the 6 Turkish dog breeds’ samples collected.
177
ERDOĞAN et al. / Turk J Biol
3. Results The estimated values of the heterozygosity index (He and Ho) of the investigated dog breeds based on each locus and over all loci are given in Table 1. The average heterozygosity values were between 0.664 (KW) and 0.778 (KB). The mean heterozygosity values calculated for each breed were not statistically significant. The population differentiation was tested with a fixation index, with FIT, FST, and FIS values for each locus and over all loci. The results of the F-statistical analysis of the 20 loci for all investigated breeds are shown in Table 2. The lack of heterozygosity level was around 3.27% (P < 0.001) for each of the analysed breeds and 5.12% (P < 0.001) for the whole population. The genetic differentiation between
breeds, FST, was calculated to be 1.92% (P < 0.001), which was relatively low. The gene flow value, assigned for the number of individuals migrating between populations for each generation and calculated from all loci, was 12.8. The gene flow between the population pairs and the FST estimations are shown in Table 3. After 1000 permutations between breed pairs, all FST values were found to be significantly different from 0 (P < 0.001). The gene flow occurring between populations in each generation varied between 8.4 (Akbash–KW pairs) and 62.3 (KB and KG pairs). The Reynolds genetic distance matrix is given in Table 4. The genetic distance values vary between 0.0011 and 0.0358. The smallest genetic distance value (0.0011) was
Table 1. The heterozygosity index, calculated according to locus and all loci of dog breeds (He and Ho), average heterozygosity, and allele numbers. Akbash (n = 33)
Locus Allele
He
Kangal (n = 30) Ho
Allele
He
KW (n = 15) Ho
Allele
He
KB (n = 23) Ho
Allele
He
KG (n = 9) Ho
Allele
He
TG (n = 31) Ho
Allele
He
Ho
FH2001
6
0.840 0.684
6
0.808 0.833
5
0.683 0.615
8
0.854 0.923
5
0.824 0.889
7
0.815 0.750
FH2161
5
0.789 0.818
8
0.823 0.733
7
0.708 0.539
11
0.836 0.913
5
0.758 0.667
9
0.809 0.625
FH2164
14
0.886 0.833
12
0.838 0.867
8
0.831 0.846
9
0.851 0.769
11
0.948 0.889
13
0.902 0.969
FH2247
24
0.950 0.976
26
0.957 0.967
15
0.954 0.923
21
0.953 1.000
9
0.909 0.889
23
0.950 0.969
FH2289
16
0.918 0.854
17
0.893 0.867
10
0.779 0.461
15
0.919 0.923
10
0.915 0.864
14
0.891 0.844
FH2324
17
0.917 0.758
19
0.918 0.867
11
0.877 0.692
12
0.863 0.826
12
0.954 0.901
11
0.880 0.813
FH2326
15
0.915 0.878
16
0.912 0.833
13
0.939 0.846
16
0.911 0.808
9
0.928 0.876
18
0.940 0.781
FHC2010
7
0.740 0.737
7
0.580 0.548
5
0.742 0.692
4
0.655 0.692
4
0.765 0.444
5
0.675 0.625
FHC2054
9
0.870 0.868
10
0.852 1.000
6
0.803 0.769
8
0.847 0.889
7
0.863 0.889
7
0.844 0.831
FHC2079
6
0.679 0.658
3
0.579 0.710
2
0.409 0.231
6
0.661 0.741
6
0.758 0.778
3
0.514 0.500
PEZ1
9
0.803 0.806
12
0.786 0.807
6
0.825 0.846
4
0.750 0.889
4
0.712 0.667
10
0.816 0.844
PEZ5
7
0.767 0.686
8
0.749 0.742
4
0.754 0.846
6
0.741 0.667
4
0.726 0.778
5
0.719 0.719
PEZ6
15
0.882 0.842
19
0.918 0.839
10
0.902 0.846
16
0.895 0.815
7
0.869 0.889
11
0.865 0.656
PEZ8
11
0.820 0.632
11
0.841 0.839
10
0.905 0.846
12
0.873 0.692
8
0.889 0.667
10
0.843 0.844
PEZ11
13
0.893 0.727
11
0.847 0.800
10
0.892 0.769
12
0.893 0.870
9
0.928 0.889
9
0.836 0.625
PEZ12
15
0.861 0.895
15
0.810 0.741
6
0.628 0.692
10
0.809 0.778
8
0.876 0.667
11
0.836 0.813
PEZ22
13
0.875 0.838
11
0.816 0.900
7
0.794 0.692
9
0.895 1.000
7
0.863 0.667
10
0.827 0.813
Tf
2
0.504 0.559
2
0.509 0.400
2
0.443 0.308
2
0.485 0.409
2
0.400 0.250
2
0.482 0.516
Poa-1
2
0.395 0.529
2
0.325 0.400
3
0.489 0.500
2
0.426 0.500
2
0.458 0.625
2
0.252 0.290
Poa-3
2
0.479 0.706
2
0.503 0.633
2
0.271 0.308
2
0.495 0.455
2
0.533 0.500
2
0.389 0.452
7.1
0.731 0.664
9.25
0.781 0.778
6.55
0.794 0.735
9.1
0.754 0.715
Mean estimates
178
10.4
0.789 0.764 10.85 0.763 0.766
ERDOĞAN et al. / Turk J Biol Table 2. F-statistic values and number of individuals migrating between populations in each generation. Locus
FIS = ƒ
FIT = F
FST = θ
FH2001
0.0485
0.0533
0.0050
FH2161
0.0781
0.0896
0.0124
FH2164
0.0129
0.0212
0.0084
FH2247
–0.0186
–0.0035
0.0148
FH2289
0.0693
0.0887
0.0209
FH2324
0.1033
0.1282
0.0278
FH2326
0.0966
0.1038
0.0080
FHC2010
0.0522
0.0590
0.0072
FHC2054
–0.0434
–0.0285
0.0144
FHC2079
–0.0320
0.0068
0.0376
PEZ1
–0.0455
–0.0444
0.0011
PEZ5
0.0322
0.0672
0.0362
PEZ6
0.1019
0.1037
0.0021
PEZ8
0.1177
0.1490
0.0355
PEZ11
0.1343
0.1519
0.0203
PEZ12
0.0292
0.0463
0.0176
PEZ22
–0.0027
0.0089
0.0116
0.0763
0.1058
0.0320
Poa-1
–0.2302
–0.2189
0.0091
Poa-3
–0.2124
–0.0937
0.0979
Tf
Mean estimates
0.0327 (0.016)***
0.0512 (0.016)***
Nem
0.0192 (0.003)***
12.8
ƒ, estimation of pure breeding within the population; F, estimation of total pure breeding; θ, measure of population differentiation. The standard deviation is given in parentheses. ***P < 0.001, from permutation tests in the TFPGA programme.
Table 3. The FST statistics (vertical triangle) and gene migration Nem (inverted vertical triangle) among dog breeds in Turkey. Breeds
Akbash
Kangal
KW
KG
KB
TG
Akbash
***
17.6
8.4
35.5
35.5
10.2
Kangal
0.014
***
9.0
14.5
15.4
9.8
KW
0.029
0.027
***
41.4
12.9
10.2
KG
0.007
0.017
0.006
***
62.3
11.1
KB
0.007
0.016
0.019
0.004
***
11.7
TG
0.024
0.025
0.024
0.022
0.021
***
179
ERDOĞAN et al. / Turk J Biol Table 4. The Reynolds genetic distance matrix among the 6 Turkish dog breeds. Breeds
Akbash
Kangal
KW
KG
KB
TG
Akbash
-
-
-
-
-
-
Kangal
0.0141
-
-
-
-
-
KW
0.0358
0.0320
-
-
-
-
KG
0.0020
0.0145
0.0150
-
-
-
KB
0.0076
0.0147
0.0289
0.0011
-
-
TG
0.0244
0.0269
0.0307
0.0214
0.0232
-
a structure test analysis using K = 6 to determine which breed or breeds an individual belongs to, and to group the individuals (Figure 3). 4. Discussion The observed heterozygosity values from all loci in all breeds were found to be between 0.664 (KW) and 0.778 (Kangal) (Table 1). These estimates of heterozygosity values show that the genetic variations are high in all animals investigated in this study, and that there is no statistically significant difference between the breeds for the heterozygosity values. This means that all the dogs have the same level of genetic variations, and it is difficult to differentiate the dog breeds from each other in terms of mean heterozygosity levels. It has been reported that the Ho and He values for Akbash, Kangal, and Turkish Greyhound breeds are 0.715, 0.701, and 0.710, and 0.620, 0.701, and 0.705 (22), respectively, and that the average heterozygosity values for Akbash and Kangal are 0.367 and 0.410, respectively (15).
Axis 2 (22.51%)
found between KB and KG. In terms of genetic distance analysis, the breeds Akbash and KW were genetically far away from each other. The genetic distance value between these 2 breeds was 0.0358. To investigate and show the relationship between individuals, 3-dimensional factorial correspondence analysis (3D-FCA) and the GENETIX 4.05 computer packet programme (20) were employed (Figure 2). The 3D-FCA grouped all populations into 4 clusters. The Akbash breed, Turkish Greyhound breed, and Kangal breed formed 3 of these groups, which were clearly separated and located in different parts of the 3-dimensional graph. The dogs from the Kars region (KB, KG, KW) constituted a single fourth group placed in the middle of all the other groups (Figure 2). Within this fourth group, it could be seen that there were 3 breeds: KB, KG, and KW. However, the breeds KW and KB were grouped into the same cluster but settled in different positions (Figure 2). It can be seen that the Turkish Greyhound is more pure or a closer relative compared to the other breeds in
Axis 3 (21.22%) Axis 1 (26.63%)
Figure 2. Three-dimensional factorial correspondence analysis depiction of the genetic relationships among indigenous dog breeds in Turkey.
180
ERDOĞAN et al. / Turk J Biol 1.00 0.80 0.60 0.40 0.20 0.00
AKBASH
KANGAL
KARS KARS WHITE GREY
KARS BLACK
TURKISH GREYHOUND
Figure 3. Structure analysis of dog breeds of Turkey.
The estimated Ho and He values for Akbash, Kangal, and Turkish Greyhound are 0.764, 0.766, and 0.715, and 0.789, 0.763, and 0.754, respectively (Table 1). The heterozygosity levels for Akbash, Kangal, and Turkish Greyhound are higher than those reported by Erdoğan and Özbeyaz (15) and Altunok et al. (22); for dog breeds in Japan (23), Bangladesh (24), Germany (25), and Finland (26); in Golden Retriever, Labrador Retriever, and Rottweiler (27), and Beagle and Labrador Retriever (28); and in 16 street dogs (29), 11 East Asian domestic dog breeds (30), 28 dog breeds (31), Bali street dogs (32), and Hannover hunting dogs (33). The reason for attaining higher heterozygosity values in comparison to other researchers could be the selection of dogs from different regions that were unrelated to each other, and the fact that the loci searched have multiple alleles. Population differentiation was tested with a fixation index of FIT, FST, and FIS for each locus and all loci. The F-statistics results, calculated from all 20 loci for all 6 dog breeds, are given in Table 2. The observed mean heterozygote deficiency was 3.27% for each investigated breed and 5.12% in all populations, and these values are statistically significant (P < 0.001). The frequency in homozygote genotypes in all populations is higher than that expected in the Hardy–Weinberg balance. This means that there is not random mating in populations or that there is inbreeding. The genetic differentiation among breeds or the calculated FST was 1.92% (P < 0.001). This value of FST shows that the populations have different genetic structures. The FIS, FIT, and FST values calculated from all loci were 0.085, 0.083, and 0.160, respectively, for 7 dog breeds in Turkey (15), whereas the FIS, FIT, and FST values were 0.072, 0.214, and 0.154, respectively, in 11 East Asian domestic dog breeds (30), and the values of FIS and FIT vary between 0.01 and 0.13 and the FST value varies between 0.602 and 0.975 in 5 dog breeds of Finland (26). However, Altunok et al. (22) determined that the pairwise FST values were 0.167 and 0.121 in Kangal–Akbash and Kangal–Turkish Greyhound, respectively. The calculated F-statistic values in this study are much lower than those of other reports (15,22,26,30).
For population mating at random, genes are equally related within or between individuals. In this case, FIT = FST or FIS = 0. Therefore, the significant difference in the estimates of FIT and FST indicates a departure from random mating. Avoidance of mating between relatives will cause positive FST values that exceed the negative values of FIT and FIS. Generally, if FIS is positive (FIT > FST), it could be interpreted as evidence of inbreeding (34). FIS was positive (0.0327) while FIT was 0.0512, which was greater than the 0.0192 value of FST (Table 1). The estimate of the FIT and FIS values was positive for some loci. This means that the frequencies of heterozygote genotypes are in accordance with the Hardy–Weinberg balance, but they are lower than those of expected heterozygote values in these loci in all breeds. It is possible that the selection factors have a positive effect on the frequencies of homozygote genotypes based on selected loci at individual and population levels. Neglecting the effects of migration, and assuming a low contribution of mutations to the genetic diversity between these breeds, the differences in allele frequencies may be interpreted as primarily the result of random genetic drift. The genetic differentiation (1.92%) may be seen as the result of an increased mean inbreeding coefficient FIT over a rather short period of time. We therefore consider the relatively low mean FIS value (0.0327) to be the result of a reduction of heterozygosity within the breeds studied and the relatively low mean FIT value (0.0512) as indicative of ineffective barriers to gene flow between populations. It has been reported that the genetic distance among Spanish dog breeds ranges from 0.000 to 0.051 (35). Although there was not a dramatic difference among dog breeds, the genetic distance varied between 0.013 and 0.242 for 7 dog breeds (15). The genetic distance between Akbash and Kangal Shepherd dogs has been estimated to be 0.093 and the Nem value to be 1.3 (15). The genetic distance between dog breeds ranged from 0.0836 to 0.3235 and the Nem value ranged between 0.43 and 10.83 (30), while Koskinen and Bredbacka (26) demonstrated that the genetic distance between dog breeds ranged between 0.182 and 0.291. In this study, the estimated genetic distance value ranged from 0.0011 (KB–KG) to 0.0358 (Akbash–
181
ERDOĞAN et al. / Turk J Biol
KW), and the Nem value ranged between 8.4 (Akbash– KW) and 62.3 (Akbash–KW). The expected Nem value calculated from all loci was 12.8. The calculated genetic distance among dog breeds in this study is in accordance with the findings in Spanish dog breeds (35), but lower than those in other reports. The Nem value was higher than other researchers’ findings, as well. The mean estimation of genetic differentiation (FST) among breeds was 1.92% (P < 0.001). This value is lower than that of other breeds, and the effective number of individuals exchanged between populations per generation was 12.8 (Table 3). The gene flow ranges from 8.4 to 62.3 between populations. If gene flow is higher than 1 (in an endless island model), this means that Nem is large enough to decrease the genetic differentiation between populations (36). Genetic drift could be a factor supporting genetic differentiation between breeds. If gene exchange occurs between breeds at a high frequency, the FST value gets lower, which means that genetic differentiation occurs through gene flow. These results could be interpreted to indicate that the investigated shepherd dog breeds originated from a common ancestor a long time ago and lived together in the same geographical area. Gene flow plays a very important role in populations living within the same or close geographical areas. The reason for high gene flow occurring among the investigated dog breeds could be that the dogs (except the Turkish Greyhound) live in the same geographical areas. Although the Turkish Greyhound is not considered among shepherd dogs, this breed has a high rate of gene flow with the other dog breeds (Table 3). The high gene flow between Turkish Greyhound and other dog breeds could be due to uncontrolled mating between Turkish Greyhound and other dog breeds. Four different clusters were obtained as a result of 3D-FCA applied to determine and show the relationship between individuals (Figure 2). The 3-dimensional graph shows that Akbash, Turkish Greyhound, and Kangal dogs are distinctly separated and positioned in different sectors. In between these 3 groups, the 3 Kars breeds constitute a fourth group. Populations are genetically separated into distinct clusters. Direct evidence supporting this conclusion derives from 3D-FCA analyses (Figure 2). Interestingly, despite their different exterior morphology, the Kars dog breeds do not show a clear genetic divergence.
However, KW and KB were grouped into the same cluster but in different positions. KG was also situated in the middle of other Kars dog breeds. These findings indicate that KW and KB are a variety of Kars Shepherd dogs separately, and KG is also a crossbreed of these 2 Kars Shepherd dog breeds. In conclusion, our analysis indicates that native Turkish dog breeds have different genetic structures on the basis of the analysed loci, since they are located in clearly separated clusters in 3D-FCA. These findings disprove the beliefs that Kangal, Akbash, and Kars Shepherd dogs are close relatives of each other. The results clearly show that Akbash, Kangal, and Kars Shepherd dogs are different populations. Therefore, the generalised grouping of Turkish shepherd dogs into a single breed called Anatolian or Turkish shepherd dogs is not correct. It is proper to differentiate between Kangal Shepherd and Akbash Shepherd dogs as separate breeds. To determine the genetic structure of the dogs, polymorphic biochemical systems and microsatellite loci can be employed to compare the dog populations and the average heterozygosity values, and F-statistics, the individual numbers migrating in each generation, and the FCA method can be used for successful classification of the breeds. Since Turkey is located in both Asia and Europe as a junction for migration roads, and since it is the cradle of many different cultures, further determination of genetic structure in Turkish dog breeds is important for better understanding of the genetic origin of dogs in Asia and Europe. The determination of genetic structure and genetic relatedness in Turkish dogs will also help to determine the migration paths not only for dogs but also for people or civilisations in history. Acknowledgements This research was supported by the Scientific and Technological Research Council of Turkey (TÜBİTAK TOVAG 103V024), and additional support was provided by Afyon Kocatepe University (AKÜ BAPK 041 VF 06). We are grateful to S. Rose and S. Loos for excellent technical assistance. Peter Savolainen is a Royal Swedish Academy of Sciences Research Fellow supported by a grant from the Knut and Alice Wallenberg Foundation.
References 1.
Clutton-Brock J. Evolution of domesticated animals. In: Mason IL. ed. Dog. Longman; 1984: pp. 198–211.
3.
Sablin MV, Khlopachev GA. The earliest ice age dogs: evidence from Eliseevichi I. Curr Anthropol 43: 795–799, 2002.
2.
Zeder MA. Domestication and early agriculture in the Mediterranean basin: origins, diffusion, and impact. P Natl Acad Sci-Biol 105: 11597–11604, 2008.
4.
Napierala H, Uerpmann HP. A ‘new’ palaeolithic dog from Central Europe. Int J Osteoarchaeol 22: 127–137, 2010.
182
ERDOĞAN et al. / Turk J Biol 5.
Fiennes R, Fiennes A. The Natural History of the Dog. Wedenfeld & Nicolson Press. London; 1968.
22. Altunok V, Koban E, Chikhi L et al. Genetic evidence for the distinctness of Kangal dogs. B Vet I Pulawy 49: 249–254, 2005.
6.
Savolainen P, Zhang Y, Luo J et al. Genetic evidence for an East Asian origin of domestic dogs. Science 298: 1610–1613, 2002.
7.
vonHoldt BM, Pollinger JP, Lohmueller KE et al. Genome-wide SNP and haplotype analyses reveal a rich history underlying dog domestication. Nature 464: 898–902, 2010.
23. Tanabe Y, Ôta K, Ito S et al. Biochemical-genetic relationships among Asian and European dogs and the ancestry of the Japanese native dogs. J Anim Breed Genet 108: 455–478, 1991.
8.
Pang JF, Kluetsch C, Zou XJ et al. mtDNA data indicate a single origin for dogs south of Yangtze River, less than 16,300 years ago, from numerous wolves. Mol Biol Evol 26: 2849–2864, 2009.
9.
Ding ZL, Oskarsson M, Ardalan A et al. Origins of domestic dog in Southern East Asia is supported by analysis of Y-chromosome DNA. Heredity 108: 507–514, 2012.
26. Koskinen MT, Bredbacka P. Assessment of the population structure of five Finnish dog breeds with microsatellites. Anim Genet 31: 310–317, 2000.
10. Ardalan A, Kluetsch CFC, Zhang A et al. Comprehensive study of mtDNA among Southwest Asian dogs contradicts independent domestication of wolf, but implies dog–wolf hybridization. Ecology and Evolution 3: 373–385, 2011.
27. Altet L, Francino O, Sánchez A. Microsatellite polymorphism in closely related dogs. J Hered 92: 276–279, 2001.
11. Kırmızı E. Türk çoban köpeklerinin tarihçesi. Türk Veteriner Hekimliği Dergisi 6: 39–41, 1994.
24. Kobayashi R, Miyakawa H, Tanabe Y et al. Blood protein polymorphism in Bangladesh native dogs. Report of the Society for Researches on Native Livestock 12: 269–289, 1987. 25. Lachmann C. Verteilung genetischer Polymorphismen bei einigen deutschen Hunderassen. PhD, University of Veterinary Medicine, Hannover, 1993.
28. Ichikawa Y, Takagi K, Tsumagari S et al. Canine parentage testing based on microsatellite polymorphisms. J Vet Med Sci 63: 1209–1213, 2001.
12. Nelson DD. A general classification of the native dogs of Turkey. In: The International Symposium on Turkish Shepherd Dogs, Selçuk University, Konya, Turkey; 1996: pp. 19–94.
29. Jouquand S, Priat C, Hitte C et al. Identification and characterization of a set of 100 tri- and dinucleotide microsatellites in the canine genome. Anim Genet 31: 266– 272, 2000.
13. Özbeyaz C. Kangal köpeklerinde bazı morfolojik özellikler. Lalahan Hayvancılık Araştırma Enstitüsü Dergisi 34: 38–46, 1994.
30. Kim KS, Tanabe Y, Park CK et al. Genetic variability in East Asian dogs using microsatellite loci analysis. J Hered 92: 398– 403, 2001.
14. Reed S. The history of Turkish shepherd dogs. In: The International Symposium on Turkish Shepherd Dogs, Selçuk University, Konya, Turkey; 1996: pp. 97–109.
31. Irion DN, Schaffer AL, Famula TR et al. Analysis of genetic variation in 28 dog breed populations with 100 microsatellite markers. J Hered 94: 81–87, 2003.
15. Erdoğan M, Özbeyaz C. Investigation of blood protein polymorphism and estimation of genetic distances in some dog breeds in Turkey. Turk J Vet Anim Sci 28: 583–590, 2004.
32. Irion DN, Schaffer AL, Grant S et al. Genetic variation analysis of the Bali street dog using microsatellites. BMC Genet 6: 1–13, 2005.
16. Sambrook J, Fritsch EF, Maniatis T. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Press. Cold Spring Harbor, NY, USA; 1989.
33. Lupke L, Distl O. Microsatellite marker analysis of the genetic variability in Hanoverian hounds. J Anim Breed Genet 122: 131–139, 2005.
17. Nei M. Molecular Evolutionary Genetics. Columbia University Press. New York; 1987.
34. Weir BS. Genetic Data Analysis. Sinauer Associates Press. Sunderland, MA, USA; 1996.
18. Weir BS, Cockerham CC. Estimating F-statistics for the analysis of population structure. Evolution 38: 1358–1370, 1984.
35. Jordana J, Piedrafita J, Sánchez A. Genetic relationships in Spanish dog breeds, II. The analysis of biochemical polymorphism. Genet Sel Evol 24: 245–263, 1992.
19. Hartl DL, Clark AG. Principles of Population Genetics. Sinauer Associates Press. Sunderland, MA, USA; 1989. 20. Belkhir K, Borsa P, Chikhi L et al. GENETIX 4.00 WindowsTM Software for Population Genetics. University of Montpellier Press. Montpellier, France; 1996.
36. Trexler JC. Hierarchical organization of genetic variation in the Sailfin Molly, Poecilia latipinna (Pisces: Poeciliidae). Evolution 42: 995–1005, 1988.
21. Pritchard JK, Stephens M, Donnelly P. Inference of population structure using multilocus genotype data. Genetics 155: 945– 959, 2000.
183