Post by Admin on Jul 26, 2019 17:43:54 GMT
We combined our ancient and modern samples with previously published ancient data6, 7, 8, 9, 10, 11, 13, 36, 37, 38 (Figure 1A) resulting in a dataset of 389 individuals and 1,046,317 SNPs when ancient and Lebanese samples were analyzed, and 546,891 SNPs when 2,583 modern samples from the Human Origins genotype data were included in the analysis (i.e., the small dataset was used only when a modern population other than the Lebanese was included in the test).9, 39 A pooled Lebanese sequence dataset (99 low coverage plus 4 high coverage) was used in all analyses except for the PCA and ADMIXTURE where a subset of 15 randomly selected individuals (5 from each group described in Figure S5) was used to avoid sample size bias. The ancient samples were grouped following the labels assigned by Lazaridis et al.13 on the basis of archaeological culture, chronology, and genetic clustering. We used this dataset to shed light on the genetic history of the Canaanites, resolving their relationship to other ancient populations and assessing their genetic contribution to present-day populations.
Figure 1 Population Locations and Genetic Structure
We first explored our dataset using PCA40 on present-day West Eurasian (including Levantine) populations and projected the ancient samples onto this plot (Figures 1B and S6). The Bronze Age Sidon samples (Sidon_BA) overlap with present-day Levantines and were positioned between the ancient Levantines (Natufians/Neolithic) and ancient Iranians (Neolithic/Chalcolithic). The overlap between the Bronze Age and present-day Levantines suggests a degree of genetic continuity in the region. We explored this further by computing the statistic f4(Lebanese, present-day Near Easterner; Sidon_BA, Chimpanzee) using qpDstat39 (with parameter f4mode: YES) and found that Sidon_BA shared more alleles with the Lebanese than with most other present-day Levantines (Figure S7), supporting local population continuity as observed in Sidon’s archaeological records. When we substituted present-day Near Easterners with a panel of 150 present-day populations available in the Human Origins dataset, we found that only Sardinians and Italian_North shared significantly more alleles with Sidon_BA compared with the Lebanese (Figure S8). Sardinians are known to have retained a large proportion of ancestry from Early European farmers (EEFs) and therefore the increased affinity to Sidon_BA could be related to a shared Neolithic ancestry. We computed f4(Lebanese, Sardinian/Italian_North; Sidon_BA, Levant_N) and found no evidence of increased affinity of Sardinians or Italian_North to Sidon_BA after the Neolithic (both Z-scores are positive). We next wanted to explore whether the increased affinity of Sidon_BA to the Lebanese could also be observed when analyzing functionally important regions of the genome that are less susceptible to genetic drift.
Our sequence data allowed us to scan loci linked to phenotypic traits and loci previously identified as functional variants in the Lebanese and other Levantines.41, 42, 43 Using a list of 84 such variants (Table S2), we estimated the allele frequency (AF) in Sidon_BA using ANGSD26 based on a method from Li et al.44 and calculated Pearson pairwise correlation coefficients between AF in Sidon_BA and AF in Africans, Europeans, Asians,32 and Lebanese. We found a high significant correlation between Sidon_BA and the Lebanese (r = 0.74; 95% CI = 0.63–0.82; p value = 8.168 × 10−16) and lower correlations between Sidon_BA and Europeans (r = 0.56), Africans (r = 0.55), and Asians (r = 0.53) (Figure S9). These results support population continuity in the region and suggest that several present-day genetic disorders might stem from risk alleles that were already present in the Bronze Age population. In addition, SNPs associated with phenotypic traits show that Sidon_BA and the Lebanese had comparable skin, hair, and eye colors (in general: light intermediate skin pigmentation, brown eyes, and dark hair) with similar frequencies of the underlying causal variants in SLC24A5 and HERC2, but with Sidon_BA probably having darker skin than Lebanese today from variants in SLC45A2 resulting in darker pigmentation (Table S2).
Figure 2 Admixture in Bronze Age Levantine Populations
The PCA shows that Sidon_BA clusters with three individuals from Early Bronze Age Jordan (Jordan_BA) found in a cave above the Neolithic site of ‘Ain Ghazal and probably associated with an Early Bronze Age village close to the site.13 This suggests that people from the highly differentiated urban culture on the Levant coast and inland people with different modes of subsistence were nevertheless genetically similar, supporting previous reports that the different cultural groups who inhabited the Levant during the Bronze Age, such as the Ammonites, Moabites, Israelites, and Phoenicians, each achieved their own cultural identities but all shared a common genetic and ethnic root with Canaanites.15 Lazaridis et al.13 reported that Jordan_BA can be modeled as mixture of Neolithic Levant (Levant_N) and Chalcolithic Iran (Iran_ChL). We computed the statistic f4(Levant_N, Sidon_BA; Ancient Eurasian, Chimpanzee) and found that populations from the Caucasus and ancient Iran shared more alleles with Sidon_BA than with Neolithic Levant (Figure 2A and S10). We then used qpAdm8 (with parameter allsnps: YES) to test whether Sidon_BA can be modeled as mixture of Levant_N and any other ancient population in the dataset and found good support for the model of Sidon_BA being a mixture of Levant_N (48.4% ± 4.2%) and Iran_ChL (51.6% ± 4.2%) (Figure 2B; Table S3).
In addition, the two Sidon_BA males carried the Y-chromosome haplogroups45 J-P58 (J1a2b) and J-M12 (J2b) (Tables 1 and S4; Figure S11), both common male lineages in the Near East today. Haplogroup J-P58 is frequent in the Arabian peninsula with proposed origins in the Zagros/Taurus mountain region.46 It forms the vast majority of the Y chromosomes in southwestern Mesopotamia and reaches particularly high frequencies (74.1%) in Marsh Arabs in Iraq.47 On the other hand, haplogroup J-M12 is widespread at low frequency from the Balkans to India and the Himalayas, with Albanians having the highest proportions (14.3%).48 We compiled frequencies of Y-chromosome haplogroups in this geographical area and their changes over time in a dataset of ancient and modern Levantine populations (Figure S12), and note, similarly to Lazaridis et al.,13 that haplogroup J was absent in all Natufian and Neolithic Levant male individuals examined thus far, but emerged during the Bronze Age in Lebanon and Jordan along with ancestry related to Iran_ChL. All five Sidon_BA individuals had different mitochondrial DNA haplotypes49 (Table 1), belonging to paragroups common in present-day Lebanon and nearby regions (Table S5) but with additional derived variants not observed in our present-day Lebanese dataset.
Figure 1 Population Locations and Genetic Structure
We first explored our dataset using PCA40 on present-day West Eurasian (including Levantine) populations and projected the ancient samples onto this plot (Figures 1B and S6). The Bronze Age Sidon samples (Sidon_BA) overlap with present-day Levantines and were positioned between the ancient Levantines (Natufians/Neolithic) and ancient Iranians (Neolithic/Chalcolithic). The overlap between the Bronze Age and present-day Levantines suggests a degree of genetic continuity in the region. We explored this further by computing the statistic f4(Lebanese, present-day Near Easterner; Sidon_BA, Chimpanzee) using qpDstat39 (with parameter f4mode: YES) and found that Sidon_BA shared more alleles with the Lebanese than with most other present-day Levantines (Figure S7), supporting local population continuity as observed in Sidon’s archaeological records. When we substituted present-day Near Easterners with a panel of 150 present-day populations available in the Human Origins dataset, we found that only Sardinians and Italian_North shared significantly more alleles with Sidon_BA compared with the Lebanese (Figure S8). Sardinians are known to have retained a large proportion of ancestry from Early European farmers (EEFs) and therefore the increased affinity to Sidon_BA could be related to a shared Neolithic ancestry. We computed f4(Lebanese, Sardinian/Italian_North; Sidon_BA, Levant_N) and found no evidence of increased affinity of Sardinians or Italian_North to Sidon_BA after the Neolithic (both Z-scores are positive). We next wanted to explore whether the increased affinity of Sidon_BA to the Lebanese could also be observed when analyzing functionally important regions of the genome that are less susceptible to genetic drift.
Our sequence data allowed us to scan loci linked to phenotypic traits and loci previously identified as functional variants in the Lebanese and other Levantines.41, 42, 43 Using a list of 84 such variants (Table S2), we estimated the allele frequency (AF) in Sidon_BA using ANGSD26 based on a method from Li et al.44 and calculated Pearson pairwise correlation coefficients between AF in Sidon_BA and AF in Africans, Europeans, Asians,32 and Lebanese. We found a high significant correlation between Sidon_BA and the Lebanese (r = 0.74; 95% CI = 0.63–0.82; p value = 8.168 × 10−16) and lower correlations between Sidon_BA and Europeans (r = 0.56), Africans (r = 0.55), and Asians (r = 0.53) (Figure S9). These results support population continuity in the region and suggest that several present-day genetic disorders might stem from risk alleles that were already present in the Bronze Age population. In addition, SNPs associated with phenotypic traits show that Sidon_BA and the Lebanese had comparable skin, hair, and eye colors (in general: light intermediate skin pigmentation, brown eyes, and dark hair) with similar frequencies of the underlying causal variants in SLC24A5 and HERC2, but with Sidon_BA probably having darker skin than Lebanese today from variants in SLC45A2 resulting in darker pigmentation (Table S2).
Figure 2 Admixture in Bronze Age Levantine Populations
The PCA shows that Sidon_BA clusters with three individuals from Early Bronze Age Jordan (Jordan_BA) found in a cave above the Neolithic site of ‘Ain Ghazal and probably associated with an Early Bronze Age village close to the site.13 This suggests that people from the highly differentiated urban culture on the Levant coast and inland people with different modes of subsistence were nevertheless genetically similar, supporting previous reports that the different cultural groups who inhabited the Levant during the Bronze Age, such as the Ammonites, Moabites, Israelites, and Phoenicians, each achieved their own cultural identities but all shared a common genetic and ethnic root with Canaanites.15 Lazaridis et al.13 reported that Jordan_BA can be modeled as mixture of Neolithic Levant (Levant_N) and Chalcolithic Iran (Iran_ChL). We computed the statistic f4(Levant_N, Sidon_BA; Ancient Eurasian, Chimpanzee) and found that populations from the Caucasus and ancient Iran shared more alleles with Sidon_BA than with Neolithic Levant (Figure 2A and S10). We then used qpAdm8 (with parameter allsnps: YES) to test whether Sidon_BA can be modeled as mixture of Levant_N and any other ancient population in the dataset and found good support for the model of Sidon_BA being a mixture of Levant_N (48.4% ± 4.2%) and Iran_ChL (51.6% ± 4.2%) (Figure 2B; Table S3).
In addition, the two Sidon_BA males carried the Y-chromosome haplogroups45 J-P58 (J1a2b) and J-M12 (J2b) (Tables 1 and S4; Figure S11), both common male lineages in the Near East today. Haplogroup J-P58 is frequent in the Arabian peninsula with proposed origins in the Zagros/Taurus mountain region.46 It forms the vast majority of the Y chromosomes in southwestern Mesopotamia and reaches particularly high frequencies (74.1%) in Marsh Arabs in Iraq.47 On the other hand, haplogroup J-M12 is widespread at low frequency from the Balkans to India and the Himalayas, with Albanians having the highest proportions (14.3%).48 We compiled frequencies of Y-chromosome haplogroups in this geographical area and their changes over time in a dataset of ancient and modern Levantine populations (Figure S12), and note, similarly to Lazaridis et al.,13 that haplogroup J was absent in all Natufian and Neolithic Levant male individuals examined thus far, but emerged during the Bronze Age in Lebanon and Jordan along with ancestry related to Iran_ChL. All five Sidon_BA individuals had different mitochondrial DNA haplotypes49 (Table 1), belonging to paragroups common in present-day Lebanon and nearby regions (Table S5) but with additional derived variants not observed in our present-day Lebanese dataset.
