Permafrost preservation reveals proteomic evidence for yak milk consumption in the 13th century

Domesticated yaks (Bos grunniens) are crucial to high-altitude communities in eastern Eurasia, providing milk, transport, and other resources. While proteomic evidence exists for ancient consumption of other ruminant and horse milk, yak dairy products haven't been detected previously. The domestication and dispersal of yaks from Tibet are also poorly understood due to limited zooarchaeological data. This study investigates the potential of paleoproteomics to address this gap, focusing on human dental calculus from elite individuals buried in the Khovsgol province of Mongolia, a region where yak herding continues today. The research question centers on whether paleoproteomic analysis can reveal evidence of yak milk consumption in this region and what this reveals about the history of yak domestication and resource use. The purpose is to provide the first direct evidence for yak dairy consumption, potentially shedding light on the species' spread and utilization in Mongolia. The study's importance lies in its potential to expand our understanding of past human-animal interactions in high-altitude environments and the role of yaks in the socio-economic structure of historical communities. The lack of prior archaeological or historical evidence for yak dairying in Mongolia makes this paleoproteomic approach especially relevant, offering a unique opportunity to explore aspects of human diet and resource management often overlooked by traditional archaeological methods.

Existing literature establishes the importance of yaks in high-altitude regions of eastern Eurasia, particularly Tibet and Mongolia. However, evidence for yak domestication and the early uses of yak products, including milk, in Mongolia remains scarce. Archaeological findings are limited to a single yak cranium and sparse zooarchaeological evidence, primarily focusing on the Mongol era or later periods. Historical texts also mention yak products like felt, indicating their economic significance but lacking details on dairying practices. Prior paleoproteomic studies in the region have focused on other animals, with limited success in identifying yak milk proteins due to the complex taxonomic challenges and the need for yak specific peptides. This study, therefore, aims to address this gap in our understanding by applying advanced paleoproteomic techniques to dental calculus samples from elite individuals.

Dental calculus samples were collected from 11 adult individuals excavated from the Khorig 1 and Khorig 2 cemeteries in Mongolia. Samples were processed using induced pressure-solvent extraction, filtered, and analyzed using liquid chromatography-tandem mass spectrometry (LC-MS/MS). Radiocarbon dating of the samples was conducted using Accelerator Mass Spectrometry (AMS) at the Biosit University Environmental Research Centre (SUBERC) Radiocarbon Dating Lab following a modified Langer (1971) protocol and standard AMS procedures, employing collagen extraction and the measurement of carbon-to-nitrogen elemental ratios (C/N) to assess the integrity of the sample. Protein identification was performed using Mascot software, searching against a custom-made database, including an Old Splice Screening Database (OSSD) to control for contamination. The OSSD screening database was used to assess the authenticity of the protein profiles, distinguishing between genuine ancient proteins and potential contaminants. A specific protocol was followed, employing multiple extraction protocols including a modified phosphate buffered saline (PBS) extraction, and pressure modulation. Protein identification was facilitated through the use of a combination of tryptic and semi-tryptic search parameters using the Mascot search engine (v2.6.0) coupled to a database including SwissProt, the original ancient DNA study database, and the HODM (house-developed Old Splice Screening Database). To confirm peptide identifications, NCBI's Protein BLAST search engine was employed. A data-dependent acquisition approach (DDA) was used for LC-MS/MS, setting the Q Exactive HF mass spectrometer parameters for optimal protein identification. The data were then processed to ensure data accuracy by plotting against the original database, HODM, to demonstrate consistency and to account for any minor discrepancies that can sometimes occur when using different databases. The study also accounted for potential contaminating modern proteins which might interfere with ancient protein identification and carefully considered the implications of semi-tryptic versus tryptic search settings. All data are available via ProteomeXchange (PXD024510).

Analysis of the dental calculus revealed the presence of yak-specific milk proteins (BLG-E) in two individuals, providing the earliest direct evidence globally for yak milk consumption. The radiocarbon dating of these individuals placed their lives between 1170-1270 CE and 1287-1397 CE. Furthermore, a wide range of other proteins were identified, including milk proteins from other ruminants (cattle, sheep, goat), horses, and blood proteins from ruminants and horses, indicating meat consumption. The diversity of proteins detected, especially the rare caseins from horses and the yak-specific BLG-E in two individuals, suggests exceptional preservation conditions at the permafrost burial site. A total of eight individuals presented evidence of milk, meat or both, with varying quantities of proteins found per individual. Specifically, two individuals (HOR-27 and HOR-01) showed evidence of yak milk consumption based on the presence of yak-specific peptides, particularly the BLG-E variant of beta-lactoglobulin. Other key findings included: the identification of horse milk proteins (beta-globulin, beta-casein, alpha-S2-casein) in several individuals, representing the first instance of horse milk proteins identified in ancient samples; a wider diversity of milk proteins (BLG, alpha-lactalbumin, Lysozyme C, beta-casein, alpha-S1-casein, alpha-S2-casein, kappa-casein) than previously documented in archaeological dental calculus was recovered, which is likely due to the unique preservation at the Khorig site; and the identification of blood and tissue proteins (hemoglobin and serum albumin) from ruminants and horses, providing additional information on the individuals' diets and potential animal husbandry practices.

The findings demonstrate the potential of paleoproteomics to recover previously unknown information on ancient diets and resource use, particularly in challenging environments. The detection of yak milk proteins in the 13th century in Mongolia suggests that yak dairying was a significant practice among elites during the Mongol period. This aligns with textual evidence describing the use of yak hair products, indicating the economic and cultural importance of yaks. However, the relatively late date may reflect the limitations of current analytical techniques, rather than the true timeframe for yak dairying. The excellent preservation of proteins in permafrost allows the recovery of more diverse proteins, increasing our understanding of ancient diets, even to the extent of identifying blood and tissue proteins, which is not often possible in such studies. The absence of yak-related remains from traditional archaeological investigations emphasizes the advantage of paleoproteomics in filling research gaps.

This study offers the first direct evidence of yak milk consumption in Mongolia, dating to the 13th century, a time period with sparse traditional archaeological or historical accounts. The results highlight the potential of paleoproteomics to illuminate the history of yak domestication, dispersal, and usage, significantly expanding our knowledge of past human-animal interactions in high-altitude environments. Future research should focus on expanding the geographical scope of paleoproteomic analyses to earlier time periods and different regions to better understand the long-term history of yak utilization in eastern Eurasia.

The study is limited by the relatively small sample size and the focus on elite individuals. While the findings suggest yak milk consumption was significant, it might not represent the full range of dietary practices within the wider population. Moreover, the reliance on dental calculus requires careful consideration of potential biases in preservation and protein representation. The exceptional preservation at the Khorig site may not be typical of other archaeological contexts, limiting the generalizability of the findings.