Age-driven marbling in Haimen goat meat: multi-omics dissection of lipid/amino signatures and IMF regulatory networks Authors: Zhipeng Liu, Jiafeng Lu, Wangang Zhang, Shisi Xie, Zhen Zhang, Kaiping Deng, Shang Gao, Yixuan Fan, Honghui Lu, Gang Xue, Yanli Zhang, Feng Wang Published: 2026 ABSTRACT AbstractElevated intramuscular fat (IMF) deposition is widely recognized as a hallmark of high-grade meat with improved sensory and nutritional properties. Thus, in the current study, the physicochemical and nutritional trajectories of Haimen goat meat slaughtered at different stages were systematically characterized. High-IMF (HF) and low-IMF (LF) were further dissected by lipidomics and amino-acid profiling to resolve fatty-acid and amino-acid signatures. IMF exhibited linear accretion with age. The HF group exhibited significantly higher proportions of SFA, MUFA and PUFA, concomitant with enhanced concentrations of essential and flavor-active amino acids. Proteomics analysis identified key regulatory genes related to intramuscular fat deposition including ECI1, HADH, and ACO2, highlighting the core roles of acetyl-CoA and succinyl-CoA in lipid metabolism. Overall, these results suggest that high-IMF goat meat may provide enhanced nutritional and flavor-related attributes.Keywords: Goat meat, Age, Intramuscular fat, Lipid metabolism, Meat flavor, Nutritive value, Flavor precursors Introduction Meat is a major source of dietary protein and fat worldwide. As living standards rise, consumers increasingly prioritize meat quality over quantity (Long et al., 2025; Meng, Li, Huo, Chang, & Shen, 2023). Goat meat is valued for its palatability, digestibility, and nutritional benefits (Deng et al., 2024; Kantono, Hamid, Ma, Chadha, & Oey, 2021). Haimen goats (HMGs), a renowned indigenous meat breed of China, constitute a nationally valued genetic resource for premium goat-meat production. In our preceding work, we undertook a systematic evaluation of meat-quality attributes across four anatomical locations of the HMGs (Deng et al., 2024). Additionally, previous studies have emphasized preservation and genetic regulation (Chang et al., 2025), yet age-related changes in its nutritional composition and their genetic basis remain underexplored. Addressing this gap is essential for a comprehensive understanding of goat meat quality. High-IMF goat meat, characterized by abundant marbling, is generally associated with superior sensory and nutritional attributes. IMF is shaped by non-genetic (nutrition, age, sex, anatomical) and genetic (breed, genotype) factors (Albano-Gaglio et al., 2025; Cai, Wang, Li, Yao, & Li, 2024; Gao, Wen, Dai, Li, & Zhou, 2024; Niu et al., 2024). IMF governs sensory traits (tenderness, juiciness, and flavor) and nutritional value, and higher IMF consistently improves flavor scores (Guo et al., 2022; Hou et al., 2023; Wang et al., 2023). Moreover, key quality indicators, including meat color, pH, and shear force, correlate strongly with IMF (Huang et al., 2020; Realini et al., 2021). Furthermore, the flavor and quality of meat are also influenced by various metabolic products, including peptides, amino acids and nucleotides (Ramalingam, Song, & Hwang, 2019; Zhang, Zhang, Zhou, & Zhang, 2021; Zhang, Zhang, Wang, Xing, & Zhang, 2020). Peptides may contribute to flavor development through Maillard reactions ( Yang et al., 2025), while fatty acids serve as direct precursors of volatile aroma compounds (Pan et al., 2025). Although previous studies have explored the relationship between IMF and meat quality in cattle and pigs, systematic investigations in goats remain limited. Therefore, a comparative assessment of meat quality traits between goats with divergent IMF levels can offer valuable insights into optimizing meat production and enhancing consumer acceptance. Using goat meat samples with different IMF levels collected at different ages, we traced IMF accretion dynamics and conducted a comprehensive comparison between HF and LF phenotypes. The lipidomics, amino acid profiling, and DIA proteomics were integrated to dissect the molecular basis of lipid composition, amino-acid signatures, protein abundance and post-translational modifications underlying quality divergence. The study further compared nutrient demands and energy metabolism between the two phenotypes. The resulting datasets will provide benchmark data and a theoretical framework for objective goat-meat quality assessment and precision nutrition. Materials and methods All animal experiments were conducted strictly following the ethical regulations of the Animal Experiment Management Committee of Nanjing Agricultural University (SYXK2022–0031), and were in full compliance with international laboratory animal care guidelines including EU Directive 2010/63 and the NIH Guide for the Care and Use of Laboratory Animals. Goat meat samples collection The experiment was conducted at the White Goat Breeding and Conservation Research Institute (Haimen, Nantong, Jiangsu Province, China). Twenty healthy female goats of similar body weight were allocated to four age groups (6, 12, 18 and 36 months; n = 5 per group) and were raised in separate pens. All goats were housed under uniform conditions and fed stage-appropriate diets. Prior to slaughter, goats were fasted for 12 h with water available ad libitum. After electrical stunning, exsanguination was performed by severing the jugular vein and carotid artery. Longissimus dorsi muscle samples were subsequently collected for analysis. For comparative IMF, 36-month goats (high-IMF group, HF) and 12-month goats (low-IMF group, LF) were respectively selected based on preliminary IMF content. Physical analyses pH measurement Muscle pH was measured three times for each sample at 45 min post-slaughter, using a pH-Star device (DELTA 320, Mettler Toledo, Columbus, OH, USA). The device, calibrated with standard buffers (pH 4.01 and 7.00 at 20 °C) prior to use, was subjected to interim recalibration after every six samples to maintain measurement accuracy (Mettler Toledo, Greifensee, Switzerland). Instrumental color measurement Meat color was assessed using a colorimeter (3nH-NH310, 3NH, Guangzhou, China). The measurements were conducted with an aperture diameter of 8 mm, a D65 illuminant, and a 10° standard observer. The lightness (L⁎), redness (a⁎), and yellowness (b*) were measured on the meat surface after 30 min of blooming at room temperature. The average values were obtained from three different positions for each sample. Shear force measurement At 24 h postmortem, the Longissimus dorsi muscle was excised. Following 30 min of equilibration on ice, samples were cut parallel to the muscle fiber orientation into rectangular strips of 1 cm2 cross-sectional area, weighing 1.00 ± 0.10 g (mean ± SD). The shear force was then measured using a C-LM3B tenderness meter (Tenovo, Beijing, China) with the blade oriented perpendicular to the muscle fiber direction at a constant crosshead speed of 5 mm s−1. The maximum force required to shear each sample was recorded in Newtons (N). Shear force (N) was recorded at the point when the blade had completely traversed the sample (Zhang et al., 2015). The average of the three technical replicates was computed and reported for each animal. Electrical conductivity Electrical conductivity was measured using a calibrated portable conductivity meter (LF-STAR, MATTHAUS, Germany). After equilibration at 25 °C, electrodes were placed on the fresh cut surface of the longissimus dorsi. Readings were recorded in s/m after stabilization. Each sample was measured three times, and the mean value was used. Assessment of nutritional value of goat meat Crude moisture Samples were cut, freeze-dried to constant weight, ground, and weighed. Moisture content was calculated from initial and final weights, according to the Chinese national standard “Permitted level of moisture in meat of livestock and poultry” (GB 18394–2020). Crude protein Protein content was determined using the Kjeldahl method as described in the Chinese national standard “National standard for food safety: determination of protein in food” (GB 5009.5–2016). Freeze-dried samples were digested with concentrated sulfuric acid, distilled, and titrated. Crude protein was calculated from nitrogen content. Measurement of IMF The IMF content was determined at 24 h post-slaughter using the Soxhlet extraction method according to the Chinese national standard GB 5009.6–2016 (National standard for food safety: determination of fat in food). All muscle samples were freeze-dried in advance and thoroughly ground into fine powder. Approximately 2.0 g of freeze-dried muscle powder was accurately weighed and subjected to Soxhlet extraction with petroleum ether for 6-8 h. After extraction, the samples were dried to a constant weight and reweighed. In this study, the formula for IMF content calculation was as follows: IMF (%) = [W1(g) – W2(g)] / W2(g) × 100%. W1: initial weight of freeze-dried sample before Soxhlet extraction (g); W2: constant weight of the residual sample after petroleum ether extraction and vacuum drying (g). Immunofluorescence and confocal imaging Frozen muscle samples embedded in OCT were sectioned at 10 μm using a cryostat at −20 °C. Sections were fixed in 4% paraformaldehyde, washed, and incubated with BODIPY solution (C2053M, Beyotime, China) at room temperature in the dark for 20 min. Nuclei were counterstained with 4′, 6-diamidino-2-phenylindole (DAPI, C1006, Beyotime, China). Images were acquired with a confocal microscope (Zeiss LSM710 META, Jena, Germany), and fluorescence intensity was quantified using ImageJ (v1.51). Electronic nose analysis of flavor profile The flavor profile of samples was determined using an electronic nose system (NOVA Enose3, Beijing, China). Before detection, the electronic nose instrument was fully stabilized and equilibrated at room temperature for 30 min to ensure stable baseline signals. Briefly, 2 g of each meat sample was accurately weighed and transferred into a 5 mL headspace vial, which was then sealed and incubated to allow the formation of volatile flavor compounds in the headspace. Subsequent detection was carried out by the electronic nose under optimized analytical conditions: sampling frequency of 1 Hz, intake flow rate of 200 mL/min, sensor cleaning time of 60 s, sample preparation time of 5 s, and sample injection detection time of 60 s. All measurements were performed with biological replicates, and the sensor response values were recorded for subsequent flavor characteristic analysis. Targeted fatty acid measurement and analysis The method for extracting fatty acid from the muscle was an improved version of the technique originally described by (Lee, Trevino, & Chaiyawat, 1996). The standard stock solution of fatty acid methyl ester was diluted with n-hexane to prepare the standard curve working solution, which should be prepared and used immediately. Approximately 2 g of the sample was weighed and transferred into a 2 mL centrifuge tube. Subsequently, 1 mL of a chloroform-methanol (2:1, v/v) solution and 100 mg of glass beads were added. The mixture was homogenized using a high-throughput tissue grinder at 55 Hz for 1 min, and this homogenization step was repeated twice. Following homogenization, the sample was subjected to ultrasonication at room temperature for 30 min and then centrifuged at 15,000g and 4 °C for 10 min. The entire supernatant was collected into a 10 mL glass centrifuge tube. Next, 2 mL of a 1% sulfuric acid-methanol solution was added, and the mixture was vortexed thoroughly. Esterification was performed in an 80 °C water bath for 30 min. After cooling to room temperature, 1 mL of n-hexane was added for extraction. The mixture was vortexed vigorously and allowed to stand for 2 min for phase separation. Then, 5 mL of ice-cold (4 °C) water was added for washing, followed by centrifugation at 5000g and 4 °C for 10 min. A 700 μL aliquot of the upper organic layer was accurately transferred to a new 2 mL centrifuge tube, and 100 mg of anhydrous sodium sulfate was added to remove residual water. After vortexing, the sample was diluted 20-fold with n-hexane. From this dilution, a 300 μL aliquot was transferred to another 2 mL centrifuge tube. Finally, 15 μL of a 500 ppm methyl salicylate solution (internal standard) was added. After thorough mixing, an appropriate volume of the supernatant was taken for instrumental analysis. The GC analysis was conducted using a Trace 1300 system (Thermo Fisher Scientific, USA) equipped with a Thermo TG-FAME capillary column (50 m × 0.25 mm i.d. × 0.20 μm film thickness). Helium was used as the carrier gas at a constant flow rate of 0.63 mL/min. Samples (1 μL) were injected in split mode (split ratio 8:1) at an injector temperature of 250 °C. The temperatures of the ion source and MS transfer line were set at 300 °C and 280 °C, respectively. The oven temperature program was as follows: initial temperature 80 °C held for 1 min; increased to 160 °C at 20 °C/min and held for 1.5 min; then raised to 196 °C at 3 °C/min and held for 8.5 min; finally increased to 250 °C at 20 °C/min and held for 3 min. Amino acid detection Targeted amino acid contents were determined by ultra-high performance liquid chromatography-tandem mass spectrometry (LC-MS/MS). A mixed working solution of amino acid standards was serially diluted to establish the standard calibration curve. For sample pretreatment, an appropriate amount of tissue sample was accurately weighed and mixed with 600 μL of a 10% formic acid methanol-water solution (1:1, v/v), and then transferred to a millipore tube (Jet Biofil, Guangzhou, China). The mixture was homogenized with steel beads at high frequency, and then centrifuged at 15,000g and 4 °C for 10 min. The obtained supernatant was diluted appropriately, spiked with stable isotope internal standard (Trp-d3), and fully vortexed for equilibration. Subsequently, the mixed solution was filtered through a 0.22 μm organic microporous membrane, and the filtrate was collected for subsequent UPLC-MS/MS analysis. Chromatographic separation was performed on an ACQUITY UPLC® BEH C18 column (2.1 × 100 mm, 1.7 μm; Waters Corporation, USA). The chromatographic conditions were set as follows: injection volume of 5 μL, column temperature maintained at 40 °C. mobile phase A was 50% methanol-water containing 0.1% formic acid, and mobile phase B was 10% methanol-water containing 0.1% formic acid. Gradient elution was performed under the following conditions: 0–6.5 min, 90–70% B; 6.5–7.0 min, 70–0% B; 7.0–14.0 min, 0% B; 14.0–14.5 min, 0–90% B; and 14.5–17.5 min, 90% B. The flow rate was maintained at 0.3 mL/min from 0 to 8.0 min and increased to 0.4 mL/min from 8.0 to 17.5 min. The mass spectrometer was operated in multiple reaction monitoring (MRM) mode for qualitative and quantitative analysis of targeted amino acids. Proteomics analysis Samples were prepared for relative and absolute quantification (iTRAQ) labeling and MS analysis following established protocols (Dyring-Andersen et al., 2020). Peptides were labeled with iTRAQ 8-plex reagents (Sciex), fractionated with 2D LC (Dionex Ultimate 3000, Thermo Scientific), and analyzed under DDA conditions. Bioinformatic analysis Metabolites detected in >50% of the samples were retained. Differential metabolites were defined as those with p-value <0.05 (Student's t-test) and VIP >1. Raw proteomics data were processed using Proteome Discoverer 2.1 (PD2.1, Thermo Scientific). Peptides were searched against the UniProt Capra hircus proteome database (UP000291000) with FDR ≤ 1%. Quantitative data were log2-transformed and normalized. Differential proteins were identified using two-sample t-test with permutation-based FDR correction (250 randomizations, S₀ = 0.1). Proteins with FDR ≤ 0.01 and |log2(fold change)| ≥ 1 were considered significant. Statistical analysis All statistical analyses were conducted using SPSS Statistics for Windows version 29.0 (IBM, Armonk, NY, USA). Normality was verified by the Shapiro–Wilk test, and homogeneity of variances by Levene's test. For pairwise comparisons, unpaired two-tailed Student's t-tests were applied. When more than two groups were involved, differences were evaluated by one-way analysis of variance (ANOVA) followed by Fisher's least significant difference (LSD) post-hoc test for multiple comparisons. Data were presented as Mean ± SEM, and significance was set at p < 0.05. Figures were prepared and edited using GraphPad Prism with version 10 (GraphPad Software, San Diego, CA, USA). Results Changes in meat quality attributes with age in Haimen goats The meat traits of Haimen goats were analyzed at different ages (Tab. 1). Comprehensive evaluation revealed that meat-quality traits and chemical composition of HMGs changed systematically with animal age (P < 0.05). Surface lightness (L*) declined progressively across the four age cohorts, with every pairwise comparison reaching significance (P < 0.01). Meanwhile, redness (a*) was significantly higher in the 12 M and 18 M groups compared to the 6 M and 36 M groups (P < 0.01), while yellowness (b*) was significantly lower only in the 36 M group (P < 0.01). The pH value was significantly higher in the 18 M group than in the other three age groups (P < 0.01). Shear force increased significantly with age, showing highly significant differences among all groups (P < 0.01). Similarly, electrical conductivity decreased with advancing age, with significant differences being observed between most groups (P < 0.01). Overall, age has a significant impact on the color, texture and final pH value of the goat meat. Table 1Differences in the physicochemical properties of Haimen mutton at different age stages.Traits6 M12 M18 M36 MF-ValueP-ValueL⁎46.07 ± 1.15 A43.29 ± 2.07B39.42 ± 1.02C33.15 ± 0.55D144.69P < 0.001A⁎15.25 ± 0.75C17.74 ± 0.68 A16.6 ± 0.45B14.95 ± 0.55C34.90P < 0.01B⁎6.19 ± 0.25 A5.76 ± 0.91 A5.95 ± 0.52 A4.36 ± 0.62B13.90P < 0.01pH6.23 ± 0.26B6.22 ± 0.03B6.52 ± 0.11 A6.08 ± 0.02B13.32P < 0.01Shear force/ N40.31 ± 0.14D43.45 ± 0.23 A42.95 ± 0.13B42.01 ± 0.13C569.86P < 0.01Electric conductivity/ s/m0.87 ± 0.11 A0.65 ± 0.12B0.49 ± 0.19b0.35 ± 0.06C24.58P < 0.01Data were expressed as means ± standard errors. Differences in the physicochemical properties of Haimen mutton at different age stages. Data were expressed as means ± standard errors. We also analyzed the content of various essential nutrients in the meat of the Haimen goat meat at different age stages (Tab. 2). Intramuscular moisture declined linearly with age, with 6 M retaining the highest proportion and 18 M and 36 M the lowest (P < 0.01). Crude protein content varied among groups, with the 12 M group showing the highest value and the 18 M group the lowest (P < 0.05). Crude fat content increased markedly with age, with all groups differing significantly from each other (P < 0.01). Therefore, those results suggest that goats rasied beyond 18 months tend to exhibit greater IMF deposition and improved marbling characteristics. Table 2Comparison of the main nutrient contents of Haimen goat meat at different age stages.Chemical compositions6 M12 M18 M36 MF-ValueP-ValueMoisture/ %78.45 ± 0.95 A75.34 ± 0.98B73.04 ± 1.47b72.93 ± 1.4b17.94P < 0.01Crude Protein/ %21.87 ± 0.64 A22.57 ± 0.36AB21.18 ± 0.56Ac22.31 ± 0.66ABC4.589P < 0.05Crude fat/ %5.22 ± 0.7D7.34 ± 1.15C11.26 ± 0.53B12.37 ± 0.65 A43.826P < 0.01Data were expressed as means ± standard errors. Comparison of the main nutrient contents of Haimen goat meat at different age stages. Data were expressed as means ± standard errors. Differences in IMF content between HF and LF goat meat To characterize the biological contrasts between divergent IMF phenotypes, goats exhibiting high IMF (HF,36 months of age) and low IMF (LF, 12 months of age) were as representative high- and low-IMF phenotypes. The HF group exhibited more pronounced marbling characteristics than the LF group(Fig. 1A). Therefore, we measured the fat contents in the muscles, and the IMF content in the HF group was 75% higher than that in the LF group (Fig. 1B). Meanwhile, the staining results of immunofluorescence indicated that the lipid droplet fluorescence intensity in the HF group was higher than that in the LF group (Fig. 1C). The electronic nose analyzer, equipped with 10 sensors (Tab. S1), was used to assess differences in the combined flavor profiles of groups LF and HF. As shown in Fig. 1D, the radar plot area showed no significant differences between the HF and LF groups. These results indicate that overall aroma profiles were broadly comparable between the two groups. Comparison of intramuscular fat deposition between LF and HF groups. (A). Representative images of muscle samples from the LF and HF groups, showing visible differences in intramuscular fat deposition. (B). Quantification of fat content by Soxhlet extraction in muscle tissues of the two groups. (C). Fluorescence staining of muscle sections with BODIPY (green) for lipid droplets and DAPI (blue) for nuclei. The right panel shows the relative fluorescence intensity analysis. (D). Radar plot comparing aroma-related sensor responses between LF and HF groups. Statistical significance: * for P < 0.05, ** for P < 0.01. Data are presented as mean ± SEM. Distinct fatty acid metabolic profiles in HF and LF groups To investigate the metabolic differences between the HF and LF groups, we performed LC-MS/MS-based analysis of fatty acid profiles. To better understand the metabolic differences between groups, principal component analysis (PCA) was applied for pairwise comparisons. In the comparison of longissimus dorsi between HF group and LF group, PCA explained the total variation, respectively (Fig. 2A). By comparing the differentially expressed metabolites (DEMs) among different groups, it was found that 36 metabolites were upregulated and 2 were downregulated, indicating a global accumulation of fatty acid metabolites in the HF group (Fig. 2B). Among the indicators of fatty acid composition, all parameters associated with individual fatty acid levels were higher in the HF group than in the LF group (Tab. 3). However, in the HF group, the levels of saturated fatty acids (P < 0.05), monounsaturated fatty acids (P < 0.05) and polyunsaturated fatty acids (P < 0.05) were significantly higher than those in the LF group. Interestingly, the PUFA/SFA ratio was significantly lower in the HF group compared to the LF group (P < 0.05). We found the levels of many key saturated fatty acids, such as lauric acid, myristic acid, palmitic acid, and stearic acid, were significantly increased in the HF group (Fig. 2C-D). Meanwhile, the ratio of C15/C17 was significantly higher in the HF group compared to the LF group (Fig. 2E). Besides, we compared the differences in essential fatty acids among unsaturated fatty acids, namely oleic acid, α-linolenic acid and arachidonic acid. It was found that the levels of these essential fatty acids in the HF group were significantly higher than those in the LF group (Fig. 2F-H). Meanwhile, the ratio of n-6/n-3 polyunsaturated fatty acids (PUFAs) was compared between the two groups. No statistically significant difference was observed in the n-6/n-3 ratio between the HF and LF groups (P > 0.05; Fig. 2I). To illustrate the relationship between fatty acids and meat quality, we have drawn a correlation analysis chart. Numerous fatty acids showed positive correlations with IMF, with correlation coefficients being greater than 0.3 (Fig. 2J, and Tab. S2). Particularly, a significant positive correlation was observed between arachidonic acid and protein content. In contrast, the correlations between pH and fatty acids were relatively weak. These findings indicate that the goat meat from the HF group exhibited a higher fatty acid content and superior nutritional value. Comparative analysis of fatty acid metabolic profiles between HF and LF goat meat. (A) PCA showing distinct separation of HF and LF groups based on metabolite profiles. (B) Number of DEMs between LF and HF groups. (C-E) Differential expression levels of representative saturated fatty acids, including stearic acid (C), 15:0/17:0 (D), and C12:0, C14:0, C16:0 (E). (F—H) Differential expression levels of representative unsaturated fatty acids, including oleic acid (F), α-linolenic acid (G), and docosahexaenoic acid (DHA) (H). (I) Comparison of the n-6/n-3 fatty acid ratio between LF and HF groups. (J) Correlation heatmap of meat quality traits (fat, protein, and pH) with fatty acid metabolites, showing Pearson's correlation coefficients (r). p-values are indicated in each panel; data are presented as mean ± SEM. Table 3The main effects of LF and HF on fatty acid compositions.Fatty acidsLFHFP-ValueSFA1C8:00.25 ± 0.251.35 ± 0.560.1018C10:011.35 ± 6.38 a35.52 ± 7.66 b0.0358C12:08.11 ± 4.48 a31.97 ± 7.5 b0.0211C13:00.30 ± 0.302.54 ± 1.110.0802C14:0308.43 ± 168.31 a1270.75 ± 309.18 b0.0211C15:039.86 ± 25.76 a193.83 ± 52.31 b0.0247C16:04188.41 ± 2234.39 a14,679.65 ± 2952.94 b0.0178C17:0175.33 ± 114.34 a613.59 ± 142.71 b0.0375C18:01815.84 ± 865.69 a6523.93 ± 1680.06 b0.0319C20:03.46 ± 1.41 a18.24 ± 6.10 b0.0399C21:00.19 ± 0.07 a1.18 ± 0.44 b0.0498C22:01.18 ± 0.122.46 ± 0.650.0837C23:02.76 ± 0.16 A5.40 ± 0.79 B0.0088C24:00.98 ± 0.111.79 ± 0.390.0725 Total SFAs6556.44 ± 3404.57 a23,382.2 ± 5093.32 b0.0206MUFA2C14:1 T2.28 ± 1.44 a23.38 ± 7.55 b0.0206C14:115.39 ± 9.96 a66.86 ± 15.93 b0.0208C15:1 T2.22 ± 1.1419.14 ± 7.710.0552C15:10.12 ± 0.110.53 ± 0.200.0997C16:1 T27.23 ± 14.42 A168.22 ± 39.94 B0.0077C16:1444.89 ± 245.38 a1636.54 ± 334.41 b0.0166C17:1 T12.91 ± 7.62 a60.82 ± 14.80 b0.0164C17:1193.23 ± 137.24517.83 ± 92.070.0779C18:1 N9T8.58 ± 4.8623.58 ± 5.960.0796C18:1 N7T204.84 ± 108.30616.02 ± 184.050.0831C18:1 N12211.88 ± 69.55 a457.42 ± 63.00 b0.0257C18:1 N9C7519.66 ± 4039.90 a21,617.9 ± 3030.89 b0.0191C18:1 N7620.21 ± 495.323252.18 ± 2001.580.2306C20:1 T2.82 ± 2.126.84 ± 1.260.1347C20:115.76 ± 8.23 a48.86 ± 9.30 b0.0236C22:1 N9T1.53 ± 0.570.55 ± 0.450.2093C22:1 N92.61 ± 0.38 a4.08 ± 0.46 b0.0322C24:114.75 ± 4.05 a27.15 ± 2.24 b0.0229 Total MUFAs9300.88 ± 5075.31 a28547.92 ± 5135.03 b0.0237PUFA34, 5C18:2 N6T13.88 ± 7.00 a53.68 ± 13.64 b0.0267C18:2 N6398.95 ± 111.92 a1131.25 ± 216.75 b0.0133C18:3 N63.12 ± 1.55 a10.48 ± 2.24 b0.0221C18:3 N317.25 ± 6.96 a56.91 ± 13.57 b0.0265C20:22.76 ± 1.47 a8.64 ± 1.84 b0.0314C20:3 N69.53 ± 1.84 a15.97 ± 1.85 b0.0333C20:3 N31.79 ± 0.122.45 ± 0.280.0557C20:4 N6141.34 ± 10.00 a200.48 ± 17.57 b0.0152C22:21.00 ± 0.430.16 ± 0.150.0925C20:5 N36.47 ± 0.727.19 ± 0.600.4617C22:417.02 ± 2.03 a33.52 ± 4.98 b0.0119C22:5 N64.00 ± 0.27 a6.15 ± 0.91 b0.0471C22:5 N319.08 ± 3.06 a48.71 ± 8.97 b0.0108C22:6 N35.24 ± 0.49 a8.97 ± 1.21 b0.0173Total PUFAs641.43 ± 144.71 a1584.56 ± 266.23 b0.011PUFA/SFA (%)23.59 ± 5.62 a7.17 ± 0.42 b0.0154Data were expressed as means ± standard errors. SFAs, saturated fatty acids; MUFAs, monounsaturated fatty acids; PUFAs, polyunsaturated fatty acids; USFAs, unsaturated fatty acids. Data with distinct superscripts within the same row indicate significant differences (P < 0.05).1SFA = (C22:0 + C10:0 + C16:0 + C15:0 + C17:0 + C14:0 + C13:0 + C11:0 + C8:0 + C21:0 + C18:0 + C12:0 + C24:0 + C23:0 + C20:0).2MUFA = (C14:1 T + C14:1 + C15:1 T + C15:1 + C16:1 T + C16:1 + C17:1 T + C17:1 + C18:1 N9T + C18:1 N7T + C18:1 N12 + C18:1 N9C + C18:1 N7 + C20:1 T + C20:1 + C22:1 N9T + C22:1 N9 + C24:1).3PUFA = (C18:2 N6T + C18:2 N6 + C18:3 N6 + C18:3 N3 + C20:2 + C20:3 N6 + C20:3 N3 + C20:4 N6 + C22:2 + C20:5 N3 + C22:4 + C22:5 N6 + C22:5 N3 + C22:6 N3). AI = (C12:0 + 4 × C14:0 + C16:0)/(∑MUFA+∑PUFA). TI = (C14:0 + C16:0 + C18:0)/(0.5 × ∑MUFA +0.5 × ∑n6 PUFA +3 × ∑n3 PUFA + (n3/n6)).4N3 = (C20:3 N3 + C22:6 N3 + C18:3 N3 + C22:5 N3 + C20:5 N3).5N6 = (C18:2 N6 + C18:3 N6 + C20:3 N6 + C18:2 N6T + C20:4 N6 + C22:5 N6). The main effects of LF and HF on fatty acid compositions. Data were expressed as means ± standard errors. SFAs, saturated fatty acids; MUFAs, monounsaturated fatty acids; PUFAs, polyunsaturated fatty acids; USFAs, unsaturated fatty acids. Data with distinct superscripts within the same row indicate significant differences (P < 0.05). SFA = (C22:0 + C10:0 + C16:0 + C15:0 + C17:0 + C14:0 + C13:0 + C11:0 + C8:0 + C21:0 + C18:0 + C12:0 + C24:0 + C23:0 + C20:0). MUFA = (C14:1 T + C14:1 + C15:1 T + C15:1 + C16:1 T + C16:1 + C17:1 T + C17:1 + C18:1 N9T + C18:1 N7T + C18:1 N12 + C18:1 N9C + C18:1 N7 + C20:1 T + C20:1 + C22:1 N9T + C22:1 N9 + C24:1). PUFA = (C18:2 N6T + C18:2 N6 + C18:3 N6 + C18:3 N3 + C20:2 + C20:3 N6 + C20:3 N3 + C20:4 N6 + C22:2 + C20:5 N3 + C22:4 + C22:5 N6 + C22:5 N3 + C22:6 N3). AI = (C12:0 + 4 × C14:0 + C16:0)/(∑MUFA+∑PUFA). TI = (C14:0 + C16:0 + C18:0)/(0.5 × ∑MUFA +0.5 × ∑n6 PUFA +3 × ∑n3 PUFA + (n3/n6)). N3 = (C20:3 N3 + C22:6 N3 + C18:3 N3 + C22:5 N3 + C20:5 N3). N6 = (C18:2 N6 + C18:3 N6 + C20:3 N6 + C18:2 N6T + C20:4 N6 + C22:5 N6). Comparative amino acid metabolic profiles between HF and LF groups Amino acids have many nutritional and physiological functions, which affect the nutritional value and taste of meat (Ma et al., 2024). The concentrations of amino acids across the different groups are presented in Tab. 4. A total of 16 amino acids were detected in the goat meat samples. Amino acids were further grouped according to their taste characteristics: umami amino acids (UAAs; Asp and Glu), sweet amino acids (SAAs; Ser, Gly, Thr and Ala), bitter amino acids (BAAs; His, Arg, Val, Met, Ile, Leu and Phe), and other amino acids (Pro, Tyr and Lys) according to the GB 5009.124–2016 (Xie et al., 2026). The content of essential amino acids (EAAs) was significantly higher in the HF group compared to the LF group (P < 0.05). Similarly, the levels of umami amino acids (UAAs), sweet amino acids (SAAs), and bitter amino acids (BAAs) were also elevated in the HF group. Among all detected amino acids, His, Gly, and Ala were the three most abundant. Table 4The main effects of LF and HF on amino acid compositions.Amino acidsLFHFP-ValueUmani amino acids (UAAs)Asp0.07 ± 0.010.18 ± 0.050.0515Glu0.13 ± 0.030.37 ± 0.190.2360 Sweet amino acids (SAAs)Gly2.46 ± 0.352.97 ± 0.990.6384Ala1.02 ± 0.14 B1.87 ± 0.24 A0.0090Ser0.18 ± 0.03 B0.38 ± 0.04 A0.0007Thr0.05 ± 0.01 b0.10 ± 0.01 a0.0217 Bitter amino acids (BAAs)Arg0.17 ± 0.04 b0.32 ± 0.04 a0.0165Met0.10 ± 0.020.15 ± 0.010.0682His3.45 ± 1.117.12 ± 1.390.0543Phe0.30 ± 0.050.41 ± 0.030.0575Ile0.18 ± 0.03 b0.27 ± 0.02 a0.0435Leu0.34 ± 0.05 b0.49 ± 0.02 a0.0471Val0.30 ± 0.05 b0.45 ± 0.03 a0.0320 other amino acidsTyr0.32 ± 0.06 b0.49 ± 0.03 a0.0309Pro0.10 ± 0.02 B0.17 ± 0.01 A0.0083Lys0.26 ± 0.07 b0.46 ± 0.06 a0.0442EAA1.61 ± 0.27 b2.43 ± 0.16 a0.0263Data were expressed as means ± SEM (ug/mg). Data with distinct superscripts within the same row indicate significant differences (P < 0.05). The main effects of LF and HF on amino acid compositions. Data were expressed as means ± SEM (ug/mg). Data with distinct superscripts within the same row indicate significant differences (P < 0.05). To better understand the differentially expressed amino acids (DEAs), PCA was applied for pairwise comparisons. In the comparison of longissimus dorsi between HF group and LF group, PCA explained 93.27% and 5.11% of the total variation, respectively (Fig. 3A). The relative expression of the identified significantly different amino acids was used to generate a hierarchical clustering heat map (Fig. 3B). Furthermore, a correlation heatmap was employed to visualize the relationships among different amino acids (Fig. 3C). Quantitative analysis of EAAs revealed significantly higher levels in the HF group compared to the LF group for all detected amino acids except Met, Phe, and Trp (Fig. 3D). Similarly, correlations between amino acids and meat quality traits were analyzed (Fig. 3E, and Tab. S3). The results indicated a significant positive correlation (r > 0.3) between protein content and multiple amino acids, including GABA, serine, aspartic acid, glutamine, and glutamic acid. IMF content was significantly positively correlated with proline and arginine (r > 0.3), while pH showed a significant positive correlation with asparagine (r > 0.3). Studies have shown that there may be a synergistic metabolic mechanism between fat deposition and the formation of amino acid flavor substances. These results demonstrate that high-IMF goat meat possesses enhanced amino acid composition and improved nutritional characteristics. Comparative amino acid metabolic profiles between HF and LF groups (A) PCA score plot showing the separation between the LF and HF groups. (B) Heatmap showing the z-scores of different amino acids in LF and HF groups. (C) Heatmap of the correlation between different amino acids. (D) Bar plot showing the expression levels of DEAs in LF and HF groups. (E) Correlation analysis between amino acids and parameters (pH, Protein, Fat). (r.sign for correlation direction, P for P - value, r.abs for absolute correlation coefficient). Statistical significance: * for P < 0.05, ** for P < 0.01. Data are presented as mean ± SEM. Differential protein expression associated with IMF content in goat meat To investigate the influence of high and low IMF content on muscle protein composition, a quantitative proteomic analysis was conducted. A total of 170 differentially expressed proteins (DEPs) were identified, comprising 141 significantly downregulated (FC < 1, P < 0.05) and 29 significantly upregulated (FC > 1, P < 0.05) proteins (Fig. 4A and Tab. S4). As illustrated in the heatmap (Fig. 4B), the expression profiles of these DEPs exhibited a clear divergence between the HF and LF groups, with elevated and reduced expression levels being represented in red and blue, respectively. This distinct clustering pattern reflects substantial proteomic remodeling associated with IMF content. Furthermore, KEGG pathway analysis revealed that DEPs were significantly enriched in signaling pathways related to energy metabolism and fatty acid metabolism, including the fatty acid biosynthesis, fatty acid metabolism, TCA cycle and glycolysis/gluconeogenesis signaling pathways (Fig. 4C). GSEA analysis further validated these results of KEGG analysis, showing the enrichment of DEPs in TCA cycle and oxidative phosphorylation signaling pathway (P adj < 0.001) (Fig. 4D). Integrated analysis of metabolomic and proteomic data from the LF and HF groups revealed that both differentially accumulated metabolites and DEPs were predominantly enriched in pathways related to carbohydrate metabolism and fatty acid biosynthesis (Fig. 4E). Acetyl-CoA and succinyl-CoA deserve particular emphasis (Fig. 4E), as the associated genes play a crucial role in fatty acid metabolism and should be prioritized in future breeding and production strategies. Differential protein expression associated with IMF content in goat meat Volcano plot showing DEPs between HF and LF groups. (B) Heatmap of DEPs in HF and LF groups. (C) Bubble plot of KEGG pathway enrichment. Bubble size: gene number; color: p - value; x - axis: rich factor. (D) GSEA plot for “Citrate cycle (TCA cycle)” and “Oxidative phosphorylation” pathways. (E) Schematic of fatty acid, amino acids and TCA cycle metabolism, with key enzymes and metabolites labeled. Nutritional value of marbled goat meat and its relevance to human dietary needs To directly compare the flavor and nutritional contributions of IMF and lean meat tissue, a comprehensive comparative analysis was performed. Firstly, the correlation between amino acid composition and flavor profiles was analyzed. The results revealed a significant positive correlation between key indicators, pungent (r = 0.8608, P = 0.017), green (r = 0.8606, P = 0.041), fruity (r = 0.7864, P = 0.028), waxy (r = 0.7928, P = 0.043), and oily (r = 0.8450, P = 0.024), and asparagine (Fig. 5A, and Table S5). Concurrently, pungent flavor was also found to exhibit a significant positive correlation with Lys (r = 0.2656, P = 0.011) (Fig. 5A, and Tab. S5). Furthermore, we analyzed the correlation between fatty acid composition and flavor profiles. The floral scent is significantly positively correlated with multiple fatty acids, myristic acid (r = 0.3364, P = 0.047), trans-oleic acid (r = 0.3477, P = 0.036), trans-vaccenic acid (r = 0.4671, P = 0.041), cis-6-octadecenoic acid (r = 0.2335, P = 0.034), and eicosadienoic acid (r = 0.3344, P = 0.022) (Fig. 5B, and Tab. S6). The correlation heatmap between fatty acids and amino acids reveals both antagonistic and complementary relationships (Fig. 5C), underscoring the importance of considering these interactions when planning a balanced diet. The hypo−/hypercholesterolemic (h/H) ratio reflects the balance between hypocholesterolemic and hypercholesterolemic fatty acids. We found that the LF group had a higher level than the HF group (Fig. 5D). Meanwhile, we evaluated the lipid quality indices, namely the Atherogenic Index (AI) and the Thrombogenic Index (TI), which provide a comprehensive measure of the impact of fatty acid composition on cardiovascular health. The results demonstrated that the HF group performed significantly better than the LF group, with a 31% lower AI and a 48% lower TI (Fig. 5E–F). Furthermore, by comparing the EAA profiles against human nutritional requirements (Tormasi, Benes, Konya, Berki, & Abranko, 2025), we found that the HF goat meat provided a more favorable essential amino acid profile relative to human nutritional requirements (Fig. 5G). Therefore, high-IMF goat meat exhibited more favorable flavor-related and nutritional characteristics than low-IMF goat meat. Nutritional value of marbled goat meat and its relevance to human dietary needs (A-B) Correlation heatmap between amino acids (A), metabolites (B) and sensory attributes. Color indicates Pearson's r; dotted lines show significant correlations (P < 0.05). (C) Bar plot of h/H in LF and HF groups. h/H = (C18:1 N9C + C18:2 N6 + C20:4 N6 + C18:3 N3 + C20:5 N3 + C22:6 N3)/(C14:0 + C16:0). (D) Bar plot of AI in LF and HF groups. AI = (C12:0 + 4 × C14:0 + C16:0)/(∑MUFA+∑PUFA). (E) Bar plot of TI in LF and HF groups. n3 = (C20:3 N3 + C22:6 N3 + C18:3 N3 + C22:5 N3 + C20:5 N3). n6 = (C18:2 N6 + C18:3 N6 + C20:3 N6 + C18:2 N6T + C20:4 N6 + C22:5 N6). TI = (C14:0 + C16:0 + C18:0)/(0.5 × ∑MUFA +0.5 × ∑n6 PUFA +3 × ∑n3 PUFA + (n3/n6)) (F) Heatmap between DEAs and DEMs levels. Color indicates z-score. (G) Radar plot comparing amino acid levels in LF, HF, and reference (human nutritional requirements) groups. Statistical significance: * for P < 0.05, ** for P < 0.01. Data are presented as mean ± SEM. Discussion IMF deposition is a key determinant of meat quality, yet its dynamic regulation across developmental stages and its integration with metabolic networks remain incompletely understood in goats. In this study, we systematically characterized age-dependent changes in meat quality traits in Haimen goats and, through multi-omics integration, revealed a coordinated metabolic framework underlying IMF accretion and flavor formation. Our results demonstrate that meat quality traits exhibit pronounced age-dependent variation between 6 and 36 months (Table 1, Table 2). The progressive increase in shear force with age is consistent with enhanced collagen cross-linking and reduced water-holding capacity, reflecting structural remodeling of muscle tissue during maturation (Rosa et al., 2018; Zhang et al., 2025). Meanwhile, IMF content showed a marked and continuous increase, accompanied by a strong inverse correlation with moisture content. This shift suggests a gradual replacement of water by lipid components within the muscle matrix, thereby altering tissue compactness and contributing to improved juiciness and mouthfeel (Li et al., 2024). Changes in color parameters (L* and a*) and the elevated ultimate pH observed at 18 months further indicate that postmortem metabolic processes, particularly glycolytic rate and redox status, may vary across developmental stages (Li et al., 2024). These findings collectively point to a coordinated interplay between structural properties and energy metabolism in shaping age-related meat quality. Age-related differences may alter the extent of postmortem glycolysis and lactate accumulation, thereby influencing muscle pigment stability and contributing to variations in meat color (Shen et al., 2025). Numerous comparative studies have shown that sex-related hormonal effects contribute to differences in the physicochemical properties and meat quality traits of livestock and poultry (Hao, Xie, & Li, 2025; Hoffman et al., 2022; Wegner et al., 2025). Female animals generally exhibit a greater propensity for lipid deposition than males, which is closely associated with differences in endocrine regulation and adipocyte metabolic activity (Abildgaard et al., 2025). This difference may be attributed to the influence of hormonal profiles in females on the lipolysis and synthesis capabilities of fat cells (Hu et al., 2024; Malnick, Somin, & Goland, 2013; Martinez de Morentin et al., 2014). Moreover, IMF content typically increases with advancing developmental or breeding stages in livestock and poultry, reflecting progressive metabolic adaptation and lipid accretion during growth (Bosch, Tor, Reixach, & Estany, 2012; Shao, Lu, Sun, Jiang, & Chen, 2024; Wang et al., 2025). Therefore, these findings suggest that age and sex-dependent metabolic regulation may jointly contribute to IMF accumulation and meat quality variation in Haimen goats. At the compositional level, comparative analysis between HF and LF groups revealed that IMF deposition is not merely a quantitative increase in lipid content but is associated with qualitative remodeling of metabolic profiles. The enrichment of saturated and monounsaturated fatty acids, particularly C18:1 (oleic acid), in the HF group suggests enhanced lipogenic activity and selective fatty acid accumulation. These lipid changes are known to influence flavor perception by modulating lipid oxidation and Maillard reaction pathways during thermal processing (Du et al., 2025). In addition, the HF group exhibited a more favorable fatty acid profile, with PUFA:SFA and n6:n3 ratios approaching recommended nutritional ranges, indicating that IMF enrichment may simultaneously improve both sensory and nutritional attributes (Albano-Gaglio et al., 2025; Giblin et al., 2025; Queathem et al., 2025). Beyond lipid metabolism, amino acid profiling revealed coordinated alterations linked to IMF deposition. Positive correlations between IMF content and amino acids such as arginine, proline, and serine suggest a coupling between nitrogen metabolism and lipid accumulation (Yang et al., 2025). Notably, aspartic acid and asparagine, which serve as key precursors for flavor-related Maillard reactions, were closely associated with flavor-related metabolites, indicating that IMF-rich muscle may possess an enhanced capacity for flavor generation (Luo et al., 2024; Mesias, Holgado, & Morales, 2023; Sa & House, 2024; Zhang, Zhang, Zhang, & Wang, 2026). These results highlight a previously underappreciated interaction between amino acid metabolism and lipid deposition in determining meat flavor complexity. Importantly, multi-omics integration identified a set of key metabolic genes that converge on central energy pathways. Rather than acting independently, these genes are functionally enriched in mitochondrial β-oxidation, the tricarboxylic acid (TCA) cycle, and acetyl-CoA metabolism. Enzymes such as Hydroxyacyl-Coenzyme A Dehydrogenase (HADH) and Acyl-CoA Dehydrogenase 8 (ACAD8) are critical for fatty acid β-oxidation, while Aconitase 2 (ACO2) and succinyl-CoA ligase subunits (SUCLA2/SUCLG2) participate in TCA cycle flux, collectively regulating energy production and substrate availability. Within this network, acetyl-CoA emerges as a central metabolic hub linking carbohydrate, lipid, and amino acid metabolism (Ma et al., 2025; Wang et al., 2023; Zhang et al., 2025). Its accumulation not only provides substrates for de novo lipogenesis but also integrates multiple metabolic signals, thereby coordinating IMF deposition (Fu et al., 2023; Wang et al., 2024; Yamauchi et al., 2024). This systems-level perspective supports the notion that IMF accretion is driven by metabolic reprogramming rather than isolated gene effects. Furthermore, protein glycosylation modification acts as a biomarker of skeletal muscle aging and also adversely affects meat quality (Kim et al., 2025). We also compared protein glycosylation modification levels between the two groups and found that the HF group exhibited slightly higher glycosylation levels than the LF group (Tab. 5). Although the mechanism by which glycosylation regulates nutrient absorption in humans have not yet been fully elucidated, future studies should aim to clarify its functional impact through well-designed dietary interventions. Table 5Differences in glycosylation modification levels.NamepeptideIdentified proteinGlycans peptide1-year-old_N8965104081-year-old _O89731042153-year-old _N9529119673-year-old _O9535119311 Differences in glycosylation modification levels. Collectively, our findings support a unified model in which age-dependent metabolic reprogramming enhances acetyl-CoA availability, promotes lipid deposition, and coordinates amino acid metabolism, ultimately driving IMF accumulation and improving meat quality and flavor characteristics. This integrative framework provides new insights into the molecular basis of IMF regulation in goats and offers a foundation for future studies aimed at optimizing meat quality through genetic and nutritional strategies. Conclusion In summary, this study employed integrated lipidomic and amino acidomic analyses to elucidate that IMF deposition rises with feeding age, accompanied by changes in nutritional components and formation mechanisms.. Although no significant difference in flavor was detected between lean and fatty goat muscle tissues, pronounced variations were observed in their amino acid and fatty acid profiles. Specifically, higher contents of both unsaturated and saturated fatty acids were found in high-fat goat meat, which also exhibited the highest level of essential fatty acids. Moreover, high intramuscular fat content was associated with advantageous profiles of both essential and flavor-related amino acids. Proteomic analysis further identified key regulatory genes involved in intramuscular fat deposition, including ECI1, HADH, and ACO2, with particular emphasis being placed on the functional roles of acetyl-CoA and succinyl-CoA. Future integrative multi-omics coupled with CRISPR perturbation in live-animal models will allow causal validation of metabolic checkpoints such as acetyl-CoA in IMF deposition (Fig. 6). Collectively, these findings advance our understanding of the molecular basis underlying intramuscular fat deposition in goats, provide a theoretical basis for precision breeding and nutritional regulation aimed at improving meat quality, and offer scientific support for the production of high-quality, nutrient-balanced goat products. Schematic diagram of the mechanism. CRediT authorship contribution statement Zhipeng Liu: Writing – review & editing, Writing – original draft, Visualization, Validation, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis. Jiafeng Lu: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis. Wangang Zhang: Writing – review & editing, Formal analysis. Shisi Xie: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis. Zhen Zhang: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis. Kaiping Deng: Validation, Methodology, Investigation. Shang Gao: Investigation. Yixuan Fan: Visualization, Investigation, Funding acquisition, Formal analysis. Honghui Lu: Resources, Investigation. Gang Xue: Resources, Investigation. Yanli Zhang: Writing – review & editing, Supervision, Resources, Project administration, Funding acquisition, Conceptualization. Feng Wang: Writing – review & editing, Supervision, Resources, Project administration, Investigation, Funding acquisition, Conceptualization. Funding statement This study was supported by the National Key Research and Development program of China (2021YFD1200902, 2022YFD1300201), Agriculture Research System of China (CARS-38), the PhD Scientific Research and Innovation Foundation of The Education Department of Hainan Province Joint Project of Sanya Yazhou Bay Science and Technology City (HSPHDSRF-2024-09-006). Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments The authors thank all members of Prof. Feng Wang's laboratory for their assistance with sample collection and experimental procedures. We acknowledge the Yangtze River Delta White Goat Conservation and Breeding Research Institute for providing animal resources and technical support. We are grateful to Dr. Chenchen Qiu and Dr. Guodong Ren (Sanya Institute of China Agricultural University) for their assistance with sample analysis. We also acknowledge the Sanya Research Institute of Nanjing Agricultural University, the National Experimental Teaching Demonstration Center for Animal Science at Nanjing Agricultural University, and Yazhou Bay Science and Technology City for providing access to research facilities and technical support. Footnotes Supplementary data to this article can be found online at https://doi.org/10.1016/j.fochx.2026.104117. Supplementary data Supplementary material Supplementary Table S1 Data availability All data generated or analyzed during this study are included in this article and its Supplementary Information Files. References Associated Data Supplementary Materials Supplementary material Supplementary Table S1 Data Availability Statement All data generated or analyzed during this study are included in this article and its Supplementary Information Files.