Investigating the hub genes and genetic basis regulating fatty acid metabolictraits in chicken breast muscle based on combined genomic and transcriptomic approaches

doi:10.21203/rs.3.rs-4864013/v1

Download PDF

Research Article

Investigating the hub genes and genetic basis regulating fatty acid metabolictraits in chicken breast muscle based on combined genomic and transcriptomic approaches

https://doi.org/10.21203/rs.3.rs-4864013/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background The composition of various fatty acids in meat plays a crucial role in the evaluation of the nutritional indicators of poultry and human health. Fatty acid metabolic indicators can more accurately reflect the nutritional status of meat and its acceptance by the general public. In this study, we utilized two analytical approaches, Genome-Wide Association Study (GWAS) and Weighted Gene Co-expression Network Analysis (WGCNA), to elucidate the genetic mechanisms underlying the fatty acid metabolic traits in chicken breast muscle.Initially, GWAS was conducted based on genotype data obtained by genotyping by sequencing (GBS) from 415 chickens of the Gushi-Anka F2 resource population and phenotypic data of 33 fatty acid metabolism traits in the breast muscle.

Results We identified a total of 291 significant single nucleotide polymorphisms (SNPs) associated with the content of 12 types of fatty acids such as C14:0/C12:0, C18:0/C16:0, C18:1/C18:0, etc., which were mapped on chromosomes (Chr) 1, 2, 3, 4, 5, 7, 9, and 10. The most significant SNP was Chr2:102, 676, 165 (P = 1.20E-06), which explained 2.33% of the phenotypic variation in the polyunsaturated fatty acids (PUFA)/mono-unsaturated fatty acids (MUFA) synthesis pathway. The SNP that explained the highest phenotypic variation was Chr4: 30, 150, 765 (5.85%). Interestingly, significant signals were identified for three fatty acid metabolic traits (C22:4/C18:3, C20:3/C18:3, C22:6/C18:3) within the same gene interval on Chr 9 (Chr9: 5.75 - 8.99Mb). Concurrently, WGCNA was conducted using the 613 candidate genes identified from the linkage disequilibrium regions of the aforementioned significant SNPs, as well as the transcriptome data and fatty acid content phenotypic values of the breast muscle of the Gujshi chicken from three periods, 13 hub genes (FIG4, METTL4, PLD4, IDH1, FH, SDHA, EHHADH, AKT1, PIK3CB, LPIN2, FUT9, PAK2, and ACSL3) were finally determined. Further, we constructed the functional networks involving these candidate genes and deduced the potential fatty acid metabolic pathway.

Conclusion These results enhance our understanding of the genetic regulation mechanisms of fatty acid composition in chicken muscle, providing a theoretical basis for future breeding practices targeting fatty acid composition traits.

Chicken

Breast muscle

Fatty acid metabolic

Genetic structure

GWAS

WGCNA

The ratio of various types of fatty acid composition traits plays a crucial role in assessing the nutritional indicators of poultry meat. The conversion efficiency of α-Linolenic acid, an omega-3 (ω-3) PUFAs, into EPA and DHA is not high, and it is even lower in the elderly [1, 2]. Like ω-3 PUFAs, ω-6 PUFAs also influence the normal growth and development of fetal brain functions and cardiovascular systems. They also serve as precursor substances for regulating many internal biochemical reactions, such as blood pressure and inflammatory responses [3, 4]. Currently, the primary source of these fatty acids is deep-sea fish, which are not easily accessible. Unbalanced diet and unhealthy lifestyle are the main causes of many prevalent diseases. For example, an increased ratio of ω-6 PUFAs to ω-3 PUFAs in the diet has been associated with inflammation and chronic diseases [5]. Moreover, research by Horman T et al. [6] on adult rats demonstrated that different ω-6 /ω-3 PUFAs ratios in the diet could lead to fatty acid remodeling in the brains of adult rats, potentially affecting mental health. DHA/EPA-enriched diet with a ω-6 to ω-3 ratio of 4:1 can, to a certain extent, reverse non-alcoholic fatty liver disease induced by a high-fat diet by increasing the levels of ω-6 and ω-3 PUFAs, reducing the quantity of MUFAs, and lowering the ω-6 to ω-3 ratio [7]. Therefore, in light of the escalating societal demands for improved meat quality, the content and ratio of pivotal fatty acids such as ω-6 and ω-3 PUFAs should be considered key traits for future poultry breeding selection.

Metabolic phenotypes are important surrogate indicators for many biological processes and pathways, and many complex traits are associated with metabolic phenotypes [8]. The metabolic phenotype of an individual can be viewed as the ensemble state of the concentrations of all endogenous small molecules in all body organs and bodily fluids [9]. In studies of both humans and animals, nume rous metabolic traits are used in various omics analysis methods. Zhang et al. [10] conducted a GWAS on 33 fatty acid metabolic traits in five pig populations, revealing four novel population-specific loci and three novel candidate genes. A GWAS of human blood metabolites discovered that longer-chain fatty acids serve a protective role against migraines, higher levels of the lysophosphatidylethanolamine (LPE) have been found to have a causal effect on migraines, leading to the consideration of LPE as a potential therapeutic target for migraine treatment [11]. Research on cancer has shown that some cancer metabolic phenotypes are determined by cancer tissue-of-origin and environment, and it has contributed to a better understanding of the factors that determine tumour nutrient utilisation, which in turn can help to improve cancer treatment [12].

GWAS and WGCNA are two powerful tools that have made remarkable progress in resolving disease resistance, production traits, adaptations, and stress responses in poultry [13–16]. However, with the rapid development of various omics sequencing technologies, researchers are continually refining statistical algorithms to enhance the accuracy and efficiency of detecting association signals [17–19]. By integrating multi-omics data, including genomic, transcriptomic, and metabolomic information, a more comprehensive perspective is provided for the genetic studies of livestock. These advancements can be applied to the development of new breeding technologies and strategies to achieve precise improvement of desirable traits in livestock.

This study conducted GWAS using the metabolic phenotypes of fatty acid composition traits in the Gushi × Anka F2 resource population (F2 resource population), aiming to detect the genetic locis and key genes affecting the variation of fatty acid phenotypes in chicken muscle. Building on this, WGCNA was performed using data on the content of fatty acid composition traits and gene expression levels from the transcriptome of Gushi chicken breast muscle at 14, 22, and 30 weeks of age, in order to detect the phenotypic correlations of fatty acid metabolism traits and hub genes. The ultimate goal is to construct the metabolic pathways and network of fatty acids in chicken muscle. These results enhance our understanding of the genetic regulation mechanisms of fatty acid composition in chicken muscle, providing a theoretical basis for future breeding practices targeting fatty acid composition traits.

Establishment of the Experimental Population

The F2 resource population used for GWAS in this study was established by the Key Laboratory of Livestock and Poultry Resources (Poultry) Evaluation and Utilization, Ministry of Agriculture and Rural Affairs, P.R.China, for the analysis of the genetic mechanisms of superior traits in local chickens and molecular breeding. The resource population consists of 4 cross-bred families (Anka cocks mated with Gushi hens) and 3 reciprocal families (Gushi cocks mated with Anka hens) [20]. Each resource group possesses a comprehensive pedigree and phenotype recordings. The results demonstrated significant segregation in traits such as fatty acid composition, growth, slaughter, and meat quality [21, 22]. All chickens were maintained in cages under uniform environmental conditions with unrestricted access to feed and water, in accordance with the National Research Council's recommendations (1994) [23].

Phenotypic Measurements

Utilizing gas chromatography, phenotypic measurements for 21 distinct fatty acid contents were procured from 439 individuals within the F2 resource group. The categories of Saturated Fatty Acids (SFA), MUFA, and PUFA include 8, 4, and 9 types of fatty acids respectively, Unsaturated Fatty Acids (UFA) is the sum of MUFA and PUFA. We determined the ratios of UFA to SFA (UFA/SFA), PUFA to MUFA (PUFA/MUFA), MUFA to SFA (MUFA/SFA), and PUFA to SFA (PUFA/SFA). Lastly, we calculated all possible products/substrates ratios in fatty acid met abolic pathway, including C14:0/C12:0, C16:0/C14:0, C17:0/C15:0, C18:0/C16:0, C16:1/C16:0, C18:1/C18:0, C20:0/C18:0, C18:1/C16:1, C16:2/C16:1, C18:2/C18:1, C18:2/C16:2, C18:3/C18:2, C20:2/C18:2, C20:3/C18:2, C20:4/C18:2, C20:3/C18:3, C20:4/C18:3, C20:4/C20:2, C20:4/C20:3, C22:3/C20:3, C22:4/C20:3, C22:4/C20:4, C22:4/C22:3, C22:6/C22:4, C22:4/C18:2, C22:4/C18:3, C22:6/C18:3, C22:6/C20:3. Before the association analysis, 33 fatty acid metabolic traits were log₁₀-transformed owing to the relationship of log₁₀ (a/b) = − log₁₀ (b/a), so that the association of a ratio and of its inverse would obtain identical results.

Transcriptome profile and phenotypic data

The mRNA expression data of Gushi chicken breast muscle transcriptome expression profile constructed by our research group were processed and analyzed, with the sample accession of mRNA expression data of PRJNA516810 [24]. Breast muscle tissues from Gushi chickens at 14, 22, and 30 weeks of age were used, with three samples at each age point for transcriptome sequencing. In total, 16,755 known genes were ultimately used for subsequent analysis. The same breast muscle samples as the transcriptome sequencing, were sent to Suzhou PANOMIX Biomedical Tech Co., LTD and performed the fatty acid targeting analysis by Trace 1310-ISQ 7IQS gas-mass spectrometer (Thermo, USA) to obtain the contents of fatty acid composition in breast muscle of Gushi chickens at 14, 22, and 30 weeks of age. Test standard according to the PRC National Standard “Meat and meat products—determination of fatty acids (GB/T 9695.2–2008).” Ultimately, we obtained the content of 49 types of fatty acids in the breast muscle tissue of Guishan chicken. The fatty acids content was expressed as the percentage of the breast muscle by fresh weight (µg/g).

Genotyping, imputation, and quality control

Genomic DNA was extracted from blood samples by the Qiagen DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany). All DNA samples were digested with a combination of EcoRI and MseI for double-digest [25]. A total of 768 chickens from the F2 resource population were genotyped using the Illumina HiSeq X Ten platform (PE150) [26]. Filtered paired reads were aligned to the chicken reference genome Gallus_gallus-6.0 (released in 2018) using Bowtie2 (version 2.3.0) [27]. In order to obtain clean reads, sequencing fragments with more than 50% low-quality bases or more than 5% N bases were removed. The quality control report for the filtered reads was generated by the FastQC software [28]. A total of 7, 258 million clean reads and 6, 071 million good barcode reads were obtained after 768 samples were sequenced.

The quality control procedurres were carried out using VCFtools (version 0.1.13), SNPs were identified using the TASSEL GBS analysis pipeline (version 5.2.31) [29, 30], and SNPs with call rates > 0.30 and minor allele frequencies > 0.05, genotypes with a quality above 98 (minGQ ≥ 98) and depth ≥ 5, consistent with Hardy–Weinberg equilibrium; and max missing rate < 0.40 samples with genotyping were retained for further statistical analyses. we identified 323, 306 SNPs ultimately for GWAS.

Single trait GWAS

Population structure is the main origin of confounding effects in genetic analyses. Principal component analysis (PCA) was performed using genome-wide complex trait analysis (GCTA) software to assess population structure, and the first two principal components (PCs) were then calculated and used as covariates in the mixed model [31]. The variances explained by PC1 and PC2 were 18.7% and 11.3%, respectively. A genomic relationship matrix was constructed with the 323, 306 SNPs using GCTA software and used as random effects in the mixed model.

GWAS of 32 fatty acid metabolic traits was carried out in the GCTA program using a mixed linear model (MLM).

The following MLM were used:

y = Wα + βx + u + e

Where y is the phenotypic value of each trait; W is a matrix of covariates (fixed effects) controlling for population structure (first two PCs), sex, and batch effects; α is a vector containing the corresponding coefficients of the intercept; β is the SNP effect and x is the carrier of the SNP genotype; u is a vector of random effects with a covariance structure, obeying a normal distribution as u ~ N (0, KVg), where K is known of the genetic relationship matrix; e is a random error vector.

The genome-wide significance thresholds were calculated using a valid number of independent SNPs using the Bonferroni correction. Genome-wide independent markers were calculated using PLINK [32] -indep-pairwise with a window size of 25 SNPs, a step size of 5 SNPs, and an r² threshold of 0.1. The genome-wide significance threshold is 1.90E-06 (0.05/26, 430; -log₁₀(P) > 5.72) and the Chrwide suggestive threshold is 3.78E-05 (1/26, 430; -log₁₀(P) > 4.42), based on 26, 430 independent SNPs markers. Manhattan and Q-Q plots were drawn from GWAS results using the CMplot package within the R software.

WGCNA

The Haploview [33] was utilized to determine the significantly associated SNPs within candidate genomic regions linked to fatty acid metabolism traits, resulting in a GWAS candidate gene set. This set was then intersected with transcriptome spectrum data from three age points, leading to the final acquisition of 613 candidate genes for WGCNA by the WGCNA package within the R software [34]. Firstly, the Pearson's correlation coefficient is calculated for each pair of genes, establishing a Pearson's correlation coefficient matrix. The pickSoftThreshold function is used to select the optimal soft threshold for constructing a gene co-expression network. Expression correlation values are used to convert the adjacency matrix into a Topological Overlap Measure (TOM), and hierarchical clustering is performed based on TOM. Ultimately, modules with a minimum gene count of 32 are identified using the dynamic tree cut algorithm. Subsequently, genes within the modules are subjected to Gene Ontology (GO) term and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis using the Clusterprofiler package, where ajusted P-value (P) less than 0.05 are considered to indicate significantly enriched genes [35–38].

WGCNA for fatty acid metabolic traits

Based on the fatty acid metabolic phenotypes used in the GWAS, we selected fatty acids with the same carbon chain length from the targeted analysis data of fatty acid composition in Gushi chicken breast muscle at 14, 22, and 30 weeks of age to obtain metabolic traits, including C14:0/C12:0, C16:0/C14:0, C18:0/C16:0, C20:0/C18:0, C22:0/C20:0, C24:0/C22:0, C16:1/C16:0, C18:1N7/C16:1, C18:1N12/C18:1N7, C18:1N9C/C18:0, C20:1/C18:1N9C, C22:1N9/C20:1, C24:1/C22:1N9, C18:2N6/C18:1N9C, C20:2/C18:2N6, C22:2/C20:2, C20:3N3/C18:3N3, C20:5N3/C20:3N3, C22:5N3/C20:5N3, C22:6N3/C22:5N3, C18:3N6/C18:2N6, C20:3N6/C18:3N6, C20:3N6/C20:2, C20:4N6/C20:3N6, C22:4/C20:4N6, C22:5N6/C22:4, totaling 26 metabolic traits for WGCNA (Table S1).

Module-trait relationships and hub gene identification

We associated the module eigengene (ME) with fatty acid metabolic traits and sought to identify the most significant correlations. Pearson's correlation coefficient to evaluate the relationship between each module and various traits. A correlation between module and trait was considered significant when the P < 0.05. A gene within a specific transcription module was considered to be a hub gene when the eigengene connectivity (KME) value was greater than 0.9 [39]. The genes obtained from the above the principles and whose functional annotations associated with fatty acid composition traits. The gene regulatory network in the key module was drawn by using Cytoscape software [40].

Statistical Results of Fatty Acid Metabolism Phenotypes and Genotypes

Table 1 presents a statistical summary of the phenotypic expressions of 32 different fatty acid metabolism traits within F2 resource population, detailing the number of individuals, mean values, standard deviations, and heritability. It is found that a proportion of the fatty acid composition traits demonstrate moderate heritability (0.2 ~ 0.3), whilst the majority display low genetic heritability (0 ~ 0.2). These observations suggest that the traits of fatty acid composition in chicken muscle are complex quantitative traits subject to multi-genic control with minor effects, highlighting the influence of environmental and nutritional factors acting in conjunction.We further calculated the Pearson correlation coefficients for 24 key fatty acid metabolic traits and 8 additional traits related to the synthesis pathways of ω-3 and ω-6 polyunsaturated fatty acids. Subsequently, the heatmap package within R was utilized for cluster analysis (Fig. 1). Noticeably, 4 metabolic traits, all using the C18 series fatty acids (C18:2, C18:0) as substrates, such as C20:2/C18:2, C20:3/C18:2, C20:0/C18:0, and C18:3/C18:2, exhibited a moderate correlation coefficient (r > 0.48). Within the ω-3 synthesis pathway, metabolic traits using C18:3 as a substrate, such as C20:3/C18:3, C22:6/C18:3, C20:4/C18:3, and C22:4/C18:3, demonstrated a substantial correlation (r > 0.59). Metabolic traits within the ω-6 synthesis pathway likewise presented moderate correlation coefficients (r > 0.50), particularly those involving C18:2 as a substrate, namely C18:3/C18:2, C20:4/C18:2, and C22:4/C18:2. Traits related to the ω-3 and ω-6 synthesis pathways showed stronger correlations with each other. It was revealed that traits involving same products or substrates often demonstrated higher correlations than other traits, emphasizing the intricate interplay within these life-critical metabolic pathways.

Table 1

Summary statistics for fatty acid metabolic traits in breast muscle of the F2 resource population
Trait	N	Mean ± SD	h²	Trait	N	Mean ± SD	h²
C14:0/C12:0	439	7.83 ± 6.05	0.024	C20:4/C18:2	432	0.87 ± 0.75	0.1
C16:0/C14:0	439	69.63 ± 62.47	0.148	C20:3/C18:3	430	9.09 ± 6.14	0.001
C17:0/C15:0	439	2.82 ± 3.17	0.02	C20:4/C18:3	439	12.43 ± 9.35	0.04
C18:0/C16:0	438	1.28 ± 8.42	0.001	C20:4/C20:2	425	23.71 ± 16.28	0.1
C16:1/C16:0	439	0.25 ± 2	0.12	C20:4/C20:3	429	7.06 ± 4.18	0.025
C18:1/C18:0	439	1.73 ± 0.53	0.07	C22:3/C20:3	429	1.33 ± 0.62	0.18
C20:0/C18:0	391	0.04 ± 0.06	0.07	C22:4/C20:3	438	1.42 ± 0.65	0.22
C18:1/C16:1	350	15.64 ± 25.24	0.15	C22:4/C20:4	439	0.24 ± 0.19	0.18
C16:2/C16:1	439	0.09 ± 0.32	0.033	C22:4/C22:3	439	1.13 ± 0.5	0.205
C18:2/C18:1	439	0.72 ± 0.77	0.1	C22:6/C22:4	439	1.78 ± 1.07	0.036
C18:2/C16:2	157	190.24 ± 153.74	0.001	UFA/SFA	439	2.06 ± 0.47	0.001
C18:3/C18:2	109	0.03 ± 0.06	0.013	PUFA/MUFA	439	1.75 ± 0.85	0.001
C20:2/C18:2	438	0.06 ± 0.1	0.1	MUFA/SFA	439	0.78 ± 0.23	0.23
C20:3/C18:2	439	0.17 ± 0.16	0.011	PUFA/SFA	439	1.29 ± 0.35	0.075

Summary statistics for fatty acid metabolic traits in breast muscle of the F2 resource population. N, The number of valid individuals for each fatty acid trait for analysis;h2, heritability.

GWAS of Fatty Acid Metabolic Traits

Manhattan and Q-Q plots were constructed for 32 fatty acid metabolic traits (Fig. 2, Table 2, Fig. S1 (1–3), and Table S2). We identified 21 locis surpassing the genome-wide significance threshold and 291 locis exceeding the suggestive threshold. These locis were located on Chr 2, 3, 4, 5, 7, 9, and 10. The aforementioned 291 SNPs were significantly correlated with 12 different fatty acid metabolic traits, with traits such as C18:3/18:2, C18:0/16:0, C22:4/C20:3 primarily located on Chr 1, 2, and 3. Among the significant SNPs identified, the most significant locus was Chr2: 102676165 (P = 1.20E-06), accounting for 2.33% of the phenotypic variance within the PUFA/MUFA synthesis pathway. The locus explaining the highest amount of phenotypic variance among all loci was Chr4: 30150765 (5.85%). The results of the GWAS for the PUFA/MUFA ratio revealed that 11 significant SNPs were identified on Chr 2 (Fig. S2). Haplotype analysis showed strong linkage disequilibrium for S2_101192564, S2_101193251, and S2_101193442 (Fig. S3). Upon annotation, it was found that the missense mutation c.332A > T in the exon is located within the METTL4 causing an amino acid change from glutamic acid to valine. The Sorting Intolerant From Tolerant software predicted a score of 0.03, suggesting that c.332A > T is a potentially deleterious mutation that may alter the function of the protein encoded by the METTL4. By scanning all significant SNP linkage disequilibrium intervals and combining them with transcriptome gene expression data (excluding genes with FPKM less than 0.5), we ultimately obtained 613 candidate genes for subsequent WGCNA.

The analysis of LD on the primary Chr

GWAS results for fatty acid metabolic traits showed that significant signals primarily appeared on Chr1, 2, and 3. To more precisely locate the QTLs affecting phenotypic variation, we conducted a genome-wide LD analysis on SNPs with significant signals on Chr 1, 2, and 3 (Fig. 3). Significant signals for C18:3/18:2, C18:0/16:0, and C22:4/C20:3 were observed on Chr1, 2, and 3, respectively. We retained 38, 10, and 52 SNPs, respectively, for LD analysis that were significantly associated with these traits (-log₁₀P < 4). On Chr1, we identified a candidate genomic region of 382 kb (Chr1: 165.1-165.5 MB) (Fig. 3A), on Chr2, a candidate genomic region of 29 kb (Chr2: 87.04–87.07 MB) (Fig. 3B), and on Chr3, two candidate genomic regions (Chr3: 67.62–67.96 Mb and Chr3: 72.86–73.20 Mb) (Fig. 3C-D). To further identify key genes within these regions that impact fatty acid synthesis and breakdown, we performed GO and KEGG enrichment analyses on all genes within the candidate gene regions on Chr1, 2, and 3. We found significant enrichment in pathways and functions related to fatty acid synthesis and breakdown, including pyruvate metabolism, insulin signaling pathway, ether lipid metabolism, citrate cycle, and fumarate reductase complex.Based on these candidate genomic regions and enrichment analysis results, the FH and SDHA on Chr2, and the FUT9 and Fig. 4 on Chr3, which participate in pathways regulating fatty acid synthesis, are potential candidate genes.

GWAS of ω-3 and ω-6 Fatty Acid Metabolic Traits

ω-3 and ω-6 fatty acids, which cannot be synthesized within the human body and must be obtained externally, are essential fatty acids. Therefore, their synthesis ratio in poultry meat is of great importance. To explore the synthetic pathways of ω-3 and ω-6 fatty acids, we calculated eight metabolic traits based on the relationships between the fatty acid products and substrates in their synthesis pathways. We then constructed Manhattan plots and Q-Q plots for these eight metabolic traits (Fig. 4). The most significant locus was Chr7: 8991311 (P = 3.82E-08), which explained 2.9% of the phenotypic variation in the C20:3/C18:3 synthesis pathway. The locus explaining the highest phenotypic variation among all locis was Chr9: 8991311 (3.01%). Interestingly, significant signals were identified in the same gene interval (Chr9: 5.75Mb − 8.99Mb) for 3 metabolic traits (C22:4/C18:3, C20:3/C18:3, C22:6/C18:3).

LD Analysis of Chr9

Significant signals were identified within the same gene interval (Chr9: 5.75Mb − 8.99Mb) on Chr9 for three fatty acid composition traits (C22:4/C18:3, C20:3/C18:3, C22:6/C18:3) (Fig. 5A). The shared gene interval on Chr 9 for multiple traits indicates the presence of crucial causative mutation sites and causative genes affecting the phenotypic variation of breast muscle fatty acids within this interval. We conducted further LD analyses in the gene interval on Chr 9. A total of 31 significant SNPs were obtained for the 3 fatty acid composition traits (C22:4/C18:3, C20:3/C18:3, C22:6/C18:3). After constructing haplotypes with the aforementioned 31 SNPs, multiple linkage disequilibrium genomic intervals were obtained. By combining the related literature, we identified a candidate gene interval of 189kb (Chr9: 8.05–8.22 MB). Within the module of 8.05–8.22 MB, there are five SNPs exhibiting strong correlation (r² > 0.9) (Fig. 5B). The aforementioned gene interval is potentially the key causative region for the phenotypic variation in fatty acid composition.

WGCNA for Fatty Acid Metabolic Traits

WGCNA was applied to the 3 weeks transcriptome and phenotype value data of the 613 candidate genes identified by the aforementioned GWAS (Table S3). 5 modules were identified (turquoise, blue, brown, yellow, and green) (Fig. 6). The turquoise module showed a strong positive correlation with C18:2N6/C18:1N9C (r = 0.74, P = 0.02), C20:5N3/C20:3N3 (r = 0.7, P = 0.04), C22:4/C20:4N6 (r = 0.6, P = 0.09), and a strong negative correlation with C22:5N3/C20:5N3 (r = -0.85, P = 0.004), C20:4N6/C20:3N6 (r = -0.81, P = 0.008), C22:6N3/C22:5N3 (r = -0.63, P = 0.07). The blue module exhibited a strong positive correlation with C22:4/C20:4N6 (r = 0.65, P = 0.06), and negative correlation with C22:5N6/C22:4 (r = − 0.77, P = 0.01). The brown module was positively correlated to C14:0/C12:0 (r = 0.74, P = 0.02), C 16:1/C16:0 (r = 0.81, P = 0.008), C18:1N7/C16:1 (r = 0.64, P = 0.06), C18:1N9C/C18:0 (r = 0.85, P = 0.004), C22:5N6/C22:4 (r = 0.68, P = 0.04), and negatively correlated to C16:0/C14:0 (r = -0.65, P = 0.02), C18:0/C16:0 (r = -0.85, P = 0.004), C22:1N9/C20:1 (r = -0.8, P = 0.01), C20:2/C18:2N6 (r = -0.62, P = 0.08), C20:3N6/C18:3N6 (r = -0.75, P = 0.02). The yellow module negatively correlated with C22:5N3/C20:5N3 (r = -0.63, P = 0.07). The green module showed a strong positive correlation with C20:5N3/C20:3N3 (r = 0.64, P = 0.06), and a negative correlation with C22:2/C20:2 (r = − 0.69, P = 0.04) (Fig. 7). Additionally, the GO and KEGG enrichment analysis results of the five fatty acid composition-characteristic transcription modules revealed that the genes within these modules were enriched in pathways related to fatty acid synthesis and degradation (Fig. 8, Table S4, Table S5). Concurrently, based on the KME of the genes within the modules and the enrichment results, hub genes in each specific transcription module were identified. In the turquoise module, 4 hub genes were identified, namely Fig. 4 (KME = 0.927), METTL4 (KME = 0.908), PLD4 (KME = 0.953), and IDH1 (KME = 0.899). In the blue module, 3 hub genes, FH (KME = 0.9), SDHA (KME = 0.862), and EHHADH (KME = 0.88) were identified. In the brown module, 2 hub genes, AKT1 (KME = 0.986) and PAK2 (KME = 0.908) were identified. In the yellow module, 2 hub genes, PIK3CB (KME = 0.921) and LPIN2 (KME = 0.9) were identified. Furthermore, within the green module, FUT9 (KME = 0.946), a singular hub gene was identified, and constructed gene co-expression network for the 5 modules (Fig. 9).

Table 2

The significant SNPs and candidate genes associated with the fatty acid metabolic traits
Chr	Trait	N_snp	Range (Mb)	Top SNP	P-Value	Candidate genes
1	C22:4/C20:3	43	3589283	S1_165130710	5.9825200
2	C18:0/C16:0	13	2672751	S2_87075700	6.9170173	SDHA、FH
	C20:4/C18:3	11	4153964	44529084	4153964
	PUFA/MUFA	7	5993869	S2_102676165	5.6293463	METTL4、LPIN2
3	C18:3/C18:2	29	4095516	S3_67674637	6.4139001	Figure 4、FUT9
	C18:3/C18:2	26	3724010	S3_73155417	6.6811711
	PUFA/MUFA	3	1480380	S3_4488171	4.9616501
4	C14:0/C12:0	3	4333536	S4_71723221	5.3238315
	C18:1/C18:0	13	6745767	S4_30150765	6.0945468
5	C14:0/C12:0	6	3615137	S5_52719061	5.6818365	PLD4、AKT1
7	UFA/SFA	2	4472952	S7_26527279	5.8744797	IDH1
9	C16:2/C16:1	17	4561368	S9_13538144	5.8288706	EHHADH ACSL3 PIK3CB、PAK2
	C22:4/C18:3	16	3816640	S9_8991311	5.2042889
	C20:3/C18:3	33	6953871	S9_8991237	5.6324689
	C22:6/C18:3	14	6117360	S9_5159752	6.1153666
10	C16:2/C16:1	6	1645811	S10_1887346	5.5818529

The significant SNPs and candidate genes associated with the fatty acid metabolic traits. Chr, chromosome; Nsnp, the total number of significant SNPs associated with the traits; Top SNP, the most significantly associated SNP; P-value, the P-value of the top SNP; Var, phenotypic variance explained by the top SNPs, Range, region of the chromosome that significant SNPs covered.

Construction of Fatty Acid Metabolic Pathways and Networks

From the results of the GWAS, we pinpointed 613 candidate functional genes that could potentially influence the phenotypic variation of fatty acids in chickens. Through WGCNA, we obtained transcription modules that were significantly associated with various metabolic traits and identified hub genes. We ultimately retained the hub genes (AKT1, FUT9, LPIN2, EHADH) that exhibited significant correlations with the same fatty acid metabolic traits across both types of analysis. These genes were all enriched in biological pathways related to fatty acid metabolism, such as Glycosphingolipid biosynthesis (P = 0.03), MAPK signaling pathway (P = 0.04), Glycerophospholipid metabolism (P = 0.02), and Butanoate metabolism (P = 0.001). Simultaneously, in 3 traits (C16:2/C16:1, C22:4/C18:3, C20:3/C18:3), we located the long-chain acyl-coenzyme A synthase (ACSL) gene, known for its ability to activate long-chain fatty acids. Numerous studies have indicated that ASCL3 can regulate lipid droplet accumulation in cells, participate in fatty acid oxidation, and maintain the metabolic pathways for droplet formation [41, 42]. In conclusion, we filtered out candidate genes that demonstrated significant correlations with the same fatty acid metabolic traits in both types of analysis (Fig. 10). This allows us to gain a more intuitive understanding of potential regulatory genes in fatty acid synthesis. Based on these genes, we constructed a potential metabolic pathway and metabolic network diagram for fatty acids in chicken breast muscle.

To date, researchers have identified key locis and genes impacting phenotypic variation across multiple species through GWAS. However, quantitative trait variations within organisms are inevitably a complex process, with associated signals often appearing on multiple Chr, and phenotypic changes influencing each other [43]. In light of this complexity, researchers have formulated the "omnigenic model" building upon the polygenic hypothesis, as proposed by Evan A Boyle et al., gene regulatory networks are sufficiently interconnected such that any gene expressed in disease-relevant cells holds potential to influence the functionality of core disease-related genes. It is suggested that most heritability can likely be elucidated by effects on genes outside core pathways [44]. Therefore, in this study, we attempt to identify some core genes or pathways affecting phenotypic variation in poultry muscle fatty acid levels on a metabolic level while identifying genes beyond these core pathways to explain the heritability of fatty acid phenotypes.

The metabolic traits of C18:3/18:2, C18:0/16:0, C22:4/C20:3 exhibit significant signals mainly located on Chr 1, 2, and 3, consistent with the findings of other research studies. The extension of C18:2 to C18:3 is the first step in synthesizing various polyunsaturated fatty acids such as omega-3 and omega-6. By conducting LD analysis on significant SNPs, we identified three important candidate genomic intervals on Chr 2 and 3 (Chr2: 87.04–87.07 MB, Chr3: 67.62–67.96 Mb, and Chr3: 72.86–73.20 Mb). These gene regions may be crucial QTLs influencing polyunsaturated fatty acid synthesis. Coupled with WGCNA, we discovered that the genes FH and SDHA on Chr 2 and FUT9 and Fig. 4 on Chr 3 are hub genes in specific transcription module. These genes are also involved in fatty acid metabolic pathways such as Inositol phosphate metabolism and the citrate cycle. Both FH and SDHA participate in the tricarboxylic acid (TCA) cycle. The SDHA can dehydrogenate succinic acid to form fumarate, and the FH can catalyze fumarate to form malate thereby accelerating the TCA cycle. This speeds up the metabolism of downstream metabolites such as coenzyme A, thus affecting the rate of fatty acid synthesis. Azzimato V et al. [45] found that overexpression and interference experiments of miR-144 in mouse liver could upregulate or downregulate SDH and FH activity. By detecting NRF2, the main mediator reflecting antioxidant activity, it was found that obesity's antioxidant response could be regulated by activating FH and SDH to drive fumarate. Liu et al. [46] reported that ω-3 PUFAs can reduce the transcription and translation of TCA cycle enzymes, including IDH1, IDH2, SDHA, FH, and MDH2, thus improving the homeostasis of the TCA cycle to prevent obesity. FUT9 encodes an enzymatic fucosyltransferase involved in fatty acid degradation and sugar metabolism [46]. Whole-genome resequencing of Holstein cattle with high and low breeding values identified Fig. 4 as a promising candidate gene involved in regulating milk fatty acid traits [47].

Significant associations with PUFA/MUFA traits were detected in the region of 96.97–102.96 MB on Chr 2, where METTL4 and LPIN2, serving as hub genes in the turquoise module, are located. METTL4 is homologous to the methylation enzyme CG14906 in Drosophila and the methylation enzyme DAMT-1 in C. elegans. Several studies have indicated that METTL4 is an m6A methyltransferase functioning in fatty acid formation [48]. Recent discoveries have reported the presence of m6A in eukaryotes [49–51]. However, due to the low abundance of m6A and the lack of a systematic study of its functional pattern, m6A methyltransferase remains one of the most challenging core issues in mammalian studies. Zhang et al. [52] found that knock-down of METTL4 led to a decrease in m6A levels in mouse embryonic fibroblasts, subsequently resulting in decreased expression of METTL4 and INSR, inhibition of the INSR signalling pathway, reduced glucose uptake by cells, and suppressed lipid formation in adipocytes. Interestingly, the mutation site c.332A > T, located in the coding region of the METTL4, is a deleterious point mutation. This could potentially cause structural changes to proteins encoded by the METTL4, impacting gene function. Therefore, the impact of this mutant type and the wild-type on muscle fatty acid content demands further exploration in future research. LPIN2 can play a dual role in lipid metabolism, acting both as lysophosphatidic acid phosphatase and a transcriptional co-regulator of gene expression [53], also relating to insulin sensitivity [54]. Additionally, LPIN2, located in the yellow module, is negatively correlated with C22:5N3/C20:5N3 (r² = − 0.63, P = 0.07), suggesting that LPIN2 might have an inhibitory effect during the extension of C20:5N3 to C22:5N3.

On Chr 5, a significant association with the C14:0/C12:0 trait was found in the 51.66–54.78 MB region. The genes PLD4 and AKT1 within this genomic region were identified as hub genes. The AKT1, a hub gene in the brown module, showed strong correlation with C14:0/C12:0 in the module-trait relationship analysis (r² = 0.74, P = 0.02). Therefore, we speculate that the AKT1 may play a crucial role in the synthesis pathway where C12:0 extends to C14:0. AKT1, located in the brown module, also showed significant positive or negative correlations with several fatty acid metabolic traits, including C16:1/C16:0 (r² = 0.81, P = 0.008) and C18:1N7/C16:1 (r² = 0.64, P = 0.06). This suggests that this gene plays an important role in multiple fatty acid synthesis pathways and is a key gene in regulating the fatty acid synthesis network. PLD4 is a hub gene in the turquoise module, which is significantly associated with the traits of C20 and C22 series polyunsaturated fatty acids. Therefore, the PLD4 may play a role in the metabolism of long-chain polyunsaturated fatty acids.

The ratio of ω-3 to ω-6 fatty acids in the diet is of significant importance [55]. The study shows that excessive amounts of ω-6 PUFAs and a very high ω-6/ω-3 ratio, as is found in today's Western diets, promote the pathogenesis of many diseases, including cardiovascular disease, cancer, and inflammatory and autoimmune diseases, whereas increased levels of ω-3 PUFAs exert suppressive effects [56–58]. The GWAS results for the metabolic traits of ω-3 and ω-6 showed that multiple traits (C22:4/C18:3、C20:3/C18:3、C22:6/C18:3) displayed significant signals within the same gene region (Chr9: 5.75Mb-8.99 Mb) on Chr 9. According to LD, a 189 kb candidate interval (Chr9: 8.05–8.22 MB) was determined. Within this region, ASCL3 was discovered as a strong candidate gene. ASCL3 is part of the long-chain fatty acid coenzyme A ligase gene family, which can activate long-chain fatty acids for synthesis of cellular lipids [42]. It is responsible for the enzyme family that converts free long-chain fatty acids into acyl-coenzyme A esters. Its capacity to act as a substrate for lipid synthesis and fatty acid beta-oxidation implies that ASCL3 is integral for the elongation of long-chain fatty acids [59].

EHHADH, PIK3CB, and PAK2, the hub genes in the blue, brown, and yellow modules, are also located on Chr 9. The EHHADH, clustered in the blue module, is a bifunctional enzyme that can regulate the quantity of medium-chain dicarboxylic acids. These acids are essential regulators of all fatty acid oxidation pathways. Transcriptome analyses have revealed the involvement of EHHADH as a hub gene in lipid metabolism in bovine fetal fibroblasts and porcine preadipocytes [60, 61]. The blue module exhibits a strong positive correlation with C22:4/C20:4N6 (r² = 0.65, P = 0.06) and a strong negative correlation with C22:5N6/C22:4 (r² = − 0.77, P = 0.01). Combining this with the GWAS location results, the expression of EHHADH may promote the synthesis from C20:4N6 to C22:4, and inhibit the synthesis from C22:4 to C22:5N6. Zhao Y et al. [62] by identifying differentially expressed mRNA, miRNA, and lncRNA in the longissimus dorsi musclebetween Songliao Black pigs and Landrace pigs, founded that PIK3CB, and PLIN2 may be the key genes affecting IMF deposition. It is interesting to note, the latest research shows that ACSL4 expression is well correlated with PAK2 in hepatocellular carcinoma, and ACSL4 even transcriptionally increased PAK2 expression mediated by Sp1 [63]. Supplementation of ω-3 PUFA in healthy males resulted in a significant downregulation of the cytoskeleton-associated gene, PAK2 [64]. However, the detailed molecular regulatory mechanisms within the organism still require further in-depth study.

In this study, we first utilized reduced representation genome sequencing data from 415 individuals within an F2 resource population, alongside phenotypic data about 33 traits of breast muscle fatty acid metabolism. A GWAS identified 291 significant SNPs across 12 fatty acid metabolism traits. Notably, candidate gene intervals were located within the same QTL region on Chr 9, which may be crucial for influencing trait variability. Concurrently, using SNPs from the linkage disequilibrium regions, 613 genes were identified, and with the integration of breast muscle transcriptomic data from Gushi chickens, a WGCNA was performed. This analysis revealed 5 transcriptional modules specifically associated with the composition of fatty acid metabolic traits and 13 candidate genes. These findings facilitated the construction of a metabolic pathway and network for chicken breast muscle fatty acid metabolism.

GWAS: genome-wide association study; WGCNA: weighted gene co-expression network analysis; GBS: genotyping by sequencing; SNPs: single nucleotide polymorphisms; Chr: chromosomes; PUFA: polyunsaturated fatty acids; MUFA: monounsaturated fatty acids; SFA: Saturated Fatty Acids; UFA: Unsaturated Fatty Acids; PCA: Principal component analysis; GCTA: genome-wide complex trait analysis; PCs: principal components; MLM: mixed linear model; TOM: topological overlap measure; GO: Gene ontology; KEGG: Kyoto Encyclopedia of Genes and Genomes; ME: module eigengene; KME: the eigengene connectivity; ACSL: the long-chain acyl-coenzyme A synthase; TCA: tricarboxylic acid

Acknowledgements

Not applicable

Authors contributions

G.L., Y.Z., X.K., H.L., and S.F. conceived and designed the experiments. S.F., H.Z., H.L., and P.Y. performed the data analysis. S.F. and Y.Z. edited and revised the manuscript. X.K., Y.T., G.L., and X.L. constructed the F2 population and collected the phenotypes. Y.D. prepared the samples. All authors read and approved the final manuscript.

Funding

This study was supported by the National Natural Science Foundation of China (32072692), the Key Research Project of the Shennong Laboratory (SN01-2022-05), the National Key Research and Development Program of China (2023YFF1001100), the Key Scientifc Research Project of Higher Education of Henan Province (23A230014).

Consent for publication

Not applicable.

Availability of data and materials

All the data supporting the results of this study are included in the article and its Additional file.

Competing interests

The authors declare no competing interes.

Ethics approval and consent to participate

The animal experiments in this study were approved by the Institutional Animal Care and Use Committee (IACUC) of Henan Agricultural University under approval number 17-0118. All experiments strictly followed the guidelines of this committee.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interest.

Author details

¹ The Shennong Laboratory, Henan Agricultural University, Zhengzhou, Henan Province, 450046, PR China. ² College of Biological Engineering, Henan University of Technology, Zheng Zhou, Henan Province, 450001, PR China. ³ College of Animal Science and Technology, Northwest A&F University, Yangling, Shaanxi 712100, China.

1. Brenna JT. Efficiency of conversion of alpha-linolenic acid to long chain n-3 fatty acids in man. Curr Opin Clin Nutr Metab Care. 2002; 5:127-132.
2. Twining CW, Lawrence P, Winkler DW, Flecker AS, Brenna JT. Conversion efficiency of α-linolenic acid to omega-3 highly unsaturated fatty acids in aerial insectivore chicks. J Exp Biol. 2018; 221.
3. Cinquina V, Keimpema E, Pollak DD, Harkany T. Adverse effects of gestational ω-3 and ω-6 polyunsaturated fatty acid imbalance on the programming of fetal brain development. J Neuroendocrinol. 2023; 35:e13320.
4. Kousparou C, Fyrilla M, Stephanou A, Patrikios I. Dha/epa (omega-3) and la/gla (omega-6) as bioactive molecules in neurodegenerative diseases. Int J Mol Sci. 2023; 24.
5. Masetto Antunes M, Godoy G, Curi R, Vergílio Visentainer J, Barbosa Bazotte R. The myristic acid:Docosahexaenoic acid ratio versus the n-6 polyunsaturated fatty acid:N-3 polyunsaturated fatty acid ratio as nonalcoholic fatty liver disease biomarkers. Metab Syndr Relat Disord. 2022; 20:69-78.
6. Horman T, Fernandes MF, Tache MC, Hucik B, Mutch DM, Leri F. Dietary n-6/n-3 ratio influences brain fatty acid composition in adult rats. Nutrients. 2020; 12.
7. Liu L, Hu Q, Wu H, Wang X, Gao C, Chen G, et al. Dietary dha/epa ratio changes fatty acid composition and attenuates diet-induced accumulation of lipid in the liver of apoe(-/-) mice. Oxid Med Cell Longev. 2018; 2018:6256802.
8. Suhre K, Gieger C. Genetic variation in metabolic phenotypes: Study designs and applications. Nat Rev Genet. 2012; 13:759-769.
9. Mayr M. Metabolomics: Ready for the prime time? Circ Cardiovasc Genet. 2008; 1:58-65.
10. Zhang W, Zhang J, Cui L, Ma J, Chen C, Ai H, et al. Genetic architecture of fatty acid composition in the longissimus dorsi muscle revealed by genome-wide association studies on diverse pig populations. Genet Sel Evol. 2016; 48:5.
11. Tanha HM, Sathyanarayanan A, Nyholt DR. Genetic overlap and causality between blood metabolites and migraine. Am J Hum Genet. 2021; 108:2086-2098.
12. Mayers JR, Vander Heiden MG. Nature and nurture: What determines tumor metabolic phenotypes? Cancer Res. 2017; 77:3131-3134.
13. Deist MS, Gallardo RA, Bunn DA, Dekkers JCM, Zhou H, Lamont SJ. Resistant and susceptible chicken lines show distinctive responses to newcastle disease virus infection in the lung transcriptome. BMC Genomics. 2017; 18:989.
14. Wei W, Xiao J, Huang N, Xing C, Wang J, He X, et al. Identification of central regulators related to abdominal fat deposition in chickens based on weighted gene co-expression network analysis. Poult Sci. 2024; 103:103436.
15. Oikonomou S, Samaras A, Tekeoglou M, Loukovitis D, Dimitroglou A, Kottaras L, et al. Genomic selection and genome-wide association analysis for stress response, disease resistance and body weight in european seabass. Animals (Basel). 2022; 12.
16. Vanvanhossou SFU, Scheper C, Dossa LH, Yin T, Brügemann K, König S. A multi-breed gwas for morphometric traits in four beninese indigenous cattle breeds reveals loci associated with conformation, carcass and adaptive traits. BMC Genomics. 2020; 21:783.
17. Lyu R, Sun J, Xu D, Jiang Q, Wei C, Zhang Y. Geslm algorithm for detecting causal snps in gwas with multiple phenotypes. Brief Bioinform. 2021; 22.
18. Matushyn M, Bose M, Mahmoud AA, Cuthbertson L, Tello C, Bircan KO, et al. Sumstatsrehab: An efficient algorithm for gwas summary statistics assessment and restoration. BMC Bioinformatics. 2022; 23:443.
19. Sun J, Lyu R, Deng L, Li Q, Zhao Y, Zhang Y. Smetabf: A rapid algorithm for bayesian gwas meta-analysis with a large number of studies included. PLoS Comput Biol. 2022; 18:e1009948.
20. Han R, Wei Y, Kang X, Chen H, Sun G, Li G, et al. Novel snps in the prdm16 gene and their associations with performance traits in chickens. Mol Biol Rep. 2012; 39:3153-3160.
21. Zhang Y, Wang Y, Li Y, Wu J, Wang X, Bian C, et al. Genome-wide association study reveals the genetic determinism of growth traits in a gushi-anka f(2) chicken population. Heredity (Edinb). 2021; 126:293-307.
22. Fan S, Yuan P, Li S, Li H, Zhai B, Li Y, et al. Genetic architecture and key regulatory genes of fatty acid composition in gushi chicken breast muscle determined by gwas and wgcna. BMC Genomics. 2023; 24:434.
23. Dale N. National research council nutrient requirements of poultry – ninth revised edition (1994). Journal of Applied Poultry Research. 1994; 3.
24. Li Y, Chen Y, Jin W, Fu S, Li D, Zhang Y, et al. Analyses of microrna and mrna expression profiles reveal the crucial interaction networks and pathways for regulation of chicken breast muscle development. Front Genet. 2019; 10:197.
25. Wang Y, Cao X, Zhao Y, Fei J, Hu X, Li N. Optimized double-digest genotyping by sequencing (ddgbs) method with high-density snp markers and high genotyping accuracy for chickens. PLoS One. 2017; 12:e0179073.
26. Watson M. Illuminating the future of DNA sequencing. Genome Biol. 2014; 15:108.
27. Langmead B, Salzberg SL. Fast gapped-read alignment with bowtie 2. Nat Methods. 2012; 9:357-359.
28. Wingett SW, Andrews S. Fastq screen: A tool for multi-genome mapping and quality control. F1000Res. 2018; 7:1338.
29. Glaubitz JC, Casstevens TM, Lu F, Harriman J, Elshire RJ, Sun Q, et al. Tassel-gbs: A high capacity genotyping by sequencing analysis pipeline. PLoS One. 2014; 9:e90346.
30. Van Tassell CP, Smith TP, Matukumalli LK, Taylor JF, Schnabel RD, Lawley CT, et al. Snp discovery and allele frequency estimation by deep sequencing of reduced representation libraries. Nat Methods. 2008; 5:247-252.
31. Yang J, Lee SH, Goddard ME, Visscher PM. Gcta: A tool for genome-wide complex trait analysis. Am J Hum Genet. 2011; 88:76-82.
32. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. Plink: A tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet. 2007; 81:559-575.
33. Barrett JC. Haploview: Visualization and analysis of snp genotype data. Cold Spring Harb Protoc. 2009; 2009:pdb.ip71.
34. Langfelder P, Horvath S. Wgcna: An r package for weighted correlation network analysis. BMC Bioinformatics. 2008; 9:559.
35. Yu G, Wang LG, Han Y, He QY. Clusterprofiler: An r package for comparing biological themes among gene clusters. Omics. 2012; 16:284-287.
36. Ogata H, Goto S, Sato K, Fujibuchi W, Bono H, Kanehisa M. Kegg: Kyoto encyclopedia of genes and genomes. Nucleic Acids Res. 1999; 27:29-34.
37. Kanehisa M. Toward understanding the origin and evolution of cellular organisms. Protein Sci. 2019; 28:1947-1951.
38. Kanehisa M, Furumichi M, Sato Y, Kawashima M, Ishiguro-Watanabe M. Kegg for taxonomy-based analysis of pathways and genomes. Nucleic Acids Res. 2023; 51:D587-d592.
39. Ma L, Zhang M, Chen J, Qing C, He S, Zou C, et al. Gwas and wgcna uncover hub genes controlling salt tolerance in maize (zea mays l.) seedlings. Theor Appl Genet. 2021; 134:3305-3318.
40. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: A software environment for integrated models of biomolecular interaction networks. Genome Res. 2003; 13:2498-2504.
41. Wang C, Liu F, Yuan Y, Wu J, Wang H, Zhang L, et al. Metformin suppresses lipid accumulation in skeletal muscle by promoting fatty acid oxidation. Clin Lab. 2014; 60:887-896.
42. Quan J, Bode AM, Luo X. Acsl family: The regulatory mechanisms and therapeutic implications in cancer. Eur J Pharmacol. 2021; 909:174397.
43. Armengaud J. Proteogenomics and systems biology: Quest for the ultimate missing parts. Expert Rev Proteomics. 2010; 7:65-77.
44. Boyle EA, Li YI, Pritchard JK. An expanded view of complex traits: From polygenic to omnigenic. Cell. 2017; 169:1177-1186.
45. Azzimato V, Chen P, Barreby E, Morgantini C, Levi L, Vankova A, et al. Hepatic mir-144 drives fumarase activity preventing nrf2 activation during obesity. Gastroenterology. 2021; 161:1982-1997.e1911.
46. Liu R, Chen L, Wang Z, Zheng X, Hou Z, Zhao D, et al. Omega-3 polyunsaturated fatty acids prevent obesity by improving tricarboxylic acid cycle homeostasis. J Nutr Biochem. 2021; 88:108503.
47. Lin S, Zhang H, Hou Y, Liu L, Li W, Jiang J, et al. Snv discovery and functional candidate gene identification for milk composition based on whole genome resequencing of holstein bulls with extremely high and low breeding values. PLoS One. 2019; 14:e0220629.
48. Kweon SM, Chen Y, Moon E, Kvederaviciutė K, Klimasauskas S, Feldman DE. An adversarial DNA n(6)-methyladenine-sensor network preserves polycomb silencing. Mol Cell. 2019; 74:1138-1147.e1136.
49. Fu Y, Luo GZ, Chen K, Deng X, Yu M, Han D, et al. N6-methyldeoxyadenosine marks active transcription start sites in chlamydomonas. Cell. 2015; 161:879-892.
50. Zhang G, Huang H, Liu D, Cheng Y, Liu X, Zhang W, et al. N6-methyladenine DNA modification in drosophila. Cell. 2015; 161:893-906.
51. Greer EL, Blanco MA, Gu L, Sendinc E, Liu J, Aristizábal-Corrales D, et al. DNA methylation on n6-adenine in c. Elegans. Cell. 2015; 161:868-878.
52. Zhang Z, Hou Y, Wang Y, Gao T, Ma Z, Yang Y, et al. Regulation of adipocyte differentiation by mettl4, a 6 ma methylase. Sci Rep. 2020; 10:8285.
53. Watahiki A, Hoshikawa S, Chiba M, Egusa H, Fukumoto S, Inuzuka H. Deficiency of lipin2 results in enhanced nf-κb signaling and osteoclast formation in raw-d murine macrophages. Int J Mol Sci. 2021; 22.
54. Ramírez-Zamudio GD, Ganga MJG, Pereira GL, Nociti RP, Chiaratti MR, Cooke RF, et al. Effect of cow-calf supplementation on gene expression, processes, and pathways related to adipogenesis and lipogenesis in longissimus thoracis muscle of f1 angus × nellore cattle at weaning. Metabolites. 2023; 13.
55. Simopoulos AP. The importance of the ratio of omega-6/omega-3 essential fatty acids. Biomed Pharmacother. 2002; 56:365-379.
56. Simopoulos AP. An increase in the omega-6/omega-3 fatty acid ratio increases the risk for obesity. Nutrients. 2016; 8:128.
57. Simopoulos AP. The importance of the omega-6/omega-3 fatty acid ratio in cardiovascular disease and other chronic diseases. Exp Biol Med (Maywood). 2008; 233:674-688.
58. Gómez Candela C, Bermejo López LM, Loria Kohen V. Importance of a balanced omega 6/omega 3 ratio for the maintenance of health: Nutritional recommendations. Nutr Hosp. 2011; 26:323-329.
59. Quan J, Cheng C, Tan Y, Jiang N, Liao C, Liao W, et al. Acyl-coa synthetase long-chain 3-mediated fatty acid oxidation is required for tgfβ1-induced epithelial-mesenchymal transition and metastasis of colorectal carcinoma. Int J Biol Sci. 2022; 18:2484-2496.
60. Li G, Yang R, Lu X, Liu Y, He W, Li Y, et al. Rna-seq analysis identifies differentially expressed genes in the longissimus dorsi of wagyu and chinese red steppe cattle. Int J Mol Sci. 2022; 24.
61. Zhu J, Yang Z, Hao W, Li J, Wang L, Xia J, et al. Characterization of a read-through fusion transcript, bcl2l2-pabpn1, involved in porcine adipogenesis. Genes (Basel). 2022; 13.
62. Zhao Y, Chen S, Yuan J, Shi Y, Wang Y, Xi Y, et al. Comprehensive analysis of the lncrna-mirna-mrna regulatory network for intramuscular fat in pigs. Genes (Basel). 2023; 14.
63. Wu D, Zuo Z, Sun X, Li X, Yin F, Yin W. Acsl4 promotes malignant progression of hepatocellular carcinoma by targeting pak2 transcription. Biochem Pharmacol. 2024; 224:116206.
64. Schmidt S, Willers J, Riecker S, Möller K, Schuchardt JP, Hahn A. Effect of omega-3 polyunsaturated fatty acids on the cytoskeleton: An open-label intervention study. Lipids Health Dis. 2015; 14:4.

No competing interests reported.

Additionalfile1TableS1.xlsx
Additional file 1:Table. S1. The dynamic change pattern of fatty aicd metabolic traits in breast muscle of Gushi chickens at 14, 22, and 30 weeks.
Additionalfile2Fig.S11.tif
Additional file 2:Fig. S1 (1-3). The Manhattan and Q-Q plots for 32 fatty acid metabolic traits. The solid and dashed lines represent genome-wide significance (0.05/N) threshold and chromosome-wide suggestive threshold (1/N), respectively.The corresponding traits of abbreviations are noted in the explanation of Figure 2.
Additionalfile2Fig.S12.tif
Additional file 2:Fig. S1 (1-3). The Manhattan and Q-Q plots for 32 fatty acid metabolic traits. The solid and dashed lines represent genome-wide significance (0.05/N) threshold and chromosome-wide suggestive threshold (1/N), respectively.The corresponding traits of abbreviations are noted in the explanation of Figure 2.
Additionalfile2Fig.S13.tif
Additional file 2:Fig. S1 (1-3). The Manhattan and Q-Q plots for 32 fatty acid metabolic traits. The solid and dashed lines represent genome-wide significance (0.05/N) threshold and chromosome-wide suggestive threshold (1/N), respectively.The corresponding traits of abbreviations are noted in the explanation of Figure 2.
Additionalfile3TableS2.xlsx
Additional file 3:Table. S2. GWAS results for 13 fatty acid composition traits (P < 3.78E-05)
Additionalfile4Fig.S2.tif
Additional file 4: Fig. S2. The Manhattan and Q-Q plots for the PUFA/MUFA traits. 11 significant SNPs were identified on Chr 2.
Additionalfile5Fig.S3.tif
Additional file 5: Fig. S3. Haplotype analysis showed strong linkage disequilibrium for S2_101192564, S2_101193251, and S2_101193442. The missense mutation c.332A > T in the exon is located within the METTL4 causing an amino acid change from glutamic acid to valine.
Additionalfile6TableS3.xls
Additional file 6:Table. S3. The 613 genes identified by integrating the results of the GWAS for fatty acid metabolic traits and the transcriptome profile data of Gushi chicken breast muscle muscl.
Additionalfile7TableS4.xls
Additional file 7:Table. S4. The lists of 30 enriched GO terms of 5 modules.
Additionalfile8TableS5.xls
Additional file 8:Table. S5. The lists of enriched KEGG pathways of 5 modules.

Download PDF

Version 1

posted

You are reading this latest preprint version

Investigating the hub genes and genetic basis regulating fatty acid metabolictraits in chicken breast muscle based on combined genomic and transcriptomic approaches

Status:

Version 1

Abstract

Figures

Background

Materials and methods

Establishment of the Experimental Population

Phenotypic Measurements

Transcriptome profile and phenotypic data

Genotyping, imputation, and quality control

Single trait GWAS

y = Wα + βx + u + e

WGCNA

WGCNA for fatty acid metabolic traits

Module-trait relationships and hub gene identification

Results

Statistical Results of Fatty Acid Metabolism Phenotypes and Genotypes

GWAS of Fatty Acid Metabolic Traits

The analysis of LD on the primary Chr

GWAS of ω-3 and ω-6 Fatty Acid Metabolic Traits

LD Analysis of Chr9

WGCNA for Fatty Acid Metabolic Traits

Construction of Fatty Acid Metabolic Pathways and Networks

Discussion

Conclusion

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1