Evidence for Early European Neolithic Dog Dispersal: New Data on South-Eastern European subfossil dogs from Prehistory and Antiquity Ages

Objectives The history of dog domestication is still under debate, but doubtlessly, it is a process of an ancient partnership between dogs (Canis familiaris) and humans. Although data on ancient DNA dog diversity are scarce, it is clear that several regional dog populations had been formed in Eurasia up to the Holocene. During the Neolithic Revolution and the transition from hunter-gatherer to farmer societies, followed by civilization changes in the Antiquity period, the dog population structure also changed. This process was due to replacement with newly formed dog breeds. Methods In this study we have presented for the first time mitochondrial data about South-Eastern Europe (the Balkans) ancient dog remains from the Early Neolithic (8 000 years BP) to the Late Antiquity ages (up to 3th century AD). A total of 25 samples were analyzed using the mitochondrial D-loop region (HVR1). Results The results have shown the presence of A (70%) and B (25%) clades throughout the whole investigated period. In order to clarify the position of our results within the ancient dog population in Eneolithic Eurasia, we performed phylogenetic analysis with the available genetic data sets. This data revealed a similarity of the Bulgarian dogs’ structure to that of ancient Italian dogs (A, B, and C clades), which suggests a new prehistoric and historic Mediterranean dog population. A clear border can be seen between South-European genetic dog structure, on the one hand, and, on the other hand, Central-West (clade C), East (clade D) and North Europe (clades A and C). This corresponds to genetic data for European humans during the same period without admixture between dog populations. Conclusions Our data have shown for the first time the presence of clade B in ancient Eurasia. This is not unexpected as the B haplogroup is widely distributed in extant Balkan dogs and wolves. The presence of this clade both in dogs and in wolves on the Balkans may be explained with hybridization events before the Neolithic period. The spreading of this clade across Europe together with the A clade is related to the possible dissemination of newly formed dog breeds from Ancient Greece, Thrace and the Roman Empire.


59
In recent years, it has been increasingly assumed that the dog's domestication was a very 60 early process that began at the end of the Pleistocene. The dog's domestication is apparently E, and F have regional geographic distribution [10]. For example, in West Eurasia, the sub-76 haplogroup A1 is with a frequency of about 70%, while clades B and C are present with 77 frequencies of 20% and 10%, respectively [7,9]. In the Middle East, this proportion is almost 78 the same but with more worldwide presented sub-haplogroups [11]. 79 Analysis of ancient dog DNA is still scarce and does not include many geographic regions. 80 Despite this, available data reveal quite different population structure concerning present day 81 dogs. These alterations are due to the emergence of the first civilizations as well as the early 82 historical and modern human migration. For example, pre-Columbian American dogs were   The third most studied geographic regions are South-Eastern Asia and the Pacific Islands, 118 including Australia (Fig 1a). Characteristic of these regions is the Neolithic pottery Lapita  Interestingly, the sub-clade B2 is different from West Eurasian dog's B1 mytotypes [9].

124
In conclusion, dogs are a very interesting object, because of the direct evidence of historical 125 human migration processes. However, there are still many uninvestigated geographic regions.

126
White spots of ancient dog genetic diversity are, for example, the Mediterranean region, 127 including South Europe, the Near East, and the Fertile Crescent, and also other domesticated 128 centers in Central and East Asia (China and India), etc.

129
Our study tries to enrich data for ancient South-Eastern Europe -the Balkan Peninsula fragment of HVR1, which is not declarative concerning haplogroup assignment (S1Table).

145
All analyzed sequences in this study belonging to clade A and B are equally dispersed in 146 investigated regions in Bulgaria as well as in historical time. In contrast, the highly dispersed 147 in ancient East Europe clade D was found only in one sample from the Late Antiquity period 148 (S1 Table).  Table). All haplotypes were separated into four main clades: A, B, C and D. There 155 is a clear geographical differentiation between the distribution of C and D clades, which are 156 specific for Central, West and East Europe. Clade A is split into many sub-clades, i.e. A2 for 157 South-Eastern Asia and the Pacific, A1 for Siberia and Central Asia, and pre-Columbian 158 America, A4 and A6 for North Europe (Scandinavia), and A and B clade are characteristic for 159 the Mediterranean region (Italy, Bulgaria and Israel). Similarly, PCA analysis grouped all 160 ancient samples into five distinct clades (Fig 1c).  The question of dog origin is geographically, genetically, and archaeologically complex.  Table). Otherwise, our

251
In modern dogs, clade B is widely distributed with a frequency of over 20% [7,9]. This clade 252 consists of two sub-clades, B1 and B2. The B1 sub-clade is disseminated worldwide with a 253 frequency of about 21%, while the B2 sub-clade is with regional distribution mainly in Eastern Trakia motorway, Yambol district, and Academic, Plovdiv) (S1 Table and