Skip to main content
Download PDF

Serum metabolic profile and metabolome genome-wide association study in chicken

Journal of Animal Science and Biotechnology volume 14, Article number: 69 (2023) Cite this article

Abstract

Background

Chickens provide globally important livestock products. Understanding the genetic and molecular mechanisms underpinning chicken economic traits is crucial for improving their selective breeding. Influenced by a combination of genetic and environmental factors, metabolites are the ultimate expression of physiological processes and can provide key insights into livestock economic traits. However, the serum metabolite profile and genetic architecture of the metabolome in chickens have not been well studied.

Results

Here, comprehensive metabolome detection was performed using non-targeted LC–MS/MS on serum from a chicken advanced intercross line (AIL). In total, 7,191 metabolites were used to construct a chicken serum metabolomics dataset and to comprehensively characterize the serum metabolism of the chicken AIL population. Regulatory loci affecting metabolites were identified in a metabolome genome-wide association study (mGWAS). There were 10,061 significant SNPs associated with 253 metabolites that were widely distributed across the entire chicken genome. Many functional genes affect metabolite synthesis, metabolism, and regulation. We highlight the key roles of TDH and AASS in amino acids, and ABCB1 and CD36 in lipids.

Conclusions

We constructed a chicken serum metabolite dataset containing 7,191 metabolites to provide a reference for future chicken metabolome characterization work.

Meanwhile, we used mGWAS to analyze the genetic basis of chicken metabolic traits and metabolites and to improve chicken breeding.

Background

Metabolomics studies small molecules (< 1,000 Da) present in biological samples [1]. Essential for growth and health in organisms, metabolites are the final products of gene transcription and protein expression, and are affected by both internal and external factors [2, 3]. Generally regarded as a bridge between genes and phenotypes [4, 5], the combination of metabolomics with genomics and transcriptomics has proven to be powerful in analyzing metabolic diversity and pathways [6, 7]. For example, metabolome genome-wide association studies (mGWAS) in crops such as tomato, corn, and wheat have revealed that many metabolite-associated loci control the effects of fruit color, crop yield, and enzyme activity on the metabolism of specific substances [4, 8, 9]. Metabolite GWAS has found effective therapeutic targets for metabolic diseases, such as human kidney disease and type 2 diabetes [10].

Growth in chickens is determined by quantitative traits regulated by multiple genes [11,12,13,14]. Traditional genome-wide association studies (GWAS) can identify SNPs associated with phenotypes but have limited ability to analyze the mechanisms underlying these phenotypes [15, 16]. Investigating metabolic phenotypes (metabotypes)—which are determined by the interaction of genetics and environment—instead of traditional complex phenotypes may help to solve this problem [17]. Currently, the mGWAS approach in livestock has been limited to small-sample comparisons so there is a need to characterize metabotypes in population samples.

Blood contains a variety of substances required for maintaining normal physiological functioning; this makes it a powerful tool for assessing the nutritional and health status of humans and animals. In agricultural animal studies, the association of serum and plasma metabolites with disease [18], meat quality traits [19], feed intake [20] and growth traits [21] has been reported. Serum and plasma are now commonly used as biological samples in metabolomic studies because of their easy accessibility and ability to reflect the overall metabolic characteristics of an organism [22, 23].

In this study, we aim to examine serum metabolome of chickens using a non-targeted metabolomics approach and construct a metabolite dataset for chickens. At the same time, the metabolic phenotype was used for genome-wide association analysis to analyze its genetic model and identify genes related to metabolite synthesis and metabolic pathways. This could greatly improve our understanding of chicken serum metabolic profiles and metabolic phenotypes, providing a strong foundation for future studies on the mechanisms and localization of chicken economic traits.

Materials and methods

Advanced intercross line

We created the AIL in this study by crossing individuals from two distinct chicken lines, namely a quality chicken Line A03 (HQLA) and a native Chinese breed Huiyang Bearded chicken (HB). Detailed feeding patterns, as well as F0–F2 mating schemes, were described in a previously published article [24]. Later, AIL generations (F3 to F16) were established from birds in the F2 population and reproduced using random mating [13, 25].

Serum sample collection and processing

Metabolomics was used to study the serum of 508 12-week-old chickens (266 hens and 242 cocks) of the F16 generation. A serum sample was obtained by centrifugation at 2,000 × g for 10 min after blood samples from chickens were left at room temperature. These samples were frozen with liquid nitrogen and then stored at −80 ℃ for later analyses.

All frozen serum samples were initially thawed on ice and vortexed, and 400 μL cold methanol/acetonitrile mixed extract (1:1, v:v) was used for metabolome extraction and protein removal for each 100 μL serum [26]. The supernatant (200 μL) was rotated and dried for analysis. Dried supernatants were then reconstituted in 50 µL of water with 50% methanol (T3 sample) and 94% acetonitrile (Amide sample).

Metabolite analysis by LC–MS/MS

A Vanquish UHPLC system was coupled to a Q-Exactive HF-X Hybrid Quadrupole-Orbitrap Mass spectrometer (Thermo Fisher Scientific, Waltham, Massachusetts, USA) for non-targeted metabonomics detection. Chromatographic separation was performed using a reverse-phase ACQUITY UPLC HSS T3 column (100 Å, 1.8 μm, 100 mm × 2.1 mm, Waters, Milford, Massachusetts, USA) at 40 ℃ with mobile phases of water containing 0.1% formic acid (A1) and methanol (B1) and HILIC-phase ACQUITY UPLC BEH Amide column (130 Å, 1.7 μm, 100 mm × 2.1 mm, Waters) at 40 °C with mobile phases of 50% acetonitrile with 10 mmol/L ammonium acetate (A2) and 95% acetonitrile with 10 mmol/L ammonium acetate (B2), pH 9. One microliter of pretreated sample was injected. The T3 flow rate was 0.2 mL/min, and gradient elution was performed as follows: the system was equilibrated with A1 for 7 min followed by linear increases to 98% B1 over 26 min and maintained at 98% for 5 min. The amide flow rate was 0.2 mL/min, and gradient elution was performed as follows: the system was equilibrated with 98% B2 for 9.9 min followed by linear reduced to 2% B2 over 20 min and maintained at 2% for 2 min. The mass spectrometer was operated in positive ion mode with a spray voltage of 3,500 V, a capillary temperature of 350 °C, a sheath gas flow rate of 30 arb, an auxiliary gas flow rate of 11 arb, and a probe heater temperature of 220 °C.

Samples were scanned using the full-MS mode, with the resolution of the full scan set at 120,000 and a scan range of m/z = 70–1,050. To collect sufficient MS/MS information for metabolite identification, Quality Control (QC) samples underwent segmented secondary scans. These consisted of four sections: m/z = 70–160, m/z = 150–260, m/z = 250–410 and m/z = 400–1,050. Using full MS-dd MS2 scan mode, full MS resolution, 60,000; dd-MS2 resolution, 15,000; top N, 15; isolation window, 1 m/z; stepped NCE at 20, 30, and 40. The final four-mode data were obtained using T3-pos, T3-neg, Amide-pos and Amide-neg representations.

Metabolite identification and classification

XCMS is a software based on R language which is often used for LC–MS data pre-processing analysis [27] and in our study it was used for peak extraction, peak alignment, etc. The parameters were the maximum allowable deviation in m/z for continuous scanning: ppm = 20, the range of peak widths: peakwidth = c (5, 34), the the signal-to-noise ratio threshold: snthresh = 4, the prefiltering step in the first step: prefilter = c (3, 15,000), the inclusiveness of the grouping: bw1 = 15 and bw2 = 5 to obtain the initial MS feature. One-MAP (One-step Metabolomics: A Smart Cloud Platform for Metabolites Identification and Biomarkers Discovery, www.5omics.com) was used to identify the metabolites of segmented secondary QC sample data. Three classes of databases of One-MAP were used for metabolite identification: a standard database, which structurally identified metabolites through direct comparison of their chromatographic and fragmentation behavior with 1,500 standards [28], a KEGG database, and an integrated database.

We used ClassyFire (https://cfb.fiehnlab.ucdavis.edu/), an online metabolite classification software, to classify the identified metabolites into substances [29]. We also combined HMDB and KEGG databases to classify metabolites by internal and external sources: Among the small-molecule metabolites, those clearly attributed to plant and drug sources were classified as exogenous metabolites, while the rest were considered endogenous metabolites (vitamins and hormones require specific analysis), and all conventional lipids were considered endogenous metabolites.

Statistical analysis

The metabolite data were log10-transformed to improve normality for statistical analysis. Coefficient of variation (CV) values were calculated for each metabolite and expressed as S/A, where S and A represent the standard deviation and mean of the metabolites in the population, respectively. Pearson's correlations between metabolites and statistical significance were estimated using R (http://www.r-project.org/). Metabolite pathway analysis was performed using the online metabolite pathway enrichment software MetaboAnalyst 5.0 [30]. Gene function enrichment was performed using the Metascape software [31].

Genotypic information

DNA was extracted from blood samples using the DNeasy Blood & Tissue Kit (Qiagen 69506, Hilden, GER), evaluated using a NanoDrop spectrophotometer (Thermo Fisher Scientific), and examined on 1% agarose gels. All samples were quantified using a Qubit 2.0 fluorometer (Invitrogen, Carlsbad, California, USA) and then diluted to 40 ng/mL in a 96-well plate. Libraries were constructed using the Tn5 method, and final libraries were sequenced on two lanes of an MGISEQ-2000 (MGI, Shenzhen, Guangzhou, CHN) to generate 2 × 100 bp double-end reads or on one lane of a BGISEQ-500 (MGI, Shenzhen, Guangzhou, CHN) to generate 2 × 100 bp double-end reads [32].

In summary, low-coverage sequencing data from more than 1,000 samples were mapped to the GRCg6a reference genome and the BaseVar + STITCH pipeline was used to impute SNPs [32]. A subset of 962,660 SNPs that tagged all other SNPs with MAF > 5% at LD r2 > 0.95 were used for subsequent analysis.

mGWAS and heritability estimation

To integrate genomic and metabolomic data, metabolome genome-wide association studies (mGWAS) were conducted in which each metabolite (n = 2,935) was considered a phenotype and examined for its association with each SNP (960 K). The mixed linear model approach was used for genome-wide association analysis based on marker SNPs, as implemented in the GCTA (1.93.2) package [33]. We estimated SNP heritability using the GREML module in the software package GCTA (1.93.2) metabolites heritability estimation. Heritability was estimated using a mixed model as follows:

$${\varvec{y}}\boldsymbol{ }={\varvec{Xb}} +{\varvec{Wu}}\boldsymbol{ }+\boldsymbol{ }{\varvec{e}}$$

with var (y) = WW'σu2 + Iσe2, where y is the vector of the metabolite phenotypes, b is a vector of the fixed effects (sex and batch), with its incidence matrix X, u is the vector of additive values based on the genotype data, and e is a random residual error. W is a genomic additive relationship matrix, σu2 is the additive variance, and σe2 is the residual variance. Variance components were estimated by genome-based restricted maximum likelihood (GREML) using the reml program in GCTA [34].

SNP annotation and candidate genes

We first used SnpEff software [35] to annotate functional genes for 10,061 SNPs that reached the significance threshold. The genes obtained from mGWAS analysis of the same metabolite species were enriched for functions using Metascape software [31] to screen candidate genes related to the regulation, synthesis, and metabolism of the species, and the SNPs annotated to the respective gene were considered key SNPs.

Results

Serum metabolic profiling and non-targeted metabolite dataset

In this study, non-targeted LC–MS/MS metabolomics was performed using serum sampled from 12-week-old chickens of the F16 generation of an advanced intercross line (AIL). A total of 7,191 metabolites and 4,204 endogenous metabolites were detected and identified (Fig. 1A and B, Table S1). Of these metabolites, three levels of databases were used for identification, the standard databased identified 934 metabolites of which were 800 endogenous; the KEGG database identified 2,782 metabolites of which were 1,322 endogenous; the Integrated database, identified 4,953 metabolites in total, of which 3,199 were endogenous (Fig. 1C). The identified compounds covered a wide range of biochemicals, including amino acids, benzenoids, carbohydrates, lipids, organic acids, organic heterocyclic compounds, organonitrogen, peptides and nucleosides, phenylpropanoids and polyketides, others, and exogenous compounds (Fig. 1D, Table S1).

Fig. 1
figure 1

Chicken serum metabolic profiles based on QC samples. A Venn diagram of all metabolites identified in the four detection modes. The same metabolites were detected by multiple modes with a count of 1. B Venn diagram of endogenous metabolites identified in the four detection modes. The same metabolites were detected by multiple modes with a count of 1. C Identification of metabolites using three classes of database. SD, Standards Database; KD, KEGG Database; ID, Intergrated Database; T, Total metabolites; E, Endogenous metabolites. D Classification of all metabolites

Full size image

We used the above metabolites to construct a non-targeted dataset of chicken serum metabolites using a systematic and automated approach and homemade software (OSI/SMMS) [28]. The dataset contains basic metabolite information such as molecular formula, m/z, retention time, primary and secondary spectral scores, and internal and external source information. For each metabolite, the first- and second-level mass spectra could be viewed, and metabolites with a combined score higher than 0.5, which is highly reliable, were retained in the dataset. Our dataset was able to help identify many more metabolites in a separate batch of muscle samples, than the traditional online database (Fig. S1).

Metabolomic characterization of the AIL F16 population

A total of 508 chicken serum samples from the F16 generation were analyzed using a non-targeted LC–MS/MS method. After quality control, the total number of metabolic features extracted in each detection mode was 27,818, 20,274, 14,569 and 11,054 (T3-pos, T3-neg, Amide-pos and Amide-neg, respectively). Using the chicken serum metabolite dataset matching identification, 1,238, 711, 518, and 468 metabolites were identified in four modes for a total of 2,525 metabolites (Fig. 2A, Table S2). The metabolites were consisted of amino acids (7.52%), benzenoids (6.34%), carbohydrates (2.53%), exogenous compounds (9.31%), lipids (49.94%), organic acids (4.32%), organic heterocycles (8.32%), organonitrogen (1.47%), peptides and nucleosides (3.84%), phenylpropanoids and polyketides (3.09%), and other metabolites (3.33%) (Fig. 2B, Table S2). These metabolites are involved in multiple important metabolic pathways supporting key cellular processes.

Fig. 2
figure 2

Metabolomic characterization of chicken AIL population. A Venn diagram of metabolites identified in the four detection modes. The same metabolites were detected by multiple modes with a count of 1. B Pie chart showing the percentage of each type of metabolite detected across all four modes. C Distribution of the values of the coefficient of variation (CV) across the four detection modes, with the mean CV indicated by a horizontal line. D Distribution of the broad-sense heritability (h2) of metabolites across the four detection modes, with the dashed red line representing h2 = 0.2. E Heat map displaying paired Pearson's correlations (r) between metabolites detected under the T3 modes, with metabolites classified according to a hierarchical clustering analysis based on correlations. Metabolites included in heat map had an r > 0.8, P < 0.05. F Heat map displaying paired Pearson’s correlations (r) between metabolites detected under the Amide modes, with metabolites classified according to a hierarchical clustering analysis based on correlations. Metabolites included in heat map had an r > 0.8, P < 0.05

Full size image

Principal component analysis (PCA) of all features showed no population stratification of metabolites detected in the same mode (Fig. S2). The levels of metabolite accumulation in the samples varied considerably, which allowed for efficient analysis of their genetic structure. In the AIL population, the mean genetic coefficients of variation (CV) of these metabolites were 66.7%, 58.0%, 67.8% and 69.5% across the four modes (Fig. 2C, Table S3). The substances with the maximum mean CV among the different modes, with amino acids (81.9% in T3-pos), benzenes (72.6% in T3-neg), organonitrogen (131.7% in Amide-pos mode) and 95% in peptides and nucleotides in Amide-neg mode. 5,6-Dihydroxyindole had the highest CV value (857.9%) of all metabolites.

The broad sense heritability (h2) distribution of metabolic traits showed that more than 7.8% of the metabolites showed heritability above 0.2 (Fig. 2D, Table S4). In Amide-neg mode, the compound with the highest heritability was 2'-O-Methyluridine, which had an h2 of 0.56. Under the other three modes (T3-pos, T3-neg and Amide-pos), ceramide (Cer 36:3) had the highest heritability, with an h2 of 0.69, 0.71 and 0.59, respectively. These results show that genetic factors affect the heritability of the metabolites.

Metabolite correlation analysis was performed using Pearson's correlation coefficient, and heat maps were plotted by screening metabolites with high metabolite correlation coefficients (r > 0.8, P < 0.05; Fig. 2E and F). The results showed that lipids can be divided into several distinct clusters, while different metabolite species were also present in the clusters. This indicates that the overall similarity of metabolites of the same species is higher, and that metabolites in the same metabolic pathway are more likely to cluster.

Metabolome genome-wide association study (mGWAS)

The AIL population was sequenced using a low-coverage whole-genome sequencing strategy (LCS), and approximately 960 K SNPs were obtained after quality control for genomic coverage ranging from 0.11 to 2.60 × , the mean coverage was 0.91 ± 0.23 × (mean ± SD). Of the 2,935 metabolites (362 metabolites present in at least two detection modes), 253 metabolites had mGWAS signals, with 10,061 SNPs reaching the significance threshold [−log10 (P) > 7.29] (Fig. 3A, Table S5). These SNP loci were widely distributed throughout the genome, mainly concentrated on chr1, chr2, chr3, chr7, and chr17 (8,152/10,061 = 81.0%) (Fig. 3A, Fig. S3, Table S5). The mGWAS signal was identified at different locations in the genomes of different compound species (Fig. 3B, Fig. S4). The number of lipid-related SNPs (3,217) was the highest among all categorized metabolites, followed by amino acids, peptides, and nucleosides (Fig. 3C). For each metabolite reaching the threshold SNPs chromosome location, of which, 78.7% (199/253) metabolites had signal on only one chromosome, 21.3% (54/253) metabolites had signal on multiple chromosomes, and L-tyrosine methyl ester had signal on 7 chromosomes (Fig. 3D). All SNPs that reached the significant threshold corresponded annotated to 1,689 genes, mainly pathways including lipid metabolism, neuronal development, and regulation of hormone levels (Fig. 3E).

Fig. 3
figure 3

Overview of metabolites and SNPs identified in the mGWAS. A Chromosomal distribution of all SNPs, where the significance threshold is −log10(P) > 7.29. B Chromosomal distribution of SNPs in different metabolites classes, where the significance threshold is −log10(P) > 7.29. C The number of SNPs mapped by different classes of metabolites. Petides and Nucle., Petides and Nucleosides; Phenylpro. and Polyke., Phenylpropanoids and polyketides. D Statistics on the number of chromosomes distributed by the same metabolite SNPs. E Functional enrichment of all SNP-annotated genes

Full size image

Identification of candidate genes

Based on metabolite and gene function annotation information, we identified a number of genes associated with multiple pathways of metabolite synthesis, metabolism, and regulation (Table 1). In this study, 26 amino acids and their derivatives had GWAS signals; 1,622 SNPs were significantly associated with these compounds, mainly on chr3 and chr6 (Fig. 3B, Table S5). The main metabolic pathways involved included glycine, serine, and threonine metabolism, glyoxylate and dicarboxylate metabolism, and arginine and proline metabolism (Fig. S5). Eight genes containing SNPs significantly associated with metabolites were related to amino acid synthesis and metabolism. For example, the SNP (3:107,175,865) was significantly associated with glycine (P = 1.64E−14, Fig. 4A), and the annotated gene for this SNP locus was TDH (encoding L-threonine dehydrogenase); trans-3-aminocyclopentane-1-carboxylic acid (1:23,028,514, P = 5.38E−11) and homoarginine (1:23,005,471, P = 1.04E−11) showed a significant association with two SNPs located in a gene on chromosome 1, encoding a bifunctional enzyme that catalyzes the first two steps of the mammalian lysine degradation pathway (aminoadipate-semialdehyde synthase, AASS; Fig. 4A).

Table 1 Summary of genes associated with metabolite synthesis, metabolism and regulation in the mGWAS
Full size table
Fig. 4
figure 4

Mapping of SNP-annotated genes associated with metabolites. A Annotated gene display of amino acid analogues corresponding to SNPs. B Annotated gene display of lipid analogues corresponding to SNPs. C Annotated gene display of organic heterocyclic analogues corresponding to SNPs. Key SNPs are marked with a red diamond, and the letters a, b, c and d following metabolite names are the detection modes, T3-pos, T3-neg, Amide-pos, Amide-neg, respectively

Full size image

A total of 123 lipids had mGWAS signals, mainly glycerophospholipids, steroids, sphingolipids, and other substances (Fig. S6). A total of 3,217 SNPs of lipids reached the significance threshold, with signals mainly from chr1, chr2 and ch7 (Fig. 3B). The SNPs were annotated to 19 genes related to lipid metabolism (Table 1), mainly involving fatty acid metabolism, medium-chain fatty acid metabolic processes, and linoleic acid metabolic processes. The SNP (2:20,748,311) located within the ATP- binding cassette subfamily B member 1 gene (ABCB1) was associated with 1-hexadecanoyl-2-(9Z-octadecenoyl)-sn-glyce ro-3-phosphoserine (PC(16:0/18:1(9Z))) (P = 7.58E−09) and PA 37:10 (P = 1.87E−20, Fig. 4B); 8(R)-hydroxy-(5Z,9E,11Z,14Z)-eicosatetraenoic acid (8R-HETE) is a metabolite of arachidonic acid, which reached a significance threshold SNP (6:18,456,793) located in the intergenic region of ALOX5 and CYP2C23a genes (P = 9.44E−12, Fig. 4B). Arachidonic acid can be oxygenated by a variety of enzymes, including lipoxygenases (ALOX5, etc.), cyclooxygenases, and cytochrome P450s (CYP2C23a), and can be converted to a complex mixture of oxygenated products as a result of lipid peroxidation. The three SNPs (1:11,301,220, P = 9.05E−13; 1:11,358,208, P = 1.33E−18; 1:11,370,522, P = 4.19E−14; 1:11,301,220, P = 1.31E−12), which were significantly associated with the sphingolipid metabolite Cer were located within the CD36 molecule gene (CD36), the most important transmembrane glycoprotein mediating the uptake of oxidized LDLs (Fig. 4B).

In addition to these amino acids and lipids, oxypurinol an inhibitor of xanthine oxidase, a metabolite of allopurinol. XDH (Xanthine dehydrogenase) is a key enzyme involved in purine degradation and is related to hydroxypurinol. The SNP located in XDH (3:4,488,553) was significantly associated with oxypurinol (P = 8.79E−09, Fig. 4C). The SNP (7:5,826,779) in bilirubin was detected in both T3-pos (P = 3.70E−19) and T3-neg (P = 6.92E−34), reached a significance threshold, and was located within UDP glucuronosyltransferase family 1 member A1 gene (UGT1A1), which is associated with cholesterol synthesis (Fig. 4C).

Discussion

Chicken metabolite dataset and metabolite identification

This study established the first large-scale serum metabolomic profile in chickens. A total of 7,191 metabolites were identified using non-targeted metabolomics; these were integrated to form a chicken serum metabolite public reference dataset that can provide a reference for future chicken metabolomics studies.

The identification of unknown metabolites is an urgent problem in metabolomics research. In this study, more than 70,000 MS features were obtained by peak extraction and filter correction; however no more than 5% of the metabolites could be identified because of the limitations of the current metabolite database, which greatly limited the subsequent tests. Metabolite structure identification based on metabolic reaction networks can largely identify unknown metabolites [36]; alternatively, a widely targeted metabolomics strategy combining the high resolution and wide coverage of non-targeted technologies with the accuracy benefits of targeted MRM technologies could improve the efficiency of metabolomic assays [37]. Once the method is established, all samples can be assayed using a triple quadrupole liquid mass spectrometry instrument to obtain more accurate quantitative information [38]. A variety of widely targeted metabolomics strategies have been published, such as the full scan and ddMS2 mode combination method [39], SWATH-MS method [40], and multiple ion monitoring enhanced product ion method (MIM-EPI) [41]. We obtained all the information on non-targeted metabolomics of chicken serum, and we will continue to select and validate MRM ion pairs to establish a widely targeted metabolomics method for chicken serum in the future.

The reproducibility of high-throughput metabolomics metabolite detection was the basis for the subsequent experiments [42]. In the present study, only peaks that were present in more than three samples and responded to signals exceeding 15,000 (prefilter = c (3, 15,000)) were retained in the peak extraction stage. Also, during subsequent data processing, the mass spectrometry features present in more than 50% of the QC samples were retained, and the missing mass spectrometry features in more than 80% of the samples were removed. Finally, only the mass spectral features with CV values < 50% in the QC samples were analyzed to avoid interference caused by false-positive peaks to the maximum extent possible. With the subsequent metabolomic detection experiments on the remaining serum samples of the AIL population, metabolites that were not reproducible in both batches will be further screened to obtain more reliable analytical conclusions.

Heritability of animal metabolites

Metabolites with moderate levels of heritability have the potential to serve as biomarkers for genetic selection. Plant secondary metabolites have higher heritability than primary metabolites, and flavonoid metabolites can have heritability greater than 0.7 [6]. The literature suggests that in humans, approximately 50% of phenotypic variation in metabolite levels is genetically induced, but heritability estimates vary across metabolite classes [43]. Compared to plants, which are protected against external stimuli by secondary metabolites [44], animals are exposed to more diverse environmental stimuli, such as living environment, diet, and drugs [45, 46]. In our study, the heritability of peptides and nucleosides was higher than that of other metabolite classes in all four detection modes, this suggests their potential as biomarkers for genetic selection. A total of 34.2% of the metabolite heritability was zero, mainly influenced by environmental factors; in contrast, a study of plasma metabolites in beef cattle found that 22 of 33 metabolites had zero or negligible heritability, the non-heritable status of these metabolites may be used as a guide to animal management [17]. In our study, 7.8% of metabolite heritability was greater than 0.2. Cer 36:3 reached high heritability levels (> 0.59 in all three assay modes), and the heritability of hypoxanthine (0.38) was moderate. Studies have shown that the heritability values of long-chain polyunsaturated fatty acids in pork are usually above 0.50 [47] and that hypoxanthine is one of the highest heritability metabolites among plasma metabolites in young healthy pigs [46].

mGWAS and candidate genes

This study reported mGWAS signals for 253 metabolites representing the largest chicken serum metabolite association study to date. Metabolites and their associated loci broadly cover representatives of all major metabolic pathways, providing a comprehensive picture of how genetic variation affects blood metabolic homeostasis in chickens. We identified two candidate genes, TDH and AASS associated with amino acid metabolites, and another two candidate genes, ABCB1, CD36 associated with lipids (Table 2) [48,49,50,51,52,53,54]. Amino acids are essential for animal growth and development. In the chicken feeding process, different types and proportions of amino acids are usually added to the diet to suit the growth needs of chickens at different stages [55]. Lipids are important biomolecules in animals and have a variety of biological functions, including energy storage, signal recognition, and immunity [56]. Lipid metabolism is closely related to the maintenance of the dynamic energy balance and physiological functions in broilers [18].

Table 2 Summary of key gene and metabolite functions in the mGWAS
Full size table

Traditional GWAS have identified a large number of SNPs and candidate genes related to economic traits in chickens; for example, the end of chromosome 1 contains a QTL with a sizable effect on growth traits such as chicken body weight [13, 24], and a large number of SNPs related to chicken meat quality traits [57], egg-laying traits [58] and fat deposition have been identified [59]. However, identifying which of these many SNPs represent the main causative mutations has proven difficult. Higher-resolution association analysis using simple metabolic phenotypes instead of complex comprehensive phenotypes can provide more refined localization. For instance, the amount of fat deposited in livestock and poultry is an important economic factor because it is associated with meat quality and feed conversion rate (FCR). Here, we identified 123 lipids with mGWAS signals, of which the annotated genes PLK3 and SLC16A1 (located in this QTL region) were associated with abdominal adipogenesis in chickens [11, 60], and ITGA8 is a core gene associated with epigenetic energy in chickens [61] (Table S5). A previous study reported that multiple causative mutations cumulatively contribute to this major QTL, which may be explained by the genes playing their respective roles in different metabolic pathways affecting chicken growth traits.

This AIL population is a very valuable experimental resource, and our GWAS analysis in F9 not only reproduced the results of F2, but also further increased the localization accuracy [13]. This is the first mGWAS in an AIL line, and the related metabolic phenotypes were also obtained for the first time. The significant SNPs will be of great breeding value, and we believe that the repeated experiments in subsequent generations can be used to further reproduce and confirm the current results and further improve the precision.

Conclusions

We performed a large-scale serum non-targeted metabolomic assay on a chicken AIL population to provide the first comprehensive characterization of the chicken serum metabolic profile. Containing 7,191 metabolites, the first non-targeted in-house metabolite database of chickens was established. The mGWAS analysis was performed with the SNP dataset obtained from low-coverage sequencing and metabotypes, and a total of 10,061 SNPs for 253 metabolites were reported as genome-wide significant [−log10 (P) > 7.29]. GWAS loci were mainly concentrated in chr1, chr2, chr3, chr7 and chr17. A large number of candidate genes related to the synthesis, metabolism, and regulation of this class of metabolites were identified in amino acids, lipids, and organic heterocycles. This study provides a comprehensive picture of how genetic variation affects blood metabolic homeostasis in chickens and provides a foundation for future studies on the use of metabolic phenotypes to understand complex economic traits in animals.

Availability of data and materials

The datasets produced and/or analyzed during the current study are available from the corresponding author on reasonable request.

Abbreviations

AIL:

Advanced intercross line

CV:

Coefficient of variation

GWAS:

Genome-wide association studies

HB:

Huiyang bearded chicken

HQLA:

High-quality chicken Line A03

LCS:

Low-coverage sequencing

mGWAS:

Metabolome genome-wide association study

MIM-EPI:

Multiple ion monitoring enhanced product ion

OSI/SMMS:

Systematic and automated approach and homemade software

PCA:

Principal component analysis

QC:

Quality control

References

  1. Lichtenberg S, Trifonova OP, Maslov DL, Balashova EE, Lokhov PG. Metabolomic laboratory-developed tests: Current status and perspectives. Metabolites. 2021;11(7):423. https://doi.org/10.3390/metabo11070423.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  2. Guo H, Guo H, Zhang L, Tang Z, Yu X, Wu J, et al. Metabolome and transcriptome association analysis reveals dynamic regulation of purine metabolism and flavonoid synthesis in transdifferentiation during somatic embryogenesis in cotton. Int J Mol Sci. 2019;20(9):2070. https://doi.org/10.3390/IJMS20092070.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  3. Fan S, Shahid M, Jin P, Asher A, Kim J. Identification of metabolic alterations in breast cancer using mass spectrometry-based metabolomic analysis. Metabolites. 2020;10(4):170. https://doi.org/10.3390/metabo10040170.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  4. Chen W, Wang W, Peng M, Gong L, Gao Y, Wan J, et al. Comparative and parallel genome-wide association studies for metabolic and agronomic traits in cereals. Nat Commun. 2016;7:12767. https://doi.org/10.1038/ncomms12767.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  5. Liu M, Tang L, Liu X, Fang J, Zhan H, Wu H, et al. An evidence-based review of related metabolites and metabolic network research on cerebral ischemia. Oxid Med Cell Longevity. 2016;2016:9162074. https://doi.org/10.1155/2016/9162074.

    Article  CAS  Google Scholar 

  6. Shi T, Zhu A, Jia J, Hu X, Chen J, Liu W, et al. Metabolomics analysis and metabolite-agronomic trait associations using kernels of wheat (Triticum aestivum) recombinant inbred lines. Plant J. 2020;103(1):279–92. https://doi.org/10.1111/tpj.14727.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. Tohge T, Fernie AR. Combining genetic diversity, informatics and metabolomics to facilitate annotation of plant gene function. Nat Protoc. 2010;5(6):1210–27. https://doi.org/10.1038/nprot.2010.82.

    Article  CAS  PubMed  Google Scholar 

  8. Chen J, Hu X, Shi T, Yin H, Sun D, Hao Y, et al. Metabolite-based genome-wide association study enables dissection of the flavonoid decoration pathway of wheat kernels. Plant Biotechnol J. 2020;18(8):1722–35. https://doi.org/10.1111/pbi.13335.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  9. Zhu G, Wang S, Huang Z, Zhang S, Liao Q, Zhang C, et al. Rewiring of the fruit metabolome in tomato breeding. Cell. 2018;172(1–2):249–61.e12. https://doi.org/10.1016/j.cell.2017.12.019.

    Article  CAS  PubMed  Google Scholar 

  10. Montemayor D, Sharma K. mGWAS: Next generation genetic prediction in kidney disease. Nat Rev Nephrol. 2020;16(5):255–6. https://doi.org/10.1038/s41581-020-0270-0.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. 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. 2021;126(2):293–307. https://doi.org/10.1038/s41437-020-00365-x.

    Article  CAS  PubMed  Google Scholar 

  12. Allais S, Hennequet-Antier C, Berri C, Salles L, Demeure O, Le Bihan-Duval E. Mapping of QTL for chicken body weight, carcass composition, and meat quality traits in a slow-growing line. Poult Sci. 2019;98(5):1960–7. https://doi.org/10.3382/ps/pey549.

    Article  CAS  PubMed  Google Scholar 

  13. Wang Y, Cao X, Luo C, Sheng Z, Zhang C, Bian C, et al. Multiple ancestral haplotypes harboring regulatory mutations cumulatively contribute to a QTL affecting chicken growth traits. Commun Biol. 2020;3(1):472. https://doi.org/10.1038/s42003-020-01199-3.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. Cao X, Wang Y, Shu D, Qu H, Luo C, Hu X. Food intake-related genes in chicken determined through combinatorial genome-wide association study and transcriptome analysis. Anim Genet. 2020;51(5):741–51. https://doi.org/10.1111/age.12980.

    Article  CAS  PubMed  Google Scholar 

  15. Curtis RE, Goyal A, Xing EP. Enhancing the usability and performance of structured association mapping algorithms using automation, parallelization, and visualization in the GenAMap software system. BMC Genet. 2012;13:24. https://doi.org/10.1186/1471-2156-13-24.

    Article  PubMed  PubMed Central  Google Scholar 

  16. Osgood JA, Knight JC. Translating GWAS in rheumatic disease: Approaches to establishing mechanism and function for genetic associations with ankylosing spondylitis. Briefings Funct Genomics. 2018;17(5):308–18. https://doi.org/10.1093/bfgp/ely015.

    Article  CAS  Google Scholar 

  17. Li J, Akanno EC, Valente TS, Abo-Ismail M, Plastow GS. Genomic heritability and genome-wide association studies of plasma metabolites in crossbred beef cattle. Front Genet. 2020;11:538600. https://doi.org/10.3389/fGene.2020.538600.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Guo Y, Balasubramanian B, Zhao ZH, Liu WC. Heat stress alters serum lipid metabolism of Chinese indigenous broiler chickens-a lipidomics study. Environ Sci Pollut Res. 2021;28(9):10707–17. https://doi.org/10.1007/s11356-020-11348-0.

    Article  CAS  Google Scholar 

  19. Beauclercq S, Nadal Desbarats L, HennequetAntier C, Collin A, Tesseraud S, Bourin M, et al. Serum and muscle metabolomics for the prediction of ultimate pH, a key factor for chicken-meat quality. J Proteome Res. 2016;15(4):1168–78. https://doi.org/10.1021/acs.jproteome.5b01050.

    Article  CAS  PubMed  Google Scholar 

  20. Karisa BK, Thomson J, Wang Z, Li C, Montanholi YR, Miller SP, et al. Plasma metabolites associated with residual feed intake and other productivity performance traits in beef cattle. Livest Sci. 2014;165:200–11. https://doi.org/10.1016/j.livsci.2014.03.002.

    Article  Google Scholar 

  21. Bovo S, Mazzoni G, Galimberti G, Calo DG, Fanelli F, Mezzullo M, et al. Metabolomics evidences plasma and serum biomarkers differentiating two heavy pig breeds. Animal. 2016;10(10):1741–8. https://doi.org/10.1017/S1751731116000483.

    Article  CAS  PubMed  Google Scholar 

  22. Jia Z, Han T, Lin Q, Qu W, Jia T, Liu M, et al. Toxicity and its mechanism study of Arecae semen aqueous extract in Wistar rats by UPLC-HDMS-based serum metabolomics. Evid Based Complement Alternat Med. 2020;2020:2716325. https://doi.org/10.1155/2020/2716325.

    Article  PubMed  PubMed Central  Google Scholar 

  23. Das L, Murthy V, Varma AK. Comprehensive analysis of low molecular weight serum proteome enrichment for mass spectrometric studies. ACS Omega. 2020;5(44):28877–88. https://doi.org/10.1021/acsomega.0c04568.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  24. Sheng Z, Pettersson ME, Hu X, Luo C, Qu H, Shu D, et al. Genetic dissection of growth traits in a Chinese indigenous × commercial broiler chicken cross. BMC Genomics. 2013;14:151. https://doi.org/10.1186/1471-2164-14-151.

    Article  PubMed  PubMed Central  Google Scholar 

  25. Wang Y, Bu L, Cao X, Qu H, Zhang C, Ren J, et al. Genetic dissection of growth traits in a unique chicken advanced intercross line. Front Genet. 2020;11:894. https://doi.org/10.3389/fgene.2020.00894.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. Rodrigues RR, Gurung M, Li Z, García Jaramillo M, Greer R, Gaulke C, et al. Transkingdom interactions between Lactobacilli and hepatic mitochondria attenuate western diet-induced diabetes. Nat Commun. 2021;12(1):101. https://doi.org/10.1038/s41467-020-20313-x.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  27. Smith CA, Want EJ, O’Maille G, Abagyan R, Siuzdak G. XCMS: Processing mass spectrometry data for metabolite profiling using nonlinear peak alignment, matching, and identification. Anal Chem. 2006;78(3):779–87. https://doi.org/10.1021/ac051437y.

    Article  CAS  PubMed  Google Scholar 

  28. Zhao X, Zeng Z, Chen A, Lu X, Zhao C, Hu C, et al. Comprehensive strategy to construct in-house database for accurate and batch identification of small molecular metabolites. Anal Chem. 2018;90(12):7635–43. https://doi.org/10.1021/acs.analchem.8b01482.

    Article  CAS  PubMed  Google Scholar 

  29. Djoumbou Y, Eisner R, Knox C, Chepelev L, Hastings J, Owen G, et al. ClassyFire: Automated chemical classification with a comprehensive, computable taxonomy. J Cheminf. 2016;8(1):61. https://doi.org/10.1186/s13321-016-0174-y.

    Article  Google Scholar 

  30. Pang Z, Chong J, Zhou G, de Lima Morais DA, Chang L, Barrette M, et al. MetaboAnalyst 5.0: Narrowing the gap between raw spectra and functional insights. Nucleic Acids Res. 2021;49(W1):W388–96. https://doi.org/10.1093/nar/gkab382.

  31. Zhou Y, Zhou B, Pache L, Chang M, Khodabakhshi AH, Tanaseichuk O, et al. Metascape provides a biologist-oriented resource for the analysis of systems-level datasets. Nat Commun. 2019;10(1):1523. https://doi.org/10.1038/s41467-019-09234-6.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  32. Yang R, Guo X, Zhu D, Tan C, Bian C, Ren J, et al. Accelerated deciphering of the genetic architecture of agricultural economic traits in pigs using a low-coverage whole-genome sequencing strategy. GigaScience. 2021;10(7):giab048. https://doi.org/10.1093/gigascience/giab048.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. Yang J, Lee SH, Goddard ME, Visscher PM. GCTA: A tool for genome-wide complex trait analysis. Am J Hum Genet. 2011;88(1):76–82. https://doi.org/10.1016/j.ajhg.2010.11.011.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. Visscher PM, Yang J, Goddard ME. A commentary on “common SNPs explain a large proportion of the heritability for human height” by Yang et al. (2010). Twin Res Hum Genet. 2010;13(6):517–24. https://doi.org/10.1375/twin.13.6.517.

    Article  PubMed  Google Scholar 

  35. Cingolani P, Platts A, le Wang L, Coon M, Nguyen T, Wang L, et al. A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w1118; iso-2; iso-3. Fly (Austin). 2012;6(2):80–92. https://doi.org/10.4161/fly.19695.

    Article  CAS  PubMed  Google Scholar 

  36. Shen X, Wang R, Xiong X, Yin Y, Cai Y, Ma Z, et al. Metabolic reaction network-based recursive metabolite annotation for untargeted metabolomics. Nat Commun. 2019;10(1):1516. https://doi.org/10.1038/s41467-019-09550-x.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  37. Chu C, Du Y, Yu X, Shi J, Yuan X, Liu X, et al. Dynamics of antioxidant activities, metabolites, phenolic acids, flavonoids, and phenolic biosynthetic genes in germinating Chinese wild rice (Zizania latifolia). Food Chem. 2020;318:126483. https://doi.org/10.1016/j.foodchem.2020.126483.

    Article  CAS  PubMed  Google Scholar 

  38. Matsuda F, Okazaki Y, Oikawa A, Kusano M, Nakabayashi R, Kikuchi J, et al. Dissection of genotype-phenotype associations in rice grains using metabolome quantitative trait loci analysis. Plant J. 2012;70(4):624–36. https://doi.org/10.1111/j.1365-313X.2012.04903.x.

    Article  CAS  PubMed  Google Scholar 

  39. Xuan Q, Hu C, Yu D, Wang L, Zhou Y, Zhao X, et al. Development of a high coverage pseudotargeted lipidomics method based on Ultra-high performance liquid chromatography-mass spectrometry. Anal Chem. 2018;90(12):7608–16. https://doi.org/10.1021/acs.analchem.8b01331.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Zha H, Cai Y, Yin Y, Wang Z, Li K, Zhu ZJ. SWATHtoMRM: Development of high-coverage targeted metabolomics method using SWATH technology for biomarker discovery. Anal Chem. 2018;90(6):4062–70. https://doi.org/10.1021/acs.analchem.7b05318.

    Article  CAS  PubMed  Google Scholar 

  41. Chen W, Gong L, Guo Z, Wang W, Zhang H, Liu X, et al. A novel integrated method for large-scale detection, identification, and quantification of widely targeted metabolites: application in the study of rice metabolomics. Mol Plant. 2013;6(6):1769–80. https://doi.org/10.1093/mp/sst080.

    Article  CAS  PubMed  Google Scholar 

  42. Ghosh T, Philtron D, Zhang W, Kechris K, Ghosh D. Reproducibility of mass spectrometry based metabolomics data. BMC Bioinf. 2021;22(1):423. https://doi.org/10.1186/s12859-021-04336-9.

    Article  CAS  Google Scholar 

  43. Hagenbeek FA, Pool R, van Dongen J, Draisma HHM, Jan Hottenga J, Willemsen G, et al. Heritability estimates for 361 blood metabolites across 40 genome-wide association studies. Nat Commun. 2020;11(1):39. https://doi.org/10.1038/s41467-019-13770-6.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  44. Salih AM, Al-Qurainy F, Nadeem M, Tarroum M, Khan S, Shaikhaldein HO, et al. Optimization method for phenolic compounds extraction from medicinal plant (Juniperus procera) and phytochemicals screening. Molecules. 2021;26(24):7454. https://doi.org/10.3390/molecules26247454.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  45. Menni C, Zhai G, Macgregor A, Prehn C, Römisch-Margl W, Suhre K, et al. Targeted metabolomics profiles are strongly correlated with nutritional patterns in women. Metabolomics. 2013;9(2):506–14. https://doi.org/10.1007/s11306-012-0469-6.

    Article  CAS  PubMed  Google Scholar 

  46. Dervishi E, Yang T, Dyck MK, Harding JCS, Fortin F, Cheng J, et al. Heritability and genetic correlations of plasma metabolites of pigs with production, resilience and carcass traits under natural polymicrobial disease challenge. Sci Rep. 2021;11(1):20628. https://doi.org/10.1038/s41598-021-99778-9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  47. Ntawubizi M, Colman E, Janssens S, Raes K, Buys N, De Smet S. Genetic parameters for intramuscular fatty acid composition and metabolism in pigs. J Anim Sci. 2010;88(4):1286–94. https://doi.org/10.2527/jas.2009-2355.

    Article  CAS  PubMed  Google Scholar 

  48. Alexander PB, Wang J, McKnight SL. Targeted killing of a mammalian cell based upon its specialized metabolic state. PNAS. 2011;108(38):15828–33. https://doi.org/10.1073/pnas.1111312108.

    Article  PubMed  PubMed Central  Google Scholar 

  49. Davis AJ, Austic RE. Dietary protein and amino acid levels alter threonine dehydrogenase activity in hepatic mitochondria of Gallus domesticus. J Nutr. 1997;127(5):738–44. https://doi.org/10.1093/jn/127.5.738.

    Article  CAS  PubMed  Google Scholar 

  50. Ryan WL, Barak AJ, Johnson RJ. Lysine, homocitrulline, and homoarginine metabolism by the isolated perfused rat liver. Arch Biochem Biophys. 1968;123(2):294–7. https://doi.org/10.1016/0003-9861(68)90137-9.

    Article  CAS  PubMed  Google Scholar 

  51. Neumann J, Rose-Sperling D, Hellmich UA. Diverse relations between ABC transporters and lipids: An overview. Biochim Biophys Acta-Biomembr. 2017;1859(4):605–18. https://doi.org/10.1016/j.bbamem.2016.09.023.

    Article  CAS  PubMed  Google Scholar 

  52. Van Helvoort A, Smith AJ, Sprong H, Fritzsche I, Schinkel AH, Borst P, et al. MDR1 P-glycoprotein is a lipid translocase of broad specificity, while MDR3 P-glycoprotein specifically translocates phosphatidylcholine. Cell. 1996;87(3):507–17. https://doi.org/10.1016/s0092-8674(00)81370-7.

    Article  PubMed  Google Scholar 

  53. Park WJ, Park JW, Merrill AH, Storch J, Pewzner-Jung Y, Futerman AH. Hepatic fatty acid uptake is regulated by the sphingolipid acyl chain length. Biochim Biophys Acta. 2014;1841(12):1754–66. https://doi.org/10.1016/j.bbalip.2014.09.009.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  54. Ritter JK, Crawford JM, Owens IS. Cloning of two human liver bilirubin UDP-glucuronosyltransferase cDNAs with expression in COS-1 cells. J Biol Chem. 1991;266(2):1043–7. https://doi.org/10.1016/S0021-9258(17)35280-8.

    Article  CAS  PubMed  Google Scholar 

  55. He W, Li P, Wu G. Amino acid nutrition and metabolism in chickens. Adv Exp Med Biol. 2021;1285:109–31. https://doi.org/10.1007/978-3-030-54462-1_7.

    Article  PubMed  Google Scholar 

  56. Ottaviani E, Malagoli D, Franceschi C. The evolution of the adipose tissue: A neglected enigma. Gen Comp Endocrinol. 2011;174(1):1–4. https://doi.org/10.1016/j.ygcen.2011.06.018.

    Article  CAS  PubMed  Google Scholar 

  57. Sun Y, Zhao G, Liu R, Zheng M, Hu Y, Wu D, et al. The identification of 14 new genes for meat quality traits in chicken using a genome-wide association study. BMC Genomics. 2013;14:458. https://doi.org/10.1186/1471-2164-14-458.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  58. Zhao X, Nie C, Zhang J, Li X, Zhu T, Guan Z, et al. Identification of candidate genomic regions for chicken egg number traits based on genome-wide association study. BMC Genomics. 2021;22(1):610. https://doi.org/10.1186/s12864-021-07755-3.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  59. Abdalla BA, Chen J, Nie Q, Zhang X. Genomic insights into the multiple factors controlling abdominal fat deposition in a chicken model. Front Genet. 2018;9:262. https://doi.org/10.3389/fgene.2018.00262.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  60. Ma X, Sun J, Zhu S, Du Z, Li D, Li W, et al. MiRNAs and mRNAs analysis during abdominal preadipocyte differentiation in chickens. Animal (Basel). 2020;10(3):468. https://doi.org/10.3390/ani10030468.

    Article  Google Scholar 

  61. Pezeshkian Z, Mirhoseini SZ, Ghovvati S. Identification of hub genes involved in apparent metabolizable energy of chickens. Anim Biotechnol. 2022;33(2):242–9. https://doi.org/10.1080/10495398.2020.1784187.

    Article  CAS  PubMed  Google Scholar 

Download references

Acknowledgements

We are grateful to Dingming Shu, Hao Qu and Chenglong Luo in State Key Laboratory of Livestock and Poultry Breeding, Guangdong Key Laboratory of Animal Breeding and Nutrition, Institute of Animal Science, Guangdong Academy of Agricultural Sciences for sample collection. We are also grateful to Jingqiang Zhang in State Key Laboratory of Agrobiotechnology, College of Biological Sciences, China Agricultural University for metabolome detection.

Funding

This study was supported by National Natural Science Foundation of China (No. 32172719, U2002205, 32272862).

Author information

Author notes

  1. Jing Tian and Xiaoning Zhu contributed equally to this work.

Authors and Affiliations

  1. State Key Laboratory of Agrobiotechnology, College of Biological Sciences, China Agricultural University, Beijing, 100193, China

    Jing Tian, Xiaoning Zhu, Hanyu Wu, Yuzhe Wang & Xiaoxiang Hu

  2. National Research Facility for Phenotypic and Genotypic Analysis of Model Animals (Beijing), China Agricultural University, Beijing, 100193, China

    Hanyu Wu, Yuzhe Wang & Xiaoxiang Hu

Authors
  1. Jing Tian
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Xiaoning Zhu
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Hanyu Wu
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Yuzhe Wang
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Xiaoxiang Hu
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

XXH and YZW designed the project. JT performed the experiments. XNZ administered the bioinformatic analysis. HYW provided study reagents and instrumentation. JT and YZW wrote and revised the manuscript. All authors read and approved the final manuscript.

Corresponding authors

Correspondence to Yuzhe Wang or Xiaoxiang Hu.

Ethics declarations

Ethics approval and consent to participate

All animals used in this study were cared for, and experiments conducted using procedures, that complied with the requirements of the Animal Welfare Committee of Agrobiotechnology of China Agricultural University (approval SKLAB-2014–06-07).

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Supplementary Information

Additional file 1: Fig. S1.

In-house non-targeted database of chicken serum metabolites. Fig. S2. Principal component analysisof all features on four detected modes. Fig. S3. Number of GWAS signaling metabolites on each chromosome. Fig. S4. Manhattan plot showing GWAS signals of another 7 classes of metabolites. Fig. S5. Analysis of the amino acid metabolites pathwayin the metabolites of GWAS signal using MetaboAnalyst. Fig. S6. Classification of lipids in the metabolites of GWAS signal using MetaboAnalyst.

Additional file 2: Table S1.

Segmented QC samples identified to metabolites under four detection modes. Table S2. Non-targeted metabolites of chicken serum. Table S3. The statistical results of coefficient of variation. Table S4. The statistical results of broad-sense heritability. Table S5. All significant SNP loci of GWAS signaling metabolites > 7.29.

Additional file 3: Table S6.

Metabolic phenotype QTL intervals.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Tian, J., Zhu, X., Wu, H. et al. Serum metabolic profile and metabolome genome-wide association study in chicken. J Animal Sci Biotechnol 14, 69 (2023). https://doi.org/10.1186/s40104-023-00868-7

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s40104-023-00868-7

Keywords

Journal of Animal Science and Biotechnology

ISSN: 2049-1891