Integrative transcriptomic and proteomic analyses of different muscles reveal the molecular mechanism of pig psoas major muscle as a high eating and nutritional quality meat

Abstract

Background

Psoas major muscle (PMM) is different from longissimus dorsi muscle (LDM) and semimembranosus (SM) and is a representative red muscle in pig’s carcass. However, scarce researches focus on the mechanism of pork quality of PMM, especially by multi-omics analysis.

Results

In this study, we investigated the mechanisms underlying the meat quality differences between PMM and LDM, SM in Chinese indigenous Saba pigs, integrating physicochemical characteristics, transcriptomics, and proteomics analyses. The results showed that PMM had appropriately lower fat content, better amino acid (AA) compositions, and greater tenderness compared to LDM and SM. Totals of 2,320 differentially expressed genes (DEGs) and 1,562 differentially expressed proteins (DEPs) were identified, of which 22 DEGs/DEPs were co-expression in both PMM vs. LDM and PMM vs. SM groups at the two-omics levels. Some identified AAs and lipids metabolism-related pathways played a crucial role in the meat quality differences among muscles from different parts. Ten genes, including ACADM, ETFB, MYL3, COX5A, CS, HADH, DLD, NDUFS8, NDUFS3, and NDUFAB1 were identified as potential key genes specific to the regulation of synthesis and metabolism of AAs and lipids in skeletal muscle.

Conclusion

These findings provide holistic insights into the molecular mechanisms of pork quality, and serve as beneficial foundation for the genetic improvement of pork quality and pork grading.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-025-12151-2.

Introduction

Pork is one of the most widely consumed meats in the world and serves as a primary source of high-quality protein for humans. With the improvement of living standards, consumers’ attention to pork quality is increasing, which further highlights the significance of improving pork quality. Meat quality traits, such as tenderness, intramuscular fat (IMF) content, and amino acid (AA) composition, crucially determine the eating and nutritional quality of pork [1, 2]. But as we know, improving pork quality through genetic selection is challenging due to its complex trait components and the difficulty of measuring these traits in live pigs using conventional breeding strategies [3]. Therefore, researchers are turning to more effective genetic improvement approaches, such as molecular breeding based on omics research [4, 5].

Research and practice have shown that pork from different anatomical locations exhibited different quality characteristics due to their distinct tissue compositions [6, 7]. Therefore, investigating the differences in pork quality across anatomical locations and elucidating their underlying molecular mechanisms are crucial for pork grading and quality improvement. It is well known that longissimus dorsi muscle (LDM) is widely used in the meat quality-related studies as a representative white muscle in pig’s carcass. However, psoas major muscle (PMM), a representative red muscle and a highly valued, tender, and expensive cut of meat [8, 9], has received relatively little research attention regarding pork quality, particularly the molecular mechanisms underlying its eating and nutritional properties. Additionally, semimembranosus (SM), as a major muscle of the hind limb, has a distinct function compared to LDM, which is primarily involved in postural support, and PMM, which is responsible for lumbar flexion and stabilization. SM has been shown to possess greater oxidative capacity and a higher percentage of type I fibers compared to LDM [10]. Therefore, analyzing the differences among LDM, PMM, and SM is of great importance.

It is widely recognized that Chinese indigenous pig breeds are renowned for their exceptional meat quality, characterized by superior tenderness, juiciness, flavor, higher IMF content, and more favorable AA composition [11, 12], which make them excellent breed materials for meat quality research and high-quality pork production. In recent years, several studies have investigated the differences in meat quality traits, primarily focusing on the comparisons between Chinese indigenous and Western commercial pig breeds. For example, Shaziling pigs have been found to have better meat color, water-holding capacity, tenderness, and IMF content compared to Yorkshire pigs [13]. Similarly, the meat AA components of Guizhou Xiang pigs were found to be superior to those of Large White pigs [12]. However, the molecular mechanisms underlying these superior quality traits of pork are still unclear. Saba pig is a typical Chinese indigenous pig breed with delicious and high nutritional value meat. It was listed as a vital breed for resource conservation by Ministry of Agriculture of China in 2006. But to date, the superior meat quality attributed to Saba pigs remains primarily based on organoleptic evaluation, and lacks comprehensive research and concrete evidence.

With the rapid advancement of high-throughput sequencing technologies and bioinformatics, multi-omics joint analyses have been extensively used to explore the molecular regulatory mechanism of agricultural complicated traits in various animal species [14, 15]. RNA-seq has proven highly effective in identifying key regulatory genes and elucidating the molecular mechanisms underlying meat quality [16, 17]. However, as the functional roles of genes primarily manifest through proteins, relying solely on transcriptomic data is insufficient to fully reveal the biological function of genes [18]. Proteomics, a large-scale investigation of protein structure and function, plays a crucial role in uncovering information relevant to specific biological reactions. It is an indispensable technique for identifying proteins that regulate metabolic pathways associated with meat quality [19, 20]. Integrating transcriptomics and proteomics data provides a more comprehensive understanding of the mechanisms underlying meat quality. In a study by Wei et al. [21] the transcriptomic and proteomic profiles of fast- and slow-muscle in pigs were compared, leading to the identification of key genes that were altered in both transcriptomic and proteomic levels. These genes were found to play pivotal roles in regulating skeletal muscle fiber formation. Similar multi-omics approaches have been extensively employed in other animal species to study meat quality, such as sheep, yaks, and chickens [15, 22, 23]. However, the analysis of meat quality involves complex interactions beyond just differences in the transcriptome and proteome. To fully understand meat quality traits, it is necessary to tightly integrate multi-omics discoveries with physicochemical characteristics. Unfortunately, there is currently a lack of research on integrating physicochemical characteristics with multi-omics techniques in pigs for the assessment of meat quality, especially for muscles from different anatomical locations.

In the present study, we aimed to examine fat content, protein content, shear force, and AA composition in three different muscles (PMM, LDM, and SM) of Saba pigs. Based on the comparison of physicochemical indicators of PMM with LDM and SM, an integrated approach that involved both transcriptomic and proteomic analyses was employed in order to identify specific genes and proteins potentially responsible for the variations in meat quality between PMM and LDM, SM. By elucidating the underlying molecular mechanisms, it would contribute to a better understanding of meat quality variation in Saba pigs and lay a solid foundation for further pork grading and genetic improvement.

Materials and methods

Animals and samples acquisition

The animal experiment was carried out in adherence to the guidelines set forth in the regulations for the Administration of Affairs Concerning Experimental Animals (Ministry of Science and Technology, China, revised in June 2004). The experimental protocol was approved by the Animal Ethics Committee of Yunnan Agricultural University (No: 202310003).

In this study, a total of 30 unrelated purebred Saba pigs (15 males and 15 females) were housed at a national-level Saba pig conservation farm (Chuxiong City, Yunnan Province, China). The pigs were raised from 90 days of age, with an initial weight of approximately 20 kg (average 19.70 ± 3.59 kg). All animals were housed under identical pen conditions (temperature 13–25 °C, humidity 65%−75%, air velocity 0.4–0.7 m/s) and fed the same diet. The animal diets were formulated in two phases: 20–60 kg (crude protein 14.73%, digestible energy 12.97 MJ/kg) and 60–100 kg (crude protein 13.42%, digestible energy 12.84 MJ/kg). The pigs were allowed free access to feed and water until they reached a weight of approximately 100 kg (average 101.50 ± 7.93 kg) at the age of 320 days, after which they were slaughtered. Prior to slaughter, the pigs were fasted overnight but given access to water. The slaughtering process followed the guidelines set by the Chinese Ministry of Science and Technology in 2006. It involved electrical stunning, exsanguination, dehairing, peeling, evisceration, and splitting down the midline. Immediately after slaughtering, three meat samples weighing 10 g, 50 g, and 100 g were sequentially collected from the left half-carcass’s PMM and SM at the mid-section, as well as the LDM between the 1 st to 3rd last thoracic vertebrae. These samples were designated for transcriptomic and proteomic sequencing, nutrient analysis (including total protein and fat content, AA composition), and shear force measurement, respectively. The muscle samples intended for transcriptomic and proteomic sequencing were quickly aliquoted, frozen in liquid nitrogen, and stored at −80 °C until further processing. The samples for nutrient composition and shear force analyses were stored at −20 °C for further testing. For proteomic and transcriptomic analyses, samples of 12 pigs (six of each sex) were randomly selected from the 30 experimental pigs.

Measurement of quality trait and AA indexes of muscle samples

The total protein and fat content, AA composition, and shear force values in PMM, LDM, and SM of the 30 pigs were measured. The total protein content in the muscle was determined using approximately 2 g of meat sample, following the Kjeldahl method as described in National Food Safety Standard of China: Determination of Protein in Food (GB 5009.5–2016) with K9840 semi-auto Kjeldahl Analyzer (Hanon, Jinan, China). Similarly, the total fat content was measured using approximately 5 g of meat samples, in accordance with the acid hydrolysis method specified in the National Food Safety Standard of China: Determination of Fat in Food (GB 5009.6–2016). For each meat sample, the total protein and total fat content were measured twice in valid independent determinations (the deviation between the two results was less than 10%), and the mean value of the two measurements was used as the final content.

The measurement of muscle shear force was conducted according to the method specified in the Determination of Meat Tenderness - Shear Force Method (NY/T 1180–2006). The meat sample with an initial core temperature of 0–4 °C was first heated in a water bath until the core temperature reached 70 °C, followed by natural cooling back to 0–4 °C. Subsequently, a cylindrical sample with a diameter of 1.27 cm and a length of 2.5 cm was excised parallel to the muscle fiber direction using a circular sampler. The shear force value was then measured using a meat tenderness instrument (C-LM4, Beijing, China). Each sample was measured three times with acceptable precision (relative deviation < 10%), and the mean value was calculated.

The concentration of AA in the muscle was quantified using approximately 50 mg of meat sample, following the method described in the National Food Safety Standard of China - Determination of Amino Acids in Food (GB 5009.124–2016), with measurements conducted using the LA8080 automatic AA analyzer (Hitachi, Tokyo, Japan). Each meat sample was independently subjected to two valid determinations (with the difference between the two results not exceeding 12%), and the average of the two determinations was taken as the AAs content of the sample.

RNA extraction and transcriptome analysis

The total RNA was extracted from each muscle sample (0.1 g) using Trizol Reagent (Invitrogen Life Technologies, Carlsbad, CA). The RNA concentration and purity were assessed using NanoDrop NC2000 (Thermo Scientific, Waltham, Massachusetts, USA), and the RNA integrity was assessed using the RNA 6000 Nano Kit of the Agilent 2100 Bioanalyzer system (Agilent Technologies Inc, California, USA) and 1% agarose gel electrophoresis. Only samples with an RNA integrity number (RIN) higher than 8.0 were used for subsequent analysis. The cDNA library construction and sequencing were carried out using the Illumina NovaSeq 6000 platform with paired-end (PE) 150 technology.

The raw reads were processed to obtain clean reads by removing unqualified reads. First, reads with quality scores lower than 20 were filtered out, and adapter sequences at the 3’ end were removed using FASTP (v0.22.0). These clean reads were then aligned to the reference genome (Sus scrofa 11.1) (https://www.ncbi.nlm.nih.gov/) using HISAT2 (v2.1.0). The expression levels of genes were estimated by counting the aligned reads and normalizing them as FPKM. To identify differentially expressed genes (DEGs), the relative expression levels were calculated using the statistical package DEseq2 (v1.38.3).

Protein extraction and proteome analysis

The proteome of muscle samples (1 g per sample) was analyzed using the 4D label-free proteomics method. First, each frozen muscle sample was pulverized into powder in liquid nitrogen. The powder was then dissolved in SDT lysis buffer, and the protein content was quantified using the BCA Protein Assay Kit from Bio-Rad, USA. Subsequently, trypsin was used to digest the proteins in the samples. The resulting peptides were desalted using peptide desalting spin columns and then dried under vacuum. Finally, the concentration of the peptides was determined by measuring the OD₂₈₀ using a NanoDrop One device (Thermo Fisher Scientific, MA, USA) for further analysis.

Liquid chromatography-tandem mass spectrometry (LC-MS/MS) was applied to perform mass spectrometric analysis. Peptides of each sample were analyzed by the Q Exactive™ Plus mass spectrometer equipped with an Easy-nLC 1200 Nanoflow liquid chromatography system (Thermo Fisher Scientific, MA, USA). Data-dependent acquisition (DDA) mode was used for mass spectrometry analysis with a Bruker timsTOF mass spectrometer. The MS data were analyzed for data interpretation and protein annotation against the database with UniProt-Sus scrofa (pig) (https://www.uniprot.org/9823) in MSFragger 3.4 software. The database search results underwent filtering based on a false discovery rate (FDR) threshold of less than 0.01. This filtering occurred at multiple levels, including peptide, peptide-spectral matching, and protein. For label-free quantification analysis, Ion Quantitation (v1.1.0) was employed. Statistical analysis of protein quantitation results was performed using an unpaired t-test. To summarize, the search results were carefully filtered to meet specific criteria, followed by statistical analysis to identify differentially expressed proteins (DEPs).

Weighted gene co-expression network analysis (WGCNA) of proteomic data

The weighted co-expression network was constructed by WGCNA (v3.3.0) R package with the normalized whole dataset according to the protocol to recognize co-expressed genes and proteins [24]. First, the weighted adjacency matrix was created based on the expression matrix. This matrix was then transformed into a topological overlap matrix (TOM) by applying a soft thresholding power β of 0.9. Secondly, according to the TOM dissimilarity, hierarchical clustering was performed to classify proteins with similar co-expression profiles into modules. The parameters used for clustering were a minimum module size of 30 and a merge cut height of 0.25. Subsequently, the Spearman correlation coefficients between module proteins and the phenotypes were calculated. Finally, proteins that showed a significant correlation (P ≤ 0.05) and a Spearman correlation coefficient greater than 0.7 with a particular trait were identified. Hub proteins, characterized by their high connectivity within the module, were presumed to play a pivotal role in its functionality. To elucidate the protein relationships, we constructed the protein-protein interaction (PPI) network of the identified module that exhibited significant associations with muscle parts.

Functional enrichment analysis and the PPI network construction

The GO and KEGG pathway enrichment analyses of DEGs, DEPs, hub genes, and hub proteins were implemented using the clusterProfiler (v3.2.11) package in R software. The single-gene gene set enrichment analysis (GSEA) was performed on the key factors using the clusterProfiler package and GSEABase (v1.58.0) in R software. The PPI network was constructed using the STRING database (https://cn.string-db.org/) and visualized using Cytoscape software (v3.9.1).

Validation of key genes with real-time quantitative PCR (RT-qPCR)

To confirm the RNA-seq results, ten DEGs were randomly selected from the identified DEGs/DEPs for validation by RT-qPCR. The PMM, LDM, and SM of six Saba pigs (three repetitions for each muscle sample) from the same samples as those used for sequencing were used to detect the expression levels of the DEGs. Total RNA from the muscle samples was reversely transcribed into cDNA using PrimeScript™ RT Reagent Kit with gDNA Eraser (Takara, Dalian, China) according to the manufacturer’s instructions. The initial reaction was conducted under the following conditions: 42 °C for 2 min, followed by 4 °C for 18 min. Subsequently, a second reaction was carried out at 37 °C for 15 min, then 85 °C for 5 s, and 4 °C for 18 min. qPCR was performed using TB Green^® Premix Ex Taq™ II (Tli RNaseH Plus) (Takara, Dalian, China) on Mx3000P qPCR System (Agilent Technologies, CA, USA) with the thermal profile: 95 °C for 30 s (initial denaturation); 40 cycles of 95 °C for 5 s and 60 °C for 30 s; dissociation. The gene-specific qPCR primers are listed in Table S1. Each experiment was performed in triplicates, and relative expression of mRNA was calculated through the 2^−ΔΔCt method, the pig GAPDH gene was used as the internal control for normalization.

Parallel reaction monitoring (PRM) analysis

To further validate the 4D label-free results, the identified DEPs were validated using the PRM method. The peptide digests were subjected to chromatographic separation and targeted mass spectrometry analysis utilizing the Easy nLC 1200 (Thermo Scientific, Waltham, MA, USA) and Q Exactive HF-X (Thermo Scientific, Waltham, MA, USA), respectively. Then, the PRM analysis was performed on the mass spectrometry raw files using Skyline (v4.1) software (SCIEX, Toronto, Canada). A |fold change| >1 and P < 0.05 were considered significant.

Statistical analysis

Meat quality traits were shown as mean ± standard deviation (SD). The statistical significance of meat quality traits was assessed using one-way analysis of variance (ANOVA) with Duncan’s multiple comparison test by SAS (v9.2) software. The transcriptome and proteome profile differences of PMM with those of LDM and SM were analyzed by two comparisons, PMM vs. LDM and PMM vs. SM. Genes with |log₂-fold change| >1 and a P < 0.05 were considered as DEGs. Proteins with a fold change > 1.5 or < 1/1.5 and P < 0.05 were considered DEPs. GO and KEGG functional terms with a P < 0.05 were considered significant enrichment. The relative mRNA expression levels between the two muscle samples within a comparison were compared by paired t-test using SAS software, and P < 0.05 was considered significant and P < 0.01 was extremely significant.

Results

Meat quality traits and AA compositions of the muscle samples

The meat quality traits of PMM, LDM and SM are shown in Table 1. The fat content and shear force of PMM were significantly lower than those of LDM (P < 0.01) and SM (P < 0.05). No significant difference was observed in total protein content among the three muscles (P > 0.05).

Table 1.

Meat quality traits of three muscle samples in Saba pigs

Items	PMM	LDM	SM
Number	30	30	30
Total protein content (g/100 g)	22.54 ± 0.94a	22.25 ± 1.33a	21.94 ± 1.22a
Fat content (g/100 g)	5.28 ± 1.77bB	9.12 ± 4.44aA	7.67 ± 3.37aAB
Shear force (N)	3.66 ± 1.44bB	6.62 ± 2.99aA	6.01 ± 1.94aAB

Data in the same row marked with the same superscript letters indicate no significant difference at the given P threshold, while those marked with different superscript letters indicate a significant difference (lowercase and uppercase letters represent the thresholds of P < 0.05 and P < 0.01, respectively). PMM: psoas major muscles; LDM: longissimus dorsi muscles; SM: semimembranosus muscles

As shown in Tables 2 and 17 AAs were detected in the three muscles. Among them, the levels of 12 AAs, Asp, Thr, Ser, Glu, Ala, Cys, Val, Ile, Leu, Phe, Lys, and Arg were significantly higher in PMM compared with LDM (P < 0.01) and SM (P < 0.05 or P < 0.01). Moreover, PMM also exhibited the highest content of essential AAs (EAAs), non essential AAs (NEAAs), flavor AAs (FAAs), and overall total AAs (TAAs). Notably, the EAA/TAA and EAA/NEAA ratios of the three muscles in Saba pigs were all above 40% and 60%, with PMM exhibiting the highest ratios.

Table 2.

The amino acid composition of three muscle samples in Saba pigs

AAs (g/100 g)	PMM	LDM	SM
Aspartic acid (Asp)	1.73 ± 0.07aA	1.54 ± 0.11bB	1.61 ± 0.13bB
Threonine (Thr)	0.87 ± 0.03aA	0.79 ± 0.05bB	0.80 ± 0.07bB
Serine (Ser)	0.69 ± 0.03aA	0.59 ± 0.04bB	0.61 ± 0.06bB
Glutamic acid (Glu)	2.69 ± 0.13aA	2.21 ± 0.21bB	2.34 ± 0.25bB
Glycine (Gly)	0.76 ± 0.03a	0.73 ± 0.06a	0.75 ± 0.11a
Alanine (Ala)	1.06 ± 0.04aA	0.99 ± 0.06bB	1.01 ± 0.09bAB
Cysteine (Cys)	0.18 ± 0.02aA	0.15 ± 0.02bB	0.15 ± 0.02bB
Valine (Val)	0.96 ± 0.04aA	0.84 ± 0.09bB	0.82 ± 0.15bB
Methionine (Met)	0.43 ± 0.03a	0.41 ± 0.08a	0.43 ± 0.03a
Isoleucine (IIe)	0.88 ± 0.04aA	0.75 ± 0.08bB	0.75 ± 0.15bB
Leucine (Leu)	1.58 ± 0.06aA	1.42 ± 0.08bB	1.46 ± 0.12bB
Tyrosine (Tyr)	0.58 ± 0.04cB	0.61 ± 0.06bB	0.66 ± 0.04aA
Phenylalanine (Phe)	0.78 ± 0.03aA	0.72 ± 0.04bB	0.74 ± 0.06bB
Lysine (Lys)	1.70 ± 0.07aA	1.56 ± 0.09bB	1.57 ± 0.15bB
Histidine (His)	0.75 ± 0.05bB	0.80 ± 0.07aAB	0.82 ± 0.06aA
Arginine (Arg)	1.17 ± 0.05aA	1.08 ± 0.07bB	1.10 ± 0.10bAB
Proline (Pro)	0.70 ± 0.04aA	0.64 ± 0.05bB	0.68 ± 0.07aAB
Essential amino acid (EAA)	7.21 ± 0.29aA	6.49 ± 0.40bB	6.57 ± 0.69bB
Non essential amino acids (NEAA)	10.29 ± 0.44aA	9.35 ± 0.70bB	9.72 ± 0.82bAB
Total amino acids (TAA)	17.50 ± 0.73aA	15.84 ± 1.04bB	16.29 ± 1.45bB
Flavor amino acids (FAA)	8.50 ± 0.35aA	7.49 ± 0.54bB	7.79 ± 0.71bB
EAA/NEAA	70.07 ± 0.01aA	69.41 ± 0.03abA	67.59 ± 0.04bA
EAA/TAA	41.20 ± 0.00aA	40.97 ± 0.01aAB	40.33 ± 0.01bB
FAA/TAA	48.57 ± 0.00aA	47.29 ± 0.01bB	47.82 ± 0.01bAB

Transcriptomic analysis

A total of 60.59 Gb of high-quality clean reads were obtained from the transcriptomic analysis. These reads were mapped to the Sus scrofa genome, with mapping ratios of 93.56%, 94.13%, and 94.29% for PMM (1,256,775,890 reads), LDM (1,308,188,927 reads), and SM (1,222,444,507 reads), of which 94.13% (PMM), 92.54% (LDM), and 92.86% (SM) were identified as unique matches to the Sus scrofa genome (Table S2). These results indicated that the sequencing quality was of a high standard, allowing for reliable quantitative analysis. A total of 17,180, 17,011, and 16,814 genes were identified in PMM, LDM, and SM, respectively. Furthermore, 1,462 DEGs (903 up-regulated and 559 down-regulated) in the PMM vs. LDM group and 858 DEGs (553 up-regulated and 305 down-regulated) in the PMM vs. SM group were identified (Figs. 1A–B). Hierarchical clustering of gene expression profiles clearly differentiated the PMM group from both LDM and SM groups (Figs. 1C–D). Four hundred and thirty-seven DEGs were identified as shared in both PMM vs. LDM and PMM vs. SM groups. More detailed information on the DEGs is listed in Table S3.

Fig. 1.

Differentially expressed genes (DEGs) analyses. A Volcano plot of DEGs in PMM vs. LDM comparison. B Volcano plot of DEGs in PMM vs. SM comparison. C Heat map of the DEGs in PMM vs. LDM comparison clustering analysis. D Heat map of the DEGs in PMM vs. SM comparison clustering analysis. E The top ten GO enrichment terms of overlapped DEGs between the two comparison groups. F The top ten KEGG enrichment terms of shared DEGs between the two comparison groups

GO enrichment analysis showed that the 437 shared DEGs were significantly enriched in 274 GO terms (Table S4). These genes primarily involved in biological processes, such as oxidative phosphorylation, respiratory electron transport chain, and myofibril assembly. Regarding cellular components, they were mainly localized to structures including the respirasome, mitochondrial respiratory chain complex I, and the NADH dehydrogenase complex. Additionally, they were associated with molecular functions, including NADH dehydrogenase (ubiquinone) activity, NADH dehydrogenase (quinone) activity, and NADH dehydrogenase activity (Fig. 1E). KEGG pathway analysis identified 33 significantly enriched pathways (Table S4), particularly those related to lipid metabolism, such as PPAR signaling pathway, adipocytokine signaling pathway, and fatty acid (FA) biosynthesis (Fig. 1F).

Proteomics analysis

Totals of 31,223 unique peptides were confidently identified in all muscle samples (Table S5). In total, 3,541, 3,231, and 3,317 proteins were detected in PMM, LDM, and SM, respectively (Table S6). Furthermore, the peptide lengths mainly ranged from 7 to 20 AAs, as shown in Fig. S1A. Data quality assessment demonstrated normal distribution patterns (Fig. S1B) and consistent within-group reproducibility (Fig. S1C). A total of 935 DEPs were identified in the PMM vs. LDM comparison, with 760 up- and 175 down-regulated proteins in PMM (Fig. 2A). In the PMM vs. SM comparison, there were 627 DEPs, consisting of 548 up- and 79 down-regulated proteins in PMM (Fig. 2B). Clustering analysis of the DEPs revealed distinct expression profiles between the PMM vs. LDM and PMM vs. SM groups (Figs. 2C–D). The detailed information about each DEP is listed in Table S7. Moreover, 450 proteins were found to be shared in both comparisons.

Fig. 2.

Differentially expressed proteins (DEPs) analyses. A Volcano plot of DEPs in PMM vs. LDM comparison. B Volcano plot of DEPs in PMM vs. SM comparison. C Heat map of the DEPs in PMM vs. LDM comparison clustering analysis. D Heat map of the DEPs in PMM vs. SM comparison clustering analysis. E The top ten GO enrichment terms of overlapped DEPs between the two comparison groups. F The top ten KEGG enrichment terms of shared DEPs between the two comparison groups

GO enrichment analysis revealed that the 450 shared DEPs were significant association with 305 GO terms (Table S8). The most prominently enriched GO terms included oxidation-reduction process, ATP metabolic process, and cellular respiration (Fig. 2E). Similarly, the shared DEPs were significantly enriched in 48 KEGG pathways (Table S8). Notably, these pathways included several key pathways related to AA and lipid metabolisms, such as 2-oxocarboxylic acid metabolism, valine, leucine and isoleucine degradation, biosynthesis of AAs, oxidative phosphorylation, FA degradation, and FA metabolism (Fig. 2F).

WGCNA, key proteins identification, and analysis of proteomic data

As shown in Fig. 3A, when β = 0.9, the corresponding soft-threshold power = 8, which was used as the criterion to construct a scale-free network. According to the principle of a minimum module size of 30 and a merge cut height of 0.25, five co-expression protein modules were identified (Fig. 3B). Spearman correlation analysis between module eigengenes and muscle samples showed that the blue module (containing 645 proteins) was strongly positively correlation with PMM (r = 0.78, P = 2e-08) (Fig. 3C). Moreover, module membership (MM) in the blue module was highly association with gene significance (GS) (r = 0.7, P < 4.3e-96) (Fig. 3D). Given these results, the blue module was selected as the key module for further analysis.

Fig. 3.

The results of WGCNA. A Scale independence and mean connectivity of various soft-thresholding values. B Cluster dendrogram of all filtered genes enriched based on the dissimilarity measure and the cluster module colors. C The intersection results of three muscles and module eigengene of genes. D Association between the module membership and gene significance within the blue module. E The top ten GO enrichment terms of DEPs in the blue module. F The KEGG enrichment terms related to amino acid and lipid metabolism of DEPs in the blue module. G PPI network based on the shared proteins

GO and KEGG enrichment analyses of the blue module proteins revealed significant enrichment of AAs and lipid metabolism-related pathways (Table S9), such as degradation of valine, leucine, and isoleucine, arginine and proline metabolism, citrate cycle (TCA cycle), FA degradation, and PPAR signaling pathway (Fig. 3E-F). Additionally, 233 overlapping proteins were identified at the intersection of the 450 shared DEPs common to both the PMM vs. LDM and PMM vs. SM comparisons and the proteins within the blue module. Based on the overlapping protein PPI network (Fig. 3G) and degree centrality, the top three hub proteins were identified, including cytochrome c oxidase subunit 5 A (COX5A, degree = 25), cytochrome c1 (CYC1, degree = 23), and NADH: ubiquinone oxidoreductase subunit b3 (NDUFB3, degree = 23). The three hub proteins were also significantly up-regulated in PMM compared to LDM and SM (Table S7). GSEA revealed that COX5A, CYC1, and NDUFB3 were all significantly enriched in energy metabolism and muscle development-related pathways, including oxidative phosphorylation, Rap1 signaling pathway, and Hippo signaling pathway (Fig. S2A-C, Table S10).

The KEGG enrichment analysis in AA and lipid metabolism of deps

Analysis of enriched pathways of DEPs in the comparisons between PMM and LDM, PMM and SM, as well as the overlapping DEPs in both groups, revealed seven common pathways related to AA metabolism. These included: degradation of valine, leucine, and isoleucine, biosynthesis of AAs, butanoate metabolism, cysteine and methionine metabolism, arginine biosynthesis, tryptophan metabolism, and metabolism of alanine, aspartate, and glutamate. Within these pathways, 54 DEPs were identified, all exhibiting higher expression levels in PMM compared to both LDM and SM. Among these DEPs, three DEPs with high degrees were identified by PPI network analysis (Fig. 4A), including citrate synthase (CS) (degree = 32), hydroxyacetyl-CoA dehydrogenase (HADH) (degree = 29), and dihydrolipoamide dehydrogenase (DLD) (degree = 28), which likely played critical roles in the regulation of AA compositions. GSEA further indicated that CS, HADH, and DLD were significantly associated with the regulation of oxidative phosphorylation at high expression levels (Fig. 4C-E, Table S11).

Fig. 4.

The KEGG enrichment analysis DEPs. A PPI analysis of the DEPs of the KEGG enrichment terms related to amino acid metabolism. B PPI analysis of the DEPs of the KEGG enrichment terms related to lipid metabolism. C – E the single-gene GSEA of hub proteins related to amino acid metabolism for (C) CS, (D) HADH, and (E) DLD. F – H, the single-gene GSEA of hub proteins related to lipid metabolism for F NDUFS8, G NDUFS3, and (H) NDUFAB1

Besides, seven common pathways related to lipid metabolism were identified, including thermogenesis, citrate cycle (TCA cycle), FA metabolism, FA degradation, PPAR signaling pathway, FA biosynthesis, and the adipocytokine signaling pathway. These pathways collectively contained 110 DEPs, all exhibiting significantly higher expression in PMM compared to both LDM and SM. Furthermore, based on the DEPs interaction network (Fig. 4B), the top three DEPs with high degree centrality were identified: NADH dehydrogenase (ubiquinone) iron sulfur protein 8 (NDUFS8) (degree = 78), NADH: ubiquinone oxidoreductase core subunit S3 (NDUFS3) (degree = 70), and NADH: ubiquinone oxidoreductase-FAB1 (NDUFAB1) (degree = 70). Additionally, GSEA further revealed positive correlations between the expression of NDUFS8, NDUFS3, and NDUFAB1 and key pathways including apelin signaling pathway, thermogenesis, and oxidative phosphorylation (Fig. 4F-H, Table S11).

Correlation between the transcriptomic and proteomic results

Integrated transcriptome and proteome analyses identified 96 DEGs/DEPs in the PMM vs. LDM comparison. Among these DEGs/DEPs, 94 exhibited concordant expression trends at both omics levels, while two showed discordant trends. Similarly, in the PMM vs. SM comparison, 31 DEGs/DEPs were identified, with 28 showing concordant trends and three showing discordant trends (Table S12). Furthermore, out of the 127 DEGs/DEPs that were expressed in both omics levels, 22 genes/proteins were common in both comparisons. Interestingly, these shared genes/proteins were upregulated in PMM compared to both LDM and SM at both omics levels.

Nine-quadrantal analysis revealed a weak correlation between gene expression in the transcriptomics and proteomics data for both PMM vs. LDM comparison (Fig. 5A) and PMM vs. SM comparison (Fig. 5B), with Pearson correlation coefficients of 0.4515 and 0.2037, respectively. Within quadrants 3 and 7, a synchronous correlation between changes in gene expression and protein levels, indicating coordinated alterations in transcription and translation levels.

Fig. 5.

Correlation analysis of the transcriptomes and proteomes. A correlation between transcripts (y-axis) and proteins (x-axis) in the PMM vs. LDM comparison. B correlation between transcripts (y-axis) and proteins (x-axis) in the PMM vs. SM comparison. C the top ten GO enrichment terms of shared DEGs/DEPs in both comparisons. D the top ten KEGG enrichment terms of shared DEGs/DEPs in both comparisons. E PPI analysis of the DEPs of shared DEGs/DEPs in both comparisons

GO annotation analysis of these shared 22 genes/proteins revealed their involvement in FA oxidation, FA catabolism, and lipid oxidation (Fig. 5C, Table S13). KEGG enrichment analysis further indicated their association with lipid metabolism related pathways such as PPAR signaling, cholesterol metabolism, and apelin signaling pathway. Additionally, pathways related to AA metabolism, including valine, leucine, and isoleucine degradation, were also identified (Fig. 5D, Table S13). These enrichment results suggested that the shared DEGs/DEPs primarily function in AA and lipid metabolism, consistent with the functional analysis of the overall DEPs. PPI network analysis of the shared DEGs/DEPs revealed tight associations, with acyl-CoA dehydrogenase medium chain (ACADM) (degree = 4) emerging as a critical node in the network. Following ACADM, electron transfer flavoprotein subunit beta (ETFB) (degree = 3) and myosin light chain 3 (MYL3) (degree = 3) were also identified as important nodes in the network (Fig. 5E).

Functional enrichment analysis demonstrated that the shared DEGs and DEPs were predominantly enriched in lipid metabolism (PPAR signaling pathway, adipocytokine signaling pathway, and FA biosynthesis), energy metabolism (oxidative phosphorylation), and muscle regulation (cardiac muscle contraction, hypertrophic cardiomyopathy, and dilated cardiomyopathy) related functional metabolic pathways. These findings suggested that these pathways exerted a critical role in the formation of meat quality differences in PMM with LDM and SM at both transcript and protein levels.

Validation of DEGs and deps

Ten randomly selected genes from the shared 22 DEGs/DEPs exhibited consistent expression trends at both mRNA and protein levels in PMM vs. LDM and PMM vs. SM comparisons. As depicted in Fig. 6, except for a few genes (MYBPC1 in PMM vs. LDM, CD36, SLC16A1, and ATP5ME in PMM vs. SM), the majority of genes showed significantly higher relative mRNA expression levels in PMM compared to LDM and SM (P < 0.05 or P < 0.01). The RT-qPCR results were in broad agreement with the RNA-seq data, thus providing further support for the reliability and accuracy of our RNA-seq findings. Similarly, PRM analysis results of the 22 DEPs were largely consistent with 4D label-free (Table S14), further validating the reliability and reproducibility of the proteomics results derived from 4D label-free in our study.

Fig. 6.

RT-qPCR validation of ten co-expressed DEGs in both PMM vs. LDM and PMM vs. SM groups. A in the PMM vs. LDM comparison. B in the PMM vs. SM comparison. (* 0.01 < P ≤ 0.05, ** 0.001 < P ≤ 0.01, *** P ≤ 0.001)

Discussion

Pork quality is a complex trait influenced by various factors, such as genetics, age, sex, nutrition, and muscle parts [25, 26]. In Saba pigs, PMM, LDM, and SM all exhibited high-quality protein and fat content. Notably, PMM demonstrated significantly lower fat content than LDM and SM. It is consistent with the results that the fat content in PMM of Duroc pigs and Chinese Huai pigs is lower than that in LDM [27]. The lower fat content in PMM aligns with the prevailing notion of healthy diets that emphasize low-fat intake [28]. Furthermore, PMM exhibited significantly lower shear force values than LDM and SM, indicating greater tenderness.

As the basic units of protein composition, AAs possess a vital role in promoting animal and human health [29]. The content and proportion of AAs are crucial factors that determine pork nutritional value and taste [2, 30]. Our study revealed that all the three muscles of Saba pigs, PMM, LDM, and SM, contained 17 AAs and offered a rich and complete AA profile, with Glu being the most abundant. EAAs present in pork serve as the primary source of AAs for human consumption [29]. FAAs found in pork, such as Asp, Thr, Ser, Glu, Gly, Ala, and Pro, are directly related to the meat flavor. Their composition and concentration critically determine taste characteristics while contributing to fresh aroma development and reducing undesirable notes (e.g., saltiness, sourness) [31]. Notably, PMM exhibited significantly higher levels of FAAs, EAAs, NEAAs, and TAAs compared to LDM and SM, indicating dual advantages in flavor potential and nutritional quality. Besides, all the three muscles in Saba pigs exhibited AA profiles meeting or exceeding the FAO/WHO thresholds for high-quality protein [32–34]. This highlighted the high nutritional value of Saba pork. Particularly, PMM demonstrated the most favorable ratios (EAA/TAA: 41.20%; EAA/NEAA: 70.07%), suggesting its potential as a premium protein source for targeted nutritional applications. Thus, it can be inferred that PMM might possess a higher nutritional value, more delicious flavor, and greater health benefits for humans.

Integrative analysis of transcriptomic and proteomic data provides powerful insights into molecular mechanisms governing porcine meat quality traits. This approach brings several advantages, such as high resolution, accurate quantification, and extensive coverage [35, 36]. Our study addressed the limited research on multi-omics integration across different parts of muscle by combining meat quality assessment with transcriptome/proteome profiling of PMM, LDM, and SM in Saba pigs. A total of 437 DEGs and 450 DEPs were found to be shared in both PMM vs. LDM and PMM vs. SM groups. They were significantly enriched in pathways related to lipid metabolism, such as PPAR signaling pathway and regulation of lipolysis in adipocytes. Furthermore, integrated multi-omics analysis identified 22 overlapping DEGs/DEPs exhibiting concordant upregulation in PMM at both transcript and protein levels relative to LDM and SM. Among these, three DEGs/DEPs, namely ACADM, ETFB, and MYL3, were identified as potential hub genes associated with meat quality. ACADM, as a member of the ACAD family, plays a pivotal role in the oxidative metabolism process of medium-chain AAs and FAs [37]. Previous research has shown that ACADM was a key gene in the PPAR signaling pathway, a regulatory pathway of lipid metabolism [38]. High expression of ACADM was believed to promote mitochondrial FA oxidation, lipid oxidation, FA transport, and FA decomposition, which were associated with IMF content of pig LDM [39]. Furthermore, a previous report also found that the intragenic variant in ACADM exon 11 led to higher expression of ACADM, which might promote FA oxidation [40]. Notably, higher ACADM mRNA levels were found in muscles of low-fat cattle and pigs, suggesting its role in enhancing lipid metabolism [39, 41]. What’s more, study in humans has shown that ACADM levels were also associated with AA metabolism, making it a key prognostic indicator in colorectal cancer patients [42]. Therefore, ACADM may have a significant impact on the lower fat content and superior AA compositions observed in PMM. Our findings on ACADM mRNA levels, which were linked to FA oxidation and fat deposition, align with the previous reports in Chinese indigenous Tibetan pigs, Diannan Small-Ear pigs [39], and Wannanhua pigs [43], suggesting that this difference might be a conserved feature related to lipid metabolism in skeletal muscle. Further investigation into the specific mechanisms underlying ACADM’s role in regulating AA and lipid metabolism could provide valuable insights into improving the meat eating and nutritional quality in livestock production. ETFB plays a crucial role in the oxidative phosphorylation pathway by accepting electrons from multiple mitochondrial dehydrogenases and transferring them to the primary mitochondrial respiratory chain. It was reported that the upregulation of ETFB expression led to increased oxidative stress in muscle tissues [44, 45]. Furthermore, the expression of ETFB is regulated by PPARα, which is closely associated with FA oxidation and metabolism, suggesting an important role for ETFB in these processes [46]. MYL3, a member of the myosin light chain family, binds calcium ions and is primarily involved in skeletal muscle contraction and development [47]. MYL3 also plays a significant role in modulating fat deposition and muscle fiber structure. Lower MYL3 expression levels were associated with increased IMF deposition in bovine skeletal muscle [48]. Additionally, MYL3 was predominantly found distributed in type I or IIa muscle fibres, making it a potential molecular marker for muscle toughness [49]. In addition, previous research has demonstrated that MYL3 could serve as a biomarker for assessing the extent of oxidative damage during the maturation process of dry-cured ham, primarily due to the susceptibility of its sequence’s Met and Pro residues to oxidation [50]. Therefore, the difference in MYL3 expression levels among the three muscles may underlie the observed variance in shear force (tenderness). In conclusion, the coordinated upregulation of ACADM, ETFB, and MYL3 in PMM might account for the mechanism resulting in its lower lipid content, greater tenderness and better AA composition compared with LDM and SM.

Furthermore, COX5A was also identified as a key protein associated with PMM pork quality by WGCNA of the proteomic data from the three muscles. COX5A, a nuclear-encoded subunit of the terminal oxidase involved in mitochondrial electron transport, plays a crucial role in the oxidative phosphorylation process and energy metabolism. Mitochondrial oxidative phosphorylation is primarily responsible for providing the main energy required for cell life activities [51], which is a key process in the post-mortem conversion of muscle to meat [52]. A recent study has identified a higher abundance of COX5A in more tender meat of cattle [53]. Additionally, COX5A might indirectly modulate the development and functional maintenance of adipocytes by preserving mitochondrial function and cell viability [54]. Moreover, it was reported that PMM exhibited a higher level of oxidation and global phosphorylation than LDM in pigs at postmortem [55]. In conclusion, COX5A may be a latent significant factor associated with meat quality, potentially affecting tenderness and flesh color through the oxidative phosphorylation process after death. However, no specific studies have investigated the molecular regulatory mechanism and function of COX5A on meat quality postmortem.

The production of proteins relies on the efficiency of gene posttranscriptional and posttranslational regulatory mechanisms, with the final product being the result of a variety of post-translational modifications [56]. The field of proteomics is thus more closely aligned with phenotypic characteristics. In our study, we observed that DEPs exhibited broader functional pathway enrichment compared to DEGs, which was generally in agreement with phenotypic characterization. The results of transcriptomic and proteomic analyses suggested that the differences in meat quality between PMM and LDM, SM were primarily attributed to variations in fat content and AA composition. Functional analysis of the DEPs and shared DEPs in the two comparison groups identified that these DEPs were significantly enriched in signal pathways related to AA metabolism (such as biosynthesis of AAs, butanoate metabolism, and cysteine and methionine metabolism) and lipid metabolism (such as FA degradation, PPAR signaling pathway, and FA biosynthesis). Within AA metabolism pathways, the proteins CS, HADH, and DLD were found to play critical roles in regulating AA compositions. CS contributes to meat quality by participating in the terminal oxidative degradation of nutrients [57]. Study has also observed a strong association between CS activities and tenderness, as well as beef flavor [58]. HADH encodes short-chain L-3-HADH and is essential for catalyzing the FA oxidation in mitochondria [42]. DLD, a key mitochondrial oxidoreductase, has a vital role in regulating energy metabolism [59]. Therefore, we speculated that these proteins/genes might affect AA profiles by regulating energy metabolism in skeletal muscle. Of course, these assumptions need further study to validate. For lipid metabolism pathways, three subunits of mitochondrial complex I emerged as key regulators. NDUFS8 is an important subunit of NADH dehydrogenase in mitochondrial complex I. A recent study discovered that NDUFS8 expression was associated with differentiation, self-renewal, and resistance to apoptosis in myoblasts, and was significantly higher in myoblasts derived from slow-type muscle compared to those from fast-type muscle [60]. This may explain PMM’s higher slow-fiber proportion, mitochondrial density, and superior meat quality [61, 62]. NDUFS3 may affect energy metabolism by promoting ATP synthesis and maintaining activity of mitochondrial respiratory chain complexes I and III [63]. NDUFAB1, as a mitochondrial acyl carrier protein, serves a multifaceted role within the mitochondria, such as regulating mitochondrial FA synthesis [64]. In addition, study has discovered that NDUFAB1 might act as a promising mitochondrial target for preventing obesity and insulin resistance by enhancing mitochondrial metabolism [65].

In general, our study highlighted the importance of post-translational modifications and protein regulation in driving meat quality differences across anatomical locations. AA and lipid metabolism and the key regulatory proteins, were likely significant factors contributing to these differences. The analysis of meat quality and genes in the different muscle parts of Saba pigs could not only provide a novel marker-assisted selection (MAS) for muscle-type regulators, but also offer scientific data for carcass grading. The candidate genes identified in this study primarily regulate AA and lipid metabolism through influencing mitochondrial activity, and may synergistically improve the overall quality of pork. However, additional studies are still necessary to clarify the specific mechanism underlying the interaction among genes, proteins, and meat quality. On the one hand, further comparative studies involving Western commercial breeds and other Chinese indigenous pig breeds are warranted to validate the uniqueness and universality of the differences observed in this study. On the other hand, functional validations are required to verify how changes in genes and metabolic pathways contribute to the variation. Besides, the potential benefits of applying these candidate genes for MAS to improve skeletal muscles quality still need further research and validation.

Conclusion

This study found that the edible quality and nutritional value of PMM were superior to those of LDM and SM in Saba pigs, which was characterized by appropriate IMF content, higher content of FAAs, EAAs, higher EAA/TAA and EAA/NEAA ratios, and better tenderness. Integrated transcriptomic and proteomic analysis revealed the molecular basis for these quality differences. Ten essential genes and proteins associated with pork quality were identified, including ACADM, ETFB, MYL3, COX5A, CS, HADH, DLD, NDUFS8, NDUFS3, and NDUFAB1, which exerted significant regulatory effects on the synthesis and metabolism of AAs and lipids. Our results highlighted the significant influence of AA biosynthesis, butanoate metabolism, cysteine and methionine metabolism, FA oxidation, FA catabolism, and the PPAR signaling pathway on pork quality. These findings offered valuable insights into the molecular mechanisms governing pork quality complexity, laying a foundation for pig breeding programs focused on meat quality improvement and pork grading system enhancement.

Abbreviations

AA

Amino acid

DEGs

Differentially expressed genes

DEPs

Differentially expressed proteins

EAAs

Essential amino acids

FA

Fatty acid

FAAs

Flavor amino acids

GSEA

Gene set enrichment analysis

IMF

Intramuscular fat

LDM

Longissimus dorsi muscle

MAS

Marker-assisted selection

NEAA

Non essential amino acids

PMM

Psoas major muscle

PPI

Protein-protein interaction

PRM

Parallel reaction monitoring

RT-qPCR

Real-time quantitative PCR

SM

Semimembranosus

TAAs

Total amino acids

WGCNA

Weighted gene co-expression network analysis

Authors’ contributions

YL: Writing – review & editing, Writing – original draft, Methodology, Investigation, Data curation, Conceptualization, Validation, Formal analysis, Visualization. YX: Writing – review & editing, Writing–original draft, Methodology, Investigation, Data curation, Validation, Formal analysis. XY: Writing – review & editing, Investigation, Validation, Formal analysis, Visualization. LX: Writing – review & editing, Investigation, Validation, Formal analysis. CG: Writing – review & editing, Formal analysis, Investigation, Visualization. JH: Writing – review & editing, Investigation, Validation, Visualization. SY: Writing – review & editing, Investigation, Validation, Visualization. QC: Writing – review & editing, Formal analysis, Visualization, Investigation. ML: Writing – review & editing, Conceptualization, Methodology, Investigation, Validation, Supervision, Resources, Formal analysis. SX: Writing – review & editing, Conceptualization, Methodology, Investigation, Supervision, Funding acquisition, Project administration, Resources.

Funding

This work was supported by Yunnan Swine Industry Technology System Program (2023KJTX016), and Yunnan Province Important National Science & Technology Specific Project (YNWR-CYJS-2018-056).

Data availability

The RNA-Seq raw data and proteomic data have been uploaded to the NCBI Sequence Read Archive (SRA) and iProX database, with the accession numbers PRJNA1223630 ([https://www.ncbi.nlm.nih.gov/sra/?term=PRJNA1223630](https:/www.ncbi.nlm.nih.gov/sra/?term=PRJNA1223630)) and PXD060800 ([https://www.iprox.cn//page/SCV017.html? query=PXD060800](https:/www.iprox.cn/page/SCV017.html? query=PXD060800)), respectively.

Declarations

Ethics approval and consent to participate

The animal trial procedures in our study were approved by the Animal Ethics Committee of Yunnan Agricultural University (No: 202310003).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.