Northern China stands as one of the earliest independent centers for cereal crop domestication globally, following closely behind the Near East. This region features two distinct agricultural systems: rainfed rice cultivation in the southern Yangtze River Basin and dryland millet farming in the north. The northern region encompasses a vast geographical area, including the Central Plain in the Yellow River (YR) basin, the cradle of the YR civilization since the Neolithic period, and other major river systems such as the West Liao River (WLR) basin in northeast China and the Amur River (AR) basin. The WLR played a distinct role in the adoption and spread of millet farming, with both foxtail and broomcorn millets cultivated there and in the lower YR basins since at least 6000 BCE. Over the following millennia, millets domesticated in northern China spread across East Eurasia. Both the YR and WLR basins boast rich archaeological cultures heavily reliant on millet farming. By the Middle Neolithic (roughly 4000 BCE), complex societies had emerged in both regions, exemplified by the Hongshan culture in the WLR and the Yangshao culture in the YR. The development of these complex societies coincided with population growth, cultural innovation, and the potential dispersal of major language families: Sino-Tibetan from the YR and Transeurasian from the WLR, although the latter's genealogical unity is debated. Unlike the YR, where crop cultivation dominated from the Middle Neolithic onward, the WLR's reliance on crops fluctuated due to climatic changes and shifts in archaeological cultures. Paleobotanical and isotopic evidence reveals a gradual increase in millet consumption from the Xinglongwa to Hongshan and Lower Xiajiadian cultures, followed by a partial shift to nomadic pastoralism in the Upper Xiajiadian culture. This shift, often attributed to climate change, raises the question of the role of human migration. The AR region, adjacent to the WLR, maintained a mixed subsistence strategy of hunting, fishing, and animal husbandry alongside some cultivation. The extent of interaction between YR and WLR societies and their impact on millet farming’s dispersal across northern China remains largely unknown. The scarcity of ancient human genomes has hindered our understanding of prehistoric migrations and contacts in this region. This study addresses this gap by presenting a genetic analysis of 55 ancient human genomes from various archaeological sites across northern China, providing insights into past human migration and admixture events, and their correlation with subsistence strategies.

Numerous studies have explored the origins and spread of agriculture in East Asia, highlighting the importance of millet cultivation in northern China. Research on the Yangshao and Hongshan cultures has revealed the development of complex societies in the YR and WLR basins, respectively. Archaeobotanical studies have provided valuable data on millet domestication and its spread throughout the region. Isotopic and paleobotanical analyses have shed light on dietary changes and the influence of climate on subsistence strategies. Archaeological evidence suggests links between cultural innovations, population growth, and the potential dispersal of language families from these regions. However, the limited availability of ancient human genomic data has restricted our understanding of the genetic processes underlying these cultural and demographic shifts. Existing studies primarily focus on specific aspects of the region's prehistory, lacking a comprehensive spatiotemporal analysis of ancient genomes across northern China. This study addresses the need for a large-scale genomic investigation to clarify the genetic connections between different regions and cultures over time, integrating archaeological and linguistic data to provide a more complete picture of human migration and cultural exchange in ancient northern China.

The study employed a comprehensive approach involving ancient DNA analysis and various population genetic methods. Initially, 107 ancient individuals from 19 archaeological sites across the AR, WLR, and YR basins (spanning ~2300 km and six millennia) were screened using shallow shotgun sequencing. Fifty-five individuals with sufficient DNA preservation were then sequenced to a higher depth (×0.03–7.53 autosomal coverage). Rigorous quality control measures were implemented to ensure data authenticity, including assessment of postmortem chemical damage, estimation of modern human contamination (generally below 4% for mitochondrial and 5% for nuclear estimates), and haploid genotype production using a combination of Affymetrix “HumanOrigins” and “1240k-Illumina” SNP datasets. The resulting ancient genomes were compared with published ancient and present-day human genome data. Principal Component Analysis (PCA) was used to visualize the genetic relationships among ancient and modern populations both in the context of Eurasian populations and East Asian populations more specifically. ADMIXTURE analysis was performed to infer ancestral components and assess population structure. Outgroup-f4 statistics were calculated to quantify the genetic relationships between populations, and qpWave/qpAdm framework was used to perform admixture modeling to assess the relative contributions of source populations to the ancient individuals in this study. A genetic continuity test was utilized to evaluate the direct ancestral relationship between ancient and present-day AR populations. Furthermore, the study considered archaeological and linguistic data to provide a broader context for the genetic findings and to explore the correlation between genetic changes and shifts in subsistence strategies. Specifically, the methods involved in screening for the ancient DNA, extracting it, preparing it and sequencing it are described. The various analysis techniques used are also described in the paper.

The study yielded several key findings regarding the genetic history of northern China. Firstly, ancient populations from the Amur River (AR) region exhibited remarkable genetic stability over time, with both ancient and present-day AR populations displaying similar genetic profiles. This contrasts sharply with the dynamic genetic profiles observed in the Yellow River (YR) and West Liao River (WLR) basins. In the YR basin, a gradual increase in genetic affinity with present-day southern Chinese and Southeast Asian populations was observed over time, correlating with the intensification of rice farming in the region. The WLR basin displayed more complex patterns of genetic change, with populations showing increased YR affinity during the Late Neolithic (linked to increased millet farming) and increased AR affinity in the Bronze Age (potentially connected to the adoption of pastoralism). The Middle Neolithic WLR region displayed notable spatial genetic heterogeneity, with some populations exhibiting a predominantly YR-related profile and others showing greater AR affinity. Admixture modeling suggested that Middle Neolithic WLR populations were a mixture of AR and YR-related ancestries. In the Late Neolithic, the WLR displayed increased YR affinity. By the Bronze Age, however, a partial shift toward an AR-related genetic profile was observed, correlating with the increasing importance of pastoralism in the region. The transition in the WLR from Middle to Late Neolithic periods was accompanied by a change in subsistence strategy, from a greater reliance on millet farming in the Middle Neolithic to a mixed agriculture/pastoralism economy in the Late Neolithic and Bronze Age. Genomic data suggested this shift was correlated with migration and admixture. In the YR, genetic data supported long-term genetic connection between YR populations over time, alongside a notable genetic contribution from southern Chinese and Southeast Asian populations, potentially reflecting the northward spread of rice farming. Genetic analysis also supported a long-term genetic connection between YR populations and present day Tibetans, though other models were also plausible. Furthermore, the study observed a clear geographical gradient in genetic affinity with AR and YR populations in the Middle Neolithic. The authors used qpAdm to model the ancestry proportions in these populations. The use of admixture modeling and f-statistics allowed for quantifying and visualizing genetic relationships and migrations. The genetic continuity test provided statistical support for the observed trends, supporting the idea that migration was a significant factor in shaping the genetic landscape of the region.

The findings of this study significantly advance our understanding of the complex interplay between subsistence changes, human migration, and genetic diversity in ancient northern China. The contrasting genetic trajectories of the AR, YR, and WLR regions highlight the influence of environmental factors and subsistence practices on population dynamics. The genetic changes observed in the YR and WLR are strongly correlated with shifts in subsistence strategies, supporting the hypothesis that changes in food production systems were significant drivers of migration and admixture. The observed patterns provide strong support for the idea that demographic changes influenced the spread of agricultural practices and potentially, the spread of languages. The study sheds light on the spatial genetic heterogeneity of northern China during the Neolithic, suggesting that migration and admixture were not uniformly distributed. The study's findings highlight the importance of integrating genomic data with archaeological and linguistic evidence to reconstruct the complex history of human populations. The temporal and spatial genetic variations in the study fuel the debate about the connection between language families and migrations in the region. However, the authors note that their data are limited and call for further research to obtain a more complete picture of the historical interactions in the region.

This study provides compelling evidence for a strong link between subsistence changes and human migration in ancient northern China. The genetic patterns observed in the AR, YR, and WLR regions reveal a dynamic interplay of genetic stability, admixture, and geographic gradients, influenced by environmental and cultural factors. The integration of genomic data with archaeological and linguistic insights offers a new perspective on the prehistoric population dynamics and the spread of cultural practices in this crucial region. Further studies incorporating additional ancient genomes from geographically diverse populations and employing more sophisticated analytical methods are warranted to further refine our understanding of the intricate genetic history of northern China and its connections with other regions of East Asia.

While this study represents a significant advancement in our understanding of ancient northern China, it also has limitations. The sample size, though substantial, may not fully represent the genetic diversity of all cultures and populations in the region. The lack of ancient genomes from some key regions, such as those involved in the early spread of rice farming from Southern China, restricts the ability of the study to define the exact origins of the observed admixtures. The authors acknowledge the limitations of inferring direct causal relationships between genetic changes and specific cultural or environmental shifts, emphasizing the need for further research using a combination of genomic and non-genomic data.