Employing pigs to decipher the host genetic effect on gut microbiome: advantages, challenges, and perspectives

ABSTRACT

The gut microbiota is a complex and diverse ecosystem comprised of trillions of microbes and plays an essential role in host’s immunity, metabolism, and even behaviors. Environmental and host factors drive the huge variations in the gut microbiome among individuals. Here, we summarize accumulated evidences about host genetic effect on the gut microbial compositions with emphases on the correlation between host genetic kinship and the similarity of microbial compositions, heritability estimates of microbial taxa, and identification of genomic variants associated with the gut microbiome in pigs as well as in humans. A proportion of bacterial taxa have been reported to be heritable, and numerous variants associated with the diversity of the gut microbiota or specific taxa have been identified in both humans and pigs. LCT and ABO gene have been replicated in multiple studies, and its mechanism have been elucidated clearly. We also discuss the main advantages and challenges using pigs as experimental animals in exploring host genetic effect on the gut microbial composition and provided our insights on the perspectives in this area.

KEYWORDS:

• Gut microbiome
• host genetics
• heritability estimates
• genome-wide association study
• pigs

Introduction

Pigs are not only important farm animals that provide a major source of meat for human consumption, but also serve as a biomedical model for the studies of human diseases. Just as that of other mammals, the gastrointestinal tract of pigs inhabits ~10¹⁴ of bacterial cells that have been estimated to be 2 to 10 times greater than the number of host cells^1,2. The collective genes of the bacterial population that resides in the pig gut have been estimated to be about 17.2 million in our previous study³, which is more than 680 times than the number of genes in the pig genome (~25,000 genes). Gut microbiota forms a complex and diverse community of ecosystem and plays an essential role in host physiological processes, such as immunity and diseases⁴, metabolism⁵, and even behaviors⁶. Various environmental factors including diets, housing, stress, and the utilization of medicines, and host factors, such as host genetics, sex, and age can influence the diversity, composition, and functional capacities of the gut microbiota^7,8.

In recent years, more and more studies have been involved in the discussion whether and how host genetics affects the compositions of the gut microbiome. Several host genes have been reported to be associated with the compositions or diversity of gut microbial community in humans and pigs^9–11. However, the studies have also suggested that environmental factors dominate over host genetics in shaping human gut mcirobiome^8,12. It is difficult to determine how much is the contribution of host genetics in shaping the gut microbiota in humans because environmental factors, such as diets, medicine, and living conditions, that significantly affect the gut microbial compositions are difficult to control for the study cohorts.

Pigs are always raised in uniformed commercial farm houses with high density. Diets, the utilization of medicines, e.g., antibiotics, and farm conditions are artificially provided and easily controlled. More importantly, microbial samples can be obtained from different gut locations that can’t be achieved in humans. These provide an excellent opportunity to study the effect of host genetics on the gut microbiota by using pigs as experimental animals. In this review, we first describe the variations of microbial compositions of pigs in different gut locations and at different ages and then summarize environmental and host factors influencing pig gut microbiome. Especially, we discuss the recent advances in deciphering the effects of host genetics on the gut microbial compositions and envisage the advantages, challenges, and perspectives on using pigs as experimental animals for investigating host genetic effect on the gut microbial composition.

Gut microbiota compositions of pigs compared to that of humans

Gut microbial compositions markedly vary between individuals (inter-individual heterogeneity) and among gut locations and ages (gut location- and age-heterogeneity). Fully understanding the compositions of pig gut microbiome in different intestinal locations and at different ages is important to both pig health managements and production performance improvements by adjusting the gut microbiota. A recent study presented a collection of cultured bacteria from pig gastrointestinal tract and identified distinct taxonomic groups that we did not know before and showed the diversity of metabolic functions¹³. This suggested that a more comprehensive and deep understanding of pig gut microbiota are further needed. The dynamic changes of a healthy microbial ecosystem with ages and the spatial distribution of gut microbiota in pig gastrointestinal tract have yet to be qualitatively or quantitatively defined.

A meta-analysis using 20 publically available data sets from 16S rRNA gene sequencing studies of pig gut microbiota indicated that gut locations that were sampled had the largest effect on the composition and structure of the pig gut microbiota¹⁴. Zhang et al. (2018)¹⁵ investigated the spatial heterogeneity of mucosal and luminal microbiome across intestinal tract and revealed the distinct differences in microbial compositions among ileum, cecum, and colon and between luminal content and mucosa. From the small intestine to the large intestine, the abundances of both luminal and mucosal Bacteroidota were increased, while luminal Firmicutes and Proteobacteria and mucosal-attached Proteobacteria were remarkably decreased¹⁵. The study on microbial compositions in five gut locations of 13 castrated male Iberian pigs found that Lactobacillus and Clostridium were the two most abundant genera in the small intestine, while Prevotella was predominant in the colon¹⁶. In our previous studies, we determined the spatial heterogeneity of gut microbial compositions among five gut locations (ileum and cecum luminal content, ileal and cecal mucosa, and feces) at the age of 240 days in F₆ and F₇ pigs from a mosaic population by 16S rRNA gene sequencing¹¹. We observed the obvious variations in microbial compositions at different gut locations. For examples, at the family level, Leuconostocaceae were nearly specific to luminal content, while Pseudomonadaceae were specific to mucosa¹¹. Enterobacteriaceae and Clostridiaceae were at least ten times more abundant in the ileum than in the other locations. Christensenellaceae had nearly ten times more abundance in feces than in luminal content and mucosa samples. With the metagenomic sequencing data of six ileum contents, 20 cecum contents and 472 feces samples from healthy pigs, a significant difference was found in the relative abundances of the same bacterial species in different gut locations. There were seven species specifically having high relative abundance (Ranking in the top 20) in ileum contents, while three and six bacterial species particularly showed high abundances in cecum contents and feces samples, respectively³. Yang et al. (2016)¹⁷ also compared the microbial compositions at different gut locations and identified eight bacterial species that were enriched in the cecum content, five species had the higher abundance in the jejunum and ileum contents, and eight species were successively decreased their abundances from anterior to posterior of intestine¹⁷. These reports indicated the variations of the gut microbial compositions in different gut locations (Figure 1a) and suggested that limiting the analysis of the gut microbiota to feces are far from enough to understand its complexity and elucidate the factors that determine it⁷.

Figure 1. Microbial compositions in different gut locations and the factors influencing gut microbial compositions in pigs. (a) the main bacterial species residing in the ileum, cecum, and feces. The data from Chen et al. (2021). (b) host and environmental factors influencing pig gut microbiota.

Distinct gut microbial compositions at different ages/growth stages have been reported in many studies^3,11,18. A meta-analysis with the 16S rRNA gene sequencing data from 16 studies observed dynamic changes in gut microbial compositions following the ages. The top three genera in relative abundances were Bacteroides, Escherichia, and Lactobacillus in lactating pigs; Prevotella, Lactobacillus, and Faecalibacterium in growing pigs; and Prevotella, Lactobacillus, and Streptococcus in finishing pigs¹⁸. We previously also determined the gut microbial compositions at the ages of 25, 120, and 240 days in F₆ and F₇ pigs of a mosaic population by 16S rRNA gene sequencing^11,19. The richness of gut microbiota was significantly increased from 25 to 240 days of age, and the topological features of phylogenetic co-occurrence networks were increased along with the ages¹⁹. Odoribacteraceae and Rikenellaceae were at least ten times more abundant in feces at the age of 25 days than at any other ages¹¹. Wang et al. (2019) investigated the pig gut microbial composition and structure at 16 time points from birth to market with 18 pigs during lactation, nursery, growing, and finishing stages, identified 19 phyla that existed in pig gut microbiome of the lifetime with Firmicutes and Bacteroidota being the most abundant, and observed an overall increased alpha-diversity of pig gut microbial community and distinct shifts in microbiome structure along different growth stages²⁰.

Because of the high similarity of omnivorous and gastrointestinal structure between human and pig, pigs have been treated as an important animal model for studying gut microbiome. A previous report identified that 78% of the KO functional pathways identified in the pig gut metagenome are present in the human gut metagenome, while 96% of the KOs found in the human gut metagenome are present in the pig gut metagenome²¹. We previously compared the gut microbial compositions between humans and pigs with the gene catalogs of human and pig gut microbiome^3,22 and found that the percentage of common bacterial genera shared between human and pig (87.24% for human gut microbiota) was higher than that shared between human and mouse (70.00% for human gut microbiota). The similar result was also observed in F₆ and F₇ pigs of the mosaic population by 16S rRNA gene sequencing, in which microbiota composition of pig feces was more similar to that of human than of mice feces¹¹. Notably, the common bacterial genera should show different abundances in human and pig gut microbiome. For example, at the genus level, Bacteroides and Eubacterium were most abundant in human feces. However, the most abundant genera in pig feces were Lactobacillus and Clostridium³.

Effects of non-genetic factors on pig gut microbial compositions

As a complex ecosystem, immense diversity of the gut microbial composition in intra- and inter-individuals have been reported. At some extreme states, natural variations in the gut microbiota can deteriorate to dysbiosis when stress conditions rapidly decrease the diversity of the gut microbiota and promote the expansion of specific bacterial taxa^23,24. Besides host ages and gut locations, the non-genetic factors influencing gut microbial compositions mainly include the diets, the use of antibiotics and probiotics, health state, housing environments, and host gender.

Diets

Diets have profound effects on the composition of the gut microbiota and are one of the most important factors that change the gut microbiota. Most of the studies about diet effect on the gut microbiota have been performed in humans and mice, especially about high-fiber or high-fat dietary patterns on the gut microbiota^25–27. In pigs, the study using a reciprocal cross-fostering model revealed that the introduction of solid feed and weaning are more important determinants of gut bacterial succession in piglets than breed and nursing mother²⁸. More studies have focused on the effects of specific dietary nutrients on the composition of gut microbiota, such as protein²⁹, fibers^30,31, vitamins^32,33, sodium butyrate³³, and some specific nutrient factors, e.g., soy isoflavone³² and methylsulfonylmethane³⁴. For example, low crude protein diet increased interaction network complexity, although it did not alter the microbial diversity³⁵. Generally, production performances had always been phenotyped when investigating the effects of dietary nutrients on the composition of gut microbiota because these studies were mainly designed to investigate the improvement of pig production performances by the additive of specific nutrient factors through regulating gut microbiome.

Utilization of antibiotics and probiotics

Over the past few decades, antibiotics had been administered to pigs broadly in the pig industry for disease treatment and growth promotion. A study investigated the effect of performance-enhancing antibiotics on the gut microbiome and found that the use of antibiotics increased the abundance of Escherichia coli in the ileum and Lachnobacterium spp. in all gut locations of the medicated pigs^36,37. Metagenomic analysis indicated the increased abundances of microbial functional genes relating to energy production and conversion and the increased diversity and abundances of antibiotic resistance genes in the antibiotic-fed pigs³⁶. Another feeding experiment of antibiotics cocktail also found significant alteration of the microbial compositions in the ileum and feces with a reduced abundances of Bifidobacterium, Lactobacillus, and Ruminococcus, and an increased abundance of Escherichia coli ³⁸.

In recent years, considering the residual effect and antibiotic resistance in bacteria in the use of antibiotics, as an alternative, probiotics has been getting increased attention in pig production³⁹. Probiotics show a strong ability to modulate the gut microbiota composition in healthy individuals by significantly reducing disease-related bacterial taxa⁴⁰. Among probiotics used in pig production, Lactobacillus has been the most commonly used probiotics, improving growth performance and feed conversion efficiency³⁹ and regulating immune system in pigs, such as beneficial effects of Lactobacillus plantarum on the modulation of gut microbiota and immune response in weaned pigs⁴¹.

Health state

Host health state such as environmental stressors, diseases, and infections also affects the gut microbial composition. For example, the diversity, composition, and functional capacity of fecal microbial community were different in humans or animal models with Crohn’s disease^42,43, ulcerative colitis⁴⁴, and acute diarrhea⁴⁵. The similar conditions were also observed in pigs. Gut microbiota dysbiosis was associated with post-weaning diarrhea in pigs⁴⁶. Porcine deltacoronavirus (PDCoV) infection reduced the diversity and richness of gut bacteria and significantly changed the structure and abundance of the gut microbiota from the phylum to genus⁴⁷. Heat stressed pigs showed the decrease in Lactobacillus johnsonii and Lactobacillus reuteri and the increase in Clostridium sensu stricto 1 in the colon microbiota⁴⁸.

Housing environments

Environmental factors, such as farm conditions and hygiene, also have significant effects on the gut microbial compositions. In our previous study, we compared the gut microbiome between free-living wild boars and Duroc pigs from two farms. Distinct gut microbial compositions were observed between wild boars and two Duroc pig populations³. Although genetic differences existed between wild boars and Duroc pigs, distinct gut microbial compositions were also found between two Duroc pig populations fed with the similar formula diet from different farms. This suggested that many other factors including environmental exposures, hygiene, temperature, and certain other unknown factors potentially affected the gut microbiota. The study by Wen et al. (2021) also indicated that environmentally enriched housing conditions affected pig gut microbiota in early life, and the relative abundance of several bacterial species involved in the production of short chain fatty acids was increased⁴⁹.

Host gender

An increasing number of studies have indicated profound effect of host gender on the gut microbiome. For examples, compared to normal male mice, sexually mature castrated male mice showed a similar gut microbial composition to female mice^50,51. Our previous study indicated an obvious influence of host gender on the gut microbial composition. Veillonellaceae, Roseburia, Bulleidia, and Escherichia had higher abundance in boars, while Treponema and Bacteroides were over-represented in gilts⁵². Castration significantly changed the fecal microbiota composition of the boars toward that of gilts⁵². A bi-directional interaction exists between gut microbiota and sex hormones. Host sex hormones, e.g., estrogens, can influence gut bacterial gene expression and composition⁵³, whereas the gut microbiota can reactivate estrogens from their inactive glucuronides⁵⁴.

Host genetic effect on the gut microbiome

Besides the non-genetic factors mentioned above, to what extent host genetics contributes to the inter-individual variations in the gut microbial compositions attracts great interest to researchers working on gut microbiome. In recent years, more and more studies have focused on this subject in both humans and animals, especially in humans. However, compared to environmental factors, the extent of host genetics shaping the composition of gut microbiome has not been consensus. The studies have demonstrated that environmental factors dominate over host genetics in shaping human gut microbiota in both cohorts of 1,046 healthy individuals with several distinct ancestral origins⁸ and 8,208 Dutch individuals¹². The overall microbiome heritability was only 1.9% ~ 8.1%⁸ or only about 6.6% of taxa are heritable¹². On the other hand, since 2014, more than dozen of the studies have investigated the associations of host genetic variants on the gut microbiome in humans (Table 1), Drosophila^55,56, and mice⁵⁷ and reported numerous genomic loci/variants associated with the diversity of gut microbiota and the abundance of individual taxa. More recently, the study using 36 healthy black and white females with similar age, weight, habitual diets, and health status indicated that individuality and ethnicity account for roughly 70% to 88% and 2% to 10% of taxonomic variation, respectively, which dominated the effects of a short-term diet intervention in shaping gut and oral microbiomes and gut viromes⁵⁸. However, the urine and plasma metabolites showed the tendency to more strongly associate with diets⁵⁸. Here, we summarize the existing evidences about host genetic effects on the gut microbiome in pigs with emphases on the correlations between host genetic kinship and the dissimilarity of gut microbial compositions, heritability estimates of specific taxa, and identification of genomic variants, or microbial quantitative locus (miQTL) associated with the gut microbial compositions by genome-wide association studies (GWAS). Because the higher similarity of gut microbial compositions between pigs and humans than between humans and mice^11,21, this has made pigs as an important model to study the host genetic effect on the gut microbiome. To facilitate the comparison and discuss the progress in this subject between humans and pigs, the advances in human genetic effect on gut microbiome investigated by GWAS are also discussed in detailed as a benchmark because more promising advances have been made in the studies of humans.

Table 1. Summary of identification of host genetic variants associated with gut microbiome by genome-wide association study.

Host species	Cohort	Sample type	Sample size	Microbiome data type	Trait	Significance threshold	Significant variants	Key Genes	Reference
Human	Dutch Microbiome Project	Stool	7,738	Shotgun metagenomic sequencing	207 microbial taxa and 205 functional pathways	Genome-wide: 5 × 10^−8; Study-wide: 1.89 × 10^−10	37 SNPs at 24 independent loci	LCT, ABO	Lopera-Maya et al., 2022^Citation9
	FR02 study	Stool	5,959	Shotgun metagenomic sequencing	2,801 taxa	Genome-wide: 5 × 10^−8; Study-wide: 3.8 × 10^−11	567 independent SNP-taxon associations	LCT, ABO, MED13L	Qin et al., 2022^Citation59
	Healthy Life in an Urban Setting cohort	Stool	4,117	16S rRNA gene sequencing	401 taxa, 169 pathways	Genome-wide: 5 × 10^−8; Study-wide: 1.9 (6.8 for pathways) ×10^−10	51 SNPs associated with gut microbial taxa	Discussed candidate genes around significant signals	Boulund et al., 2022^Citation60
	MiBioGen consortium (24 cohorts)	Stool	18,340	16S rRNA gene sequencing	α-diversity indices, 211 taxa for mbQTL and 177 taxa for mbBTL analysis	Genome-wide: 5 × 10^−8	31 loci (20 mbQTL and 11 mbBTLs)	LCT	Kurilshikov et al., 2021^Citation61
	Flemish Gut Flora Project and two German cohorts	Stool	Flemish: 2,223; FoCus: 950; PopGen: 717	16S rRNA gene sequencing	α-diversity indices, 139 taxa	Genome-wide: 2.5 × 10^−8; Study-wide: 1.57 × 10^−10	2 variants achieved study-wide significance, 11 associations at genome-wide significance.	RAPGEF1, LINC01787, MCM6/LCT	Hughes et al., 2020^Citation62
	healthy Japanese adults	Stool	1,068	16S rRNA gene sequencing	α-diversity indices, 21 core genera	Genome-wide: 8.9 × 10^−8, suggestive: 1 × 10^−5	303 loci at the suggestive level, 5 loci at the genome-wide significance	Discussed candidate genes around significant signals	Ishida et al., 2020^Citation63
	healthy Israeli adults, LifeLines-DEEP cohort	Stool	1,046, 1,539	16S rRNA gene sequencing, Shotgun metagenomic sequencing	β-diversity, 336 taxa	P < 5 × 10^−8	limited evidence for the association of any individual SNP with microbiome β -diversity, 43 loci with special taxa, but none remained statistically significant, 7 loci replicated	LCT	Rothschild et al., 2018^Citation8
	LifeLines-DEEP cohort, 500 Functional Genomics cohort, Maastricht IBS cohort	Stool	1,514	Shotgun metagenomic sequencing	219 taxa, 636 MetaCyc pathways and 661 Gene Ontology terms	Genome-wide: 5 × 10^−8	9 loci with microbial taxa and 33 loci with microbial pathways and gene ontology terms	LCT	Bonder et al., 2016^Citation64
	PopGen and FoCus cohorts	Stool	1,812	16S rRNA gene sequencing	β-diversity, 40 OTUs and 58 taxa	Meta-analysis: 5 × 10^−8; single cohort: 5 × 10^−4	42 loci for β-diversity, and 40 loci for 22 taxa	VDR	Wang et al., 2016^Citation10
	Longitudinal GEM Project	Stool	1,561	16S rRNA gene sequencing	166 taxa	Genome-wide: 5 × 10^−8	58 SNPs associated with the relative abundance of 33 taxa	CNTN6, DMRTB1, SALL3, UBR3	Turpin et al., 2016^Citation65
	TwinsUK	Stool	1,126	16S rRNA gene sequencing	β-diversity, 20 heritable taxa	Genome-wide: 5 × 10^−8	3 loci linked to β-diversity, and 28 loci for bacterial taxa	LCT, ALDH1L1	Goodrich et al., 2016^Citation66
	Hutterite individuals	Stool	127	16S rRNA gene sequencing	116 taxa for winter, 104 taxa for summer	genome-wide FDR of 20%	37 associations	Discussed candidate genes around significant signals	Davenport et al., 2015^Citation59
	Individuals of HMP subjects	Stool	93	Shotgun metagenomic sequencing	α- and β-diversity, 615 microbiome abundance traits	Q-value<0.1	83 associations	LCT, HLA-DRA, TLR1	Blekhman et al., 2015^Citation67
Pig	F6 and F7 generations of a mosaic population constructed with four western pig breeds and four Chinese indigenous pig breeds	Feces at the ages of 25, 120 and 240 days, cecal and ileal content at days 240, cecal and ileal mucosal scrapings at days 240	Total 5,110 samples from 1,430 pigs	16S rRNA gene sequencing	55 ~ 375 taxa for Quantitative model and 499 ~ 3,762 taxa for binary model	Genome-wide: 5 × 10^−8; Study-wide: 1.5 × 10^−12	1,527 genome-wide significant signals, 6 signals exceeding the experiment-wide discovery threshold	ABO	Yang et al., 2022^Citation11
	Erhualian, Bamaxiang	Cecum content from Erhualian and stool from Bamaxiang	256 cecum content and 244 stool samples	16S rRNA gene sequencing	1,411 OTUs and 153 taxa in fecal samples, 1,177 OTUs and 168 taxa in cecum contents	Genome-wide: 6.65 (6.84) × 10^−8; suggestive: 1.33 (1.37) × 10^−12	40 significant associations for feces samples, 34 significant associations for cecum content	Discussed candidate genes around significant signals	Chen et al., 2018^Citation68
	Duroc pig population	Stool	390	16S rRNA gene sequencing	2 diversity indexes, 31 bacterial and 6 commensal protist genera	P < 0.05, 0.01, and 0.001	Immune genes are key regulators for multiple microbes	five regulator genes and dozen target genes	Reverter et al., 2021^Citation69
	Duroc pigs	Stool	514	18S and ITS ribosomal RNA gene sequencing	α-diversity, 492 fungal and 227 protist ASVs	FDR<0.05	174 SNPs related with the abundance of protist community	Discussed candidate genes around significant signals	Ramayo-Caldas et al., 2020^Citation70
	Duroc×Yorkshire × Landrace or Duroc × Landrace × Yorkshire	Stool	1,205–1,283 per time point	16S rRNA gene sequencing	α-diversity, 334 OTUs associated with growth and fatness	P < 1 × 10^−5	207 significant associations	Discussed candidate genes around significant signals	Bergamaschi et al., 2020^Citation71

Correlation between host genetic kinship and the dissimilarity of gut microbial compositions

There have still been controversies about the correlation between host genetic kinship and the dissimilarity of gut microbial compositions. The study in humans found that memberships with the closer degree of relatedness in families showed a higher overall similarity in gut microbiome⁷². Microbiotas were more similar within twin pairs compared to unrelated individuals, and monozygotic twin pairs had slightly more similar microbial compositions compared to dizygotic twin pairs in the Twins UK population⁷³. The variation of a given taxon abundance was more frequently explained by host genetic similarity, while the presence of a given taxon more depended on a shared home environment⁷¹. However, Rothschild et al. (2018) found no significant association between genetic kinship and microbiome composition⁸. This controversial result should be caused by the fact that environments dominate over host genetics in shaping human gut microbiota⁸. Compared to humans, livestock and experimental animals are more easily controlled to a uniform condition by raising in the same farm house and being provided the same diets. This could eliminate part of the influence of environmental factors on gut microbiome and highlight the effect of host genetics. For examples, in cattle, sire breed composition primarily explains the differences in gut microbiota structure among breed groups of preweaning calves that were maintained in the same pasture and accessed to an identical diet⁶⁴. In pigs, the diversity of gut microbiota was compared among full-sibs, half-sibs, and unrelated pigs in two pig populations of Erhualian and Bamaxiang and found that full-sibs and half-sibs had a higher similarity of microbial compositions than unrelated pigs in both Weighted and Unweighted Unifrac analysis in two populations raised in the same farm⁵⁷. The investigation on the relationship between genetic relatedness based on genome-wide SNP identity-by-state and gut microbiota dissimilarity based on Bray-Curtis distance found significantly negative correlations between genetic similarity and microbiota dissimilarity in the F₆ and F₇ generations of a mosaic population¹¹. The analysis was restricted to full-sib littermates raised in the same environment, which avoided the interferences of environmental and maternal factors to the results¹¹. In mice, bacterial communities retained strain specificity in eight progenitor mouse strains of the collaborative cross, although it became more similar after cohabitation of different mouse strains, indicating the interaction of host genetics and environmental factors in shaping gut microbiota⁶². Notably, compared to mice, pigs are more suitable as a model for the studies on the role of host genetics in the gut microbiome because pigs showed a higher similarity of the gut microbial composition to humans than mice¹¹. The pairwise overlap of genes in the gut microbiome was significantly higher for pigs and humans (9.49%) than for mice and humans (1.50%). More similar gut microbiome and easily controlled environments have made pigs as an important model to investigate the relationship between host genetic kinship and the similarity of gut microbial compositions.

Heritability estimates

Heritability has been generally used to predict genetic influence on phenotype and always shows the proportion of phenotypic variation in a trait (such as body weight) that is measured across a population and can be explained by genetics rather than environmental effects. In the studies of gut microbiome, the degree to which host genotype versus environment influences the gut microbiome has become increasingly more interesting. Quantifying the heritability of the gut microbiome is to discern genetics from environmental factors that structure the microbiome and to potentially identify members of the gut microbial community influenced by host genotypes. There are two conventional methods to quantify heritability of microbial taxa. One method is twin-based studies, which estimates heritability (h²) of microbial taxa by comparing the variation in microbial taxon abundances owing to host genetics. The other method is standard statistical approaches, such as additive genetic model and unique environment (ACE) model. In both methods, the abundance of each gut microbe is treated as a continuously varying quantitative trait⁶¹. Heritability predicts the host genetic effect on the abundance of a gut microbe. This heritability estimate has been performed more in the studies of human gut microbiome. For instance, Goodrich et al. (2014)⁷³ firstly identified heritable gut microbial taxa in the TwinsUK population and reported that 5.3% of the taxa showed a heritability estimate greater than 0.2 in stool samples of 416 pairs of dizygotic and monozygotic twins. This percentage was increased to 8.8% in their further study with larger sample size of 1,126 twin pairs⁷⁴. The most heritable taxon was the family Christensenellaceae, which was confirmed in different cohorts^55,74,75. Further study in Canadian Genetic Environmental Microbial Project cohort identified that more than half of the heritable taxa could be compared between TwinsUK population and Canadian Genetic Environmental Microbial Project cohort⁵⁵. Davenport et al. (2015)⁷⁶ identified 15 heritable taxa in stools from different sampling seasons in Hutterites. Most of these heritable bacteria belong to the phyla Proteobacteria or Firmicutes. In mice, the study in HMDP strain mice revealed that host genetics explained a substantial amount of the variation (up to 0.5 or more) for many common taxa when experimental mice were maintained under controlled conditions⁷⁰. O’Connor et al. (2014)⁷⁷ observed the influence of host genetic background in both microbial community structure and individual taxa and identified a wide range of heritability for taxa at the genus level (0.26–0.86) in a collaborative cross. In pigs, Camarinha-Silva et al. (2017)⁷⁸ estimated the heritability of gut bacterial genera using 207 Pietrain pigs as experimental animals and identified 8 out of 49 bacterial genera with heritability ranging from 0.32 to 0.57. Unlike the twins’ study, pig populations are composed of complex pedigree. Visscher et al. (2008)⁷⁹ suggested that for samples from the cohorts with complex pedigree relationships, liner mixed model is more suitable for the heritability estimate. In our previous study, with a liner mixed model, we identified 81 taxa having h²>0.15 in fecal samples of Bamaxiang pigs and 67 taxa in cecum luminal samples of Erhualian pigs. Among them, 31 taxa had h²>0.15 in both types of samples⁵⁷. In the F₆ and F₇ generations of a Mosaic population, we estimated the heritability of the abundances of individual taxa/OTUs and found that the total heritabilities of gut microbiota were ≤11.8%¹¹. The total heritability for fecal samples was higher than that for luminal content samples (10.2% vs. 5.7%)¹¹. Based on congruent estimates in F₆, F_7, and across F₆ and F₇, a total of 55 most likely heritable taxa/OTUs were identified, including OTUs belonging to Christensenellaceae and Blautia that were concordance with the identifications obtained in human studies¹¹. Interestingly, some bacterial taxa, such as Christensenellaceae^11,63,73, Lachnospiraceae^57,70,76, Rikenellaceae^57,70,Oscillospira^70,73,80, Ruminocaceae^11,63,70,74, and Blautia^11,57,74 were identified to be heritable in multiple host species in different studies.

It should be noted that conventional heritability measurements tend to consider the interaction between host and gut microbiome as unidirection, in which the host regulates microbe colonization⁶¹. However, more comprehensive views have thought that both host and microbiome play a role in heritability. A community heritability (H ²_C) which treats the host as part of an ecosystem and measures the extent to which variation in “whole-community” phenotype is attributable to the host genetic variation has been recommended to predict host genetic influence on the gut microbiome⁶⁰.

Identification of miQTL influencing gut microbial composition by GWAS

Identifying host genomic loci (genetic variants) or key genes associated with the gut microbiome directly by applying GWAS (most popular method to date) or quantitative trait locus (QTL) mapping using quantitative measures or presence/absence of individual taxa and diversity indices and functional pathways of the gut microbial community as the complex traits^{55,58,67,70,74,76,77} is another method evaluating the effect of host genetics on the gut microbiome. As a mainstream approach, GWASs have been performed broadly to evaluate the effect of host genetics on the gut microbiome in different host species, such as humans (Table 1)^{9,56,74,81,82}, pigs^11,57,65, cattle^63,80, and poultry⁸³. Here, we focused on the advances in the studies of pig gut microbiome. As the benchmarks, the studies about microbial GWASs on human gut microbiome are also summarized in details because more promising advances have been made in human studies with large sample sizes from several thousands to tens of thousands of feces samples. Blekhman et al. (2015)⁵⁸ first performed the microbial GWAS in 93 individuals from the Human Microbiome Project and reported the association between the abundance of Bifidobacterium in feces and LCT locus. Davenport et al. (2015) performed a GWAS in humans to identify genomic loci associated with gut bacterial taxa in 127 samples collected from different seasons with 16S rRNA gene sequencing and identified 37 associations at genome-wide Q value <0.2⁷⁶. Since then, more than dozen studies have investigated host genetic effects on human gut microbiome (Table 1). More recently, a fecal microbial GWAS in 4,117 individuals from a single multiethnic cohort identified ethnicity-specific associations between host genomic variants and gut microbiota⁸⁴. The study in human gut microbiome with the largest sample size so far (stool samples from 18,340 individuals in 24 cohorts) identified 31 genomic loci associated with the gut microbiome at the threshold of 5 × 10⁻⁸, and this study replicated the finding of LCT genes influencing the abundance of Bifidobacterium⁷⁹. However, the studies in humans described above have relied on stool samples. Fecal microbiota is mostly from the colon and luminal microbes⁸⁵. The microbiota farther up the gastrointestinal tract may be rarely detectable in the stool.

Less studies about host genetic effect on the gut microbiome by GWAS have been reported in pigs, and the sample sizes used in each study were also significantly smaller (Table 1). To our knowledge, until now, only five studies have reported the GWAS for identifying genomic loci for pig gut microbial composition. Yang et al. used pigs from the F₆ and F₇ generations of a mosaic population that was constructed with four Western pig breeds and four Chinese indigenous pig breeds as experimental animals. Whole-genome resequencing of 1,430 pigs from two generations identified more than 30 million host genomic variants (i.e., one variation per 100bp on the genome). A total of 5,110 samples collected at different ages and from different gut locations, including feces at the ages of 25, 120, and 240 d, cecal and ileal contents at days 240, and cecal and ileal mucosal scrapings at days 240 (only for F₇ pigs) from these experimental pigs were performed high throughput sequencing of the V3-V4 region of 16S rRNA gene. Based on the F₆ and F₇ experimental pig populations with high genetic variations and gut microbiota data, genome-wide association analysis was carried out on 8,490 bacterial taxa. A total of 1,527 miQTL that significantly affected the abundance or absence/presence of 846 bacterial taxa were identified at the threshold of 5 × 10⁻⁸. Among 1,527 genome-wide significant signals, 6 signals exceeded the experiment-wide significant threshold (P < 1.5 × 10⁻¹²)¹¹. Another study from our laboratory collected 256 cecum content samples from Erhualian pigs and 244 stool samples from Bamaxiang pigs and acquired the 16S rRNA gene sequencing data from these samples. All these 500 pigs were genotyped with 1.4 million genome-wide SNPs. GWAS for 1,411 OTUs and 153 taxa in fecal samples and 1,177 OTUs and 168 taxa in cecum contents identified 40 and 34 significant associations, respectively⁵⁷. In both studies described above, core taxa (existed in >95% experimental pigs) were treated as quantitative traits, and taxa identified in 20–95% pigs were considered both quantitative and binary features. Another representative study used 1,205–1,283 feces samples per time point from Duroc × Yorkshire × Landrace or Duroc × Landrace × Yorkshire, performed the GWAS for the α-diversity indices and 334 OTUs associated with growth and fatness, and identified a total of 207 significant associations between host genetic variants and microbial taxa⁶⁵.

Candidate or causative genes influencing the gut microbial composition

About the genes influencing gut microbiome, many GWASs have discussed candidate genes isolated from the defined intervals around significant signals (variants). For examples, Blekhman et al. (2015)⁵⁸ first observed the significant correlation between SNPs in the LCT gene and the abundance of Bifidobacterium. Goodrich et al. (2016)⁷⁴ replicated the association between LCT and Bifidobacterium and identify another association between ALDH1L1 and SHA-98. Wang et al. (2016)¹⁰ identified multiple genetic loci, including vitamin D receptor (VDR) gene associated with overall microbial variation and the abundance of individual taxa. Turpin et al. (2016)⁵⁵ identified the SNPs nearest to UBR3 associated with Rikenellaceae, the SNPs in CNTN6 associated with Faecalibacterium, the variants in DMRTB1 associated with Lachnospira, and the SNPs in SALL3 associated with Eubacterium. The study reported by Qin et al. (2022) not only replicated the association between the variants at the LCT locus and the abundance of Bifidobacterium, but also identified the correlations between the abundance of Faecalicatena lactaris and ABO gene and between the abundance of Enterococcus faecalis and the variants in the MED13L locus. Functional classifications have always been performed to annotate the functional categories of these isolated candidate genes near to strong signals from GWAS and suggested that these candidate genes are mainly associated with metabolism, gut homeostasis, immunity functions and response, signal transduction, and so on. In pigs, our previous study identified six signals exceeding the experiment-wide significant threshold (P < 1.5×10⁻¹²) in one cohort and the experiment-wide replication threshold(P < 0.007) in the other cohort. These six signals mapped within 3,037 bp from each other in the 272.8–273.1Mb interval on chromosome 1 and significantly affect the abundance or presence/absence of the OTUs from p-75-a5 or Erysipelotrichaceae as well as genus p-75-a5. Fine-mapping identified a 2.3-kb deletion in ABO gene is the causative mutation for this miQTL¹¹. Another study based on gene co-association network analysis in a Duroc pig population identified five regulator genes and dozen target genes regulating gut microbial communities⁸⁶.

Unfortunately, until now, the genes or genomic loci that could be replicated in multiple studies are few in both humans and pigs. To the best of our knowledge, only LCT and ABO locus have been found to be associated with gut microbes in different cohorts. Among them, LCT locus and its association with the abundance of Bifidobacterium has been most commonly replicated in multiple studies of human gut microbiome^{8,9,56,58,67,74,82}. As for ABO gene, at least three studies with the large scale of sample sizes reported the significant associations of ABO gene with human or pig gut microbiome^9,82. But different from LCT locus, ABO gene has been reported to be correlated with different bacterial taxa in different studies. In 7,738 participants of the Dutch Microbiome Project, ABO locus was associated with Bifidobacterium bifidum, Collinsella aerofaciens, Collinsella, Coriobacteriaceae, and Coriobacteriales depending on the secretor status by host FUT2 genotype⁹. In the participants of the FR02 study, ABO gene was associated with Faecalicatena lactaris, which was also largely driven by secretor status⁸². In pigs, ABO gene was associated with the OTUs belonging to p-75-a5 or Erysipelotrichaceae, as well as the genus p-75-a5. However, no evidence showed that the associations in pigs were independent on the secretor status¹¹.

Elucidating how host genes influence the gut microbial composition is important to regulate gut microbiome for host health or pig production performance. However, to our knowledge, until now, only the mechanisms of LCT and ABO genes affecting the abundance of gut bacterial taxa have been elucidated. LCT encodes lactase that hydrolyzes lactose, the main sugar in mammalian milk, to produce D-glucose and D-galactose. Mutations in the translated region of the LCT gene cause lactase deficiency that is a severe gastrointestinal disease characterized by watery diarrhea in infants fed with breast milk or other lactose-containing formulas⁸⁷. Dairy intake does not affect the abundance of Bifidobacterium in individuals producing lactase through adulthood. However, in lactase deficiency or intolerant individuals, dairy diet significantly increases the gut Bifidobacterium abundance, and heterozygote individuals had an intermediate increase in Bifidobacterium abundance^67,82. CAZyme profile analysis further revealed that, compared to Bifidobacterium spp. that are affected by LCT gene, Bifidobacterium dentium that is not affected by LCT harbored more genes encoding CAZyme families with preferred fiber/plant-related substrates, whereas Bifidobacterium spp. affected by LCT gene harbor more milk oligosaccharide-targeting CAZyme families (GH129 and GH112)⁸².

ABO gene encodes the alpha 1–3-N-acetylgalactosaminyltransferase which adds N-acetylgalactosamine to glucoprotein of intestinal mucosa and blood cell. In pigs, the blood group A gene encodes A transferase, which transfers GalNAc to the galactose residue of the acceptor H substrates, including the heavily glycosylated mucins which can be used as carbon source by intestinal bacteria⁸⁸, while the group O gene is nonfunctional due to a 2.3-kb deletion encompassing part of intron 7 and the whole exon 8. And H substrates in group O individuals remain without additional modifications^89–91. Different mechanism of ABO gene affecting the abundance of gut bacterial taxa has been reported between humans and pigs. In humans, the effect of ABO gene on the gut microbes depends on secretor status (FUT2 genotypes)^9,82, but was not associated with fiber intake⁸². In pigs, our study did not identify any evidences for an effect of FUT2 variants on the abundance of the bacterial taxa affected by ABO gene¹¹. By integrating multi-omics data of metabolome, cultureomics, metagenome, and transcriptome, we systematically illustrated how the ABO genotypes affect the abundance of Erysipelotrichaceae species:

The 2.3-kb deletion of the ABO gene (in type O individuals) results in the absence of N-acetylgalactoamine transferase activity of the encoded protein, and therefore, the inability to add N-acetylgalactoamine (GalNAc) to the highly glycosylated mucins in intestinal mucus. This led to the reduced intestinal concentrations of GalNAc in individuals with OO genotype, and thereby, reduced the growth of some Erysipelotrichaceae species used GalNAc as carbon source (Figure 2). Furthermore, functional GalNAc import and catabolic pathway analyses found that, compared to bacterial taxa that were not affected by ABO locus, the bacterial taxa affected by ABO gene (Erysipelotrichaceae species) had more complete GalNAc import and catabolic pathways, and the genes involved in GalNAc import and catabolism tended to cluster together to form coregulated gene operons in these Erysipelotrichaceae species¹¹.

Figure 2. The mechanism of a 2.3-kb deletion in ABO gene affecting erysipelotrichaceae species.

Gut microbiota conversely shapes the host epigenetics through its components and derived metabolites

The gut microbiota can regulate gene expressions of both local intestinal cells and peripheral tissues through its components and derived metabolites under a potent mechanism of epigenetic regulation⁹². Microbial biosynthesis or metabolism influences the availability of chemical donors for DNA or histone modifications. DNA/histone methyltransferases and histone acetyltransferases generally depend on methyl and acetyl donors, respectively, for their catalytic activity. As an additional source, the gut microbiota can synthesize biological compounds that serve as epigenetic substrates, cofactors or regulators of epigenetic enzyme activity⁹³. For instance, the gut microbiota, such as Bifidobacterium and Lactobacillus species, generates folate and other B vitamins (B2, B12) that donate methyl groups for DNA/histone methylation^94,95. The gut microbiota also regulates epigenetic modifying enzyme expression and/or activity. For example, short-chain fatty acids (SCFAs) that are exclusively produced by commensal microbes through fermentation of complex non-digestible carbohydrates and represent an important kind of epigenetically relevant molecules can inhibit deacetylase activity of histone deacetylases (HDACs) that is generally associated with increased expression of target genes⁹⁶. The researchers are also aware of the importance of the gut microbiome on the regulation of host epigenetics in pigs. The study in piglets found that early microbial colonization affects DNA methylation of genes related to intestinal immunity and metabolism in preterm pigs⁹⁷.

Advantages and challenges of employing pigs to decipher the effect of host genetics on gut microbiota composition

Advantages

Intensive farming has been popularly used in the pig industry, which reduces the number of pig farms and expands the scale of farming. It has also achieved the industrialization of pig farming through transforming production methods from quantity expansion to quality-benefit. The efficiency of pig production has been greatly improved. The indoor intensive housing for pig farming always keeps pigs of similar ages in groups in common environments. The similar diets (formula feeds in general) are supplied in place for pig populations, which mix a variety of different feed ingredients according to a certain proportion, and meet the energy and nutrition requirements for each growth stage of pigs. Ventilation systems have always been used to regulate moisture and temperature of farm house. This kind of rearing model would benefit for the study investigating the host genetic effect on the gut microbiome in pigs because most environmental factors that also influence the gut microbial compositions are shared in common and can be controlled easily by man-made factors. More importantly, the formula feeds provided to all pigs are similar and meet the energy and nutrition requirements according to pig nutrition requirement criterion. This significantly reduces the diet effects on the gut microbiome. Furthermore, the healthy status of pigs and the use of medicines (e.g., antibiotics and probiotics) can be registered easily by workers. All these proceedings significantly reduce the variates influencing the gut microbial compositions. As an example, in our previous study¹¹, all F₆ and F₇ pigs used in the experiments were reared under standard and uniform housing and feeding conditions and fed twice per day with formula diets containing 16% crude protein and 3,100 kJ digestible energy.

Another advantage using pigs as experimental animals is that the pedigrees of experimental pig cohorts can be recorded clearly for all pigs, even in multiple generations. Pedigree information is important to evaluate the host genetic effect on the gut microbiome, especially in the heritability estimation. Moreover, pigs consisting of large pedigrees with existing rich phenotypes should be important resource populations for high-throughput sequencing in the search of the causative variants influencing complex traits⁵⁹. And the large pedigree design is also particularly useful in the presence of locus heterogeneity among families⁶⁶.

In addition, litter size of sows is large, which facilitates to adjust the maternal effect on the gut microbial compositions in the studies evaluating the genetic effect on the gut microbiome. At birth or during the nursing, vertical transmission of maternal microbiota to offspring supports postnatal growth and development⁹⁸. The nursing mother influences the gut microbial composition of offspring during early development via breastfeeding⁹⁹, which plays important roles for health of the animals at later stages of life¹⁰⁰. Sows also can significantly affect the gut microbiome of piglets as observed in other mammals. As a multiparous mammal, more than ten offspring at one birth of a given sow provide excellent materials for evaluating host genetic effect on the gut microbiome, e.g., heritability estimates. It can circumvent the challenge to distinguish between microbes that have been transmitted vertically (maternal effects) or selected genetically because the rates of microbial vertical transmission are not expected to differ among piglets in a farrow and the environments are identical, differences in microbial similarity are attributed to genetic factors alone.

To our knowledge, all existing studies in humans evaluating host genetic effect on the gut microbiome or identifying genetic variants associated with gut microbial taxa by GWAS were carried out with feces samples and also in most of the studies of pigs (Table 1). As mentioned above, distinct gut microbial compositions in different gut locations have been reported in many studies^15–17, and limiting the analysis of the gut microbiota to feces are far from enough to understand its complexity and elucidate the factors that determine it⁷. The conditions are similar to the gut microbial compositions at different ages^18,19, and the samples collected at different ages were needed to systematically and deeply evaluate host genetic effect on the dynamic changes of gut microbes. The heterogeneity of miQTL effect should exist for the samples from different gut locations and ages. For example, the miQTL identified at the ABO locus affects Erysipelotrichaceae species in cecum content and mucosa samples in day 120 and 240 feces, but not in day 25 feces and in ileum contents at day 240¹¹. As an advantage using pigs as experimental animals, microbiota samples can be obtained from the same population at different ages, and lumen content and mucosa samples also can be obtained from different gut locations when pigs are slaughtered.

Elucidating the mechanisms of host variants affecting microbial taxa is important for the regulation of gut microbial compositions. To our knowledge, except LCT gene in humans and a 2.3-kb deletion of ABO gene in pigs, no other host variants have been elucidated its mechanism affecting the gut microbes. Multi-omics method that generally includes metagenome of the gut microbiome, metabolome of feces, intestinal contents, and other target tissues, transcriptome of microbes, intestinal tissues and other target tissues, and culturomics are useful to elucidate the mechanism of causative mutations affecting the gut microbes. Using pig as experimental animals, many different tissue types, e.g., intestinal biopsy, muscle, blood, and so on can be harvested from the same individuals that would facilitate the elucidations of mechanism of host genes in affecting the gut microbial compositions.

Main challenges

Compared to that in humans, less GWASs have been performed to identify host genetic variants associated with the diversity of gut microbiota and the abundance of specific taxa in pigs. Several studies have identified some significant associations in various populations^11,65,86 (Table 1). However, similar to the conditions observed in humans, the replication of significant associations is low between studies, although the standard and uniform housing and feeding conditions significantly reduce the variates influencing the gut microbial compositions compared with that in humans, which should benefit the identification of genomic loci associated with gut microbial taxa in a study. The significant differences in farming conditions, managements, and workers provided to experimental pigs have still existed across the studies from different laboratories and groups, which led to the heterogeneity of microbial compositions and high interindividual variability across pig populations. Furthermore, to avoid the effect of maternal and early colonization on the gut microbial compositions, it should be best that experimental pigs should be from the same farm. However, huge expense and manpower are required when large sample size is used, especially, when microbial samples are collected from different gut locations, experimental pigs would need to be slaughtered for sample collection.

Another challenge, we think, resulting in the low replication between studies is the bias introduced in quantifying the abundances of bacterial taxa in samples by 16S rRNA gene sequencing (and also by shotgun metagenomic sequencing,¹⁰¹ which has been met commonly in earlier studies of microbial community structure, including sample processing pipeline, DNA extraction (kit and protocol), PCR amplification (different efficiency of amplifications for bacterial taxa)¹⁰², the choice of which hypervariable region(s) of 16S rRNA gene to sequence¹⁴, sequencing technology, and data processing^103,104. As an example of the effect of technique bias on microbial GWAS, Qin et al. (2022) reported that the association of LCT locus with Bifidobacterium in a single cohort of 5,959 individuals was stronger than that in an integrative data of 18,473 individuals from 28 different cohorts, suggesting the importance of standardized methodology^56,82.

Perspectives

As we have discussed above, there are only a few studies about host genetic effect on the gut microbiome in pigs, and a small number of genomic variants associated with pig gut microbes have been reported. Compared to the current studies in humans^9,56,82, the sample sizes used in the studies of pigs are small (Table 1). As indicated by Lopera-Maya et al.⁹ (2022), more than 50,000 individuals are required to identify an effect size similar to that of LCT and ABO for bacterial taxa present in >20% of the samples in humans. Therefore, in the future studies, we should take advantage of easily obtaining large sample sizes in pigs even from one farm under uniformed conditions. Furthermore, the protocols of experiments, such as DNA extraction and sequencing, and bioinformatic analyses to obtain the gut microbial composition data should be standard to facilitate to combine the datasets from different studies (meta-analysis of GWAS). In addition, following the decrease of sequencing cost, instead of 16S rRNA gene sequencing, metagenomic sequencing data has begun to be used in the GWASs of gut microbiome just like two current studies in humans^9,82 and the previous study by Bonder et al. (2016)⁶⁷.

Obtaining the associations between genetic variants and microbial taxa is far from enough. To those genomic loci that are significantly associated with gut microbes with large effect size (e.g., achieving study-wide significance level), further fine-mapping should be performed to identify causative genes and mutations, and the mechanisms will also need to be elucidated using multi-omics data. Based on this, we can further investigate the effect of the interaction between host gene and gut microbiota on complex traits or pig production performances. Another method evaluating and verifying the effect of host gene or genomic locus on the gut microbiome is to inactivate the gene(s) of interest through genetic modification and define its effect on the gut microbiome. With the rapid development of new revolutionary technologies of genetic modification, such as CRISPR-Cas9 and transcription activator-like effector nuclease (TALEN), it is possible for the first time to precisely modify the porcine genome as never before. However, to our knowledge, there is no relevant report about investigating host genetic effects on the gut microbiota using these technologies. In mice, the knock-out of genes at least 30 genomic loci has been shown to produce distinct gut microbiomes or altered bacterial colonization engraftment compared to wild-type mice¹⁰⁵.

Finally, the gut microbiota community is a complex ecosystem. Beside bacteria, there are large numbers of viruses, fungi, and archaea. It is interesting to identify host genes/variants associated with viruses, fungi, and archaea. However, to our knowledge, there are few studies about host genetic effects on these microbes because bigger challenges exist for these kinds of microbes compared to bacteria owning to the low coverage of sequences in tested samples and the lack of universal marker genes (except fungi). Large sample size and deep sequencing are required for further GWAS analysis.

Acknowledgments

We are grateful to Hui Yang, Jinyan Wu and the colleagues in the State Key Laboratory of Pig Genetic Improvement and Production Technology, Jiangxi Agricultural University, and Professor Michel Georges and Dr. Carole Charlier in the Unit of Animal Genomics, GIGA-Institute and Faculty of Veterinary Medicine, University of Liege for working together on the host genetic effects on pig gut microbiome.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by The National Natural Science Foundation of China (32272831, 31790410, 31772579) and National pig industry technology system (CARS-35).