Comparative Transcriptomic and Metabolomic Profiling of Ovaries from Two Pig Breeds with Contrasting Reproductive Phenotype

Abstract

Although numerous quantitative trait loci (QTLs) and candidate genes have been implicated in litter size in certain pig breeds, the genetic basis underlying the pronounced differences in reproductive capacity among breeds remains incompletely understood. To elucidate the underlying mechanisms responsible for the heterogeneity in reproductive capacity, we performed integrated transcriptomic and metabolomic analyses on ovarian tissues from two pig breeds with contrasting litter sizes: Diannan Small-ear (DSE) pigs and Yorkshire (YK) pigs. YK pigs exhibited significantly higher total born piglets. Transcriptome analysis revealed that upregulated DEGs in YK ovaries were enriched in ovarian steroidogenesis, retinol metabolism, vitamin digestion/absorption, and folate biosynthesis. In contrast, DSE pigs showed enrichment in inflammatory and immune-related pathways. Furthermore, integrative transcriptomic and metabolomic analysis revealed that upregulated DEGs in YK pigs, such as STAR and COL3A1, and concurrently elevated metabolites (e.g., L-threonine, L-asparagine, L-proline, L-methionine, arachidonic acid, and progesterone) were jointly enriched in three key pathways: protein digestion and absorption, mineral absorption, and aldosterone synthesis and secretion. These genes and metabolites are implicated in granulosa cell and oocyte proliferation, maturation, and protection against oxidative damage. Our findings provide a theoretical foundation for future strategies aimed at improving reproductive performance through targeted modulation of key genes and metabolites.

1. Introduction

In the global livestock industry, pork remains one of the primary sources of dietary protein worldwide, and sow productivity is closely linked to the profitability and efficiency of large-scale pig farming operations [1]. Reproductive performance, particularly litter size, is a complex phenotype shaped by a combination of genetic and environmental factors [2]. Numerous genes have been implicated in regulating reproductive capacity [3]. As the primary female reproductive organ, the ovary is fundamental to sexual reproduction, producing oocytes, secreting key hormones, and thereby regulating both reproductive function and endocrine balance to support ovulation, fertility, and overall reproductive health [4,5]. The oocyte-secreted members of the TGF-β superfamily, such as bone morphogenetic proteins (BMPs), glial cell-derived neurotrophic factor (GDNF), and activin/inhibin, have been shown to play crucial roles in regulating follicular development and facilitating cell–cell communication within the follicle [6]. Consequently, variations in the expression patterns of critical genes and changes in metabolite levels within the ovary may help explain the differences in litter size observed across pig breeds.

Earlier studies using genome-wide association studies (GWAS) and whole-genome sequencing (WGS) have primarily focused on identifying quantitative trait loci (QTLs) and candidate genes associated with pig litter size, leveraging advances in molecular technologies to elucidate the genetic basis of this trait [7]. Several fecundity-related genes, such as GRID2, PALB2, SOX9, and GDF9, have been reported [8,9]. Moreover, ovarian transcriptomic profiling of pig breeds with contrasting litter size performance (Tibetan vs. Yorkshire) has revealed numerous candidate genes (CYP11A1, INHBA, ARRB2, HSD17B1, MIF, and SCARB1) potentially influencing reproductive performance [10]. Using ovaries from Xiang pigs with high and low reproductive performance, transcriptomic analysis revealed that STAR, COX3, HSD3B, SCARB1, and CYP11A1 were also associated with litter size in this breed [11]. Moreover, metabolomic profiling identified androstenedione, dehydroepiandrosterone, progesterone, and carnosine as key metabolites playing critical roles in the ovaries of Chenghua pigs across different developmental stages [12]. However, integrated transcriptomic and metabolomic analyses to elucidate ovarian functional variation across pig breeds are still scarce. Here, we systematically integrated transcriptomic and metabolomic analyses to identify potential candidate genes and metabolites underlying the divergent reproductive phenotypes between the two pig breeds. This integrative approach holds significant promise for uncovering the key drivers of these differences and advancing swine reproductive health.

The DSE pig is an indigenous Chinese breed originating from Xishuangbanna, Yunnan Province. It is characterized by its small stature, early sexual maturity, strong environmental adaptability, and superior meat quality [13]. In terms of reproductive performance, however, DSE pig populations exhibit considerable variation in litter size, with an average of 4.95 ± 2.13 piglets born per litter [14]. In contrast, YK pigs demonstrate significantly higher prolificacy, averaging 12.07 ± 3.40 piglets per litter [15]. This disparity is likely attributed to breed-specific regulatory mechanisms in the expression of genes associated with litter size. To identify key candidate genes underlying this trait, ovarian tissues were collected from both YK and DSE pigs exhibiting high and low litter sizes, respectively. Transcriptomic and metabolomic profiles were comprehensively analyzed using RNA-seq and untargeted metabolomics, respectively. These integrated data provide valuable insights into the molecular mechanisms underlying the divergent reproductive capacities and litter sizes observed across pig breeds.

2. Materials and Methods

2.1. Experimental Animals and Sample Collection

DSE and YK pigs were selected as experimental animals in December 2024. All animals were housed under standardized, controlled conditions. Sows were (non-pregnant stage) individually confined in standard stalls and maintained at a constant temperature of 22–25 °C. The dietary formulation is presented in Table S1. Total number born (TNB) records from the third parity were collected from a total of 6 DSE sows and 6 YK sows. Following the comparison of litter size differences, sows were humanely euthanized by electrical stunning followed by rapid exsanguination. Ovarian tissues were immediately harvested and snap-frozen in liquid nitrogen. These samples were subsequently used for histological examination, transcriptomic sequencing, and metabolomic profiling. All experimental procedures were conducted in strict accordance with the ethical guidelines approved by the Ethics Committee of Hunan Agricultural University (Approval No. 541234-12). All experimental procedures involving live animals were conducted in strict compliance with international standards for animal welfare, including the ARRIVE guidelines (https://arriveguidelines.org/) (accessed on 27 December 2024) for the reporting of experiments using live animals.

2.2. Hematoxylin and Eosin (HE) Staining

Ovary tissues from sows were fixed in 4% buffered paraformaldehyde (Biosharp, Hefei, China) for more than 24 h. Following fixation, samples were dehydrated through a graded ethanol series using a Leica ASP200S (Wetzlar, Germany) automated tissue processor. The dehydrated tissues were then embedded in paraffin, and 4 μm-thick sections were cut with a YD-355AT rotary microtome (Jinhua, China). Sections were subsequently stained with hematoxylin and eosin (H&E). Histological examination was carried out under a Zeiss light microscope (ZEISS, Oberkochen, Germany), and morphometric analysis was performed using Image-Pro Plus software (version 6.0). Follicles were classified based on morphological features as described by previous study [12]. Secondary follicles in three sections of each ovary were counted [16].

2.3. Enzyme-Linked Immunosorbent Assay (ELISA)

Serum estradiol (E2) and follicle-stimulating hormone (FSH) concentrations were quantified according to the manufacturer’s instructions. Hormone concentrations were measured using commercial ELISA kits for porcine E2 (CUSABIO, Shanghai, China) and FSH (CUSABIO, Shanghai, China) Among them, the detection range of E2 is 40–1000 pg/mL, and the coefficient of variation CV is less than 15%; the detection range of FSH is 20–700 mIU/mL, and the coefficient of variation CV is less than 15%. And the R-values for the standard curves were greater than 0.99 for these assays. Serum samples were diluted twofold, and technical replicates were performed in triplicate for each sample.

2.4. The Process of RNA Extraction

Total RNA was isolated from ovarian tissues using TRIzol^® reagent (GlpBio, Montclair, CA, USA). Approximately 50 mg of tissue was processed for each RNA extraction. RNA integrity and concentration were rigorously assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA). Only samples meeting the quality criteria (total RNA ≥ 1 μg and concentration ≥ 30 ng/μL) were used for library construction. Strand-specific sequencing libraries were prepared with the Illumina^® Stranded mRNA Prep Kit following the manufacturer’s protocol and subsequently sequenced on an Illumina NovaSeq 6000 (San Diego, CA, USA) platform in paired-end mode (2 × 150 bp).

2.5. Analysis of Sequencing Data

Raw sequencing data were subjected to quality control to remove reads containing adapter sequences or low-quality bases, ensuring the reliability of downstream analyses. Adapter trimming and quality filtering were performed using Trim Galore (v0.4.1) [17]. Read quality metrics before and after filtering were generated with fastp (v0.23.4) [18]. Clean reads were then aligned to the Sus scrofa reference genome using HISAT2 (v2.1.0) [19]. To minimize noise from lowly expressed genes, we implemented a filtering criterion to exclude genes with very low expression levels. Specifically, only genes with at least 5 counts in 50% of the samples were retained for downstream analysis. Gene-level expression was quantified as transcripts per million (TPM), with read counts obtained via FeatureCounts (v2.0) [20]. Principal component analysis (PCA) and heatmap visualization were performed using the “ggplot2 (v2.0)” and “ComplexHeatmap (v2.15.1)” packages in R (v4.2.3), respectively. KEGG pathway enrichment was performed with KOBAS (v3.0).

2.6. Non-Targeted Metabolic Profiling of Ovary Samples

Metabolite extraction was carried out following established protocols [21,22]. Briefly, approximately 50 mg of ovarian tissue was accurately weighed and homogenized in 400 µL of ice-cold methanol: water containing L-2-chlorophenylalanine as an internal standard. Samples were incubated at −10 °C, then homogenized using a high-throughput tissue grinder (Wonbio-96C, Wanbo, Shanghai, China) at 50 Hz for 7 min, followed by ultrasonication at 50 kHz for 25 min at 5 °C. Protein precipitation was achieved by storing the samples at −20 °C for 25 min. After centrifugation at 12,000× g for 15 min at 4 °C, the supernatants were carefully collected and transferred into autosampler vials for LC-MS/MS analysis.

Chromatographic separation was performed on a Thermo Fisher Scientific UHPLC-Q Exactive HF-X system (Waltham, MA, USA). A 2–3 µL aliquot of supernatant was injected onto an HSS T3 column (100 mm × 2.1 mm, 1.9 µm; Waters) maintained at 40 °C. The mobile phase consisted of (A) 0.1% formic acid in water and (B) 0.1% formic acid in acetonitrile, delivered at a flow rate of 0.5 mL/min. All samples were kept at 4 °C in the autosampler throughout the analytical run. Metabolite detection was conducted in both positive and negative electrospray ionization (ESI) modes, with spray voltages set to 3.5 kV (positive) and −2.5 kV (negative).

Metabolite identification was based on accurate mass (m/z), retention index (RI), and MS/MS fragmentation patterns, cross-referenced against the Human Metabolome Database (HMDB; http://www.hmdb.ca/) (accessed on 18 June 2025), METLIN (https://metlin.scripps.edu/) (accessed on 18 June 2025), and the Majorbio database. OPLS-DA score plots were generated using the “ropls (v1.28.2)” R package.

For multi-omics integration, KEGG pathway enrichment analyses of DEGs and SDMs were first performed independently. Overlapping pathways enriched in both datasets were identified, and spearman correlation analyses were conducted between DEGs and SDMs co-annotated to the same pathways. Associations between DEGs and SDMs were assessed using Mantel tests implemented in the “linkET (v1.0.5)” R package.

2.7. Quantitative Real-Time PCR (qRT-PCR) Validation

To confirm the RNA-seq results, qRT-PCR was performed on a Roche LightCycler^® 480 II system (Roche, Basel, Switzerland). Total RNA was isolated using the protocol described above and reverse-transcribed into cDNA. Gene expression levels were quantified using the 2^−ΔΔCt method, with β-actin serving as the reference gene for normalization. Each biological replicate was analyzed in technical triplicate, and all primer sequences are provided in Table S2.

2.8. Statistical Analysis

Differences between the two groups were assessed using an unpaired Student’s t-test. Prior to analysis, the normality of data distribution was assessed using the Shapiro–Wilk test, and homogeneity of variances was evaluated by Levene’s test. Results are expressed as mean ± standard deviation (SD). DEGs were identified using DESeq2 (v1.40.2), using a significance threshold of |log2 fold change| ≥ 2 and a false discovery rate (FDR) < 0.05. Significantly differential metabolites (SDMs) were defined by a variable importance in projection (VIP) score > 1 (from OPLS-DA) and a p-value < 0.05 (Student’s t-test). A significance threshold of p < 0.05 was applied for functional enrichment analyses of both DEGs and SDMs. Correlation between DEGs and SDMs was considered statistically significant if the absolute spearman correlation coefficient exceeded 0.4 (|r| > 0.4) and the p-value was less than 0.05. Data visualization was performed using R (v4.2.3) and GraphPad Prism (v9.3).

3. Results

3.1. Differences in Reproduction Performance and Overview of Transcriptomic Data

By comparing litter size, we found that YK pigs (12.6 ± 2.9) had a significantly higher total number of piglets compared to DSE pigs (8.5 ± 0.9) (Figure 1A, p < 0.05). Subsequently, we performed H&E staining on ovarian sections to examine morphological differences. H&E staining revealed that ovaries from YK pigs contained numerous follicles (primary follicles and secondary follicles), whereas ovaries from DSE pigs exhibited only progressively enlarging antral follicles, with fewer primary and secondary follicles (Figure 1B). Moreover, we found that E2 and FSH levels were significantly higher in YK pigs than in DSE pigs (Figure 1C,D, p < 0.05). Histological analysis of HE-stained ovarian sections revealed that the proportion of normal SFs was significantly greater in YK sows compared to DSE sows (Figure 1E, p < 0.05).

RNA sequencing was performed to uncover key functional genes and signaling pathways in the ovaries of the two pig breeds. Following adapter trimming, removal of low-quality reads, and exclusion of reads containing ambiguous bases (N), we obtained between 37,994,326 and 45,416,260 high-quality clean reads across the 12 RNA libraries (Table S3). The average Phred quality score (Q30) for the sequencing data was 96.91 ± 0.06%, and the GC content of all samples fell within the typical range of 49% to 52%, collectively indicating high reliability and quality of the sequencing results.

3.2. Identification and Functional Analysis of DEGs in the Ovary Between Two Pig Breeds

To investigate molecular differences underlying ovarian function between DSE and YK pigs, we identified differentially expressed genes (DEGs) in ovarian tissues. A total of 35,682 expressed genes were retained for downstream analysis. As shown in Figure 2A, the first two principal components (PC1 and PC2) accounted for 47.23% and 19.99% of the total variance, respectively, collectively explaining 67.22% of the total variation. Samples from the DSE and YK groups were clearly separated along the PC1 axis, indicating distinct transcriptional profiles between the two groups. Comparative analysis yielded 1341 DEGs, with 852 up-regulated and 489 down-regulated in DSE relative to YK sows (Figure 2B; Table S4). Hierarchical clustering of these DEGs clearly separated the samples into two well-defined groups (Figure 2C). KEGG pathway enrichment analysis was then performed to infer the functional implications of these transcriptional changes in ovarian physiology (Figure 2D). In YK pigs, the 489 down-regulated DEGs were significantly enriched in pathways related to ovarian steroidogenesis (e.g., STAR, LHCGR, DHCR7, TM7SF2, and SC5D), retinol metabolism (CYP2A19, RDH16, RDH11, RDH12, and ALDH1A1), vitamin digestion and absorption (SCARB1, SLC19A3, and SLC5A6), and folate biosynthesis (TPH2, GCH1, and CBR1). Conversely, the 852 up-regulated DEGs in DSE pigs were predominantly associated with inflammation- and immunity-related signaling, including cytokine–cytokine receptor interaction (CCL5, CCL17, IL11, TNFSF13, CD40LG, and GDF15), T cell receptor signaling (CD8A, CD3D, and CD40LG), inflammatory mediator regulation of TRP channels (ADCY8, PLCB2, and KNG1), and Th17 cell differentiation (RXRG, CD3D, and CD3E) (Figure S1). The KEGG pathways enriched in the two breeds differed significantly, reflecting that the up- and down-regulated genes may play key roles in modulating ovarian reproductive performance. Finally, we randomly chose six DEGs for qRT-PCR validation. The expression patterns observed in the qRT-PCR assay aligned well with those from the RNA-seq data (Figure S2), confirming the accuracy and reliability of the transcriptomic findings.

3.3. Identification and Functional Analysis of SDMs in the Ovary Between Two Pig Breeds

To gain deeper insights into the metabolic differences underlying ovarian function between DSE and YK pigs, an untargeted metabolomics analysis was conducted to explore variations in their metabolite profiles. Clear metabolic distinctions between the DSE and YK groups were evident (Figure 3A,B). In total, 397 SDMs were identified, of which 90 were upregulated and 307 were downregulated (Figure 3C; Table S5). The heatmap illustrating the clustering pattern of the top 100 SDMs clearly delineated distinct metabolic profiles between the DSE and YK pigs (Figure 3D). Furthermore, among these metabolites, YK pigs showed a higher relative abundance of organic acids and derivatives, lipids and lipid-like molecules, and organic oxygen compounds compared to DSE pigs (Figure 3E). This suggests that the observed metabolic differences may play a pivotal role in shaping the reproductive phenotype.

Next, KEGG pathway enrichment analysis of the 307 SDMs enriched in YK pigs revealed significant associations with several pathways, including protein digestion and absorption, linoleic acid metabolism, D-amino acid metabolism, mineral absorption, and aldosterone synthesis and secretion (Figure 4A). Notably, three of these pathways, protein digestion and absorption, mineral absorption, and aldosterone synthesis and secretion, were also significantly enriched among the upregulated DEGs in YK pigs, suggesting potential coordinated regulation at both the transcriptomic and metabolomic levels (Figure 4B). Specifically, the SDMs upregulated in YK pigs, namely L-threonine, L-asparagine, L-proline, and L-methionine, were jointly enriched in both the protein digestion and absorption and mineral absorption pathways (Figure 4C). Additionally, arachidonic acid and progesterone were co-enriched in the aldosterone synthesis and secretion pathway. On the other hand, in DSE pigs, the 90 SDMs enriched in this group were primarily annotated to the KEGG pathways of neuroactive ligand-receptor interaction, cushing syndrome, and inflammatory mediator regulation of TRP channels, pathways that also overlapped with the KEGG enrichment results of DEGs in DSE pigs (Figure 4D). Key metabolites contributing to this enrichment include palmitoyl ethanolamide, arachidonoyl ethanolamide, and 17-hydroxyprogesterone (Figure 4E).

3.4. Correlation Analysis Between DEGs and SDMs Enriched in the Same Pathway

Subsequently, we performed a correlation analysis between DEGs and SDMs enriched in the pathways of protein digestion and absorption, mineral absorption, and aldosterone synthesis and secretion. Notably, arachidonic acid exhibited strong positive correlations with multiple genes, including ATP1B1, ATP2B3, COL3A1, STAR, and COL14A1 (Figure 5). Furthermore, COL3A1 showed strong positive correlations with all the aforementioned metabolites except progesterone. In contrast, SLC1A1 was negatively correlated with progesterone, and MME exhibited a negative correlation with L-lysine.

4. Discussion

Investigating ovarian status across different reproductive phenotypes and the underlying regulatory mechanisms will provide deeper insights into strategies for improving ovarian function. In this study, we performed integrated transcriptomic and metabolomic analyses of ovaries from two pig breeds that differ significantly in TNB. By conducting functional enrichment analyses of differentially expressed genes and metabolites, we identified key genes and metabolites potentially contributing to the observed differences in reproductive performance.

Through histological analysis of H&E-stained ovarian sections, we observed that YK pigs exhibited a significantly higher number of follicles compared to DSE pigs. E2, primarily secreted by developing follicles, especially granulosa cells, plays a pivotal role throughout follicular development, from the early to the antral stage [23]. E2 enhances follicular sensitivity to FSH by upregulating FSH receptor expression via a positive feedback mechanism, thereby supporting sustained follicular growth and maturation [24]. Consistent with the histological findings, ELISA results revealed significantly higher circulating E2 levels in YK pigs than in DSE pigs. This suggests that differences in the underlying genetic and metabolic regulatory networks may contribute to the observed disparities in follicular development and ovarian function between the two pig breeds.

Ovarian steroidogenesis, retinol metabolism, vitamin digestion and absorption, and folate biosynthesis are all critical for proper ovarian function. These pathways were significantly enriched in LK pigs, highlighting several candidate genes of particular interest. The transport of cholesterol into mitochondria, mediated by the steroidogenic acute regulatory protein (STAR), is widely regarded as the rate-limiting step in de novo steroid hormone synthesis. Previous work using in situ hybridization demonstrated that STAR mRNA expression is markedly upregulated in theca cells during antral follicle development [25]. Furthermore, studies in cultured porcine theca cells have shown that luteinizing hormone (LH) induces STAR mRNA expression, and this response is potentiated by physiological concentrations of insulin [26]. Regarding retinol metabolism, prior studies have shown that FSH upregulates the expression of ADH1 and ALDH1 genes and elevates total retinoid and retinoic acid levels in mouse ovaries in vivo [27]. Moreover, all-trans retinoic acid (ATRA) and 9-cis retinoic acid (9cRA) have been reported to enhance granulosa cell proliferation, support steroid hormone production, and facilitate oocyte maturation during folliculogenesis [28,29]. Notably, retinoic acid has been shown to induce the expression of the differentiation marker LHR (luteinizing hormone receptor) in mouse ovarian follicles in a dose-dependent manner, thereby promoting both granulosa cell differentiation and oocyte development [30]. Studies on retinol metabolism suggest that the upregulated genes CYP2A19, RDH16, RDH11, RDH12, and ALDH1A1 in the ovaries of YK pigs may play critical roles in follicular development and maturation. For folate biosynthesis, Research has shown that supplementing the in vitro maturation (IVM) medium with folic acid alters the redox balance within porcine oocytes, modulates DNA methylation patterns, and enhances oocyte quality [31,32]. It has been reported that folic acid exerts antioxidant effects by boosting intracellular glutathione levels, thereby improving subsequent embryonic developmental potential. Furthermore, folic acid has been demonstrated to exhibit potent antioxidative activity through suppression of the ERK1/2/NOX4/ROS signaling axis [33]. As previously noted, folic acid is closely associated with improved cytoplasmic maturation of oocytes and a marked reduction in reactive oxygen species (ROS) [34,35]. These findings collectively suggest that folic acid supplementation may enhance oocyte developmental competence, likely by reinforcing antioxidative defenses.

Integrated analysis of upregulated DEGs and SDMs in YK pigs revealed that L-asparagine, L-threonine, L-proline, and L-methionine were co-enriched in both the protein digestion and absorption and mineral absorption pathways. Previous study has reported that L-asparagine, L-arginine, L-serine, gamma-linolenic acid, pentadecanoic acid, and alpha-linolenic acid are closely associated with diminished ovarian response in older patients [36]. L-asparagine is a non-essential amino acid and important for T cell function [37]. It was demonstrated that L-proline supplementation during porcine IVM improves oocyte quality by enhancing mitochondrial content and function [38]. Research has shown that treatment with 0.5 mM L-threonine significantly promotes granulosa cell viability at 0.5, 1, and 2 h post-exposure [39]. The ovarian follicle serves as a fundamental functional unit in mammals, playing a critical role in both steroid hormone production and the development and maturation of the oocyte. Earlier research has indicated that methionine enhances follicular growth and estrogen biosynthesis in rats across the estrous cycle, thereby supporting improved embryo implantation during early pregnancy [40]. However, the functional role of L-asparagine in ovarian physiology in vitro remains underexplored, warranting further investigation into its effects on ovarian cell function.

Notably, we found that arachidonic acid exhibited strong positive correlations with multiple genes, including ATP1B1, ATP2B3, COL3A1, STAR, and COL14A1. Previous research has demonstrated that arachidonic acid specifically modulates the physiological properties, gene expression patterns, and steroid hormone production in bovine granulosa cells [41]. In human study, investigations using ovarian granulosa cells have shown that arachidonic acid does not influence cell proliferation but exerts a protective effect against apoptosis induced by saturated fatty acids [42]. The aforementioned findings suggest that arachidonic acid exerts protective effects on ovarian granulosa cells. However, research on its role in other species remains limited. Future studies could further investigate the specific actions of arachidonic acid on ovarian function across different species. Among the DEGs upregulated in YK pigs, COL3A1 exhibited strong positive correlations with the majority of SDMs within the three co-enriched pathways. Studies have highlighted the involvement of COL3A1 in facilitating cell proliferation and organogenesis, with mutations in this gene linked to developmental delays or severe clinical phenotypes [43,44]. The expression of COL3A1 is regulated by multiple signaling cascades, including the transforming growth factor-beta 1 (TGF-β1) pathway, the Wnt/β-catenin pathway, and the p38 mitogen-activated protein kinase (MAPK) pathway [45,46]. As a key structural component of the extracellular matrix (ECM), COL3A1 contributes to a supportive microenvironment that enhances the survival and proliferation of various cell types. Notably, COL3A1 has been detected in the cytoplasm of ovarian granulosa cells and shown to be positively associated with their proliferative activity [47].

While our study provides a detailed molecular portrait of ovarian function in two genetically distinct pig breeds, DSE and YK, we acknowledge a key limitation that the findings are confined to these two breeds. Future studies should expand this integrative omics approach to include a broader panel of breeds with varying prolificacy to determine whether the identified gene-metabolite networks are conserved or breed specific. Such validation would strengthen the generalizability of our findings and their potential for translational application in selective breeding programs.

This study successfully identified key genes and metabolites involved in ovarian function. Future research should further validate their roles through in vitro cell-based assays, particularly focusing on critical metabolites such as arachidonic acid and the gene COL3A1, to elucidate their specific effects on ovarian somatic and germ cells.

5. Conclusions

Overall, our study reveals that YK and DSE pigs, differing in litter size, exhibit distinct ovarian functions, with pathways such as ovarian steroidogenesis, retinol metabolism, vitamin digestion and absorption, and folate biosynthesis significantly enriched in YK pigs. Integrated analysis identified key DEGs (STAR and COL3A1) and SDMs (L-asparagine, L-threonine, L-proline, L-methionine, and arachidonic acid) co-enriched in protein and mineral absorption pathways, which likely contribute to granulosa cell and oocyte proliferation, maturation, and protection against oxidative damage. These findings offer a potential molecular and metabolic basis for targeted strategies to enhance sow reproductive performance.