Effect of host genetics and gut microbiome on fat deposition traits in pigs

Fat deposition affects meat quality, flavor, and production in pigs. Fat deposition is influenced by both genetics and environment. Symbiotic microbe with the host is an important environmental factor to influence fat deposition. In this study, the fat deposition traits were measured in 239 individuals obtained from Tongcheng pigs × Large White pigs resource population. The interactions between genetics and gut microbiome in fat deposition traits were investigated through whole-genome sequencing and cecum microbial 16S ribosomal RNA sequencing. The results showed that the percentage of leaf fat (PL) and intramuscular fat content (IMF) were significantly influenced by host genetics–gut microbiome interaction. The effects of interactions between host genetics and gut microbiome on PL and IMF were 0.13 and 0.29, respectively. The heritability of PL and IMF was estimated as 0.71 and 0.89, respectively. The microbiability of PL and IMF was 0.20 and 0.26, respectively. Microbiome-wide association analysis (MWAS) revealed Anaeroplasma, Paraprevotella, Pasteurella, and Streptococcus were significantly associated with PL, and Sharpea and Helicobacter exhibited significant association with IMF (p < 0.05). Furthermore, Paraprevotella was also identified as a critical microbe affecting PL based on the divergent Wilcoxon rank-sum test. Overall, this study reveals the effect of host genetics and gut microbiome on pig fat deposition traits and provides a new perspective on the genetic improvement of pig fat deposition traits.

Introduction

Fat deposition is closely related to carcass and meat quality traits in pigs. Fat deposition in different parts of the body has different effects. Subcutaneous backfat thickness is an important target trait in pig breeding. There is a negative correlation between backfat thickness and carcass lean percentage. In the viscera, fat could support and protect organs, and it affects dressing percentage and feed consumption in pigs. In addition, intramuscular fat content (IMF) is an important factor determining the tenderness, juiciness, flavor, meat quality characteristics, and consumer acceptance (Gao and Zhao, 2009). IMF is positively correlated with juiciness and flavor of meat (Fortin et al., 2005). Thus, fat deposition is an important economic trait of pigs, which affects meat quality and production in pig. Furthermore, the physiological traits and biochemical indices of pigs are similar to humans, thus, pig can be used as a model animal for studying human obesity (Hishikawa et al., 2005).

Fat deposition traits are a type of quantitative traits with moderate-to-high heritability. This quantitative trait is influenced by heredity and environment. The gut microbiota that are symbiotic with the host are irreplaceable environmental factors. The gut microbiome is known as the second genome of humans (Zhu et al., 2010), which inhabits 10 times more gut microbial cells than human cells (Savage, 1977). Pig has richer gut microbes and is considered to be the main model animal of human obesity and disease research (Hishikawa et al., 2005; Merlotti et al., 2017). A reference gene catalog of pig gut microbiome with a total of 7.69 million genes has been established by metagenome sequencing of 278 pig fecal microbiome (Xiao et al., 2016).

The influence of gut microbe on host phenotype, the ratio of microbial relative abundance variance to the phenotype variance, is defined as “microbiability” (Difford et al., 2018). Microbiability has been used for the studies in pigs (Camarinha-Silva et al., 2017; Tang et al., 2020), cattle (Difford et al., 2018), and chicken (Wen et al., 2019b). According to the microbiability of traits, cecal microbiota has a greater contribution to fat deposition than other large intestine microorganisms (Wen et al., 2019b; Tang et al., 2020). Moreover, the Treponema in cecum has been reported to be associated with feed conversion (Quan et al., 2018). The Prevotellaceae ucg-001 in cecum is positively correlated with backfat thickness and IMF (Tang et al., 2020). Besides, the cecum exhibits high capabilities of degradation and digestion, whose metabolites can provide energy for the host (Yang et al., 2016). Therefore, the cecum could be considered as the representative of gut segments for fat deposition research.

However, the effect of interactions between host genetics and gut microbiome on fat deposition traits remains largely unclear. In our study, the resource population was obtained by crossing a Chinese local breed (Tongcheng pigs) and a commercial breed (Large White pigs). We measured fat deposition traits (backfat thickness at shoulder (SBFT), loin (LBFT), rump (RBFT), average backfat thickness (ABFT), percentage of leaf and caul fat (PL, PC, and PLC), and intramuscular fat content (IMF)). We performed whole-genome sequencing and cecal microbial 16S ribosomal RNA sequencing of 239 pigs to determine whether there are any interactions between the host genetics and gut microbiome. Then, we estimated heritability and microbiability for the detected interactions between host genetic–gut microbiomes and further screened candidate microbes influencing traits. This study provides a new perspective and opportunity for the genetic improvement of pig fat deposition traits.

Materials and methods

Resource population, fat deposition traits measured, and sample collection

The resource population is an advanced generation intercross population. F1 was obtained by crossing 36 Tongcheng sows (a Chinese local breed) with 11 Large White boars (a commercial breed). The F1 sows and boars were subsequently intercrossed to produce F2. Then, intercross was performed in each generation until F9 population was produced. In this study, a total of 239 individuals from 9 to 10 intercross generations including 119 castrated boars and 120 castrated sows were used for further analysis. All the pigs were provided by Yunzhi Farm (Tongcheng Country, Hubei province, China) with coincident feeding and management conditions. All individuals were healthy, and they were not administered any antibiotics before being slaughtered.

The phenotypes of 239 pigs were collected, including live body weight (BW) before slaughter, carcass weight (CW), backfat thickness at the thickest shoulder (SBFT), loin (LBFT), rump (RBFT), weight of leaf fat, weight of caul fat, and intramuscular fat content (IMF). Backfat thickness was measured using Vernier caliper. The average backfat thickness (ABFT) was an average of three measurements of backfat thickness at the shoulder, loin, and rump. The weight of leaf and caul fat was measured using a scale with an accuracy of 0.01 kg. The percentage of leaf fat (PL) was calculated as the weight of leaf fat divided by CW. Similarly, percentage of caul fat (PC) was calculated as the weight of caul fat divided by CW. The percentage of leaf and caul fat (PLC) was sum of PL and PC. A near-infrared spectroscopy analyzer was used to measure the IMF of the longissimus dorsi samples. The longissimus dorsi samples were collected from the penultimate 3rd to 4th thoracic vertebrae, about 20–30 cm long. The gut contents of 239 individuals were sampled from cecum after pigs were slaughtered, snap-frozen in liquid nitrogen, and stored at −80°C until sequencing. Spleen tissues were collected into 50 ml centrifuge tubes containing 75% alcohol by volume and stored at −20°C.

All the animal experiments in this study were approved by the Animal Care and Use Committee of Huazhong Agricultural University.

Whole-genome sequencing and analysis

Genomic DNA samples were extracted from spleen of 239 pigs to construct the whole-genome sequencing libraries. Each library was sequenced using 150 bp paired-end reads with a HiSeq X5 instrument (Illumina). To minimize mapping error, low-quality reads were removed using FastQC software. The clean reads from each pig were aligned to the porcine reference genome using the Burrows-Wheeler Alignment tool (BWA ver 0.7.15) with default parameters. We subsequently used the Picard toolkit to sort the alignment results and remove potential PCR duplicate reads. The resultant alignments were indexed using SAMtools (ver 1.6) and processed with the Genome Analysis Toolkit (GATK, ver 3.7). To call variants, we set a minimum quality score as 20 based on the bases and mapped reads. The single nucleotide polymorphisms (SNPs) of each pig were combined to obtain a common dataset of single nucleotide polymorphism (SNP) data, and the dataset was processed by GATK. Finally, the software PLINK (ver 1.90) was used for quality control of the obtained dataset with the following filtering criteria: SNP call rate >80%, minor allele frequency >1%, and Hardy–Weinberg equilibrium p-value <10^–6.

16S ribosomal ribonucleic acid sequencing and analysis

First, the total DNA of cecal contents was extracted to construct sequencing library, and 30 ng genomic DNA samples and their corresponding fusion primers were used to prepare PCR reaction system. PCR reaction was performed to amplify V3-V4 region of 16S ribosomal RNA (rRNA) with the primers of 338R (ACTCCTACGGGAGGCAGCAG) and 806R (GGACTACHVGGGGTWTCTAAT). A 468 bp segment was amplified by PCR with the number of tags of 50,000. The PCR products were purified by using Agencourt AMPure XP magnetic beads and dissolved in the Elution Buffer. Agilent 2100 Bioanalyzer was used to detect the fragment range and concentration of the library. According to the size of the inserted fragment, HiSeq platform was selected for pair-end sequencing. FLASH (Fast Length Adjustment of Short reads, ver 1.2.11) was used to splice the sequences. Based on the sequence overlapping relationship, the paired reads obtained from pair-end sequencing were assembled into tags. The assembled tags were clustered into operational taxonomic units (OTUs) using software USEARCH (ver 7.0.1090). Clustering analysis was performed by using UPARSE under 97% similarity to obtain representative sequence of each OTU. Afterward, OTU representative sequences were aligned against the Greengene database by ribosomal database project (RDP) classifier (ver 2.2) software to obtain the annotation at the levels of phylum, class, order, family, and genus.

Analysis of interactions between host genetics and gut microbiome

Construction of genomic relationship matrix

A total of 14,139,625 filtered SNPs were used to estimate the genomic relationship matrix (GRM) using the HIBLUP software¹ according to the following equation:

$$G={\mathit{MM}}^{{}^{\prime }}/2\sum _{i=1}^n{p}_i(1-p_i)$$

In which M indicates m (the number of individuals) × n (the number of loci) matrix, and p_i is the frequency of the coded allele.

Construction of microbial relationship matrix

All the OTUs were used to construct the microbial relationship matrix (MRM) by R script (Wen et al., 2019b) based on the following equation:

$$m_{\mathit{ij}}=\frac{1}{N}\sum _{o=1}^N\frac{\left(X_{\mathit{io}}-X_o\right)(X_{\mathit{jo}}-X_o)}{{\mathrm{\sigma }}_o^2}$$

In which m_ij is the estimation of microbial similarity in cecum of individual i and individual j; X_io and X_jo indicate the relative abundance of OTU o in the cecum of individual i and individual j; X_o represents the average relative abundance of OTU o in cecum of the whole population; ${\mathrm{\sigma }}_o^2$ is the variance of OTU o relative abundance; and N is the total OTU count in cecum.

Interactions between host genetics and gut microbiome

The following multiple random effects model was established to estimate variance components of the target traits using HIBLUP software (see text footnote 1).

$$y=\mathrm{\hspace{0.33em}1}\mu +Z_{1}g+Z_{2}m+Z_{3}a+e$$

In which y is the n × 1 vector of the fat deposition traits; μ is the overall mean; g ∼ N (0,$G{\mathrm{\sigma }}_g^2$) is the q × 1 vector of host genetic random effect, where G and ${\mathrm{\sigma }}_g^2$ are the GRM and host genetic variance; m ∼ N (0,$M{\mathrm{\sigma }}_m^2$) is the q × 1 vector of gut microbiome random effect, where M and ${\mathrm{\sigma }}_m^2$ are the MRM and gut microbiome variance; a ∼ N (0,$A{\mathrm{\sigma }}_a^2$) is the q × 1 vector of interactions between host genetics and gut microbiome random effect, where A and ${\mathrm{\sigma }}_a^2$ are the GRM × MRM and variance of interactions between host genetics and gut microbiome; e ∼ N (0,${\mathrm{\sigma }}_e^2$)is an n × 1 vector of residual effect, where ${\mathrm{\sigma }}_e^2$ is the residual variance; Z₁, Z₂, and Z₃ are, respectively, the corresponding incidence matrices of g, m, and a.

Estimation of heritability of fat deposition traits

A total of 14,139,625 filtered SNPs were used to estimate the principal components and heritability based on GRM using genome-wide complex trait analysis (GCTA) software [ver 1.93.1; (Yang et al., 2011)] according to the following equation:

$$y=Kc+g_1+\epsilon$$

where y is a vector of the phenotype; c is a vector of fixed covariates, including sex effect and body weight (BW); K is the matrix corresponding to c; and g₁ is a vector of the total effects of all SNPs following g₁ ∼ N (0,$G^{\prime }{\mathrm{\sigma }}_{g_1}^2$) where G′and ${\mathrm{\sigma }}_{g_1}^2$ are the GRM and genetic variance, respectively; and ε is the residual effect. The G′ estimation equation is as follows:

$$g_{\mathit{jk}}=\frac{1}{N}\sum _{i=1}^N\frac{\left(X_{\mathit{ij}}-2p_i\right)(X_{\mathit{ik}}-2p_i)}{2p_i(1-p_i)}$$

In which g_jk is the genetic similarity between individual j and individual k; X_ij and X_ik represent the number of reference alleles in individual j and individual k; p_i is the reference allele frequency; and N is the SNP number.

Estimation of microbiability of fat deposition traits

The microbiability is referred to the ratio of cecum microbial variance to phenotypic variance, and it was calculated by the following equation:

$$y=Kc+m_1+\mathrm{\epsilon }$$

where y and ε are phenotype and residual effect vectors, respectively. m₁ is the gut microbiota effect following the m₁ ∼ N (0, M${\mathrm{\sigma }}_{m_1}^2$). The microbiability was estimated by GCTA software with the MRM substituted for GRM. In our study, c contained six covariates, namely, covariates 1–6. Covariate 1 included BW and the first five host genetic principal components (PCs); covariate 2 included BW and the first two PCs generated at SNPs which were significantly associated with PL and IMF; covariate 3 consisted of BW, the first five host genetic PCs, and the first two PCs generated at SNPs which were significantly associated with PL and IMF; covariate 4 included BW, sex, and the first five host genetic PCs; covariate 5 included BW, sex, and the first two PCs generated at SNPs which were significantly associated with PL and IMF; covariate 6 consisted of BW, sex, the first five host genetic PCs, and the first two PCs generated at SNPs which were significantly associated with PL and IMF. The first five PCs were generated by the genome relationship matrix to account for population structure.

The Bayesian information criterion (BIC) was calculated to select the model with optimal covariates. The smaller the BIC value was, the higher the model’s fitting level was. The BIC was calculated according to the following formula:

$$\mathit{BIC}=\mathit{kIn}\left(n\right)-2\mathit{In}(L)$$

where k is the number of model parameters, n is the number of samples, and L is the likelihood value.

Identification of candidate microbes with fat deposition traits

The divergent Wilcoxon rank-sum test and microbiome-wide association analysis (MWAS) were used to identify the candidate microbes for target traits.

Differences in genera, percentage of leaf fat, and intramuscular fat content between divergent groups

According to the distribution of fat deposition traits of the 239 individuals, the highest 20% (n = 48) and lowest 20% (n = 48) were formed into two extreme divergent groups. The Wilcoxon rank-sum test was used to test the significance in 56 genera with average relative abundance greater than 0.01% between the two groups. By the same method, microbes were grouped in terms of the relative abundance of the genus, and the significance in individual traits between the two divergent groups was tested by the Wilcoxon rank-sum test with the P < 0.05 as the chosen significant test criteria.

Microbiome-wide association analysis of percentage of leaf fat and intramuscular fat content

The MWAS between the fat deposition traits of 239 individuals and 56 genera whose average relative abundance was more than 0.01% was conducted using Gaston package² in R software with the following equation:

$$y=X\beta +u+\epsilon$$

y is the vector of fat deposition traits; β is a fixed effect vector, including 56 genera with average relative abundance value >0.01%; u is a random effect vector. At the same time, sex and BW are considered as covariates. F-test is used to determine the significance of the regression coefficients.

Results

Characterization of fat deposition traits and sequencing outcomes

Host fat deposition phenotype characteristics are presented in Table 1. There are three categories of fat depositions including subcutaneous fat (SBFT, RBFT, LBFT, and ABFT), viscera fat (PL, PC, and PLC), and intramuscular fat (IMF). All measured traits displayed a high coefficient of variation (15.39 to 42.69%) (Table 1). The distribution of each trait is shown in the Supplementary Figure 1. The host genetics was analyzed by the whole-genome sequencing, and the whole-genome sequencing information of the 239 individuals was analyzed subsequently. The SNPs were quality-controlled by Plink software. After quality control, a total of 14,139,625 SNPs was obtained for subsequent analysis.

TABLE 1

Table 1. Summary of fat deposition traits in the resource population.

The gut microbiome was analyzed by the cecal 16S rRNA sequencing. The 16S rRNA sequencing analysis produced a total of 31,801,522 reads from the 239 samples with an average of 133,061 reads, and 2,324 OTUs were then clustered with 97% sequencing identity. Subsequently, these 2,324 OTUs were clustered into 19 phyla, 33 classes, 56 orders, 87 families, and 144 genera (Supplementary Table 1). At the phylum level, the phyla with relative abundance at the top five were Firmicutes, Bacteroidetes, Proteobacteria, Fusobacteria, and Spirochaetes (Figure 1). At the genus level, the average relative abundance of 56 genera was greater than 0.01% (Supplementary Table 2). The average Sobs index, Chao index, Ace index, Shannon index, and Simpson index of the microbiota are 782.92, 970.75, 962.85, 4.63, and 0.04, respectively (Supplementary Table 3).

FIGURE 1

Figure 1. Phylum of relative abundance.

Influence of host genetic–gut microbiome interactions on percentage of leaf fat and intramuscular fat content

The effects of host genetics and gut microbiome on all traits were different (Table 2). The results indicated that host genetics (g) and gut microbiome (m) exhibited 0.62 and 0.20 independent effect on SBFT, 0.68 and 0.18 independent effect on LBFT, and 0.86 and 0.11 independent effect on RBFT, respectively. In addition, ABFT displayed 0.75 g effect and 0.18 m effect. Generally, the g effect was higher than the m effect in the presence of other environmental factors. RBFT exhibited the highest g effect among backfat thickness at three positions. The PC exhibited 0.40 g effect and 0.24 m effect. However, PL displayed 0.77 g effect, 0.10 m effect, and 0.13 effect of interactions between host genetics and gut microbiome (a). The multiple variance components completely explained the phenotypic variation of PL and IMF. The IMF showed 0.71 independent g effect and 0.29 a effect, but no independent m effect. The a effect of PL and IMF was estimated as 0.13 and 0.29, indicating that host genetic–gut microbiome interactions affected the formation of PL and IMF.

TABLE 2

Table 2. Host genetics and gut microbiome effects for fat deposition traits.

Heritability and microbiability of percentage of leaf fat and intramuscular fat content

The heritability of PL was estimated as 0.71 and that of IMF as 0.89. PL and IMF were mainly affected by host genetics in the resource population (Table 3). To estimate microbiability of PL and IMF, six models were established based on host genetics, population structure, sex, and BW (Supplementary Table 4). The BIC value was used to select optimal model fitting PL and IMF. The optimal model to calculate PL microbiability contained three covariates including BW, the first five host genetic principal components (PCs), and the first two PCs generated at SNPs significantly associated with PL. The microbiability of PL was estimated as 0.20 (Table 3). The optimal model to calculate IMF microbiability contained two covariates including BW and the first two PCs generated at SNPs significantly associated with IMF. The microbiability of IMF was estimated to be 0.26 (Table 3).

TABLE 3

Table 3. Heritability and microbiability of PL and IMF in the resource population.

The genera that are significant association with percentage of leaf fat and intramuscular fat content

To identify the candidate microbiome from the resource population, we tested the significance in traits or relative abundance between the highest 20% and lowest 20% microbial abundance or trait groups by Wilcoxon rank-sum test. We chose the microbes with significant difference in trait and genus relative abundance between divergent groups as candidate microbes (Figure 2) in the subsequently association analysis. PL and IMF had different candidate microbes. Campylobacter and Paraprevotella exhibited significant difference between divergent groups of PL, while PL also showed significant difference between divergent groups of Campylobacter and Paraprevotella (P < 0.05, Table 4). PL was 6.18% in the highest 20% group of Campylobacter relative abundance, while PL was 7.08% in the lowest 20% group (P < 0.05, Table 4). Then, PL was 6.25% in the highest 20% group of Paraprevotella relative abundance, while it was 6.95% in the lowest 20% group (P < 0.05, Table 4). For divergent Wilcoxon rank-sum test of IMF, Actinobacillus, Dialister, and YRC22 exhibited significant differences in relative abundance between highest 20% and lowest 20% groups (P < 0.05, Table 4). For divergent Wilcoxon rank-sum test of genus relative abundance, Anaeroplasma, Megasphaera, and Succinivibrio exhibited significant differences between highest 20% and lowest 20% IMF (P < 0.05, Table 4). However, there were no overlapping genera between divergent groups of traits and genus relative abundance (Table 4).

FIGURE 2

Figure 2. Microbial detections via Wilcoxon rank-sum tests between divergent groups. Microbial abundance in the highest 20% group with lowest 20% group traits. Displayed from the outer to the inner circle is the trait, the significance test in trait between the two groups with the highest and lowest microbial abundance (P_m–trait), the significance test in each microbial abundance between the highest and lowest trait (P_trait–m), where p-values are plotted as –log10 (p-value); the red line shows the significance threshold (P < 0.05). Each point represents a microbe, and the big point indicates the p-value passed the significance threshold. The dot on the same gray dashed line indicates that the P_m–trait and P_trait–m values for one microbe are all <0.05.

TABLE 4

Table 4. Microbiome relative abundance and trait Wilcoxon test of highest 20% and lowest 20% groups in the resource population.

Microbiome-wide association analysis of percentage of leaf fat and intramuscular fat content

The MWAS was used to investigate whether there were significant associations between genus relative abundance and traits (PL and IMF). The MWAS results are shown in Supplementary Table 5. The number of microbes significantly associated PL and IMF was 4 and 2, respectively (P < 0.05, Table 5). Notably, some genera showed high significant associations with traits. For example, Anaeroplasma was significantly associated with PL (P < 0.05), whereas Sharpea with IMF. Moreover, Paraprevotella, Pasteurella, and Streptococcus were also significantly associated with PL (P < 0.05), while Helicobacter exhibited significant association with IMF (P < 0.05).

TABLE 5

Table 5. Microbes significantly associated with PL and IMF.

We adopted two methods (divergent Wilcoxon rank-sum test and MWAS) to screen microbes affecting PL and IMF at the genus level and found that Paraprevotella was significantly associated with PL in both methods (P < 0.05). As a significant candidate genus, Paraprevotella is one of the most important butyric acid-producing bacteria, and it may affect visceral fat deposition by producing butyric acid (a short-chain fatty acid).

Discussion

Fat deposition trait is closely related to carcass and meat quality traits in pigs, and it affects economic income and feed consumption. Pig fat deposition traits are influenced by host genetics and gut microbiome. Our study found that PL and IMF were influenced by the interactions between host genetics and gut microbiome. Several studies have identified that host genetics have interacted with gut microbiota and some host genetic variations affect the gut microbiota (Li et al., 2019; Bergamaschi et al., 2020; Wen et al., 2021). In our study, the IMF showed 0.29 effect of interactions between host genetics and gut microbiome, but no independent effect of gut microbiome. The reason that the IMF showed no independent effect of gut microbiome may be that gut microbiota affects IMF by interacting with host genetics, but not alone. Then, we estimated the heritability and microbiability of PL and IMF. In this study, a total of 14,139,625 SNPs were obtained from whole-genome sequencing of 239 samples from the resource population. GCTA was used to construct the GRM and estimate heritability of PL and IMF. The heritability of PL and IMF was estimated as 0.71 and 0.89, respectively. The proportion of SNP variance has been reported to be SNP heritability of traits (Yang et al., 2017). Heritability is calculated by genome sequencing data to analyze the effect of genetics on human body weight (Yang et al., 2010). The heritability is influenced by breed (Lopez et al., 2018), sample size (Tang et al., 2020), and sequencing methods (Uemoto et al., 2017). In addition, the heritability depends on the population, because both the variation in additive and non-additive genetic factors of each trait, and the environmental variance are population-specific (Visscher et al., 2008). The study results showed that when environmental factors of the resource population were consistent, most of the variation that is observed in the present population is caused by variation in genotypes. Besides, PL and IMF of this resource population were not strongly selected, and their genetic variation was large. The pedigree information (Costa et al., 2015) and chip data (Won et al., 2018; Wurtz et al., 2018) were widely applied for quantitative trait locus (QTL) mapping. So far, no report on more than 10 million SNP markers used for estimating the heritability of quantitative traits in livestock and poultry is available. This study provides a reference for heritability estimation based on genome-wide SNP data.

In our study, heritability estimation model was consistent with microbiability estimation model. Both sex and BW were used as covariates to explore the influence of cecal microbiome on traits. To correct the effect of host genome on traits, the first five PCs generated by the GRM and the first two PCs generated at SNPs which were significantly associated with PL and IMF are considered as covariates. In the previous study, the first two PCs and the first five PCs of population structure were used as covariates to estimate microbiability (Wen et al., 2019b). In our study, the host heredity had major effect on traits. Several studies have reported that the microbiability of traits was lower than the heritability of traits in pigs (Camarinha-Silva et al., 2017; Tang et al., 2020; Khanal et al., 2021), chicken (Wen et al., 2019b), and cattle (Difford et al., 2018). In previous studies, the microbiability of the IMF was estimated as 0.13 in Enshi pigs (Tang et al., 2020), estimated as 0.03–0.06 following different models in commercial F1 sows composed of Yorkshire × Landrace or Landrace × Yorkshire (Khanal et al., 2021). The microbiability of IMF in the resource population was estimated to be 0.26, which was higher than that in the previous study. The reason for the difference in these studies’ results is that estimation of microbiability is affected by breed, environment, sample size, sequencing method, and estimation algorithm (Wen et al., 2019a). So far, few studies have described the microbiability of fat deposition traits in pigs, and estimation of microbiability of fat deposition traits in pigs needs to be further studied. Our estimated microbiability suggested that the gut microbiome had an influence on PL and IMF and provided a reference for microbiability estimation.

Pig fat deposition traits are affected by a large number of complex microbes and their metabolites which are closely related to host immune diseases, nutritional metabolism, and body behavior (Patil et al., 2020). To select microbes affecting traits, divergent Wilcoxon rank-sum test and MWAS were applied in our study. The divergent Wilcoxon rank-sum test has been used to analyze microbes in previous studies (Yang et al., 2016; Gardiner et al., 2020). MWAS has been proposed with reference to genome-wide association analysis to establish the relationship between traits and gut microbes, and it has been applied in basic research (Gilbert et al., 2016; Vollmar et al., 2020). In our study, the divergent Wilcoxon rank-sum test showed that Campylobacter and Paraprevotella were significantly associated with PL, and the MWAS results showed that Anaeroplasma, Paraprevotella, Pasteurella, and Streptococcus were significantly associated with PL. Among these genera, Streptococcus is reported as one of the most abundant bacteria in the gut microbes of Jinhua Pigs (Yang et al., 2018). Jinhua pig is an obese-type breed characterized by higher levels of intramuscular fat (Miao et al., 2009). MWAS results indicated that Helicobacter and Sharpea were significantly associated with IMF. Sharpea azabuensis has been reported to affect methane emissions in rumen, and it was viewed as one of the important driving factors of lactic acid production and utilization (Kamke et al., 2016). In previous studies, lactic acid can reduce fat synthesis and accumulation (Yonejima et al., 2013; Gan et al., 2020).

Divergent Wilcoxon rank-sum test and MWAS results indicated that Paraprevotella was significantly associated with PL, suggesting the influence of single genus on host phenotype. Paraprevotella belongs to Bacteroidetes phylum which could degrade cellulose and produce butyric acid (Gao et al., 2018; Hasan et al., 2018). Butyric acid is a short-chain fatty acid (SCFA), and SCFAs are closely related to glucose and lipid metabolism and energy metabolism (Dabke et al., 2019). It has been demonstrated that SCFAs enhance adipocyte differentiation in porcine adipose tissue (Li et al., 2014). The content of butyric acid has been reported to be positively correlated with the relative abundance of Paraprevotella in mouse feces (Fei et al., 2019). Butyric acid is a major energy source for intestinal epithelial cells and plays key functional roles in maintaining intestinal homeostasis and in overall health status. It can promote the development of the intestine and has antioxidant and anti-inflammatory effects (Fu et al., 2019). Our data revealed a significant negative correlation between the relative abundance of Paraprevotella and PL. Paraprevotella may affect the intestinal homeostasis by influencing the anabolism of butyric acid, thereby affecting the visceral fat deposition in pigs, but the mechanism of Paraprevotella on affecting fat deposition in pig needs to be further studied.

Conclusion

This study revealed that PL and IMF were influenced by the host genetics–gut microbiome interaction. Through divergent Wilcoxon rank-sum test and microbiome-wide association analysis, we screened out Paraprevotella which were significantly associated with PL. Overall, this study reveals the effect of host genetics and gut microbiome on pig fat deposition traits and provides a reference for the genetic improvement of pig fat deposition traits, which can affect pig fat deposition by altering the gut microbiome.

Data availability statement

The genotyping data in PLINK binary format is available from the Figshare database (https://figshare.com/s/c0ece21cf32f35833569). The 16S ribosomal RNA sequencing data are available in the Sequence Read Archive (SRA) repository (SRA accession number PRJNA868520). Further information on the original data can be directed to the corresponding author.

Ethics statement

This animal study was reviewed and approved by the Animal Care and Use Committee of Huazhong Agricultural University.

Author contributions

YW collected the samples, performed the data analysis and interpretation, and wrote the manuscript. PZ performed the data analysis and wrote the manuscript. MF, ZL, and TW measured pig phenotype and collected samples. ZW and XL participated in constructing mathematical models. BL and XZ conceived and designed the study and manuscript revision. All authors contributed to manuscript revision, read, and approved the submitted version.

Funding

This study was supported by the National Natural Science Foundation of China (31930104) and the Fundamental Research Funds for the Central Universities (2662022DKPY008).

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmicb.2022.925200/full#supplementary-material

Supplementary Figure 1 | Distribution of fat deposition traits. (A–H) Distribution of backfat thickness at shoulder (SBFT), backfat thickness at loin (LBFT), backfat thickness at rump (RBFT), average backfat thickness (ABFT), percentage of leaf fat (PL), percentage of caul fat (PC), percentage of leaf fat, and caul (PLC), and intramuscular fat content (IMF).

Footnotes

1. ^ https://www.hiblup.com/
2. ^ https://cran.r-project.org/web/packages/gaston

References

Bergamaschi, M., Maltecca, C., Schillebeeckx, C., McNulty, N. P., Schwab, C., Shull, C., et al. (2020). Heritability and genome-wide association of swine gut microbiome features with growth and fatness parameters. Sci. Rep. 10:10134. doi: 10.1038/s41598-020-66791-3

PubMed Abstract | CrossRef Full Text | Google Scholar

Camarinha-Silva, A., Maushammer, M., Wellmann, R., Vital, M., Preuss, S., and Bennewitz, J. (2017). Host genome influence on gut microbial composition and microbial prediction of complex traits in pigs. Genetics 206, 1637–1644. doi: 10.1534/genetics.117.200782

PubMed Abstract | CrossRef Full Text | Google Scholar

Costa, E. V., Diniz, D. B., Veroneze, R., Resende, M. D. V., Azevedo, C. F., Guimaraes, S. E. F., et al. (2015). Estimating additive and dominance variances for complex traits in pigs combining genomic and pedigree information. Genet. Mol. Res. 14, 6303–6311. doi: 10.4238/2015.June.11.4

PubMed Abstract | CrossRef Full Text | Google Scholar

Dabke, K., Hendrick, G., and Devkota, S. (2019). The gut microbiome and metabolic syndrome. J. Clin. Invest. 129, 4050–4057. doi: 10.1172/JCI129194

PubMed Abstract | CrossRef Full Text | Google Scholar

Difford, G. F., Plichta, D. R., Lovendahl, P., Lassen, J., Noel, S. J., Hojberg, O., et al. (2018). Host genetics and the rumen microbiome jointly associate with methane emissions in dairy cows. PLoS Genet. 14:e1007580. doi: 10.1371/journal.pgen.1007580

PubMed Abstract | CrossRef Full Text | Google Scholar

Fei, Y., Wang, Y., Pang, Y., Wang, W., Zhu, D., Xie, M., et al. (2019). Xylooligosaccharide modulates gut microbiota and alleviates colonic inflammation caused by high fat diet induced obesity. Front. Physiol. 10:1601. doi: 10.3389/fphys.2019.01601

PubMed Abstract | CrossRef Full Text | Google Scholar

Fortin, A., Robertson, W. M., and Tong, A. K. W. (2005). The eating quality of Canadian pork and its relationship with intramuscular fat. Meat Sci. 69, 297–305. doi: 10.1016/j.meatsci.2004.07.011

PubMed Abstract | CrossRef Full Text | Google Scholar

Fu, X., Liu, Z., Zhu, C., Mou, H., and Kong, Q. (2019). Nondigestible carbohydrates, butyrate, and butyrate-producing bacteria. Crit. Rev. Food Sci. Nutr. 59(sup1), S130–S152. doi: 10.1080/10408398.2018.1542587

PubMed Abstract | CrossRef Full Text | Google Scholar

Gan, Y., Tang, M. W., Tan, F., Zhou, X. R., Fan, L., Xie, Y. X., et al. (2020). Anti-obesity effect of Lactobacillus plantarum CQPC01 by modulating lipid metabolism in high-fat diet-induced C57BL/6 mice. J. Food Biochem. 44:e13491. doi: 10.1111/jfbc.13491

PubMed Abstract | CrossRef Full Text | Google Scholar

Gao, B. B., Wang, R. J., Peng, Y., and Li, X. B. (2018). Effects of a homogeneous polysaccharide from Sijunzi decoction on human intestinal microbes and short chain fatty acids in vitro. J. Ethnopharmacol. 224, 465–473. doi: 10.1016/j.jep.2018.06.006

PubMed Abstract | CrossRef Full Text | Google Scholar

Gao, S. Z., and Zhao, S. M. (2009). Physiology, affecting factors and strategies for control of pig meat intramuscular fat. Recent Pat Food Nutr. Agric. 1, 59–74. doi: 10.2174/2212798410901010059

CrossRef Full Text | Google Scholar

Gardiner, G. E., Metzler-Zebeli, B. U., and Lawlor, P. G. (2020). Impact of intestinal microbiota on growth and feed efficiency in pigs: A review. Microorganisms 8:1886. doi: 10.3390/microorganisms8121886

PubMed Abstract | CrossRef Full Text | Google Scholar

Gilbert, J. A., Quinn, R. A., Debelius, J., Xu, Z. Z., Morton, J., Garg, N., et al. (2016). Microbiome-wide association studies link dynamic microbial consortia to disease. Nature 535, 94–103. doi: 10.1038/nature18850

PubMed Abstract | CrossRef Full Text | Google Scholar

Hasan, S., Junnikkala, S., Peltoniemi, O., Paulin, L., Lyyski, A., Vuorenmaa, J., et al. (2018). Dietary supplementation with yeast hydrolysate in pregnancy influences colostrum yield and gut microbiota of sows and piglets after birth. PLoS One 13:e0197586. doi: 10.1371/journal.pone.0197586

PubMed Abstract | CrossRef Full Text | Google Scholar

Hishikawa, D., Hong, Y. H., Roh, S., Miyahara, H., Nishimura, Y., Tomimatsu, A., et al. (2005). Identification of genes expressed differentially in subcutaneous and visceral fat of cattle, pig, and mouse. Physiol. Genomics 21, 343–350. doi: 10.1152/physiolgenomics.00184.2004

PubMed Abstract | CrossRef Full Text | Google Scholar

Kamke, J., Kittelmann, S., Soni, P., Li, Y., Tavendale, M., Ganesh, S., et al. (2016). Rumen metagenome and metatranscriptome analyses of low methane yield sheep reveals a Sharpea-enriched microbiome characterised by lactic acid formation and utilisation. Microbiome 4:56. doi: 10.1186/s40168-016-0201-2

PubMed Abstract | CrossRef Full Text | Google Scholar

Khanal, P., Maltecca, C., Schwab, C., Fix, J., and Tiezzi, F. (2021). Microbiability of meat quality and carcass composition traits in swine. J. Anim. Breed. Genet. 138, 223–236. doi: 10.1111/jbg.12504

PubMed Abstract | CrossRef Full Text | Google Scholar

Li, F., Li, C., Chen, Y., Liu, J., Zhang, C., Irving, B., et al. (2019). Host genetics influence the rumen microbiota and heritable rumen microbial features associate with feed efficiency in cattle. Microbiome 7:92. doi: 10.1186/s40168-019-0699-1

PubMed Abstract | CrossRef Full Text | Google Scholar

Li, G., Yao, W., and Jiang, H. (2014). Short-chain fatty acids enhance adipocyte differentiation in the stromal vascular fraction of porcine adipose tissue. J. Nutr. 144, 1887–1895. doi: 10.3945/jn.114.198531

PubMed Abstract | CrossRef Full Text | Google Scholar

Lopez, B. I. M., Song, C., and Seo, K. (2018). Genetic parameters and trends for production traits and their relationship with litter traits in Landrace and Yorkshire pigs. Anim. Sci. J. 89, 1381–1388. doi: 10.1111/asj.13090

PubMed Abstract | CrossRef Full Text | Google Scholar

Merlotti, C., Ceriani, V., Morabito, A., and Pontiroli, A. E. (2017). Subcutaneous fat loss is greater than visceral fat loss with diet and exercise, weight-loss promoting drugs and bariatric surgery: A critical review and meta-analysis. Int. J. Obes. 41, 672–682. doi: 10.1038/ijo.2017.31

PubMed Abstract | CrossRef Full Text | Google Scholar

Miao, Z. G., Wang, L. J., Xu, Z. R., Huang, J. F., and Wang, Y. R. (2009). Developmental changes of carcass composition, meat quality and organs in the Jinhua pig and Landrace. Animal 3, 468–473. doi: 10.1017/S1751731108003613

PubMed Abstract | CrossRef Full Text | Google Scholar

Patil, Y., Gooneratne, R., and Ju, X. H. (2020). Interactions between host and gut microbiota in domestic pigs: A review. Gut Microbes 11, 310–334. doi: 10.1080/19490976.2019.1690363

PubMed Abstract | CrossRef Full Text | Google Scholar

Quan, J., Cai, G., Ye, J., Yang, M., Ding, R., Wang, X., et al. (2018). A global comparison of the microbiome compositions of three gut locations in commercial pigs with extreme feed conversion ratios. Sci. Rep. 8:4536. doi: 10.1038/s41598-018-22692-0

PubMed Abstract | CrossRef Full Text | Google Scholar

Savage, D. C. (1977). Microbial ecology of the gastrointestinal tract. Annu. Rev. Microbiol. 31, 107–133. doi: 10.1146/annurev.mi.31.100177.000543

PubMed Abstract | CrossRef Full Text | Google Scholar

Tang, S., Xin, Y., Ma, Y., Xu, X., Zhao, S., and Cao, J. (2020). Screening of microbes associated with swine growth and fat deposition traits across the intestinal tract. Front. Microbiol. 11:586776. doi: 10.3389/fmicb.2020.586776

PubMed Abstract | CrossRef Full Text | Google Scholar

Uemoto, Y., Sato, S., Kikuchi, T., Egawa, S., Kohira, K., Sakuma, H., et al. (2017). Genomic evaluation using SNP- and haplotype-based genomic relationship matrices in a closed line of Duroc pigs. Anim. Sci. J. 88, 1465–1474. doi: 10.1111/asj.12805

PubMed Abstract | CrossRef Full Text | Google Scholar

Visscher, P. M., Hill, W. G., and Wray, N. R. (2008). Heritability in the genomics era–concepts and misconceptions. Nat. Rev. Genet. 9, 255–266. doi: 10.1038/nrg2322

PubMed Abstract | CrossRef Full Text | Google Scholar

Vollmar, S., Wellmann, R., Borda-Molina, D., Rodehutscord, M., Camarinha-Silva, A., and Bennewitz, J. (2020). The gut microbial architecture of efficiency traits in the domestic poultry model species Japanese Quail (Coturnix japonica) assessed by mixed linear models. G3 Genes Genomes Genet. 10, 2553–2562. doi: 10.1534/g3.120.401424

PubMed Abstract | CrossRef Full Text | Google Scholar

Wen, C. L., Yan, W., Sun, C. J., Ji, C. L., Zhou, Q. Q., Zhang, D. X., et al. (2019b). The gut microbiota is largely independent of host genetics in regulating fat deposition in chickens. ISME J. 13, 1422–1436. doi: 10.1038/s41396-019-0367-2

PubMed Abstract | CrossRef Full Text | Google Scholar

Wen, C. L., Sun, C. J., and Yang, N. (2019a). [The concepts and research progress: From heritability to microbiability]. Yi Chuan 41, 1023–1040. doi: 10.16288/j.yczz.19-130

PubMed Abstract | CrossRef Full Text | Google Scholar

Wen, C., Yan, W., Mai, C., Duan, Z., Zheng, J., Sun, C., et al. (2021). Joint contributions of the gut microbiota and host genetics to feed efficiency in chickens. Microbiome 9:126. doi: 10.1186/s40168-021-01040-x

PubMed Abstract | CrossRef Full Text | Google Scholar

Won, S., Jung, J., Park, E., and Kim, H. (2018). Identification of genes related to intramuscular fat content of pigs using genome-wide association study. Asian Australas. J. Anim. Sci. 31, 157–162. doi: 10.5713/ajas.17.0218

PubMed Abstract | CrossRef Full Text | Google Scholar

Wurtz, K. E., Siegford, J. M., Ernst, C. W., Raney, N. E., Bates, R. O., and Steibel, J. P. (2018). Genome-wide association analyses of lesion counts in group-housed pigs. Anim. Genet. 49, 628–631. doi: 10.1111/age.12713

PubMed Abstract | CrossRef Full Text | Google Scholar

Xiao, L., Estelle, J., Kiilerich, P., Ramayo-Caldas, Y., Xia, Z., Feng, Q., et al. (2016). A reference gene catalogue of the pig gut microbiome. Nat. Microbiol 1:16161. doi: 10.1038/nmicrobiol.2016.161

PubMed Abstract | CrossRef Full Text | Google Scholar

Yang, H., Huang, X., Fang, S., Xin, W., Huang, L., and Chen, C. (2016). Uncovering the composition of microbial community structure and metagenomics among three gut locations in pigs with distinct fatness. Sci. Rep. 6:27427. doi: 10.1038/srep27427

PubMed Abstract | CrossRef Full Text | Google Scholar

Yang, H., Xiao, Y. P., Wang, J. J., Xiang, Y., Gong, Y. J., Wen, X. T., et al. (2018). Core gut microbiota in Jinhua pigs and its correlation with strain, farm and weaning age. J. Microbiol. 56, 346–355. doi: 10.1007/s12275-018-7486-8

PubMed Abstract | CrossRef Full Text | Google Scholar

Yang, J. A., Benyamin, B., McEvoy, B. P., Gordon, S., Henders, A. K., Nyholt, D. R., et al. (2010). Common SNPs explain a large proportion of the heritability for human height. Nat. Genet. 42, 565–569. doi: 10.1038/ng.608

PubMed Abstract | CrossRef Full Text | Google Scholar

Yang, J. A., Lee, S. H., Goddard, M. E., and Visscher, P. M. (2011). GCTA: A tool for genome-wide complex trait analysis. Am. J. Hum. Genet. 88, 76–82. doi: 10.1016/j.ajhg.2010.11.011

PubMed Abstract | CrossRef Full Text | Google Scholar

Yang, J., Zeng, J., Goddard, M. E., Wray, N. R., and Visscher, P. M. (2017). Concepts, estimation and interpretation of SNP-based heritability. Nat. Genet. 49, 1304–1310. doi: 10.1038/ng.3941

PubMed Abstract | CrossRef Full Text | Google Scholar

Yonejima, Y., Ushida, K., and Mori, Y. (2013). Effect of lactic acid bacteria on lipid metabolism and fat synthesis in mice fed a high-fat diet. Biosci. Microbiota Food Health 32, 51–58. doi: 10.12938/bmfh.32.51

PubMed Abstract | CrossRef Full Text | Google Scholar

Zhu, B. L., Wang, X., and Li, L. J. (2010). Human gut microbiome: The second genome of human body. Protein Cell 1, 718–725. doi: 10.1007/s13238-010-0093-z

PubMed Abstract | CrossRef Full Text | Google Scholar