Post by Admin on May 8, 2023 17:55:05 GMT
On the basis of the broad distribution of SBB and TAK individuals along the PC1 axis, we find that each site is highly heterogeneous. To statistically compare the genetic diversity of each site and the Xiongnu as a whole (“all Xiongnu”), we applied the Brown-Forsythe test for the null hypothesis that two groups have equal PC1 variances (data file S1D). TAK is highly heterogeneous to the level indistinguishable from the variance across all Xiongnu individuals (P = 0.622), while SBB shows less diversity than all Xiongnu (P = 0.030). We also compared the genetic diversity of the previously published Xiongnu sites with at least five sampled individuals (14). The early Xiongnu site Salkhityn Am (SKT; n = 11) in northern Mongolia exhibits a high degree of heterogeneity similar to all Xiongnu (P = 0.411). The late Xiongnu site of Uguumur Uul (UGU; n = 5) located further to the east shows a more intermediate level of diversity with higher contributions of western Eurasian ancestry, but is still statistically comparable to the all Xiongnu (P = 0.185); however, the small sample size for this site provides limited statistical resolution. Still, UGU shows much higher diversity than the preceding EIA groups (P = 0.001 and 0.015 for SlabGrave and Chandman_IA, respectively). In contrast, another late Xiongnu site that is located further to the northeast, Il’movaya Pad (IMA; n = 8), has predominantly eastern Eurasian ancestry. Its genetic diversity is much lower than all Xiongnu (P = 0.002), and is comparable to those of the preceding EIA groups (P = 0.975 and 0.398 for SlabGrave and Chandman_IA, respectively) (Fig. 4). We note that the low genetic diversity at the IMA site may be an artifact of biased sampling as only eight individuals were analyzed among 300 circular grave and square tomb burials at the site (67). By repeating PC coordinates calculation 100 times for each individual with genotype data downsampled to 10,000 SNPs, we found that the results were not affected by higher uncertainty in PC coordinates in low-coverage individuals (fig. S5).
The comparable level of genetic diversity of all Xiongnu individuals broadly and local communities suggests that the dynamic demographic processes that constituted the highly heterogeneous Xiongnu empire also occurred at local scales. High genetic diversity within the cemeteries of TAK, SBB, and UGU confirm the coresidence of individuals with diverse genetic backgrounds within a single local community and that this continued even into the late Xiongnu period, centuries after the political formation of the Xiongnu and the associated demographic processes of genetic admixture that were already in progress during the initial early Xiongnu period. Overall, the high genetic diversity found among the Xiongnu during all periods prevents any meaningful attempt to define a “representative” Xiongnu genetic profile, as it is instead population-level genetic heterogeneity spanning nearly the entire breadth of Eurasian genetic diversity that most characterizes the Xiongnu Empire.
Genetic diversity and archaeological signifiers of social status
To better understand how Xiongnu genetic diversity might be structured by social status or social group affiliation, we examined archaeogenetic data from the aristocratic elite cemetery of TAK (Fig. 3D) and the local elite cemetery of SBB (Fig. 3E) separately. At TAK, two complete square tomb complexes have been excavated (THL-82 and THL-64), allowing genetic diversity within discrete mortuary units to be investigated. For the richest tomb complex, THL-82, which contained an older adult female buried in a square tomb (TAK001) flanked by two satellite graves of adult males (TAK008 and TAK009), we identified no genetic relatedness between them. While the genetic profile of the high-status female strongly differed from that of the two low-status males (Fig. 3D), the latter were genetically similar to each other and had much higher levels of western Eurasian ancestry (Fig. 2B and data file S2C). That is, these two males are collectively modeled with 86.8% earlyXiongnu_west ancestry, while TAK001 requires only 9.3% of this ancestry component. In contrast to the high western ancestry of the low-status males, the high-status women (TAK001) shows a high level of eastern ancestry, represented by Khovsgol_LBA, not SlabGrave1.
For the other square tomb complex, THL-64, which contained an adult female buried in a square tomb (TAK002) along with two satellite graves of adolescents, we determined that both adolescents were males. Unlike THL-82, however, the two low-status young males were genetically dissimilar (Fig. 2B and data file S2C), with one (TAK003) having a very high level of western Eurasian ancestry (Chandman_IA and Gonur1_BA) and the other (TAK004) having a high level of eastern Eurasian ancestry (SlabGrave1). The high-status female likewise had a high degree of eastern Eurasian ancestry (Fig. 3D) deriving from two sources (SlabGrave1 and Han_2000BP).
Last, we analyzed the three satellite graves associated with THL-25, for which the square tomb has not been excavated. Two of the three individuals yielded sufficient DNA for analysis: a child (TAK005) and an adult male (TAK006). We determined that both were unrelated males and that they were highly genetically dissimilar (Fig. 3D). The child had a very high level of western Eurasian ancestry (Chandman_IA), while the adult male had the highest eastern Eurasian ancestry (SlabGrave1 and Han_2000BP) observed at the TAK site. Thus, we find very high genetic diversity within individual tomb complexes at the TAK cemetery. Although both high-status females had relatively high levels of eastern Eurasian ancestry, the low-status satellite males exhibited extremely high genetic heterogeneity ranging from very high levels of western Eurasian ancestry to very high levels of eastern Eurasian ancestry. If the low-status males were retainers or servants of the high-status females, it suggests that they were drawn from diverse parts of the Xiongnu empire and possibly beyond.
At the SBB site, we found lower overall genetic diversity, and specifically, no individuals with very high levels of western Eurasian ancestry (Fig. 3E). However, the individuals with the highest levels of western Eurasian ancestry were both adult males (SBB001 and SBB010), although they derived their western ancestry from slightly different sources (Chandman_IA for SBB001 and Chandman_IA and Gonur1_BA for SBB010). As at TAK, the highest-status graves belonged to females (SBB002, SBB003, SBB007, and SBB008), whose modeled ancestries all included SlabGrave1, with other minor ancestry contributions. The genetic determination of SBB007 as female was particularly noteworthy because the grave goods included horse-riding equipment, a gilded iron belt clasp, and a Han-painted lacquer cup, which have been assumed in other contexts to be accouterments associated with male horse-mounted warriors. Similarly, SBB010, an adult male, was buried with a bone tube case containing an iron needle, indicating that sewing implements were not exclusively associated with women. We also determined the genetic sex of three children (SBB004, SBB005, and SBB006) and one adolescent (SBB009) whose sex was uncertain. SBB005 and SBB006 were females, while SBB004 and SBB009 were males. SBB009, an adolescent 11 to 12 years old, was buried with a child-sized bow similar to the bow buried with the adult male SBB001 and the older adolescent SBB011, corroborating accounts of males in Xiongnu society learning to wield bows at a young age (68), likely by a young male’s early teens as in the case of SBB009, but likely not in very early childhood, as evidenced by the absence of such equipment in the grave of SBB004, a child 4 to 6 years old.
Examining spatial relationships of burials at SBB, we did not find a significant correlation between spatial proximity and genetic ancestry profiles (Fig. 3E). To represent the similarity of the genetic ancestry profiles of two individuals, we used the Euclidean distance between two points defined in the space of the top two PCs. We could not reject the null hypothesis that the spatial proximity between two burials and genetic distance was unrelated, with a P value of 0.146 provided by the Mantel test. Nor could we support a hypothesis that the individuals in the five-grave cluster in SBB (consisting of SBB001, SBB005, SBB006, and SBB007; SBB011 did not produce analyzable genome-wide data) were more similar in their genetic profiles than the others. To determine this, we replaced the geographic distance of each individual pair with 1 for within-cluster pairs and 0 for the others, respectively (Mantel test P value = 0.3085). However, we did find that the relatives were placed significantly closer to each other (Mantel test P value = 0.0025), based on the two pairs of related individuals buried next to each other: the pair SBB001 and SBB005 and the pair SBB005 and SBB007. These were the only genetically related individuals identified at either SBB or TAK. Although the adult male SBB001 and the female child SBB005 were second-degree relatives, they were genetically dissimilar to one another, with the male having a much higher level of western Eurasian ancestry (Chandman_IA). SBB005 was also a second-degree relative of SBB007, the high-status woman buried with the gilded belt clasp and Han lacquer cup. Both females shared a minor ancestry component modeled as Han_2000BP. Previous investigations of genome-wide genetic relatedness among the Xiongnu have identified 10 other cases of kinship pairs (14). Among these, all pairs were buried within the same site or at closely neighboring sites, and most pairs are genetically similar. However, one pair of second-degree maternally related males at the site of Tamiryn Ulaan Khoshuu (TUH001 and TUH002) in north-central Mongolia likewise showed a relatively high degree of genetic dissimilarity (14), with one male having substantially more western Eurasian ancestry than the other, although the difference was not as great as between SBB001 and SBB005. Such extended families containing higher degrees of genetic diversity may have been relatively common among the Xiongnu, but denser sampling of cemeteries will be necessary to identify them. Beyond detecting genetic relatedness (69, 70), we also examined runs of homozygosity (ROH) blocks (71). We found that SBB005 had long ROH blocks totaling 55.1 centimorgan (cM), with the longest block extending 40.7 cM (fig. S6), suggesting that she is the offspring of a pair of second-degree relatives. Because consanguinity can distort estimates of genetic relatedness, the degree of genetic relatedness between her (SBB005) and SBB001 and SBB007 may be slightly overestimated. Nevertheless, the overall patterns of genetic diversity, heterogeneity, and genetic relatedness at SBB suggest that some local elite families were highly genetically diverse, with marriages occurring between genetically heterogeneous individuals that created complex networks of extended kinship.
The comparable level of genetic diversity of all Xiongnu individuals broadly and local communities suggests that the dynamic demographic processes that constituted the highly heterogeneous Xiongnu empire also occurred at local scales. High genetic diversity within the cemeteries of TAK, SBB, and UGU confirm the coresidence of individuals with diverse genetic backgrounds within a single local community and that this continued even into the late Xiongnu period, centuries after the political formation of the Xiongnu and the associated demographic processes of genetic admixture that were already in progress during the initial early Xiongnu period. Overall, the high genetic diversity found among the Xiongnu during all periods prevents any meaningful attempt to define a “representative” Xiongnu genetic profile, as it is instead population-level genetic heterogeneity spanning nearly the entire breadth of Eurasian genetic diversity that most characterizes the Xiongnu Empire.
Genetic diversity and archaeological signifiers of social status
To better understand how Xiongnu genetic diversity might be structured by social status or social group affiliation, we examined archaeogenetic data from the aristocratic elite cemetery of TAK (Fig. 3D) and the local elite cemetery of SBB (Fig. 3E) separately. At TAK, two complete square tomb complexes have been excavated (THL-82 and THL-64), allowing genetic diversity within discrete mortuary units to be investigated. For the richest tomb complex, THL-82, which contained an older adult female buried in a square tomb (TAK001) flanked by two satellite graves of adult males (TAK008 and TAK009), we identified no genetic relatedness between them. While the genetic profile of the high-status female strongly differed from that of the two low-status males (Fig. 3D), the latter were genetically similar to each other and had much higher levels of western Eurasian ancestry (Fig. 2B and data file S2C). That is, these two males are collectively modeled with 86.8% earlyXiongnu_west ancestry, while TAK001 requires only 9.3% of this ancestry component. In contrast to the high western ancestry of the low-status males, the high-status women (TAK001) shows a high level of eastern ancestry, represented by Khovsgol_LBA, not SlabGrave1.
For the other square tomb complex, THL-64, which contained an adult female buried in a square tomb (TAK002) along with two satellite graves of adolescents, we determined that both adolescents were males. Unlike THL-82, however, the two low-status young males were genetically dissimilar (Fig. 2B and data file S2C), with one (TAK003) having a very high level of western Eurasian ancestry (Chandman_IA and Gonur1_BA) and the other (TAK004) having a high level of eastern Eurasian ancestry (SlabGrave1). The high-status female likewise had a high degree of eastern Eurasian ancestry (Fig. 3D) deriving from two sources (SlabGrave1 and Han_2000BP).
Last, we analyzed the three satellite graves associated with THL-25, for which the square tomb has not been excavated. Two of the three individuals yielded sufficient DNA for analysis: a child (TAK005) and an adult male (TAK006). We determined that both were unrelated males and that they were highly genetically dissimilar (Fig. 3D). The child had a very high level of western Eurasian ancestry (Chandman_IA), while the adult male had the highest eastern Eurasian ancestry (SlabGrave1 and Han_2000BP) observed at the TAK site. Thus, we find very high genetic diversity within individual tomb complexes at the TAK cemetery. Although both high-status females had relatively high levels of eastern Eurasian ancestry, the low-status satellite males exhibited extremely high genetic heterogeneity ranging from very high levels of western Eurasian ancestry to very high levels of eastern Eurasian ancestry. If the low-status males were retainers or servants of the high-status females, it suggests that they were drawn from diverse parts of the Xiongnu empire and possibly beyond.
At the SBB site, we found lower overall genetic diversity, and specifically, no individuals with very high levels of western Eurasian ancestry (Fig. 3E). However, the individuals with the highest levels of western Eurasian ancestry were both adult males (SBB001 and SBB010), although they derived their western ancestry from slightly different sources (Chandman_IA for SBB001 and Chandman_IA and Gonur1_BA for SBB010). As at TAK, the highest-status graves belonged to females (SBB002, SBB003, SBB007, and SBB008), whose modeled ancestries all included SlabGrave1, with other minor ancestry contributions. The genetic determination of SBB007 as female was particularly noteworthy because the grave goods included horse-riding equipment, a gilded iron belt clasp, and a Han-painted lacquer cup, which have been assumed in other contexts to be accouterments associated with male horse-mounted warriors. Similarly, SBB010, an adult male, was buried with a bone tube case containing an iron needle, indicating that sewing implements were not exclusively associated with women. We also determined the genetic sex of three children (SBB004, SBB005, and SBB006) and one adolescent (SBB009) whose sex was uncertain. SBB005 and SBB006 were females, while SBB004 and SBB009 were males. SBB009, an adolescent 11 to 12 years old, was buried with a child-sized bow similar to the bow buried with the adult male SBB001 and the older adolescent SBB011, corroborating accounts of males in Xiongnu society learning to wield bows at a young age (68), likely by a young male’s early teens as in the case of SBB009, but likely not in very early childhood, as evidenced by the absence of such equipment in the grave of SBB004, a child 4 to 6 years old.
Examining spatial relationships of burials at SBB, we did not find a significant correlation between spatial proximity and genetic ancestry profiles (Fig. 3E). To represent the similarity of the genetic ancestry profiles of two individuals, we used the Euclidean distance between two points defined in the space of the top two PCs. We could not reject the null hypothesis that the spatial proximity between two burials and genetic distance was unrelated, with a P value of 0.146 provided by the Mantel test. Nor could we support a hypothesis that the individuals in the five-grave cluster in SBB (consisting of SBB001, SBB005, SBB006, and SBB007; SBB011 did not produce analyzable genome-wide data) were more similar in their genetic profiles than the others. To determine this, we replaced the geographic distance of each individual pair with 1 for within-cluster pairs and 0 for the others, respectively (Mantel test P value = 0.3085). However, we did find that the relatives were placed significantly closer to each other (Mantel test P value = 0.0025), based on the two pairs of related individuals buried next to each other: the pair SBB001 and SBB005 and the pair SBB005 and SBB007. These were the only genetically related individuals identified at either SBB or TAK. Although the adult male SBB001 and the female child SBB005 were second-degree relatives, they were genetically dissimilar to one another, with the male having a much higher level of western Eurasian ancestry (Chandman_IA). SBB005 was also a second-degree relative of SBB007, the high-status woman buried with the gilded belt clasp and Han lacquer cup. Both females shared a minor ancestry component modeled as Han_2000BP. Previous investigations of genome-wide genetic relatedness among the Xiongnu have identified 10 other cases of kinship pairs (14). Among these, all pairs were buried within the same site or at closely neighboring sites, and most pairs are genetically similar. However, one pair of second-degree maternally related males at the site of Tamiryn Ulaan Khoshuu (TUH001 and TUH002) in north-central Mongolia likewise showed a relatively high degree of genetic dissimilarity (14), with one male having substantially more western Eurasian ancestry than the other, although the difference was not as great as between SBB001 and SBB005. Such extended families containing higher degrees of genetic diversity may have been relatively common among the Xiongnu, but denser sampling of cemeteries will be necessary to identify them. Beyond detecting genetic relatedness (69, 70), we also examined runs of homozygosity (ROH) blocks (71). We found that SBB005 had long ROH blocks totaling 55.1 centimorgan (cM), with the longest block extending 40.7 cM (fig. S6), suggesting that she is the offspring of a pair of second-degree relatives. Because consanguinity can distort estimates of genetic relatedness, the degree of genetic relatedness between her (SBB005) and SBB001 and SBB007 may be slightly overestimated. Nevertheless, the overall patterns of genetic diversity, heterogeneity, and genetic relatedness at SBB suggest that some local elite families were highly genetically diverse, with marriages occurring between genetically heterogeneous individuals that created complex networks of extended kinship.