Introduction
Understanding the landscape of animal-based food components is essential to inform potential nutritional material flow and assess protection on human health and food security (Tyndall et al. Reference Tyndall, Maloney and Cole2022). The production of these substances through biological transformation is a critical aspect of One Health, particularly regarding their production processes in animals and their consumption by humans (Bizzaro et al. Reference Bizzaro, Vatland and Pampanin2022). Milk or meat food component analysis of proteins, fats, carbohydrates, solids and/or ash is widely assessed by traditional identifying methods or standard. In addition, milk and meat processing, storage, origin, breed, feed regime, gender, age, and other factors have a strong effect on the metabolome of the milk and meat of ruminants. Many factors (nutrition, feed sources, genders, management, weather, processing, handling conditions, adulteration, and related concerns) impact the integrity and security of products (Suh Reference Suh2022). Potential biomarkers and metabolic mechanisms associated with the meat and milk synthesis need more research, which could lay a strong foundation of the food quality standards and then improve the consumers’ options and producers’ initiative.
Metabolomics, also known as metabonomics, focuses on studying small molecules technology and compounds identification and quantification with the high-throughput techniques (<1500 Da) (German et al. Reference German, Hammock and Watkins2005; Plumb et al. Reference Plumb, Gethings and Rainville2023), encompasses a variety of both internal and external chemical compounds, including fatty acids (FAs), peptides, amino acids (AAs), carbohydrates, nucleic acids, vitamins, organic acids, alkaloids, polyphenols, minerals and just about any other chemicals that can be utilized, produced or consumed by specific cells or organisms. Metabolomics has contributed significantly to livestock research and industry. This includes breakthroughs in animal health, breeding, and production, showcasing its pivotal role in advancing agricultural practices. Till now, many reviews are focusing on the metabolomics technology summary (Afshari et al. Reference Afshari, Pillidge and Dias2020; Akhtar et al. Reference Akhtar, Samar and Shami2021; Munekata et al. Reference Munekata, Pateiro and López-Pedrouso2021) or metabolome database development related to livestock or food, such as a recent comprehensive Livestock Metabolome Database (LMDB, available at http://www.lmdb.ca) was released for targeted metabolomic studies (Goldansaz et al. Reference Goldansaz, Guo and Sajed2017) and a bovine and bovine rumen fluid metabolome (Foroutan et al. Reference Foroutan, Fitzsimmons and Mandal2020; Saleem et al. Reference Saleem, Bouatra and Guo2013). Lipids, encompassing a range of FAs and lipid-soluble bioactive compounds, are of paramount importance in enhancing the sensory qualities of foods, such as flavor and texture, while also extending shelf life. These effects are pivotal in food processing and shape consumer preferences (Muroya Reference Muroya2023; Wittwer et al. Reference Wittwer, Lee and Ranadheera2023; Yu et al. Reference Yu, Zhang and Jiang2024). Lipids found in bovine milk and meat possess a multitude of biological functions, having significant impacts on human health and the physical properties of food products. The emergence of metabolomics technology underscores the complex interplay between diet and health, highlighting the critical role of these lipids in both nutritional science and food technology (Wishart Reference Wishart2008; Zhang et al. Reference Zhang, Sun and Elmaidomy2023), which bring in a development of recognizing the milk and food quality through the metabolome. The application of metabolomics with selected markers is potentially useful in evaluating the genuineness of unidentified food specimens (Zhong et al. Reference Zhong, Wei and Li2022). Ruminant food metabolomics plays a key role in food chemistry, food quality, and the identification of biomarkers linked to economically valuable traits. Thus, to build the database of the ruminant foods of milk and meat metabolome (for the most common ruminant species, namely dairy cow, beef cattle, camel, buffalo, yak, reindeer, sheep, and goats, etc.) is necessary for producers (food industry and animal farmers), consumer preference, and researchers.
Correlation analyses between metabolites and compositional traits of ruminant foods (milk or meat) provide insights into the underlying biological mechanisms and aid in developing fingerprints and biomarkers for identifying food properties (Caboni et al. Reference Caboni, Murgia and Porcu2019; Munekata et al. Reference Munekata, Pateiro and López-Pedrouso2021). The core metabolic biomarkers and biological biosynthetic processes of milk and meat enhance our understanding of the interplay between conventional food production from animals and new synthetic biological technologies. In this review, we summarized the recent progress and applications of metabolomics in determining ruminant food products quality to acquire a detailed overview of the metabolite profile and its fluctuations, as well the biomarkers and indicators to reflect the food origin, adulterate, trait, quality, flavor, taste, safety, etc. The milk or meat metabolome not only directly represents food quality and safety but also serves as an indicator of the animal’s metabolic properties and health status (Lu et al. Reference Lu, Antunes Fernandes and Páez Cano2013; Ye et al. Reference Ye, Yang and Xiao2023). Thus, this review aimed to clarify the following: (1) What are the preferred metabolomics technologies in ruminant foodomics? (2) What are the most obvious weaknesses and advantages in ruminant metabolomics relative to other fields of metabolomics research? (3) What are the known or measured metabolites and biomarkers for the seven major ruminant food (cow, sheep & goat milk, camel milk, yak milk, beef, sheep & goat meat, yak meat)? (4) What is the relationship between ruminant food metabolome and animal science, as well the connection between ruminant food metabolome and human health and nutrition? and (5) What role does metabolomics play in the production of ruminant food alternatives?
Metabolomics techniques
Techniques and data acquisition
A variety of metabolomics technologies are available, such as nuclear magnetic resonance (NMR), gas chromatography–mass spectrometry (GC-MS), liquid chromatography–mass spectrometry (LC-MS), capillary electrophoresis–mass spectrometry (CE-MS), high-performance liquid chromatography with ultraviolet detection (HPLC-UV), and inductively coupled plasma mass spectrometry (ICP-MS), each with distinct advantages and disadvantages. Various analytical platforms, including NMR, HPLC-UV, LC-MS, GC-MS, ICP-MS, and CE-MS, are commonly used in metabolomics studies (Table 1). To date, one of the most exhaustive metabolomic analyses conducted involved the utilization of 5 distinct platforms in a study focusing on the metabolome of bovine ruminal fluid (Saleem et al. Reference Saleem, Bouatra and Guo2013). However, for the food-metabolomics study, the GC-MS-, LC-MS-, and NMR-based platform are the frequenters.
Table 1. Metabolomics techniques used in metabolomics studies
GC-MS is a cornerstone of metabolomic research, valued for its efficiency, reproducibility, reliability, selectivity, and robustness. It is distinguished by its exceptional sensitivity and highly consistent fragmentation patterns, making it highly effective for the precise analysis of complex metabolite profiles (Ren et al. Reference Ren, Zhang and Kong2018; Zeki et al. Reference Zeki, Eylem and Recber2020). Specifically, GC-MS based on volatile organic compounds (VOCs) is always used to detect the volatile metabolites to indicate the meat flavor, which can also be treated as volatilomics (Pavlidis et al. Reference Pavlidis, Mallouchos and Ercolini2019). With the development of new techniques, nowadays, GC × GC-TOF-MS has gained widespread application in the analysis of complex food matrices, thanks to its enhanced resolution and sensitivity (Li et al. Reference Li, Chen and Blank2021). Ion mobility spectrometry (IMS) is a potent analytical tool, and when combined with gas chromatography (GC-IMS), it serves as a rapid approach for profiling VOCs in food (Li et al. Reference Li, Chen and Blank2021). Headspace solid-phase microextraction paired with gas chromatography–mass spectrometry (HS-SPME/GC-MS) emerges as a promising approach for distinguishing between different meat species (Pavlidis et al. Reference Pavlidis, Mallouchos and Ercolini2019). To enhance the volatility and thermal stability of these analytes, strategies such as methoximation and trimethylsilylation are frequently employed in large-scale metabolomics research, facilitating their analysis via GC-MS (Atapattu and Temerdashev Reference Atapattu and Temerdashev2023; Fiehn Reference Fiehn2016). The most advantage of the GC-MS is its high sensitivity in identifying the volatile metabolites that are crucial factors for food flavor evaluation.
LC-MS has become a predominant tool in identifying food metabolites, owing to its high-resolution molecular mass determination and detailed fragmentation patterns observed in MS/MS spectra, thereby facilitating the analysis of complex mixtures with unparalleled precision (Lu et al. Reference Lu, Zhao and Bai2008). In LC-MS-based metabolomic studies, samples can be directly analyzed with minimal preparation, often requiring only filtration.
HPLC-UV is an analytical technique combining high-pressure liquid chromatography with UV light detection (Amarnath et al. Reference Amarnath, Amarnath and Amarnath2003). In metabolomics, HPLC-UV is utilized to separate, identify, and quantify metabolites in complex biological samples based on their absorption of UV light (Amarnath et al. Reference Amarnath, Amarnath and Amarnath2003). This method is especially useful for analyzing compounds with known UV absorbance characteristics, making it valuable for targeted metabolite analysis and contributing to the comprehensive profiling of metabolomes in dairy and meat products (Herzallah Reference Herzallah2009; Korchazhkina et al. Reference Korchazhkina, Jones and Czauderna2006; Zergiebel et al. Reference Zergiebel, Ueberschaar and Seeling2023).
NMR spectroscopy is a high-performance tool for the analysis of metabolome and organic compounds, which has been successfully utilized in milk and meat (Klein et al. Reference Klein, Buttchereit and Miemczyk2012; Zhu et al. Reference Zhu, Petracci and Li2020). NMR stands out in the realm of metabolomics for offering a wider spectrum of profiling data, coupled with the benefits of straightforward sample preparation and relatively swift analysis. NMR spectroscopy offers a distinct advantage over LC-MS and GC-MS: a direct and quantitative correlation exists between molar concentration and the intensity of NMR resonances (Tenori et al. Reference Tenori, Santucci and Meoni2018). On the other hand, compared with other MS-based analytical techniques, NMR requires simple sample pretreatment, which is time saving and NMR analysis is environmentally friendlier due to its reduced consumption of organic solvents. However, low resolution and sensitivity hinder its utilization in identifying novel compounds and in foodomics. The 1H-NMR metabolomic approach has been successfully applied to study the potential biomarkers of different diet (Madrid-Gambin et al. Reference Madrid-Gambin, Brunius and Garcia-Aloy2018), geographical origin (Jung et al. Reference Jung, Lee and Kwon2010), and in detection of the adulteration in Chicken, Chevon, Beef and Donkey meat (Akhtar et al. Reference Akhtar, Samar and Shami2021). Recently, three quantitative NMR metabolomics analysis methods (ultrafiltration, solvent precipitation with either acetonitrile/acetone/methanol or chloroform/methanol) with excellent protein removal, high concentrations of metabolites and high reproducibility are recommended to be used for lamb meat metabolome analysis (Samuelsson et al. Reference Samuelsson, Olivecrona and Cônsolo2021).
The choice of technique depends on the study’s goals: GC-MS is preferred for analyzing volatile compounds, LC-MS for complex mixtures with high precision, NMR for broader profiling with simpler sample preparation, and HPLC-UV for targeted analysis of UV-absorbing compounds (Table 2). Combining these techniques can improve metabolome coverage and enhance the identification of metabolites, providing a more complete understanding of the biochemical processes under study. The integration of GC-MS, LC-MS, NMR, and HPLC-UV allows researchers to capitalize on the strengths of each method, thereby increasing the breadth and depth of metabolite detection and characterization, ultimately yielding a more comprehensive metabolic profile.
Table 2. The choice of analytical platform for ruminant foods
Nowadays, relying solely on a single platform poses significant challenges in identifying compounds, and the accuracy of data prediction is also limited. Combining GC-MS and LC-MS techniques is common in metabolomics studies to broaden the scope of detected metabolites. This integration aims to improve the accuracy, precision, and comprehensiveness of identifying phenotype-related metabolites (Zeki et al. Reference Zeki, Eylem and Recber2020). These three methods collectively identified 353 metabolites, with only 65 detected by the GC-MS component. This underscores how employing diverse instruments in metabolomic profiling can augment the number of identified metabolites (Simon-Manso et al. Reference Simon-Manso, Lowenthal and Kilpatrick2013). A comprehensive milk metabolome profiling was conducted, employing chemical isotope labeling and LC-MS techniques. Specifically, dansylation labeling was utilized to target the amine/phenol submetabolome, allowing for a focused analysis of these compounds, illustrating the complexity of the milk metabolome (Mung and Li Reference Mung and Li2017).
Targeted and untargeted approaches
In metabolomic approaches utilizing GC-MS and LC-MS platforms, it is crucial to delineate between targeted and untargeted methods. Targeted approaches focus on identifying and quantifying a specific set of known metabolites, typically ranging from tens to hundreds. These may include common marker compounds pivotal in clinical or technological analyses (Chen et al. Reference Chen, Zhong and Zhu2020; Lelli et al. Reference Lelli, Belardo, Timperio and Xianquan2021). Conversely, untargeted approaches aim to gather extensive information by annotating metabolites and examining both known and unknown metabolic alterations (Guo and Huan Reference Guo and Huan2020; Lelli et al. Reference Lelli, Belardo, Timperio and Xianquan2021).
In untargeted metabolomics studies, two commonly utilized data acquisition methods exist. The first relies on full scan MS-only acquisitions to provide accurate mass measurements for individual molecules (raw mass features), facilitating multivariate statistical calculations. Subsequently, data-dependent acquisition is employed for identification purposes, generating fragmentation patterns for the metabolites with the highest signal intensity (Guo and Huan Reference Guo and Huan2020). Another untargeted metabolomics strategy involves data independent acquisition, which integrates full scan MS-only acquisition with MS/MS fragmentation for all precursor ions, either concurrently or within specific mass ranges (Guo and Huan Reference Guo and Huan2020; Wang et al. Reference Wang, Yin and Zhu2019a). A previous study systematically evaluated the advantages and disadvantages of targeted and nontargeted metabolomics approach (Lelli et al. Reference Lelli, Belardo, Timperio and Xianquan2021). It was due to the lack of standard pure compounds these days; most metabolites cannot be detected using targeted metabolomics. However, unlike targeted metabolomics, nontargeted approaches present the opportunity to discover new biomarkers, albeit with potentially lower robust accuracy due to the risk of false identification of metabolites or bias/signal drift induced by matrix effects.
Targeted metabolomics focuses only on a preselected set of known metabolites, typically ranging from tens to hundreds. This limited scope can be a drawback in food safety monitoring, where unexpected contaminants or novel metabolites may need to be detected. For example, in cases of food adulteration or contamination with unknown toxins, targeted metabolomics may miss important compounds simply because they are not part of the target list (Sarmad et al. Reference Sarmad, Viant and Dunn2023). Targeted approaches require standard pure compounds for calibration and identification, which may not always be available, especially for newly emerging contaminants. In food safety, this poses a significant challenge, as many potential contaminants or toxins may lack commercially available standards (Lelli et al. Reference Lelli, Belardo, Timperio and Xianquan2021). The inability of targeted metabolomics to identify unknown or emerging contaminants limits its effectiveness in detecting unexpected food safety issues, such as novel pesticide residues or chemical contaminants that may appear due to changes in farming or food processing practices.
Untargeted metabolomics has lower accuracy compared to targeted approaches because of the risk of false identification (Cajka and Fiehn Reference Cajka and Fiehn2016). In food safety monitoring, this lack of robustness can lead to misidentification or the inclusion of false positives, which can complicate or hinder regulatory decision-making (McGrath et al. Reference McGrath, Haughey and Patterson2018). Untargeted metabolomics generates large datasets, and interpreting these data can be challenging and time-consuming. Food safety monitoring often requires rapid response, but the complexity of data analysis in untargeted studies can delay actionable outcomes. In untargeted approaches, matrix effects and signal drift can significantly affect the results, making it difficult to achieve consistent quantification across different sample matrices (Watrous et al. Reference Watrous, Henglin and Claggett2017). This inconsistency is problematic in food safety monitoring, where reliable, quantitative information is essential for determining the level of risk posed by a contaminant. Unlike targeted metabolomics, untargeted approaches are less effective at accurate quantification of detected metabolites, particularly when the concentrations of specific contaminants are low. For food safety monitoring purposes, precise quantification is crucial for determining whether contaminant levels exceed regulatory thresholds.
Targeted metabolomics is always used for the detection of preselected set of known metabolites, typically ranging from tens to hundreds. However, in food safety monitoring, the unexpected contaminants or novel metabolites may need to be detected (Sarmad et al. Reference Sarmad, Viant and Dunn2023) as many potential contaminants or toxins may lack commercially available standards (Lelli et al. Reference Lelli, Belardo, Timperio and Xianquan2021). In the meantime, untargeted metabolomics has much more metabolites than can be detected but some of them with lower accuracy compared to targeted approaches because of the risk of false identification (Cajka and Fiehn Reference Cajka and Fiehn2016). In untargeted approaches, matrix effects and signal drift can significantly affect the results, making it difficult to achieve consistent quantification across different sample matrices (Watrous et al. Reference Watrous, Henglin and Claggett2017). Thus, these two methods can be combined to mitigate the drawbacks.
Metabolome database
Nowadays, Human Metabolome Database (HMDB) (http://www.hmdb.ca/) (Wishart et al. Reference Wishart, Guo and Oler2022) and Bovine Metabolome Database (BMDB) (Foroutan et al. Reference Foroutan, Fitzsimmons and Mandal2020) represent two of the most comprehensive databases to work on metabolomics in ruminants biology. PubChem compounds of NCBI and KEGG COMPOUND also provide a reference metabolite in further details. An online database (http://www.lmdb.ca) includes data on the analytical platform(s), experimental conditions, field of research, and animal breed used in acquiring the metabolomic data (Table 3). To enhance consistency, concentrations of all metabolites with quantitative data were converted into a standardized unit, such as μM. Currently, five levels of confidence in identification have been established with the highest confidence of validated identification (Level 1), a putative identification (Level 2), preliminary identifications (Level 3), molecular formula candidates (Level 4), and de-convoluted experimental m/z features (Level 5) (Rocchetti and O’Callaghan Reference Rocchetti and O’Callaghan2021). Moreover, a review of livestock metabolomic studies encompassing cattle, sheep, goats, horses, and pigs detected and/or quantified a total of 1070 metabolites (Goldansaz et al. Reference Goldansaz, Guo and Sajed2017).
Table 3. Metabolome database used in metabolomics studies
Ruminant milk and dairy product
Classification and function of milk
Milk, an important biofluid of animals, is often called the “perfect food,” rich in key nutrients such as proteins (Albenzio et al. Reference Albenzio, Santillo and Ciliberti2016). Milk stands as one of the most extensively consumed beverages globally, with 927 million tons produced in 2023 (OECD/FAO 2024). Comprising primarily water (85–87%), bovine milk also contains fats (3.8–5.5%), proteins (2.9–3.5%), and carbohydrates (5%) at the macronutrient level. Additionally, it harbors various bioactive compounds such as vitamins, minerals, biogenic amines, organic acids, nucleotides, oligosaccharides, and immunoglobulins at the micronutrient level (Foroutan et al. Reference Foroutan, Guo and Vazquez-Fresno2019).
Global milk production and commercial milk products is dominated by five ruminant species with 81% of total milk production coming from cows, followed by buffaloes with 15%, 4% for goat, sheep and camel milk combined (OECD/FAO 2024). Dairy cow milk is the most prevalent choice among consumers, whereas sheep, goat, and camel milk are significantly rarer in the market (Ahamad et al. Reference Ahamad, Raish and Ahmad2017; Akhtar et al. Reference Akhtar, Samar and Shami2021; Caboni et al. Reference Caboni, Murgia and Porcu2019; Foroutan et al. Reference Foroutan, Guo and Vazquez-Fresno2019). Thus, we can refer to these as ruminant milk products. Bovine milk, a biofluid rich in nutrients and chemically intricate, encompasses a multitude of diverse components. The hydrolytic AA content was found to be greater in bovine colostrum compared to human colostrum, suggesting a need for further investigation into AA metabolomics and its implications for infant formula development (Liang et al. Reference Liang, Han and Zhao2018). A Web-accessible database called the Milk Composition Database (MCDB, http://www.mcdb.ca/) was constructed based on 2355 identified metabolites in bovine milk (Foroutan et al. Reference Foroutan, Guo and Vazquez-Fresno2019).
Factors affect milk metabolome
Milk and milk products are globally consumed and renowned for their nutritional richness. Consequently, safeguarding their nutritional quality and ensuring product safety have emerged as paramount concerns in ruminant food research. Metabolites serve as indicators of milk and milk products’ quality, encompassing aspects such as nutritional value, authenticity, and safety (Suh Reference Suh2022). Till now, several metabolites such as choline, citrate, valine, hippuric acid, 2-butanone, lactate and some FAs have been used as robust biomarkers for milk quality, traceability and safety studies (Zhu et al. Reference Zhu, Kebede and McComb2021b). Milk metabolome studied is not only a sign of milk quality but also a metabolic indicator of animal performance (Sun et al. Reference Sun, Shi and Wu2017, Reference Sun, Wang and Wang2015). It was commonly known that milk composition varies with the cattle breed (i.e., Holstein, Jersey, Brown Swiss, etc.), stage of lactation, level of parity, number of viable pregnancies, and processing after milk collection (Boudonck et al. Reference Boudonck, Mitchell and Wulff2009; Yang et al. Reference Yang, Zheng and Zhao2016). The fluctuations in milk metabolites are influenced by various factors such as dietary nutrition, genetics, dairy animal species, lactation stage, as well as external factors like season, geographic origin, disease, and processing and storage conditions (Fig. 1) (Goetsch et al. Reference Goetsch, Zeng and Gipson2011; Klein et al. Reference Klein, Almstetter and Schlamberger2010). Following metabolomic analyses, potential ruminant milk biomarkers indicative of these factors was summarized (Table 4).
Figure 1. Overview of ruminant milk and dairy product classification and factors influencing the milk metabolome. (Created in BioRender. Zhang, B. (2024) BioRender.com/s41o643).
Table 4. Summary of screened potential biomarkers for ruminant milk are shown when available
Feed factor
To enhance the health attributes of dairy products, dietary feed regulation has been made to increase their contents of unsaturated FA (UFA), especially that of omega-3 polyunsaturated fatty acids (n3-PUFA). Additionally, rumen protection techniques are applied to ensure that PUFAs are not degraded by rumen microbes, thereby validating the effects of dietary supplementation (Lanier and Corl Reference Lanier and Corl2015). Untargeted metabolomics was used to investigate that both feed-derived (such as phenolic metabolites) and animal-derived compounds (such as FAs) are potential biomarkers associated with dairy cows fed different feeding regimens (Rocchetti et al. Reference Rocchetti, Gallo and Nocetti2020). Validation of these biomarkers involves using standardized analytical methods such as ultra-high-performance liquid chromatography–high-resolution mass spectrometry (UHPLC-HRMS) to ensure reproducibility and accuracy (Rocchetti et al. Reference Rocchetti, Gallo and Nocetti2020). Alternatively, grazing has been found to increase the content of n3-PUFA in goat milk. Controlled feeding comparisons and the application of consistent analytical platforms such as LC-MS have been used to validate these observations, ensuring the robustness of biomarkers like n3-PUFA (Argov-Argaman et al. Reference Argov-Argaman, Glasser and Muklada2021). Milk sourced from animals grazing on pasture is commonly (Hadaya et al. Reference Hadaya, Landau and Glasser2020), albeit not always, considered healthier. Biomarker validation here involves repeated studies to confirm the effect of grazing on milk FA content, with standardization achieved by using consistent sampling and analysis methods, such as GC-MS and LC-MS. In addition, many plants or plant extracts were shown to potentially change the milk flavor or taste by upregulate or downregulate milk-specific metabolites, such as essential oils (Sundekilde et al. Reference Sundekilde, Clausen, Lejonklev, Francesco, Luca and Peter2015), purple perilla leaf (Wang et al. Reference Wang, Sun and Tu2021a), berry extracts (Prestel et al. Reference Prestel, de Falco and Blidi2020), and bamboo leaf extract (Zhan et al. Reference Zhan, Shen and Lin2021). For instance, using LC-MS/MS, supplementation of Perilla frutescens leaf could potentially modify the milk metabolome with elevated levels of oleanolic acid, nucleotides, PE-NMe (18:1(9Z)/18:1(9Z)), and DG (18:0/20:4(5Z,8Z,11Z,14Z)/0:0), influencing pathways like pyrimidine metabolism and the biosynthesis of UFA in dairy cows (Wang et al. Reference Wang, Sun and Tu2021a). The validation of these biomarkers is performed through controlled supplementation trials, followed by standardized metabolomic analysis using LC-MS/MS to ensure reproducibility. Organic bovine milk exhibits a significantly higher concentration of beneficial FAs, including conjugated (9-cis,11-trans)18:2 linoleic acid (CLA), α-linolenic acid, linoleic acid, and total UFA, alongside a reduction in caproleic acid levels. The validation process for these biomarkers includes multivariate analysis, such as principal component analysis, partial least squares discriminant analysis, and receiver operating characteristic analysis, applied to NMR metabolomics data, which allows for consistent differentiation between organic and conventional milk profiles (Tsiafoulis et al. Reference Tsiafoulis, Papaemmanouil and Alivertis2019). These distinctions were revealed through the application of 1H-NMR and 1D TOCSY NMR techniques, highlighting the impact of agricultural practices on the nutritional composition of milk.
Animal species and varieties
The dairy animals (cows, buffaloes, sheep, goats, camel) have their respective share (81% from cow, 15% from buffaloes and the rest of 4% from goat, sheep and camel) in overall milk yield of the world and an increase has been seen over years (Bittante et al. Reference Bittante, Amalfitano and Bergamaschi2022). Bovine milk products are the main predominantly probiotic carrier in dairy foods (Ranadheera et al. Reference Ranadheera, Vidanarachchi and Rocha2017). The consumption of milk and dairy products obtained from sheep and goat is expected to increase by 26% and 53% respectively until 2030 (Bittante et al. Reference Bittante, Amalfitano and Bergamaschi2022; Pulina et al. Reference Pulina, Milan and Lavin2018). By integrating NMR with chemometrics, researchers identified 10 metabolites – carnitine, N-acetylcarbohydrates, acetate, choline, ethanolamine, citrate, creatine, lecithin, d-lactose, and d-sucrose – as reliable markers for detecting milk adulteration. This approach enhances the ability to safeguard the integrity and authenticity of milk product (Li et al. Reference Li, Yu and Zhu2017a). Yak milk has a richer composition than Holstein milk, with higher levels of fat, protein, solids-not-fat, and calcium. Metabolomics has been applied to explore these differences, demonstrating yak milk’s potential for specialized dairy products like cheese (Zhang et al. Reference Zhang, Yang and Cai2020b). Traditional fermented yak milk, produced by Tibetan herders in Gannan, contains a diverse microbial community, including Streptococcus salivarius, Lactobacillus helveticus, and Kluyveromyces marxianus. Metabolomics helps link these microbes to flavor compounds, with O2PLS analysis identifying key bacterial and fungal genera contributing to flavor. This insight supports the development of traditional fermented yak milk products with enhanced flavor profiles (Li et al. Reference Li, Yang and Tang2024a). Yak milk from the Tibetan Plateau is known for its high nutritional value, containing high protein, fat, lactose, and bioactive components than cow milk. Metabolomics has identified essential AAs, UFAs, and bioactive peptides, which contribute to antioxidant and immune-boosting effects. These qualities make yak milk ideal for functional dairy products that promote health benefits such as anti-fatigue and hypoxia resistance, particularly useful in high-altitude regions (Li et al. Reference Li, Liu and Han2023). Recent metabolomic research has highlighted the unique nutritional profiles of different animal species and varieties of milk. Yak colostrum contains high levels of inositol, glycine, and carnitine, along with a favorable essential AAs to total AA ratio, contributing to its superior nutritional value (Zhang et al. Reference Zhang, Li and Huang2024b). Yak milk also has elevated levels of creatine, lipoprotein lipase, and specific bioactive proteins that reflect its adaptation to high-altitude environments, making it particularly rich in health-promoting compounds (Li et al. Reference Li, Wang and Zhang2024b). Buffalo colostrum is rich in primary bile acids and bioactive peptides, enhancing its medicinal properties, such as antihypertensive, antioxidant, and anti-inflammatory effects, while supporting the survival of probiotic bacteria in fermented dairy products (Li et al. Reference Li, Wang and Zhang2024b; Zhang et al. Reference Zhang, Li and Huang2024b). In comparison, cow milk is characterized by higher concentrations of iminostilbene and osteopontin, which support bone health and immune function, though it has lower concentrations of functional components compared to yak and buffalo milk (Li et al. Reference Li, Wang and Zhang2024b). These insights emphasize the distinct health benefits of each milk type, offering potential for developing specialized dairy products that cater to various health needs, including infant nutrition and functional foods.
The production of various milk and dairy products from goat (Capra hircus) and sheep (Ovis aries) milk is on the rise, though their global market share remains significantly smaller in comparison to cow and buffalo milk. In certain regions, the issue of a “goaty” or “mutton” taste in milk products may arise, affecting consumer preference and acceptance (Pandya and Ghodke Reference Pandya and Ghodke2007; Watkins et al. Reference Watkins, Jaborek and Teng2021). However, goat and sheep milk and related dairy products have gained increasing attention from both consumers and the industry due to their superior digestibility and higher concentrations of bioactive substances compared to cow or human milk. Studies on goat milk, particularly involving infant formula, were based on populations that included 62 infants in a randomized controlled trial, 200 infants in a 12-month study, and 79 infants in another trial, demonstrating the nutritional adequacy of goat milk formula compared to cow milk formula (Prosser Reference Prosser2021). Sheep milk studies included various experimental models, including in vivo trials using Wistar rats to evaluate the health benefits of fermented sheep milk products, highlighting its anticancer, gastrointestinal health, and other health-promoting properties (Flis and Molik Reference Flis and Molik2021; Prosser Reference Prosser2021). Research has identified that valine and glycine are unique to goat milk, whereas talose and malic acid are distinctive markers of cow milk (Scano et al. Reference Scano, Murgia and Pirisi2014). For instance, the goat milk not only has more digestible proteins and fats but contains higher contents of short- and medium-chain FAs (MCFA), UFA, n6 FA, n3 FA, eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) than cow milk, which results in increasing particular interest by consumers, especially to infants and elder people (Li et al. Reference Li, Zhao and Zhu2017b). In the comparison between sheep and goat milk, through a GC-MS-based metabolomics approach, it was discovered that arabitol, citric acid, α-ketoglutaric acid, glyceric acid, myo-inositol, and glycine are predominantly found in sheep’s milk. Conversely, goat’s milk exhibited elevated levels of mannose-6-phosphate, isomaltulose, valine, pyroglutamic acid, leucine, and fucose, highlighting the distinct metabolic profiles between these two types of milk (Caboni et al. Reference Caboni, Murgia and Porcu2019), but there is still a lack of whole picture regarding the goat and sheep milk metabolite profiles as the animal varieties (Caboni et al. Reference Caboni, Murgia and Porcu2019). Camel milk is celebrated for its nutritional richness, containing all essential nutrients alongside compounds that may possess anticarcinogenic, antihypertensive, antioxidant, hypoallergenic, and cholesterol-lowering properties, making it a unique and beneficial addition to the diet (Buchilina and Aryana Reference Buchilina and Aryana2021). Camel milk is a specific local food used in some countries and regions of Southeast Asian, Middle East, and African continent, showing with immunomodulatory effects and good for human health, and is easily digested and well tolerated by lactose intolerance people (Ahamad et al. Reference Ahamad, Raish and Ahmad2017; Al-Awadi and Srikumar Reference Al-Awadi and Srikumar2001). The bioaccessible fraction of fermented camel milk has been shown to possess enhanced biological functionality compared to that of fermented bovine milk, highlighting its superior nutritional and health-promoting qualities (Ayyash et al. Reference Ayyash, Abdalla and Alhammadi2021).
The lipidome and FA composition in human, bovine and caprine milk were analyzed and compared using UHPLC-QTOF-MS and GC–MS. Human milk is rich in triglycerides (TG) containing linoleic acid (C18:2), sphingomyelin (SM) containing ultra-long-chain FAs and phospholipids (PLs) containing arachidonic acid (ARA)/DHA/ dihomo-γ-linolenic acid (DGLA). Caprine milk is rich in PLs, including hexosylceramide (Hexcer), Hex2Cer, SM, ceramide (Cer), and phosphatidylcholine (PC). Bovine milk is rich in PC and CL. The detailed examination of the lipid profiles of Chinese human, bovine, and caprine milk contributed valuable insights that could assist in formulating infant nutrition that is more precisely tailored to meet the dietary needs of Chinese babies (Wang et al. Reference Wang, Li and Liu2020). These findings have significant real-world applications beyond infant formula production. Understanding the unique lipid and FA profiles in these different milk types allows for better nutritional interventions for specific populations, such as elderly individuals or those with particular dietary requirements (Mollica et al. Reference Mollica, Trinchese and Cimmino2021). By leveraging the specific bioactive components found in caprine and bovine milk, it is possible to develop specialized nutritional products that address deficiencies in essential FAs or PLs, enhance cognitive function or support immune health (Eggersdorfer et al. Reference Eggersdorfer, Berger and Calder2022). The identification of FA like DHA and ARA, which are crucial for neural and visual development, is particularly useful for designing targeted nutritional supplements and functional foods (Fan et al. Reference Fan, Wang and Szeto2024). Furthermore, these insights can be applied to improve the nutritional quality of dairy products aimed at supporting growth and development in young children, as well as enhancing the health benefits of dairy-based functional foods for adults.
Geographical origin
The NMR metabolomic technique has been validated as an effective method for determining the origin of authentic products, demonstrating its applicability across both narrowly defined geographic regions, such as the Mugello valley, and broader areas encompassing large-scale distribution networks (Rocchetti and O’Callaghan Reference Rocchetti and O’Callaghan2021; Tenori et al. Reference Tenori, Santucci and Meoni2018). Using an untargeted approach that combines UPLC-Q-Exactive Orbitrap MS with multivariate statistical analysis, researchers identified 38 lipid molecules as potential indicators for determining the geographical origins of goat milk, and 19 lipid molecules for discerning its lactation stages, showcasing the method’s precision in tracing the provenance and physiological status of goat milk (Liu et al. Reference Liu, Guo and Zhao2020).
Commercial processing
Metabolomics is useful to evaluate changes caused by food processing and can be seen as a crucial tool to support academia and industry on revealing the transformation of raw animal materials into ready-to-eat products (Utpott et al. Reference Utpott, Rodrigues and Rios2022). The manufacturing and processing of milk commercially is for pasteurized beverage, Ultra-high-temperature (UHT), evaporated milk, ice cream, butter, milk powder and cheese, whey protein concentrate, paneer, ghee, traditional milk products, even soaps, lotions, and sweets, besides the popular cheeses and yoghurt (Pandya and Ghodke Reference Pandya and Ghodke2007). Fermentation and thermal processing enhance the flavor and texture of dairy foods, making them more appealing (Dos Santos Rocha et al. Reference Dos Santos Rocha, Magnani and Ramos2022). Hydroxyglutaric acid is a biomarker and unique metabolite of pasteurized goat or cow milk (Scano et al. Reference Scano, Murgia and Pirisi2014). In recent years, there are 60 marker metabolites from three categories of peptides, lipids, and nucleic acids that were detected for distinguishing between UHT and reconstituted milk by UPLC-Q-TOF-MS (Tan et al. Reference Tan, Zhang and Su2021). Applying the same technique, researchers identified seven oxylipids – 9-hydroxydecanoic acid, 12-hydroxydodecanoic acid, 2-hydroxymyristic acid, 3-hydroxytetradecanoic acid, 5-hydroxyeicosatetraenoic acid, 3-hydroxyhexadecanoic acid, and 10-hydroxyoctadecanoic acid – as effective markers for differentiating UHT milk from raw and pasteurized varieties (Zhang et al. Reference Zhang, Li and Zheng2018). Furthermore, l-carnitine, succinate, and acetate were pinpointed as biomarkers to differentiate UHT and reconstituted milk, based on comparisons with standard NMR-spectra databases (Cui et al. Reference Cui, Zhu and Su2019). Post-pasteurization, the percentage of SM in milk saw an increase, whereas fermentation into yogurt did not affect its levels (Argov-Argaman et al. Reference Argov-Argaman, Glasser and Muklada2021). Following the fermentation of brown goat milk, there was a notable increase in the contents of organic acids, peptides, medium- and long-chain FAs, and heterocyclic compounds through a comprehensive approach that integrated lipidomics and metabolomics (Jia et al. Reference Jia, Liu and Shi2021b).
Cheese, a fermented dairy delight, hosts a variety of microbial communities that evolve over time and differ based on the cheese variety and the specific starter and adjunct cultures used in its production. The milk of all ruminant species can be used to make cheese, but to improve efficiency, cheese-making procedures need to be optimized to take into account the large differences in their coagulation, curd-firming, and syneresis properties (Bittante et al. Reference Bittante, Amalfitano and Bergamaschi2022). The deployment of metatranscriptomics, metaproteomics, and metabolomics – collectively referred to as “cheesomics” – utilizes a multi-omics approach to enhance our comprehension of cheese’s microbial makeup and predict cheese characteristics such as flavor, quality, texture, and safety, as well as uncovering bioactive metabolites that may impact human health (Afshari et al. Reference Afshari, Pillidge and Dias2020). However, untargeted metabolomic approach is the most commonly used method to detect the cheese quality. For instance, the chemical fingerprints distinguishing Protected Designation of Origin (PDO) Grana Padano cheeses from non-PDO “Grana-type” cheeses have been established using UHPLC/QTOF-MS. This analysis revealed that lipids (FAs and their derivatives, PLs and monoacylglycerols), AAs, and oligopeptides, along with plant-derived compounds, emerged as the markers with the highest potential for discrimination (Rocchetti et al. Reference Rocchetti, Lucini and Gallo2018). Utilizing a synergistic metabolomics and peptidomics strategy, researchers identified potential markers to detect counterfeit cheeses, particularly those with an excessive rind content (>18%). The compounds 2-hydroxyadenine and argininic acid, along with 5-hydroxyindole acetaldehyde, were pinpointed as the most effective indicators of rind inclusion. Furthermore, the medium-chain aldehyde 4-hydroperoxy-2-nonenal emerged as a common marker indicative of both the cheese’s ripening duration and abnormal rind inclusion (Rocchetti et al. Reference Rocchetti, Michelini and Pizzamiglio2021). During the Mongolian cheese storage, it was found that the bitter AAs, bitter peptide (Phe-Ile), and organic acids (sinapic acid, butyric acid) increased accompanied with the increased contents of short-chain FAs, 2-undecanone and ethyl esters, which increased the cheese unpleasant smell and decreased the overall acceptability (Zhang et al. Reference Zhang, Zheng and Zhou2022). Metabolomic analysis of the commercial processing of milk highlights several significant findings regarding fermentation and co-fermentation, as well as cheese production. co-fermentation with Bifidobacterium adolescentis and other probiotics improved the stability of probiotic fermented beverages compared to single-strain fermentation, enhancing levels of health-promoting metabolites such as gamma-aminobutyric acid and L-malic acid even after 30 days of storage (Guo et al. Reference Guo, Sun and Wu2024a). In cheese production, various fermentation methods, including the use of probiotics like Lactobacillus plantarum and Lactobacillus helveticus, have been shown to enhance the production of AAs, Fas, and other bioactive compounds, improving both the nutritional quality and sensory properties of cheese. These probiotics also influence key metabolic pathways related to flavor and bioactivity, leading to health-promoting effects such as anti-inflammatory benefits, thereby providing opportunities for the development of functional dairy products (Ma et al. Reference Ma, Liu and Jiao2024).Thus, metabolomics is useful for assessing the cheese quality through the changing of small molecule compounds.
Dairy cow physiological condition
Diseases such as mastitis, lameness, and rumen acidosis are very common in dairy cows, beef cattle, sheep, and goat, exerting huge economic loss. Heat stress and mastitis represent significant financial challenges within the dairy industry. The metabolomics approach is helpful for better understanding the pathobiology of these diseases (Xi et al. Reference Xi, Kwok and Wang2017; Zhu et al. Reference Zhu, Tang and Lu2021a; Zwierzchowski et al. Reference Zwierzchowski, Zhang and Mandal2020). The analysis of whole raw milk from Holstein dairy cows affected by lameness, utilizing direct inject/LC-MS and NMR techniques, has proven beneficial for identifying and potentially mitigating lameness-associated pathological processes (Zwierzchowski et al. Reference Zwierzchowski, Zhang and Mandal2020). Significant variations of metabolome were found between healthy and mastitis cows by a novel metabolomics technique based on LC-MS. Milk arginine and Leu-Leu were increased in both the clinical and subclinical mastitis groups compared to healthy cows, indicating these metabolites were the potential biomarkers (Xi et al. Reference Xi, Kwok and Wang2017). Furthermore, the enrichment of the tricarboxylic acid cycle and the biosynthesis pathways of phenylalanine, tyrosine, and tryptophan were identified to elucidate the mechanisms behind the variation in the metabolome of mastitic milk, employing an untargeted 1H-NMR approach (Zhu et al. Reference Zhu, Tang and Lu2021a). NMR-based metabolomics analysis showed an increase in lactate, butyrate, isoleucine, acetate, and β-hydroxybutyrate levels, while levels of inosine and fumarate decreased in milk exhibiting high somatic cell counts (Sundekilde et al. Reference Sundekilde, Poulsen and Larsen2013b).
Milk choline, phosphocholine, N-acetylcarbohydrates, lactate, and β-hydroxybutyrate have been identified as potential markers of inflammation, exhibiting varying patterns dependent on the ambient temperature (Salama et al. Reference Salama, Contreras-Jodar and Love2020). An integrative metabolomics investigation employing LC-MS and 1H NMR spectroscopy identified several biomarkers in milk – lactate, pyruvate, creatine, acetone, β-hydroxybutyrate, trimethylamine, oleic acid, linoleic acid, lysophosphatidylcholine 16:0, and PC – that effectively indicate heat stress (Tian et al. Reference Tian, Zheng and Wang2016). Changes in concentrations of individual milk metabolites (volatile metabolites, and nonvolatile metabolites) can be related to the ruminal CH4 production pathway (van Gastelen et al., 2018). Furthermore, multi-omics studies represent an important gap revealing in livestock research (Sun et al. Reference Sun, Zhou and Wang2020; Xue et al. Reference Xue, Sun and Wu2020, Reference Xue, Xie and Zhong2022). The combination of metabolomics and proteomics was always used in dairy milk characteristics (Lu et al. Reference Lu, Antunes Fernandes and Páez Cano2013, Reference Lu, Boeren and van Hooijdonk2015).
Ruminant meat metabolomes
Meat quality is generally focused on the phenotypes, including nutrient contents, meat color, tender, juiciness, etc., which may be dependent on the subjective preference by consumers and have limited information for the food characteristics. However, the meat metabolome could show the basic fingerprint of food and help consumers and health professionals make informed decisions (Fig. 2). Metabolomics, supported by bioinformatics, identifies biomarkers in muscle and meat that enhance animal production and meat quality, benefiting both producers and consumers (Muroya Reference Muroya2023). The breed of the animal, its diet, and even the specific part of the meat influence meat quality and its metabolites. Consequently, potential biomarkers in ruminant meat indicative of these variables have been systematically compiled (Table 5).
Figure 2. Metabolomic insights into ruminant meat quality, nutritional traits and authenticity assessment. (Created in BioRender. Zhang, B. (2024) BioRender.com/m65h597).
Table 5. Summary of screened potential biomarkers for ruminant meat are shown when available
The compounds such as N′-formylkynurenine, kynurenine, and kynurenic acid (all part of tryptophan metabolism) and the oxidative stress indicator dityrosine and 3-dehydroxycarnitine are mechanistically connected to pathways associated with red meat, distinguishing them from those linked to white meat (Rombouts et al. Reference Rombouts, Hemeryck and Van Hecke2017). Ruminant meat is typically red meat, high in saturated FA (SFA), and red meats from ruminants exhibit a higher SFA to UFA ratio compared to meat products derived from monogastric animals and fish (Vahmani et al. Reference Vahmani, Mapiye and Prieto2015). However, ruminant meat, especially for beef and mutton, is high in CLA. The cis-9, trans-11 isomer stands out as the primary dietary form of CLA present in products from ruminants. This particular isomer is generated through partial biohydrogenation of linoleic acid in the rumen or via endogenous synthesis within the tissues themselves (Webb Reference Webb and Chhabi2021). Boosting the CLA content in meat could enhance its nutritional and therapeutic benefits. To elevate CLA intake in the human diet, options include either consuming more ruminant-derived foods or increasing the CLA levels in milk and meat products (Dhiman et al. Reference Dhiman, Nam and Ure2005). On the negative outcome side, due to the rumen biohydrogenation, the beef and mutton are always showing lower UFA, especially of PUFA compared to the pork, chicken, and fishes. Nowadays, the enhancement of PUFA, especially for n3-PUFA in beef and mutton are mainly through supplementing of n3-PUFA enriched feed sources but with relative lower transformation rate, or using rumen protected methods to release the n3-PUFA in gut (Ebrahimi et al. Reference Ebrahimi, Rajion and Goh2014). Till now, differences in meat between ruminant animals and monogastric animals have been conducted by GC-MS based metabolomics, such as the comparison between beef and pork (Ueda et al. Reference Ueda, Iwamoto and Kato2019). However, this study did not use fresh muscle samples. Another limitation is that the animal species and sample number are not enough to show the meat metabolome differences using same analyzed metabolomics method. Future research could address these limitations by using fresh muscle samples to ensure that the metabolomic profiles accurately reflect the native biochemical state of the meat. Moreover, increasing the number of animal species and sample sizes would provide a more comprehensive understanding of metabolome variations. Standardizing the metabolomics protocols across different species would also enhance comparability and reproducibility, leading to more robust conclusions about ruminant and monogastric meat differences.
Furthermore, the mixing of high-price meat species with low-quality/-price meat can be seen in food market or catering industry (Akhtar et al. Reference Akhtar, Samar and Shami2021). The ruminant meat is commonly in higher price than other livestock meat; thus, metabolomics can be used in adulteration of beef or mutton. NMR-based metabolomics has emerged as an effective technique for identifying unique signatures (potential biomarkers) to distinguish meats from different sources. This method holds promise for quality control applications, offering a way to differentiate between meat types – an important factor for ensuring food safety and addressing public health concerns (Akhtar et al. Reference Akhtar, Samar and Shami2021).
Beef
Beef, a staple food in numerous countries, faces increasing scrutiny due to misinformation regarding beef products. Concerns include false claims about origin, species, and production methods, highlighting the need for accurate information and transparency in the beef industry (Man et al. Reference Man, Chan and Tang2021; Visciano and Schirone Reference Visciano and Schirone2021). Metabolomics has been utilized to explore potential biomarkers associated with meat quality traits, concentrating on factors like the genetic background of the animal, sensory characteristics, feeding systems, and formulations. This research also encompasses processes such as postmortem storage and hygiene management, aiming to comprehensively understand how these elements influence meat quality (Akhtar et al. Reference Akhtar, Samar and Shami2021). The volatile compounds like 2-methylfuran-3-thiol, 3-sulfanylpentan-2-one, furan-2-ylmethanethiol, 2-propylpyrazine, 1-furan-2-ylpropan-2-one, 1H-pyrrole, 2-methylthiophene, and 2-(furan-2-ylmethyldisulfanylmethyl)furan are potential key contributors to beef-related attributes and flavor in glutathione-Maillard reaction products (Lee et al. Reference Lee, Kwon and Kim2011). Both UPLC-Orbitrap-MS and GC–MS analytical platforms were used and the geographical origins and feeding regimes could be differentiated by the potential biomarkers including AAs, several sugar metabolites, and a number of PCs and phosphatidylethanolamines (PEs), which offers a method to identify the geographical origin of beef at any point along the supply chain and could be used to develop a verifiable traceability system (Man et al. Reference Man, Chan and Tang2021). Various AAs and succinate emerge as potential biomarkers for discriminating the geographical origin of beef sourced from four countries: Australia, Korea, New Zealand, and the United States (Jung et al. Reference Jung, Lee and Kwon2010). Utilizing NMR-based metabolomics with High Resolution Magic Angle Spinning (HR-MAS) enables the classification of meat samples based on their storage time (Castejón et al. Reference Castejón, García-Segura and Escudero2015). This study represents the first metabolomic investigation of Japanese Black cattle using GC/MS analysis, comparing metabolites across different muscle and intramuscular fat (intermuscular fat) in marbled beef. MCFA implicated in triacylglycerol synthesis were exclusively detected in fat tissue. Additionally, decanoic acid, uric acid, elaidic acid, and 3-phosphoglyceric acid emerged as potential biomarkers for IMF, indicative of marbling levels. Notably, decanoic acid and glutamine were identified as potential biomarkers associated with oily flavor, wagyu beef aroma, and overall sensory evaluations (Ueda et al. Reference Ueda, Iwamoto and Kato2019). Differences in enzymatic activities (lactate dehydrogenase, malate dehydrogenase, and succinate dehydrogenase), pyruvate contents, and the number of metabolites associated with the tricarboxylic acid cycle were observed between longissimus lumborum and psoas major muscles from Chinese Jinjiang yellow cattle (Yu et al. Reference Yu, Tian and Shao2019). Yak (Bos grunniens), a special beef cattle, is mainly located in the Himalayan highlands region (Qiu et al. Reference Qiu, Zhang and Ma2012), the longissimus thoracis from locally farmed Jiulong yaks had higher concentrations of carnosine and formate and lower concentrations of mannose, inosine, threonine, alanine, valine, isoleucine, tyrosine, phenylalanine, and leucine compared to biceps femoris by an untargeted NMR metabolomic approach (Zhu et al. Reference Zhu, Petracci and Li2020). Beef from grass-fed animals is characterized by tenderness, lower total fat content, a higher omega-3/omega-6 ratio, and superior protein content compared to grain-fed counterparts (Carrillo et al. Reference Carrillo, Bai and He2021, Reference Carrillo, He and Li2016). Metabolomics has been extensively used to understand the biochemical changes in beef during processing and storage, with particular emphasis on color stability and tenderness. The metabolomic investigation of fresh beef, lamb, and venison using NMR spectroscopy highlighted the metabolite changes affecting color stability in meat under retail display conditions. It was found that the stability of fresh beef color was greater than that of lamb and venison, with metabolites such as leucine, isoleucine, valine, succinate, inosine monophosphate, and choline playing key roles in these differences (Kanokruangrong et al. Reference Kanokruangrong, Kebede and Carne2024). In another study, LC-MS-based metabolomics revealed that superchilling treatment of beef accelerated the degradation of μ-calpain and caspase 3, leading to improved tenderness. The metabolomic profiles of superchilled beef were distinguished from other treatments, indicating changes in pathways such as arginine and proline metabolism, which are associated with tenderness (Chen et al. Reference Chen, Mao and Liang2024). Additionally, metabolomics has been employed to investigate the effects of chilling regimes on the metabolome of beef, showing that superchilling improved the tenderness of beef by affecting key metabolic pathways, including ATP and its degradation products, which were influenced during the early post-mortem period (Guo et al. Reference Guo, Zhang and Wang2024b).
On the other hand, the occurrence of dark cutting meat has a notable impact on meat quality attributes and concentrations of post-mortem glycolytic metabolites in Angus Í Nellore crossbreed cattle, which was associated with mitochondrial activity and energetic metabolic pathways (Cônsolo et al. Reference Cônsolo, Rosa and Barbosa2021). Significant differences in lactic acid, alanine, methionine, fumaric acid, inosine, inosine monophosphate, creatine, betaine, carnosine, and hypoxanthine were observed based on aging type (dry and wet aging) in the beef metabolome (Bischof et al. Reference Bischof, Witte and Terjung2021).
Lamb/mutton meat
Meat flavor, one of the most important sensory characteristics and main attributes that determines consumers’ decisions to purchase a meat, is attributed to some volatile compounds (Khan et al. Reference Khan, Jo and Tariq2015). In mutton, either from sheep or goat, a specific mutton dodur exists due to the volatile medium- and short-chain FAs such as 4-methyloctanoic acid, 4-ethyloctanoic acid and 4-methylnonanoic acid (Wang et al. Reference Wang, Gao and Wang2021c). It was revealed that significant variations in FAs, aldehydes, ketones, lactones, alkaloids, flavonoids, phenolics, and drug residues among three types of goat meat: Lubei white goats, Boer goats, and Jining grey goats. This underscores how untargeted LC-MS can elucidate the subtle differences in flavors and sensory attributes among these varieties (Wang et al. Reference Wang, Sun and Hu2019b). Among the lamb breeds studied, (E)-2-hexenal was exclusively detected in Tan lambs, whereas (E)-2-nonenal and (E, E)-2,4-nonadienal were only present in Dorper lambs. Hu lambs exhibited the fewest volatile compounds. Analysis demonstrated that Dorper lambs had a higher proportion of PUFA, AAs, and volatile compounds compared to Tan and Hu lambs. However, specific PUFA derivatives in Dorper lambs were found to have a negative influence on the odor profile (Zhang et al. Reference Zhang, Zhang and Liu2020a). Using lipidomics and targeted metabolomics, hypoxanthine metabolites and volatile compounds of 1-octen-3-ol and hexanal were significantly increased by castration, which might be beneficial in lamb quality (Li et al. Reference Li, Tang and Zhao2020). NMR spectroscopy was employed to delve into the confinement odor phenomenon in lamb meat, establishing correlations between this odor and specific meat and drip metabolites. These included tyramine, formate, alanine, carnosine, urea, proline, aspartate, glutathione, and nicotinate, which are substrates or products of glucose fermentation and AA catabolism (Cônsolo et al. Reference Cônsolo, Olivecrona and Samuelsson2020). The study on Mongolian sheep highlighted the dynamic changes in metabolites during early postmortem chilled aging, with significant alterations in AAs, fatty acyls, and glycerophospholipids, which are crucial for the flavor and quality of meat. Metabolites like AAs and small peptides accumulated significantly, enhancing flavor through pathways like AA metabolism and protein digestion (Zhang et al. Reference Zhang, Sun and Su2024a).
From the view of lipids composition, low IMF leads to the decrease of flavor precursors in lamb (Bravo-Lamas et al. Reference Bravo-Lamas, Barron and Farmer2018). Lipids such as PC, PE, and TG and their structures with SFA and UFA are critical to the thermal oxidative capacity of glycerol chain-based lipids (Wu and Wang Reference Wu and Wang2019). A validated UPLC-Q-Orbitrap HRMS combined lipid screening strategy method-based lipidomics was used for identification of Tan sheep meat products with different thermal processing methods. For atherosclerosis patients, the boiled cooking method proved to be preferable due to greater losses of SM compared to Cer in meat. Conversely, the steamed approach resulted in fewer losses of PC and lysophosphatidylcholine, indicating that Tan sheep meat prepared in this manner might be more suitable for elderly and infant populations (Jia et al. Reference Jia, Li and Wu2021a).
Volatile biomarkers, including alkanes, ketones, terpenes, and 2,3-octanedione, were found in ruminant tissues could distinguish exclusive pasture diets from exclusive concentrate diets (Sivadier et al. Reference Sivadier, Ratel and Engel2009). The feed regimes affect the metabolites found in longissimus dorsi muscle of sheep, showing that total cholesterol, esterified cholesterol, choline, glycerophosphocholine, and glycerophospholipids were significantly lower in calcium soap of palm FAs and prilled fat plus lecithin diets, while glycerol and SM were significantly higher in calcium soap of palm FAs and prilled fat plus lecithin diets (Behan et al. Reference Behan, Akhtar and Loh2021). An untargeted metabolomic and lipidomic method based on UPLC-Q-TOF/MS provided a basis for differentiation of meat from pasture-fed and concentrate-fed sheep/goats (Wang et al. Reference Wang, Xu and Zhang2021d). It was also found that the contents of 1- octen-3-ol and 2,3-octanedionone in mutton from pasture-fed animals were significantly higher (Wang et al. Reference Wang, Xu and Zhang2021d). Using both untargeted and targeted metabolomics, the main increased N-acetyl-l-aspartic acid, N-acetylaspartylglutamate, acetylcarnitine, and l-carnitine, but decreased carnosine and creatinine were the main newly found grazing feeding regime associated metabolites, which might contribute to the improved lamb meat quality under artificial pasture grazing (Wang et al. Reference Wang, Wang and Zuo2021b). Incorporating P. frutescens seeds into Tan lamb diets enhances n3-PUFA levels and flavor, marked by increased volatile compounds like acetaldehyde and 1,2,4-trimethyl-benzene. This inclusion boosts beneficial lipids and metabolites such as PG(18:1/18:1), PG(18:2/18:2), and 5′-inosinic acid, while reducing lysophosphatidylcholine, guanidinosuccinic acid, 3-hydroxydecanoic acid, and 2-methylbutyroylcarnitine (Yu et al. Reference Yu, Zhang and Jiang2024). It has been summarized previously that volatile alcohols, like 1-octen-3-ol, and ketones, such as 2-heptanone and 3-hydroxy-2-butanone, contribute to the distinctive aroma of lamb. Studies indicate that compounds like 1-octen-3-ol, which imparts a “mushroom-like” note, are often appreciated by consumers for enhancing the umami aspects of cooked meat, leading to increased consumer preference. The aldehyde content, including hexanal, nonanal and heptanal, contributes significantly to the characteristic mutton odor. Compounds like (E)-2-nonenal and (E)-2-octenal impart fresh, fatty and slightly green aroma notes, which are desirable for a well-rounded lamb flavor, and the 4-methyloctanoic acid and 4-ethyloctanoic acid, which are responsible for the specific “mutton” flavor (Khan et al. Reference Khan, Jo and Tariq2015). Volatile compounds like phenylacetaldehyde and methional have been noted to provide antioxidant properties, which help in reducing oxidative stress. These antioxidants are significant for maintaining the overall nutritional quality of lamb meat and may contribute to the health benefits of reducing the risks associated with high-fat diets. Castration in lambs has been found to alter the concentration of volatile compounds, like hexanal and 1-octen-3-ol, reducing off-odors and enhancing the flavor profile, which might positively impact consumer preferences while also increasing the levels of beneficial lipids (Li et al. Reference Li, Tang and Zhao2020). Such dietary adjustments suggest significant metabolomic advantages and potential for healthier lamb meat production, as analyzed by HD-mix LC-MS/MS for lipid and metabolite profiling.
Metabolomics for ruminant food alternatives
Synthetic/artificial milk
With the gradual increase in world milk consumption and concerns about the safe production of ruminant products under modern intensive farming conditions, some problems caused by efficient intensive farming, such as animal health, environmental pollution, animal welfare, etc. are closely related to “One Health” concept contradicts (Hu et al. Reference Hu, Cheng and Tao2017). In this context, the new technology of “lab-grown food” “replace an extremely inefficient, input-intensive and waste-producing traditional agricultural sector with precise, targeted and easy-to-process systems” production system by using small land and drastically reduces water and nutrient requirements (Monbiot Reference Monbiot2020). To produce ruminant food, the most direct way to think of is to use controllable artificial conditions and the technical basis of laboratory cultivation to synthesize meat, eggs and milk. Synthetic milk has emerged as a new potential alternative to cow’s milk, unlike plant-based oat, nut, and soy milks, which are designed to replicate its taste, appearance, and mouthfeel. Synthetic milk may have a smaller carbon footprint than dairy production, cause less pollution, and apparently eliminate animal welfare concerns (George Reference George2023). Here we need enough information to understand the composition of naturally produced or real milk to make synthetic milk closer to natural milk. Metabolomics as a powerful tool can play a crucial role in understanding the real composition of milk for lab-grown milk production. The application of metabolomics has strong advantages in understanding the chemical characteristics and dynamics of dairy products, as well as potential biomarkers for differentiating the consumption of different dairy products and identifying milk quality, traceability and safety (Goldansaz et al. Reference Goldansaz, Guo and Sajed2017; Zhu et al. Reference Zhu, Kebede and McComb2021b). Applications of NMR-based metabolomics in milk research include linking milk metabolite analysis to nutritional aspects and technical quality of milk. Identification of novel metabolites through metabolomics as biomarkers or bioactive compounds (Sundekilde et al. Reference Sundekilde, Larsen and Bertram2013a). The application of metabolomics can provide a detailed understanding of milk composition, reveal changes in metabolite levels related to lactation, breed, diet, and other factors, and provide more valuable information to produce synthetic milk.
Ruminant meat alternatives
A new wave of plant-based or cell-based meat alternatives, designed to replicate the taste and nutritional profile of red meat, has garnered significant consumer interest, research focus, and media attention. Plant-based diets exclude or substantially limit the consumption of meat and animal products and are of growing interest to many due to their sustainability and health benefits (Rubio et al. Reference Rubio, Xiang and Kaplan2020; Wang et al. Reference Wang, Wang and Zuo2021b). Nowadays, the plant-based meat is an artificial processed meat that are in evolution with the developed understanding and upgrading knowledge of meat metabolites and structure (Allen and Locasale Reference Allen and Locasale2021). The clearer revealing of the meat metabolome contributes to that the artificial meat is more nearly taste and flavor close to the real animal feeding meat (Fig. 3) (Allen and Locasale Reference Allen and Locasale2021; Munekata et al. Reference Munekata, Pateiro and Rocchetti2022; van Vliet et al. Reference van Vliet, Bain and Muehlbauer2021).
Figure 3. Metabolomic strategies for the development and Quality Control of Ruminant Food Alternatives. (Created in BioRender. Zhang, B. (2024) BioRender.com/l16r397).
The metabolomics analysis found that metabolite abundances between the plant-based meat alternative and grass-fed ground beef differed by 90%. Nutrients such as DHA, niacinamide (vitamin B3), glucosamine, hydroxyproline, and antioxidants including allantoin, anserine, cysteamine, spermine and squalene was exclusively present in beef. Conversely, ascorbate (vitamin C), phytosterols, and several phenolic antioxidants such as loganin, sulfurol, syringic acid, tyrosol, and vanillic acid were only found in the plant-based meat alternatives (van Vliet et al. Reference van Vliet, Bain and Muehlbauer2021). For the future study of ruminant meat, metabolomics in targeting animal blood and meat can be used as a noninvasive technology to prediction of meat quality (Muroya Reference Muroya2023).
Spatial metabolomics and future artificial meat
The development of plant-based and cell-cultured meat has accelerated in recent years due to the advancements in metabolomics, 3D printing, and cell culture technologies. Specifically, 3D printing techniques play a crucial role in the production of ruminant meat alternatives (K. Handral et al. Reference K. Handral, Hua Tay and Wan Chan2022; Rubio et al. Reference Rubio, Xiang and Kaplan2020). To bridge the gap between traditional meat and artificial meat, spatial metabolomics provides detailed information about the localization and chemistry of small molecules in individual cells (Taylor et al. Reference Taylor, Lukowski and Anderton2021), tissues (Bourceau et al. Reference Bourceau, Michellod and Geier2022), and the host–microbe interface (Geier et al. Reference Geier, Sogin and Michellod2020). For instance, spatial metabolomics has been used to map PL distributions in the tissues of marine bivalves (Bourceau et al. Reference Bourceau, Michellod and Geier2022), revealing insights that can also be applied to understand the deposition of intramuscular fat and specific lipids in beef and mutton. Furthermore, imaging MS allows spatial metabolomics to identify the localization of limited content compounds in plants (Nakabayashi et al. Reference Nakabayashi, Hashimoto and Mori2021). Thus, the use of spatial metabolomics offers a potential bridge between traditional and synthetic food production by providing a detailed map of the food metabolites.
Ruminant food metabolites on human health
As is well known, ruminant has a unique rumen that contributes to a large number greenhouse gas production, which brings increasing attention due to its threat to the world climate change and human being living safety (McMichael et al. Reference McMichael, Powles and Butler2007). Ruminant animal-based foods not only have high social and climate costs but have strong effects on human health. For instance, the CLA that is a special ruminant FA has beneficial effects on human health (Badawy et al. Reference Badawy, Liu and Guo2023; Mcguire and Mcguire Reference Mcguire and Mcguire2000). In our previous study, we found that indoxyl sulfate in lamb meat was increased by feeding high energy diet but flavory AAs such as l-glutamine, l-serine, l-glutamate, and oleic acid were decreased (Wang et al. Reference Wang, Zhao and Zhang2022). The l-glutamine, l-serine, l-glutamate and oleic acid were good taste or health beneficial (Sales-Campos et al. Reference Sales-Campos, Souza and Peghini2013) but sulfate compounds were potentially harmful for human health and easily result in human liver and kidney diseases (Vanholder et al. Reference Vanholder, Schepers and Pletinck2014). One Health and Global Health are based on the idea that human health and animal health are interdependent as well as being linked to the health of the ecosystems of which they are part. Thus, ruminant production is a key procedure in One Health that is the collaborative effort of multiple health science professions to attain optimal health for people (Hernando-Amado et al. Reference Hernando-Amado, Coque and Baquero2019).
Furthermore, antibiotics are a class of secondary metabolites mainly produced by microorganisms. The abuse of antibiotics in livestock would easily result in antimicrobial resistance, and the antibiotic residues in animal derived food would accelerate the spread of resistant bacteria within and between these sectors and around the globe (Hernando-Amado et al. Reference Hernando-Amado, Coque and Baquero2019). The application of different analytical techniques for metabolomics have been successfully used in environmental, food or health sciences with various advantages and drawbacks (Fraga-Corral et al. Reference Fraga-Corral, Carpena and Garcia-Oliveira2022). Living systems encompass microorganisms, plants, animals, and humans, with food systems directly impacting nutrition and human health, which is from farm to human (Kim et al. Reference Kim, Kim and Yun2016). The ruminant food metabolomics are helping to interpret and connect the front-end animal production to back-end human life and open a window toward a better understanding of the complex interactions between food and human health (Scalbert et al. Reference Scalbert, Brennan and Manach2014). However, very limited research did the traceability and metabolites flow from planting to animal food, and then to human health.
The emergence of metabolomics enables the discovery of biomarkers that enhance the deposition of functional substances during nutritional interventions, thereby improving meat quality (Zhong et al. Reference Zhong, Wei and Li2022). Simultaneously, it facilitates the detection and regulation of harmful compounds, ensuring the safety of food products (Lu et al. Reference Lu, Antunes Fernandes and Páez Cano2013; Ye et al. Reference Ye, Yang and Xiao2023). By constructing a comprehensive metabolite spectrum, metabolomics surpasses traditional methods in capturing the complexity of food quality and safety. Traditional approaches to evaluating food quality and safety often focus on individual components, such as protein, fat or specific contaminants. While effective, these methods fail to address the intricate interplay of metabolites that collectively define food quality and safety (Meijer et al. Reference Meijer, Lähteenmäki and Stadler2021). In contrast, metabolomics integrates a broad array of metabolites, including AAs, FAs, organic acids, vitamins, and bioactive compounds. This holistic approach provides a more nuanced and accurate representation of food quality and safety (Liu et al. Reference Liu, Zhao and Yin2023). By aligning with the goals of modern food science, metabolomics ensures the production of high-quality, safe, and trustworthy food products for consumers. However, as metabolomics databases for ruminant-derived foods continue to expand, this field holds the potential for even greater precision and comprehensiveness in identifying and regulating food quality and safety.
Conclusions
Metabolomics has been optimally and effectively utilized in the field of ruminant food science and nutrition research, facilitating the identification of crucial metabolites including AAs, n3-PUFA, and various organic acids. These metabolites serve either as precursors or as critical determinants in assessing food quality, flavor, and taste, which in turn influence consumer preferences (Fig. 4). The development of instrumentation would make metabolite detection and quantification more sensitive, accurate, robust, automated, and comprehensive, while promising advances need to improve and replenish the relevant metabolome. The most difficulties for metabolomics analysis are potentially the new metabolites or compounds identification, and the internal relationship among these metabolites in the food generation. For many of the small molecules, such as AAs and FAs, are part of the macronutrients such as protein and fat. Many more reference spectral or chromatographic databases on food components need to be developed and the name of metabolites should be uniform with chemical name and trivial name. A specific database related to the nutritional and ruminant food relevant compounds can be routinely identified or quantified.
Figure 4. The whole workflow and key points in determining ruminant meat and milk production. (Created in BioRender. Zhang, B. (2024) BioRender.com/k75h157).
Acknowledgements
This work was supported by Hainan Provincial Natural Science Foundation of China (324QN289), the Special Youth Scientist Fund of Bayannur Research Institute, China Agricultural University (2024BYNECAU002), and the Open Subject from Key Laboratory of Animal Molecular Nutrition of Ministry of Education (KLMAN202206).
Author contributions
B.Y.Z: Investigation, Writing – original draft. J.K.W: Writing – review & editing. B.W: Conceptualization, Investigation, Supervision, Writing – review & editing.
Competing interests
The authors declare none.
Open access
