|
Post by Admin on Nov 11, 2021 4:19:29 GMT
Results Ancient DNA Sequencing We generated genome-wide genotype data from 19 ancient humans, including one Upper Paleolithic individual (dated to 14,050–13,770 BP, see Orlova, 1995, Pavlenok et al., 2019), four Early Neolithic individuals (7,320–6,500 BP), and 14 Late Neolithic to Early Bronze Age (LNBA) individuals (4,830–3,570 BP) from a total of 10 archaeological sites (Figure 1; Table S1). The radiocarbon date offsets caused by the local freshwater reservoir were estimated using the carbon and nitrogen isotopic values as described in previous studies on the same region (see STAR Methods; Schulting et al., 2014, Schulting et al., 2015). We built single- and double-stranded DNA libraries from teeth or petrous portions of the temporal bone for the studied individuals, and shotgun sequencing revealed high levels of DNA preservation with endogenous DNA contents ranging from 0.12% to 50.54% (Table S1). Subsequently, libraries were enriched for human DNA by SNP-capture targeting a set of 1.24 million variable sites (Fu et al., 2015) and sequenced to mean coverage ranging from 0.04X to 2.07X. Pseudo-haploid genotypes were called on the targeted SNPs by randomly sampling a single allele at each position, with 34k to 886k SNPs covered by our samples. Additionally, we performed deep shotgun sequencing on eight individuals with high endogenous DNA levels (12%–51%), to achieve genomic coverage that ranged from 0.1X to 1.9X and refined their diploid genotypes by genotype likelihood-based imputation, resulting in 386k to 518k SNPs overlapping with the Human Origins dataset (Table S1). Figure 1. Geographic Location, Time Period, and Genetic Profile of Studied Individuals (A) Location of the 19 newly reported and published ancient individuals relevant in this study. Newly reported individuals are shown in outlined squares or circles. (B) Ages of newly reported individuals from each site, with the x axis showing the median calibrated radiocarbon dates after correction for freshwater reservoir effect. (C) PCA of Eurasian and Native American populations. The modern individuals are shown in light gray circles, with the name of several representative populations marking the positions of West Eurasians (Sardinian), East Asians (Ami and Uyghur), Siberians (Chukchi, Eskimo Naukan, Koryak, Nganasan, Selkup), and Native Americans (Chipewyan, Mixe, Karitiana). Ancient individuals are shown in colored symbols. The individual KPT005, which showed a significant shift toward west Eurasian populations, and individuals GLZ001&GLZ002, which also represented outlier genetic profiles, are marked out with an arrow. (D) Population clustering pattern of studied individuals together with representative modern and ancient populations, when K = 16. Most Early Neolithic to Bronze Age Lake Baikal region individuals are modeled as admixture of northeast Asian (dark red), ANE (dark blue), and Nganasan component (purple). The BZK002 individual has similar genetic profile as Okunevo, with a significantly larger ANE proportion compared with Baikal individuals, while the KPT005 individual shows a large component associated to WHG (light blue). See also Figure S1 and Table S1. We determined genetic sex by comparing the coverage on the sex chromosomes with the autosomal chromosomes, which revealed four females and 15 males. All individuals revealed low modern human DNA contamination at the mitochondrial level as well as through an estimation of X chromosomal heterozygosity on male individuals, except for KAG001 that showed 9.6% nuclear contamination (Table S1). No kin relationship was found among these individuals. We finally intersected our genotypes with SNPs on the Affymetrix Human Origins array (Lazaridis et al., 2014) and combined with published genotype data from 3,014 present-day worldwide individuals and 453 ancient individuals for population genomic analysis (Table S1).
|
|
|
Post by Admin on Nov 11, 2021 21:40:52 GMT
Population Structure We first performed principal-component analysis (PCA) to understand the genetic background of the studied individuals, against modern Eurasian and Native American populations, and projected selected ancient individuals onto the PCs calculated with modern ones (Figure 1C). Most of the Lake Baikal individuals occupied the space on a “ANE-NEA” cline running between “Northeast Asian” (NEA) ancestry represented by Neolithic hunter-gathers from the Devil’s Gate in the Russian Far East (Sikora et al., 2019, Siska et al., 2017), and the ANE ancestry represented by Upper Paleolithic Siberian individuals MA1, AfontovaGora 2 (AG2), and AfontovaGora 3 (AG3) (Fu et al., 2016, Raghavan et al., 2014a), which was first described by Damgaard et al. (2018a). Our newly sequenced Upper Paleolithic genome from the Ust-Kyakhta-3 site (UKY) just south to the Lake Baikal is placed close to the Mesolithic northeastern Siberian Kolyma individual (Sikora et al., 2019) and is shifted toward Native American populations compared to the rest of the ancient Baikal individuals along PC2. All four Early Neolithic individuals cluster with published Early Neolithic groups from the same region (Shamanka_EN, Lokomotiv_EN, UstBelaya_EN) (Damgaard et al., 2018a, Flegontov et al., 2019) designated as the “Baikal_EN” population. The LNBA individuals were divided into four groups. The major “Baikal_LNBA” group included 10 individuals and clustered with published Late Neolithic to Bronze Age Baikal populations (Shamanka_EBA, Kurma_EBA, UstIda_EBA, UstIda_LN, UstBelaya_BA). These individuals were positioned in PCA closer to ANE-related individuals compared with the Early Neolithic individuals from the same region, as well as closer to the Paleo-Eskimo Saqqaq individual (Rasmussen et al., 2010). Another two individuals (GLZ001 and GLZ002) from the Glazkovskoe predmestie site, unlike the third individual from the same archaeological site (GLZ003), seemed shifted from the main cluster and showed closer genetic affinity to the Devil’s Gate and Early Neolithic Baikal individuals. One of the six individuals from the Kachug site (KPT005) was substantially displaced from the Baikal_LNBA group toward western Eurasians along PC1, not along the ANE-NEA cline but toward later Bronze Age populations, suggesting a potential introgression of the Steppe-related ancestry. Finally, an Early Bronze Age individual (BZK002) from the Bazaikha site in the Yenisei River region further to the west of the Lake Baikal was significantly displaced toward ANE-related individuals and located close to published Bronze Age individuals associated to the Okunevo culture (Damgaard et al., 2018a). Population clustering with ADMIXTURE based on worldwide populations also showed a similar clustering pattern. When selecting a K value of 16 (see STAR Methods), the published and newly sequenced individuals belonging to main Early Neolithic to Bronze Age Baikal groups all showed genetic profiles composed of a mixture of three major components that were mostly enriched in ANE-related individuals, northeast Asians, and central Siberians represented by the Uralic-speaking Nganasan population (Figure 1D). The ANE and central Siberian ancestries were both of higher proportion in most LNBA Baikal individuals than in the Early Neolithic ones, while GLZ001 and GLZ002 showed higher NEA ancestry, similar to the Early Neolithic population. The BZK002 individual presented a profile similar to the published Okunevo group (Damgaard et al., 2018b), with a much larger ANE component compared to other Lake Baikal individuals. The KPT005 individual also displayed a substantial contribution derived from European “Western Hunter-Gatherer” (WHG) ancestry, likely acquired through gene flow from the west. We estimated the runs of homozygosity (ROH) of selected individuals together with published Baikal individuals (Table S1) and did not identify an inbreeding signal in any individual. The Kolyma individual showed significantly more ROH compared with other individuals, suggesting a smaller population size in Mesolithic northeastern Siberia (Figure S1). The sharing of identity-by-descent (IBD) segments between individuals suggested a close relationship between UKY and Kolyma, supporting our analyses based on genome-wide SNP data, and also revealed that Early Neolithic and LNBA Baikal individuals shared genetic affinity with each other as well as with the older UKY and Kolyma genomes (Figure S1). Figure S1. Population Size and Relatedness of Lake Baikal Populations Revealed by ROH and IBD Segments, Related to Figure 1 and Table S1 This figure summarizes the accumulative ROH length detected in each individual (row 1), shared IBD segment length of individuals within population (row 2), and shared IBD segment length of UKY, Baikal_EN and Baikal_LNBA individuals with other population (row 3-5), respectively. The long segments (> 8Mb) and short segments (< 2Mb) are also summarized separately.
|
|
|
Post by Admin on Nov 12, 2021 0:38:14 GMT
Upper Paleolithic Baikal Ancestry Links with Non-Arctic Native Americans In the population structure analysis, we found the Upper Paleolithic UKY individual to be closely related with the northeastern Siberian Kolyma individual. This is further validated by outgroup f3 statistics (Figure 2A) where, similarly to Kolyma, UKY showed close genetic affinity with Native American and Beringian populations (Figure 2A). F4 statistics in the form of f4(Mbuti, X; Kolyma, UKY) revealed that Kolyma is more closely related to populations from northeastern Siberia and North America compared with UKY (Figure S2). We further applied f4 statistics to explore the relationship of UKY and Kolyma with Native Americans and USR1 that was described as an outgroup to all non-Arctic Native Americans (Moreno-Mayar et al., 2018). Both UKY and Kolyma were symmetrically related with non-Arctic Native Americans and USR1, while USR1 shared significantly more genetic affinity with Native American populations compared to UKY and Kolyma (Figure S2; Table S2). Figure 2. Genetic Affinity between Upper Paleolithic UKY, Kolyma, and Native Americans (A) Genetic affinity between UKY and worldwide population assessed by f3(Mbuti;X,UKY). The sampling location of UKY is shown with a green triangle. The 10 test populations with highest f3 values are shown in diamonds and other populations in circles. (B) Graphic model of the relationship among UKY, Kolyma, and Native American populations. We first find the best fitted model with only UKY or only Kolyma as described in Figure S2 and then add Kolyma on the selected model with UKY and choose the best model based on maximum f-statistics Z scores and final scores reported for each model. The lineages related with Native American population are colored orange, and the northeast Asian-related lineages are colored red. See also Figure S2 and Table S2. Figure S2. Relationship between UKY, Kolyma, and Modern-Day Populations Based on f4 Statistics and qpGraph Modeling, Related to Figure 2 (A) This figure shows the different genetic affinities between UKY, Kolyma with worldwide population, assessed by f4(Mbuti, X; Kolyma, UKY). The test populations with significant f4 values (|Z| > 3) are shown in diamonds and other populations in circles. (B) This figure shows the graphic modeling of UKY (left) and Kolyma (right) on the skeleton graph including Mbuti, AG3, Onge, Devil’s Gate, USR1, ASO and ESN described in STAR Methods. The best fitted model for each individual is selected based on the maximum f-statistics Z scores and final scores reported for each model. We also investigated their genetic composition using qpAdm modeling (Haak et al., 2015) and found that both UKY and Kolyma possessed a similar level of ANE contribution, around 30%, when modeled as two-way mixture of Devil’s Gate (representing NEA ancestry) and AG3 (representing ANE ancestry) (Table S3). Noticeably, this model did not fit well for both UKY (p = 1.45E-03) and Kolyma (p = 3.98E-08), as the Native American Karitiana population showed extra affinity with the tested individuals compared to the fitted model (Table S3). This observation suggests that UKY and Kolyma shared a certain degree of genetic drift with Native American populations that occurred after the ancestors of Native Americans diverged from ANE and NEA ancestries. We further explored the relationships among UKY, Kolyma, and ancient Native American groups using the graphic-based qpGraph modeling (Patterson et al., 2012, Reich et al., 2009). We found that both UKY and Kolyma could be modeled as mixture between a northeast Asian lineage and a sister group of the Native American clade represented by USR1, Ancient Southwestern Ontario (ASO) individuals from Canada, and Early San Nicolas (ESN) individuals from the California Channel Islands in the USA (Scheib et al., 2018; Figure S2). When UKY and Kolyma were included in the same graph, they were consistently modeled as descendants from two independent admixture events, with ancestral lineages of both deriving from the Native American-related and the northeast Asian-related clades (Figure 2B). These findings confirm the close affinity of UKY and Kolyma to Native Americans but also highlight that both lineages contributing to UKY were ancestral to the groups contributing to Kolyma. In addition, aside from the first wave migrating into the Americas through Beringia, the admixture modeling suggests that the source of the Native American ancestry was more broadly spread across Siberia during the Upper Paleolithic, as UKY was found to be ∼4,000 years older and over 3,000 km further to the southwest from Kolyma. In fact, our admixture graph indicates that this basal Native American group experienced multiple genetic contacts with northeast Asian populations giving rise to distinct ancient Siberian populations.
|
|
|
Post by Admin on Nov 12, 2021 21:13:17 GMT
Complex Transition between the Early Neolithic and Bronze Age in the Lake Baikal Region A previous study described the transition between Early Neolithic and Bronze Age populations from the Lake Baikal region as the result of a discrete admixture event of ANE ancestry into the local gene pool (Damgaard et al., 2018a). In this study, we combined the newly sequenced Baikal_EN and Baikal_LNBA individuals with published data from the same time period (Figure 1C) and analyze these two combined datasets, Baikal_EN_all (n = 19) and Baikal_LNBA_all (n = 34), to better elucidate the genetic transition that occurred in this region. Prior to analyzing the combined groups, we confirmed the similarity between the new individuals and published groups using outgroup f3 statistics. Both Baikal_EN and Baikal_LNBA groups showed the highest genetic affinity with published Early Neolithic and LNBA Baikal populations, respectively (Table S4). From outgroup f3 statistics of the combined groups, we found both of the Baikal Early Neolithic and LNBA groups to be sharing high genetic affinity with ancient and modern northeast Asian and Siberian populations (Figure S3). The LNBA Baikal population also showed a high genetic affinity with the Paleo-Eskimo Saqqaq individual. Compared to their NEA proxy, they both carried extra genetic affinity with ANE-related populations while the LNBA population more so than the Early Neolithic population, as shown by f4 statistics (Figure S3). These results revealed the existence of ANE-related ancestry in the Early Neolithic population and, at the same time, validated the previous finding that an extra ANE ancestry gene flow is responsible for the genetic shift between Early Neolithic and Bronze Age Baikal populations. Figure S3. Genetic Affinity of Combined Baikal Populations with Worldwide Population, Related to Figure 3 The outgroup f3-statistics in the form of (A) f3(Mbuti, X; Baikal_EN_all) and (B) f3(Mbuti, X; Baikal_LNBA_all) are applied to measure the genetic affinity of Early Neolithic and LNBA Baikal individuals with worldwide population. The ten population with highest f3 are shown in diamonds. Then (C) f4(Mbuti, X; Devil’s Gate, Baikal_EN_all) and (D) f4(Mbuti, X; Baikal_EN_all, Baikal_LNBA_all) are used to show the genetic difference between NEA ancestry, Early Neolithic Baikal population and LNBA Baikal population. We further applied qpAdm modeling to quantify the proportion of ANE-related ancestry in Early Neolithic and LNBA Baikal populations, Saqqaq and Nganasan. Using Devil’s Gate as the NEA proxy, the Upper Paleolithic UKY was found to be a better fit than Kolyma as the ANE-related proxy for both Baikal populations, while AG3 provided a good fit for the Early Neolithic population (Table S5). Using Devil’s Gate and AG3 as the two proxies of NEA and ANE ancestries, respectively, we estimated the ANE-related ancestry increasing from 14.3% in the Early Neolithic Baikal population to 22.7% in the LNBA population (Figure 3A; Table S3). Of note, the northeastern Siberian Kolyma individual could work as a sufficient ANE proxy for Saqqaq, as described in the study where this genome was first reported (Sikora et al., 2019) but did not provide a good fit for the Baikal populations and Nganasan. This suggests that the Baikal hunter-gatherer and Nganasan populations are more likely to have formed in central or southern Siberia while Paleo-Eskimo ancestry could have emerged in either central or northeastern Siberia.
|
|
|
Post by Admin on Nov 13, 2021 22:31:24 GMT
Figure 3. Genetic Modeling of Early Neolithic to Bronze Age Baikal Individuals and Admixture Dating (A) qpAdm modeling of Early Neolithic and LNBA Baikal populations, together with Nganasan and Saqqaq, as admixture between NEA ancestry represented by Devil’s Gate and different ANE ancestries. The error bars show the standard errors of estimated ancestry proportions. Details for the modeling are provided in Table S3. (B) Estimated dates of admixture events between ANE and NEA ancestries in Early Neolithic and LNBA Baikal population. The individual ages are the averages and standard deviations of median radiocarbon dates without correcting for freshwater reservoir effect, to be consistent with previously published individuals. The estimated admixture dates are calculated with generation time of 29 years, and the error bars show the sum of standard errors of DATES estimations and individual ages. See also Figures S3 and S4 and Tables S3, S4, and S5. Furthermore, the program DATES was used to date the admixture events between ANE and NEA ancestries in the Baikal population based on the decay of ancestry covariance (Moorjani and Patterson, 2018). We detected a recent admixture signal in the Early Neolithic population, estimated to around 21 generations ago, while the admixture signal in LNBA population was dated to 71 generations ago, although this group harbored significantly more ANE-related ancestry (Table S5). When considering the average radiocarbon date of each population and the standard errors of their admixture dates, we identified contiguous intervals for the admixture events that spanned ∼8,500–6,000 BP, considering a generation time of 29 years (Figure 3B; Figure S4; Table S1; Table S5). Assuming a dating offset of 400–500 years due to freshwater reservoir effect estimated for the newly reported individuals, the admixture timings ranged between ∼8,000 and 5,500 BP. This suggests that both Baikal populations could have been formed through an extended admixture process between local groups and northeast Asian-related populations. The Early Neolithic groups were thus found to have experienced a prolonged admixture process, in contrast to the discrete and rather abrupt event suggested earlier (Damgaard et al., 2018a). This admixture, however, did not continue substantially in the Late Neolithic and Bronze Age, as suggested by the older admixture date for the LNBA population (Figure 3B) and the relatively larger genetic variation among Early Neolithic individuals compared to the homogeneous LNBA cluster, as shown in the PCA plot (Figure 1C). Figure S4. Dating of the Admixture Events in Baikal, Okunevo Population, and the BZK002 Individual, Related to Figures 3 and 5 and Table S5 This figure shows the DATES estimation of (A) time of admixture events in Early Neolithic and LNBA Baikal population and (B) time of admixture events in Okunevo population and BZK002 with different ancestor pairs. The red cross dots show the weighted ancestry covariance in different genetic distances, and the green curves show the exponential fitting starting at 0.5 cM. Details of the results are listed in Table S5.
|
|