We successfully analysed mtDNA and Y-chromosome diversity in ancient sheep remains from Finland and found that Finnish sheep genetic diversity has been quite constant over the last 1000 years. Our aDNA results fit well with the genetic context of the modern north-European domestic sheep breeds analysed either previously [3, 12] or in the present study. Both ovine haplogroups A and B have been present in the Finnish sheep population for at least more than 700 years and no remarkable temporal changes in their frequencies have occurred. Four of the 26 ancient mtDNA sequences obtained here were assigned to haplogroup A (the frequency in different ancient cohorts varied from 0 to 28.6%) and 22 sequences to haplogroup B (frequency ranging between 71.4 and 100%). The respective haplogroup A and B frequencies in the present sample set of three Finnish native breeds are 21.9% and 78.1% according to our analysis and 16.7% and 83.3% according to . Four of the 18 ancient mtDNA haplotypes were present in the modern Finnish native breeds and 14 in the set of 10 Eurasian breeds analysed here. The affinity between modern mtDNA haplotypes and the ancient mtDNA haplotypes not found in the modern breeds is also attested by their proximity within the phylogeny and haplotype network in Figure 1 and Figure S3 [see Additional file 6: Figure S3], respectively. According to historical written records, foreign breeding animals, which were obviously ancestors of the modern Merino sheep, were imported into Finland during the 16th, 17th and 18th centuries [8, 56]. However, most of the imported individuals being rams, they did not have a major impact on the mtDNA diversity in the Finnish native sheep.
In addition, results of the Y-chromosome analysis on aDNA agree well with those on DNA from modern Finnish indigenous sheep breeds. SNP G-oY1 in the 5’-promoter region of the SRY gene on sheep Y-chromosome was detected in the three ancient ram samples. Since these ancient ram samples (OaNaa1, OaTor1, OaKök2, Table 1) were collected from different excavation sites and from two different time periods (Medieval and Post-Medieval), it is unlikely that they were close relatives. However, the number of available ancient ram samples is not sufficient to draw detailed conclusions on the temporal changes in the frequencies of G-oY1 and A-oY1. In the contemporary native Finnish breeds, the frequency of SNP G-oY1 ranges from 57 to 77% [see Additional file 4: Table S3] and the frequency of SNP G-oY1 is highest in north European, British Islands and central Russian populations , [see Figure 2 and Additional file 4: Table S3]. SNP G-oY1 is less frequent in southern and central European sheep (25% and 32%, respectively) and very low for the remaining breeds analysed to date (< 8%) [see Figure 2 and Additional file 4: Table S3]. The presence of SNP G-oY1 in modern sheep populations often reflects introgression of English breeds . However, this cannot be the case for ancient Finnish sheep. Consequently, our finding that SNP G-oY1 is present in sheep aDNA suggests that (at least in Finland) this paternal line predates the arrival of sheep in northern Europe.
Our results on radiocarbon-dating of 19 sheep remains excavated from archaeologically important sites in Finland provide essential information for Finnish archaeological research. Our research on ancient sheep material shows that previous attempts to determine the age of remains based on archeological context are fairly accurate, since only in three cases, did the dating results differ from those expected (OaPih1, OaKök2 and OaUuk2, Table 1). The oldest samples analysed here originate from the Late Iron Age from the excavation site Mulli in Raisio and from Pasvik in northern Norway. In Finnish archaeology, age determination of specimens is usually based on archaeological context. However, this can be misleading because, in Finland, cultural layers are thin and younger bone and other animal materials may have sunk to earlier cultural layers and thus, the dating of a single bone might differ from the general dating of a site . Therefore, the dating of archaeological materials by radiocarbon methods is recommended.
Here, we investigated the genetic affinities between the ancient Finnish sheep populations and the modern sheep breeds by searching for identical matches in the GenBank DNA database with the ancient haplotype sequences. Comparison of our ancient mtDNA data with those of the contemporary breeds appears to be relatively uninformative in terms of unfolding geographic origin or origins of native sheep in Finland or the dispersal of sheep husbandry to Finland. Our results confirm that the domestic sheep populations share their origins with O. orientalis populations domesticated in the Near Eastern region, from where the sheep spread around the World, and that whereas haplogroup B is common in Europe, haplogroup A is much rarer [1–3]. Ancient mtDNA haplotypes can be found in modern sheep breeds originating from geographically distant regions. Ancient Finnish sheep show a maternal genetic affinity to western and southern European breeds, but also to eastern European breeds. For example, haplotype H05 detected in the Post-Medieval sheep is also present in the two Russian (Figure 1) and the Iberian sheep breeds [see Additional file 3: Table S2] and several haplotypes of the haplogroup B detected in the Finnish ancient sheep are present in eastern, western and southern European breeds. However, four mtDNA haplotypes present in the ancient sheep samples are absent in the contemporary sheep breeds suggesting a possible loss of these haplotypes. Our results agree with a previous analysis of mtDNA D-loop polymorphisms on modern Eurasian sheep breeds in which a genetic historical influence of Russian sheep breeds in northern European sheep breeds was detected .
We assume that the fluctuations in mtDNA genetic diversity estimates obtained for the different cohorts (Tables 2 and 3) may be stochastic in nature as a result of genetic drift and sampling. Assuming neutrality for the mtDNA D-loop region, the high positive values for (θπ-θs) indicate a loss of genetic variability in terms of number of segregating nucleotide sites, occurring in the modern populations of Åland, Kainuu Grey, Olkuska sheep and the Serbian Pramenka population of Vlashko Vitoroga. These breeds experienced a genetic bottleneck during the 20th century and are classified as endangered sheep breeds . In addition, their positive Tajima’s D-values – statistically significantly different from zero for Kainuu Grey and Olkuska sheep – point towards an effect of a decline of population size on mtDNA diversity. In contrast, our within-population diversity estimates for Finnsheep, which is a large, major sheep population in Finland and having descended during the 20th century from a broad founder population, do not indicate loss of mtDNA diversity in haplogroup B. However, our Finnsheep data can be considered slightly biased because we could not detect haplogroup A in our sample set . When the diversity estimates of the ancient samples are compared with those of the modern breeds, values for (θπ-θs) and Tajima’s D in the Medieval sheep population are similar to estimates that could be obtained for an endangered modern breed having experienced a reduction in population size. In contrast, the estimates for the Post-Medieval cohort are similar to those of a modern large sheep breed with a growing population size. However, the origin and availability of archaeological materials could explain these differences in estimates rather than demographic events that occurred in the past. The Medieval samples are mainly from one region, the Turku region, while Post-Medieval samples are from a wider geographic region, from southern, northern, western and central parts of Finland. The present Iron Age samples displaying a lower level of mtDNA diversity are from the Mulli farm from Raisio and four of the samples share the same maternal ancestry. Interestingly, it appears that the same sheep ‘dam-lineage’ was raised in Mulli for a long period and no new ewes were introduced into the flock.
This aDNA study focusing on sheep biodiversity in Finland shows that population genetic analysis of ancient domestic animal populations and studies investigating changes in genetic diversity across different eras are challenging. The availability of ancient materials and the success rate of ancient DNA analysis have a decisive impact on how comprehensively a sample set represents the genetic variation of ancient animal populations. In the Finnish context, survival of unburned bones from time periods prior to the Late Iron Age is infrequent due to the acid centeracter of soils. In addition, in our study, the Medieval and Post-Medieval samples are not from the same temporal population as the samples of the modern breeds, but mainly from temporally distant generations. For example, the age difference between the oldest and youngest Medieval samples is approximately 300 years, corresponding to a time span of 100 generations. The population structures of ancient and modern cohorts are different, interfering with conclusions on temporal changes between ancient and modern populations. Moreover, the 300 years long Medieval period or Iron Age period were not socially or culturally static and our research area can be roughly divided into western, eastern and northern cultural regions with their own cultural and trade networks possibly influencing the genetic variation and structure of ancient animal populations. More ancient materials are needed to examine the archaeological questions in more detail.