The experimental protocol, including handling and blood sampling, was approved by the Ethics Committee of the Faculty of Animal Science, Hasanuddin University (Approval No. 016/UN4.12/EC/VI/2025). Blood collection was performed by trained personnel using standard restraint to minimize stress, and all birds were monitored for adverse effects throughout the procedure, in compliance with applicable institutional and national animal welfare guidelines.

The study was conducted from June to October 2025. Clinical chemistry analyses were performed at the Faculty of Animal Science, Universitas Hasanuddin (Makassar, Indonesia), and metabolomic profiling was conducted at the National Research and Innovation Agency (BRIN), Cibinong, West Java, Indonesia.

A cross-sectional experimental design was employed to characterize age-related differences in blood biochemical parameters and plasma metabolomic profiles of Alope village chickens (G. g. domesticus). Thereafter, the species is referred to as G. g. domesticus. A total of 30 clinically healthy female birds were included and stratified into three biologically relevant age groups: G1 (8 weeks), G2 (12 weeks), and G3 (16 weeks), with n = 10 birds per group. Birds were reared under a semi-intensive floor housing system with litter, provided ad libitum access to feed and drinking water, and maintained under typical tropical environmental conditions. The primary objective was to conduct an age-wise comparison of classical biochemical indices and untargeted plasma metabolomic signatures to describe metabolic maturation during growth. Blood samples were collected in the morning before feeding, without prolonged fasting, to minimize postprandial variation while avoiding fasting-induced metabolic stress.

The analytical workflow integrated clinical chemistry assays, untargeted gas chromatography–mass spectrometry (GC–MS)–based plasma metabolomics, and univariate and multivariate statistical analyses to identify age-associated metabolic patterns and putative biomarker candidates.

Birds were fed a commercially formulated grower diet commonly used for village and slow-growing chickens under semi-intensive systems. Feed and clean drinking water were provided ad libitum throughout the experimental period. The same diet was offered to all age groups to ensure that observed metabolic differences primarily reflected age-related physiological variation rather than dietary effects. The diet contained approximately 18%–19% crude protein and 2,800–2,900 kcal/kg metabolizable energy on a dry matter basis, consistent with recommended nutrient specifications for growing chickens between 8 and 16 weeks of age. Feed was offered twice daily, and minimal refusals were observed, indicating consistent intake across groups.

Blood was collected from the brachial (wing) vein of each bird in the morning before feeding. Two tubes were obtained per bird: a lithium-heparin tube for plasma intended for biochemical analysis and a dipotassium ethylenediaminetetraacetic acid (K₂-EDTA) tube for plasma intended for metabolomic analysis. Tubes were gently inverted 8–10 times, kept on wet ice, and processed within 30 min. Plasma was separated by centrifugation at 1,500–2,000 × g for 10–15 min at 4°C, aliquoted into pre-chilled microtubes, immediately frozen, and stored at −80°C with a single freeze–thaw cycle maximum. Heparinized plasma was used for biochemical analyses, whereas EDTA plasma was used for metabolomic analyses [13].

Glucose, total protein, urea, and triglycerides were quantified in lithium-heparin plasma by enzymatic colorimetry using an automated clinical analyzer (cobas c 311, Roche Diagnostics, Mannheim, Germany) with manufacturer-supplied reagents and calibrators. When automation was unavailable, identical chemistries were performed in 96-well format using a microplate spectrophotometer (Multiskan SkyHigh, Thermo Fisher Scientific, Waltham, MA, USA). Assay principles included the hexokinase/glucose-6-phosphate dehydrogenase (G6PD) method for glucose (340 nm), the biuret method for total protein (540–560 nm), the urease–glutamate dehydrogenase ultraviolet method for urea (decrease of nicotinamide adenine dinucleotide, reduced form [NADH], at 340 nm), and the glycerol-3-phosphate oxidase–peroxidase (GPO–PAP) method for triglycerides (~500–550 nm, with glycerol blank where applicable).

Although uric acid (UA) is the primary nitrogenous waste product in birds, urea was included as a supportive indicator of general nitrogen turnover and amino acid catabolism, as commonly applied in avian biochemical profiling and comparative physiological studies. Each analytical batch included multilevel calibration and two-level internal quality controls (Liquichek, Bio-Rad Laboratories, Irvine, CA, USA). Samples were analyzed in duplicate, with intra-assay coefficients of variation <5% for glucose and total protein and <7% for urea and triglycerides. Samples exceeding predefined acceptance criteria were reanalyzed. Plasma was separated within 30 min of venipuncture and subjected to a single freeze–thaw cycle before analysis [14].

Blood biochemical data were tested for normality using the Shapiro–Wilk test and analyzed using one-way analysis of variance (ANOVA), followed by Tukey’s honest significant difference post hoc test where appropriate. Results are presented as mean ± standard deviation, and statistical significance was set at p < 0.05. Raw GC–MS data were processed using MetaboAnalyst 6.0 for peak detection, deconvolution, and retention time alignment. Features with coefficients of variation >30% in QC samples or present in <50% of samples in at least one group were excluded. Data were normalized, log-transformed, and Pareto-scaled prior to analysis. Principal component analysis was used for exploratory visualization, and age-discriminant metabolites were identified using partial least squares–discriminant analysis. Model robustness was evaluated using 7-fold cross-validation and 200 permutation tests. Univariate metabolite differences were assessed using ANOVA with Benjamini–Hochberg false discovery rate correction. Metabolites with variable importance in projection >1.0 and adjusted p < 0.05 were considered discriminant, and receiver operating characteristic curve analysis was used to assess classification performance by area under the curve (AUC). Statistical analyses were conducted using SPSS version 26.0 (IBM Corp., Armonk, NY, USA), MetaboAnalyst 6.0, and GraphPad Prism 9.0 (GraphPad Software Inc., San Diego, CA, USA).

Four plasma biochemical parameters, namely glucose, total protein, urea, and triglycerides, were measured in Alope chickens across increasing ages (Figure 1). Glucose increased from 177.80 ± 10.32 mg/dL at 8 weeks to 193.30 ± 10.13 mg/dL at 12 weeks and 216.60 ± 10.78 mg/dL at 16 weeks, corresponding to relative gains of 8.72% (8→12 weeks) and 12.05% (12→16 weeks). Total protein increased from 4.01 ± 0.14 g/dL to 4.26 ± 0.14 g/dL and 4.54 ± 0.15 g/dL, representing increases of 6.23% and 6.57%, respectively. Urea concentrations increased from 14.94 ± 0.65 mg/dL to 17.50 ± 0.63 mg/dL and 19.43 ± 1.12 mg/dL (17.14% and 11.03%, respectively). Triglyceride levels increased from 108.10 ± 4.56 mg/dL to 115.80 ± 5.41 mg/dL and 127.00 ± 6.18 mg/dL, corresponding to relative increases of 7.12% and 9.67%. All biochemical parameters differed significantly among age groups (p < 0.05) and remained within published physiological reference ranges for growing chickens, indicating normal metabolic status throughout the study period. Effect size estimates (η²) indicated moderate to large age effects, particularly for glucose and triglycerides, underscoring the biological relevance of the observed age-associated changes.

Multivariate projection of the plasma metabolome revealed a clear age-dependent structure. The three age groups formed tight, non-overlapping clusters in the principal component analysis (PCA) score plot (Figure 2A), with G1 loading predominantly on negative PC1, G2 on positive PC2, and G3 on positive PC1/negative PC2, indicating marked between-group separation and limited within-group dispersion. Partial least squares–discriminant analysis (PLS-DA) analysis (Figure 2B) identified metabolites driving this separation, with several HMDB-annotated features aligning toward the G3 centroid, whereas others were associated with G1 or G2, consistent with progressive metabolic changes across growth stages.

Univariate screening supported these findings. The ANOVA bubble plot (Figure 3A) demonstrated numerous features with strong evidence of group differences, indicated by higher −log₁₀p values and larger effect magnitudes across the m/z–retention-time space. Similarly, hierarchical clustering analysis (Figure 3B) clearly separated samples by age and revealed coherent metabolite modules that increased monotonically from G1 to G3 or declined in the opposite direction, with G2 exhibiting an intermediate, transitional profile. The VIP score plot (Figure 4) ranked the metabolites contributing most strongly to class separation (VIP ≈ 1.2–1.37), with accompanying heat strips showing consistent age-related abundance gradients. Collectively, these analyses demonstrate robust, system-level remodeling of the plasma metabolome with age in Alope chickens and define a focused subset of metabolites for downstream pathway analysis and validation.

Pathway enrichment analysis was performed in an exploratory manner to contextualize age-associated metabolite differences at the pathway-level. Twelve KEGG pathways contained at least one matched metabolite; however, none reached statistical significance after false discovery rate correction (all FDR ≥ 0.35). Accordingly, pathway-level findings are reported descriptively as associative patterns rather than causal inferences. The glycine, serine, and threonine metabolism pathway exhibited the highest topology impact (0.331; hits = 2: HMDB0000167, HMDB0000139) and moderate signal intensity on the bubble plot (−log₁₀p ≈ 1–1.2). Glycerolipid metabolism (impact = 0.093; HMDB0000139) and the citrate cycle (impact = 0.090; HMDB0000094) also showed non-trivial impacts based on single metabolite hits. The pentose phosphate and glyoxylate/dicarboxylate pathways each included two hits (FDR = 0.347) but exhibited low impact scores (≤ 0.042 and 0.010, respectively). Additional single-hit associations were observed for histidine, purine, glycerophospholipid, β-alanine, and branched-chain amino acid biosynthesis pathways (FDR = 0.62–1.00). Overall, these results suggest that age-associated metabolic variation is primarily linked to amino acid metabolism, with secondary contributions from lipid and central energy-related processes.

Supervised multivariate modeling identified seven putative biomarker candidates with high discriminatory contribution (VIP ≈ 1.17–1.35) and strong single-analyte classification performance (AUC ≥ 0.985). These candidates were derived from internal model optimization and should therefore be considered hypothesis-generating rather than definitive diagnostic markers. Three metabolites characterized the youngest age group (G1): HMDB0000625, HMDB0001448, and HMDB0000177. HMDB0000139 peaked at mid-growth (G2), whereas HMDB0000094, HMDB0000167, and HMDB0000289 were enriched in the oldest group (G3), collectively reflecting age-related shifts in amino acid, purine, and carbohydrate–lipid metabolism. Group-wise boxplots demonstrated minimal overlap among age categories, supporting monotonic age-associated abundance trends. Despite strong internal discrimination, external validation using independent and preferably longitudinal cohorts is required to confirm robustness and minimize overfitting. From a practical standpoint, several of these metabolites are chemically stable and analytically accessible, supporting their potential use in future targeted assays and positioning them as discovery-level leads for age-informed metabolic indicators in village chicken production systems.

Age emerged as the dominant axis structuring metabolism in Alope chickens. Across the 8–16-week growth window, concordant changes in blood biochemistry and plasma metabolomics indicate coordinated metabolic reorganization with growth rather than isolated shifts detected by either approach alone. The progressive increases in glucose, total protein, urea, and triglycerides observed in Figure 1 are consistent with a developmental transition from early lean tissue accretion toward greater emphasis on lipid handling and a more stable energy balance at later stages, with no evidence of a metabolic plateau within the studied age range [17, 18]. Multivariate projections clearly separated age groups in the PCA score plot and identified the metabolites driving this separation in the PLS-DA biplot (Figure 2), while univariate screening and hierarchical clustering revealed coordinated metabolite modules that increased or declined with age (Figure 3) [19,20]. Pathway-level trends supported these findings, with contributions from glycine–serine–threonine metabolism, glycerolipid and glycerophospholipid metabolism, purine metabolism, the citrate cycle, and nodes linked to the pentose phosphate pathway (Figure 5, Table 1) [21]. In the avian context, enrichment labels are best interpreted as topological descriptors of shared metabolic nodes rather than literal pathway activation, consistent with the dicarboxylate–tricarboxylic acid cycle. intersections suggested here [22].

Pathway enrichment results are presented as exploratory associations. Total indicates the total number of metabolites mapped to each pathway, Hits represent the number of detected metabolites mapped to the pathway, FDR denotes false discovery rate–adjusted p value, Impact reflects pathway topology impact based on network position, HMDB = Human Metabolome Database, TCA = Tricarboxylic acid cycle.

Integration of biochemical and metabolomic layers enabled translation of these findings into stage-relevant nutritional implications. The seven putative biomarker candidates identified (Figure 6, Table 2) mapped consistently to three functional axes, namely one-carbon/redox economy, membrane-lipid turnover, and nucleotide/energy metabolism, reflecting the biochemical trajectories observed across ages. In the youngest age group, HMDB0000625 (pentose phosphate pathway), HMDB0001448 (sulfur and purine metabolism), and HMDB0000177 (histidine, β-alanine, and glycerophospholipid pathways) collectively indicate elevated requirements for NADPH generation, antioxidant and methyl-group capacity, histidine–carnosine buffering, and active membrane remodeling. From a nutritional standpoint, this stage would benefit from higher digestible protein density with balanced essential amino acids, adequate methionine and cysteine supply, sufficient glucogenic substrates to support one-carbon flux, and controlled lipid provision favoring membrane biogenesis over storage [23].

Direction indicates the age group (G1, G2, or G3) in which the metabolite showed the highest relative abundance, VIP denotes variable importance in projection score derived from partial least squares–discriminant analysis, Pathway refers to the primary metabolic pathways associated with each metabolite, AUC represents the area under the receiver operating characteristic curve indicating classification performance, HMDB = Human Metabolome Database.

At the intermediate growth stage, HMDB0000139 showed peak abundance and bridged pentose phosphate, glycine–serine–threonine, glyoxylate–dicarboxylate, and glycerolipid metabolism, suggesting a transitional redistribution of carbon toward membrane assembly and one-carbon trafficking. At this stage, diets should maintain a balanced energy-to-protein ratio, ensure adequate glycine and serine availability, and provide choline and essential fatty acids to support phospholipid turnover without promoting excessive neutral-lipid deposition [4]. In the oldest chickens, enrichment of HMDB0000094 (glyoxylate–dicarboxylate metabolism), HMDB0000167 (glycine–serine–threonine metabolism with links to branched-chain amino acids), and HMDB0000289 (purine metabolism) reflects increased mitochondrial flux and anaplerosis, reorganization of amino acid utilization as growth shifts toward maintenance, and heightened nucleotide turnover approaching physiological maturity. Nutritional strategies at this stage should emphasize slightly higher energy density with careful control of fat quality, continued amino acid balance with attention to threonine and branched-chain profiles, and micronutrients supporting redox balance and nucleotide economy, with field validation recommended prior to routine application [24].

Several limitations should be considered when interpreting these findings. First, the study focused on a single village chicken variety with a modest sample size (n = 10 per age group), limiting statistical power and generalizability for omics-based inference [25]. Second, the cross-sectional design captured population-level age differences rather than individual developmental trajectories, and the restricted age window of 8–16 weeks precludes inference beyond this period [26]. Third, the metabolomics platform provided partial chemical coverage biased toward derivatizable compounds, and pathway enrichment signals did not withstand conservative multiple-testing correction, rendering pathway associations exploratory rather than confirmatory [27]. Fourth, UA, the principal nitrogenous waste product in birds, was not measured, and endocrine indicators of growth and metabolism, including insulin, thyroid hormones, and corticosterone, were not assessed, limiting resolution of nitrogen handling and hormonal regulation [28]. Finally, environmental and management variability typical of smallholder systems was not explicitly modeled and may have influenced metabolic profiles [29].

Taken together, Figures 1–6 and Tables 1–2 delineate a coherent trajectory of metabolic maturation in Alope chickens, in which early growth is characterized by elevated one-carbon/redox and membrane synthesis demands, mid-growth reflects transitional substrate redistribution, and later growth emphasizes mitochondrial energy handling and nucleotide turnover. Although these integrated biochemical and metabolomic patterns provide a robust descriptive framework, their application to precision feeding or routine physiological monitoring should be regarded as provisional, pending validation in larger, longitudinal cohorts under controlled dietary and environmental conditions. Notably, this study establishes age-stratified biochemical reference values for an indigenous village chicken genotype, integrates field-relevant clinical chemistry with untargeted plasma metabolomics, and introduces physiological age as a metabolomics-informed construct rather than relying solely on chronological age. By linking metabolic maturation to nutrition-sensitive pathways, this work contributes to sustainable poultry production and One Health–aligned food systems in low- and middle-income settings.