Saanen dairy goats are widely recognized as one of the most productive dairy goat breeds and contribute substantially to milk production systems at both national and global levels (Irawan et al., 2026). In intensive dairy goat operations, optimizing milk yield while maintaining metabolic stability is essential for ensuring production efficiency, economic sustainability, and animal welfare (Pardo et al., 2022). However, considerable inter-individual variation in milk production is frequently observed even among animals maintained under identical feeding and management conditions (Dosseh et al., 2025). Such variability suggests that productive performance is not solely determined by nutrient supply but is also influenced by differences in metabolic efficiency and physiological adaptability.

Lactation is an energy-demanding process that requires coordinated regulation of nutrient partitioning, endocrine signaling, and systemic metabolic adaptation (J, 2022). During high milk synthesis, glucose, fatty acids, and amino acids are preferentially allocated to mammary gland activity, leading to dynamic changes in whole-body energy balance (Irawan et al., 2026). Circulating non-esterified fatty acids (NEFA) and beta-hydroxybutyrate (BHBA) are commonly used indicators of lipid mobilization and ketogenesis, whereas insulin, cortisol, and insulin-like growth factor-1 (IGF-1) reflect key aspects of glucose utilization, stress response, anabolic signaling, and mammary function (Zamuner et al., 2020); (Caputo et al., 2021); (Zheng et al., 2025). Together, these metabolic and endocrine indicators provide useful insight into lactational energy status, although they may not fully capture the complexity of biochemical networks underlying individual differences in milk yield. Conventional assessment of nutritional status in dairy goats primarily relies on production records, body condition scoring, and selected biochemical indicators (Andjelić et al., 2022). While these approaches are valuable for herd-level management, they offer limited insight into systemic metabolic coordination at the individual-animal level (Ghavipanje et al., 2021). Increasing evidence suggests that productive variability is closely related to differences in metabolic resilience, nutrient-use efficiency, and pathway-level biochemical regulation (Islamiyati et al., 2025). Therefore, integrative molecular approaches are required to characterize better metabolic phenotypes associated with variations in milk production.

Metabolomics has emerged as a powerful analytical platform for profiling low-molecular-weight metabolites that represent the downstream products of gene expression and enzymatic activity (Yusuf et al., 2026). By providing a comprehensive snapshot of metabolic processes, metabolomic analysis enables the identification of pathway-level alterations associated with physiological adaptation and production efficiency. In ruminant research, metabolomic profiling has been increasingly applied to investigate energy balance, metabolic disorders, and lactational physiology (Irawan et al., 2026). Alterations in amino acid metabolism, lipid pathways, carbohydrate utilization, and oxidative stress-related metabolites have been implicated in variability in productive performance, highlighting the potential of metabolomics within precision nutrition frameworks (Carlino et al., 2025).

Despite growing interest in precision livestock management and metabolomic technologies, comparative characterization of systemic metabolic signatures between high- and low-producing dairy goats maintained under uniform management conditions remains limited (Islamiyati et al., 2025). In the present context, precise nutritional status is considered an integrative metabolic phenotype inferred from production performance, energy metabolic biomarkers, endocrine indicators, and serum metabolomic profiles, rather than a direct measurement of individual feed or nutrient intake. Identifying distinct metabolic and endocrine patterns associated with productive differences may enhance understanding of nutrient utilization efficiency and support the development of biomarker-based strategies for metabolite-informed nutritional evaluation.

Therefore, the present study aimed to evaluate and compare production-associated metabolic and metabolomic phenotypes in lactating Saanen dairy goats classified by milk production level. By integrating milk production performance, energy metabolic biomarkers (NEFA and BHBA), hormonal profiles (insulin, cortisol, and IGF-1), and LC–HRMS-based untargeted serum metabolomics, this study sought to identify differential serum metabolites and associated metabolic pathways related to variations in milk yield. The findings are expected to provide associative biological insights into metabolic efficiency and support the development of metabolite-informed nutritional evaluation strategies in dairy goats.

This study demonstrates that variation in milk yield among lactating Saanen dairy goats maintained under similar management conditions was associated with coordinated differences in energy metabolism, endocrine response, and serum metabolomic profiles. Although more than half of the annotated serum metabolites were shared between HP and LP goats, multivariate analysis, differential metabolite testing, and pathway enrichment consistently indicated distinct serum metabolic patterns between production groups (Islamiyati et al., 2025). The main pathways involved included tryptophan-related immunometabolic signaling, one-carbon metabolism, riboflavin-linked redox regulation, carbohydrate-related metabolism, and organic acid metabolism (Lima et al., 2022); (Daddam et al., 2025). These findings suggest that differences in milk production level were associated with systemic metabolic variation rather than isolated changes in individual metabolites (Wang et al., 2024).

A central physiological difference between the production groups was observed in the energy metabolic profile. LP goats exhibited markedly higher NEFA and BHBA concentrations, indicating greater lipid mobilization and ketogenesis, which are commonly associated with a more pronounced negative energy balance during lactation (Ghavipanje et al., 2021); (Lisuzzo et al., 2022). In contrast, HP goats showed lower NEFA and BHBA despite producing more milk, suggesting that higher milk output was associated with a less pronounced lipid mobilization and ketogenesis response. This finding indicates a more stable energy metabolic profile in HP goats than in LP goats and is consistent with previous reports linking metabolic stability and oxidative balance to production performance in dairy animals (Meli et al., 2025); (Ponnampalam et al., 2024).

Endocrine profiling showed that LP goats had significantly higher cortisol concentrations, whereas IGF-1 and insulin did not differ significantly between groups. This finding indicates that the difference between HP and LP goats was more evident in stress-associated endocrine response than in anabolic hormone status. Cortisol is closely related to hypothalamic–pituitary–adrenal axis activity and may influence glucose metabolism, lipid mobilization, and stress adaptation (Knezevic et al., 2023); (Ünal & Uztimür, 2025). The detection of cortisone as an LP-associated discriminatory metabolite further supports the involvement of glucocorticoid-related metabolism in the LP phenotype. In addition, the positive correlation between cortisone and BHBA in LP goats suggests an association between glucocorticoid-related metabolism and ketone body production, although this relationship should be interpreted as correlative rather than causal (Zamuner et al., 2020).

At the pathway level, LP goats showed higher kynurenine and L-formylkynurenine, indicating that tryptophan metabolism differed between production groups. Kynurenine pathway activity has been associated with immune activation, oxidative stress adaptation, and endocrine–metabolic interactions (Tsuji et al., 2023); (Sorboni et al., 2022)). In the present study, the increase in kynurenine-related metabolites in LP goats was accompanied by higher cortisol, NEFA, and BHBA concentrations, suggesting that this pathway may be associated with the stress–energy metabolic profile observed in the LP group. The positive correlation between kynurenine and IGF-1 in LP goats further suggests a possible link between tryptophan metabolism and endocrine regulation, although this association requires further validation (Salama et al., 2021); (Chen et al., 2025).

LP goats also exhibited higher 1-methylhistidine, which may reflect increased protein turnover or dietary muscle-derived compounds. Under conditions of increased energy demand or metabolic challenge, protein mobilization may contribute to the supply of energy and glucose precursors in lactating animals (Siachos et al., 2024). In this study, 1-methylhistidine was negatively correlated with IGF-1 and insulin in LP goats, suggesting an association between this metabolite and reduced anabolic endocrine signaling. Together with higher NEFA, BHBA, and cortisol, this finding supports the interpretation that LP goats had a more catabolic metabolic profile than HP goats (Li et al., 2026).

In contrast to LP goats, HP goats were characterized by higher UDP-glucose, S-adenosylmethionine, betaine, and riboflavin. UDP-glucose is related to carbohydrate activation and glycosylation pathways, and carbohydrate utilization is particularly important during lactation because glucose supports lactose synthesis and milk volume regulation (Lin et al., 2016). In the present study, higher UDP-glucose in HP goats, together with lower NEFA and BHBA concentrations, suggests that the HP phenotype was associated with a more stable systemic energy profile and less dependence on lipid mobilization (Zamuner et al., 2020).

HP goats also showed higher S-adenosylmethionine and betaine, both of which are involved in one-carbon metabolism and methyl-group transfer reactions (Jiang et al., 2023). One-carbon metabolism is linked with methylation processes, hepatic metabolic function, antioxidant defense, and metabolic adaptation (Clare et al., 2019). In this study, higher S-adenosylmethionine and betaine in HP goats may indicate differences in methyl-donor-related metabolism between production groups. The positive correlation between S-adenosylmethionine and BHBA in the HP group may also suggest an adaptive association between methyl-donor metabolism and lipid-derived energy use, although this interpretation should be approached with caution, given the limited sample size (Ghavipanje et al., 2021).

Riboflavin emerged as another prominent HP-associated metabolite with high pathway relevance and discriminative performance. Riboflavin is a precursor of flavin cofactors involved in mitochondrial oxidative metabolism, fatty acid oxidation, and redox balance (Hampel et al., 2016); (Ponnampalam et al., 2024). In the present study, higher riboflavin levels in HP goats and their positive correlation with milk production suggest that riboflavin metabolism is associated with the high-production phenotype. This finding is consistent with the lower NEFA, BHBA, and cortisol levels observed in HP goats and may reflect a more favorable metabolic condition for sustaining milk production, although targeted validation is still required (Meli et al., 2025).

Beyond these core pathways, enrichment analysis also identified carbohydrate- and amino acid-related pathways, as well as steroid hormone biosynthesis. Although some pathways had modest impact values, their collective appearance suggests that production-level differences were associated with coordinated changes in substrate metabolism and endocrine-related pathways (Ni et al., 2025); (Zamuner et al., 2020). The steroid hormone biosynthesis signal was consistent with the observed differences in cortisol and cortisone between groups, supporting an association between endocrine response and serum metabolomic variation in LP goats (Ünal & Uztimür, 2025).

The integration of VIP scores, univariate significance, pathway involvement, and ROC analysis identified nine candidate metabolites that differentiated HP and LP goats. HP-associated metabolites included UDP-glucose, riboflavin, S-adenosylmethionine, and betaine, whereas LP-associated metabolites included kynurenine, L-formylkynurenine, 1-methylhistidine, cortisone, and methylmalonic acid. This integrated approach is commonly used to prioritize candidate biomarkers in metabolomics studies, although such candidates require further validation before practical application (Cardoso et al., 2024). Methylmalonic acid was particularly relevant in LP goats because it may reflect altered organic acid metabolism and vitamin B12-related metabolic function, which can influence energy metabolism in ruminants (Liu et al., 2022). Therefore, the LP-associated metabolite profile may indicate greater stress-associated and catabolic metabolic adaptation, whereas the HP-associated profile may indicate more stable substrate and cofactor-related metabolism (Salama et al., 2021); (Tsuji et al., 2023).

From a translational perspective, the present findings suggest that serum metabolomics can complement conventional indicators such as milk yield, NEFA, BHBA, and hormone profiles by providing additional pathway-level information on production-associated metabolic phenotypes (Ghavipanje et al., 2021); (Lisuzzo et al., 2022). A composite panel comprising UDP-glucose, riboflavin, S-adenosylmethionine, betaine, kynurenine pathway metabolites, cortisone, 1-methylhistidine, and methylmalonic acid may serve as a candidate indicator panel for identifying goats with distinct metabolic profiles under intensive management. However, because individual feed intake was not measured and the sample size was limited, the practical application of this panel for precision nutrition should be considered preliminary and requires further validation (Clare et al., 2019); (Pardo et al., 2022).

The correlation analysis further supported the physiological relevance of selected biomarker candidates. In HP goats, riboflavin was positively correlated with milk production, while S-adenosylmethionine was positively correlated with BHBA. In LP goats, kynurenine was positively correlated with IGF-1, 1-methylhistidine was negatively correlated with IGF-1 and insulin, and cortisone was positively correlated with BHBA. These findings suggest that the candidate metabolites were associated with production, endocrine, and energy metabolic parameters. However, these correlations do not imply causality and should be interpreted cautiously.

Several limitations should be acknowledged. First, individual feed intake and nutrient intake were not measured in this study. Therefore, the concept of precision nutritional status should be interpreted as an integrative metabolic and metabolomic phenotype under similar feeding and management conditions, rather than as a direct measurement of individual nutrient consumption. Second, the cohort size was relatively small (n = 10 per group), which may inflate apparent classification performance, increase the risk of overfitting in multivariate analyses such as PLS-DA, and limit generalizability across farms, diets, and lactation stages. Therefore, the multivariate classification results should be interpreted as exploratory and should be validated in larger independent datasets. Third, untargeted metabolomics provides relative abundance rather than absolute quantification; thus, targeted validation with standardized assays is necessary before defining actionable thresholds (Islamiyati et al., 2025). Fourth, serum reflects systemic metabolism and cannot directly specify tissue-level flux in the liver, adipose tissue, or mammary gland (Sun et al., 2017). Future studies should incorporate individual feed intake and nutrient intake measurements, larger multi-farm populations, longitudinal sampling across lactation phases, targeted quantification of the key metabolites identified, and integration with inflammatory markers and rumen-derived energy precursors to strengthen biological interpretation and develop more robust predictive models (Irawan et al., 2026).

Overall, milk yield divergence in Saanen dairy goats under similar feeding and management conditions was associated with distinct systemic metabolic phenotypes. HP goats were characterized by lower NEFA, BHBA, and cortisol concentrations and higher abundance of metabolites related to carbohydrate activation, one-carbon metabolism, and riboflavin-linked pathways. In contrast, LP goats showed higher lipid mobilization and ketogenesis indicators, glucocorticoid-related metabolites, kynurenine pathway metabolites, methylmalonic acid, and 1-methylhistidine. These findings indicate that serum metabolomics, when combined with conventional metabolic and hormonal indicators, may help characterize production-associated metabolic differences in lactating dairy goats and support the future development of metabolite-informed nutritional evaluation strategies (Islamiyati et al., 2025); (Irawan et al., 2026).

Lactating Saanen dairy goats with higher milk production exhibited a more favorable metabolic phenotype characterized by improved energy balance, reduced metabolic stress, and enhanced carbohydrate and redox metabolism, whereas lower-producing goats showed metabolic signatures associated with lipid mobilization, ketogenesis, protein catabolism, and stress-related pathways. The identified metabolite biomarkers provide promising indicators for distinguishing metabolic efficiency between production groups. These findings support the application of metabolomic profiling as a practical tool for the early assessment of metabolic status and for developing targeted nutritional and herd management strategies to improve milk production, animal health, and the sustainability of dairy goat production systems.

The authors declare no conflicts of interest, financial or personal, that could have influenced the work reported in this manuscript.

The authors would like to express their sincere gratitude to the Head and the laboratory staff of the Animal Production Laboratory and the Animal Nutrition and Feed Laboratory, Department of Animal Science, Faculty of Agriculture, Lambung Mangkurat University, for providing the necessary facilities and technical support throughout the study. We would also like to express our sincere appreciation to the Talent Management of the National Research and Innovation Agency for supporting the postdoctoral program.

During the writing process, none of the authors used generative AI or AI-assisted technologies.