Molecular microevolution and epigenetic patterns of the long non-coding gene H19 show its potential function in pig domestication and breed divergence

Background The domestic pig Sus scrofa domesticus originated from the wild boar S. scrofa about 10,000 years ago. During domestication, drastic morphological, physiological, and behavioral changes developed between domestic pigs and wild boars through artificial and natural selection. The long non-coding RNA (lncRNA) H19, which is located within the imprinting gene cluster H19-IGF2, plays an important role in regulating muscle development in humans and mice. This study systematically analyzed the molecular evolution of H19 and its possible epigenetic changes during pig domestication and breeding to explore the genetic and epigenetic contributions of H19 to pig domestication. Results The molecular evolution of H19 was initially analyzed on a large phylogenetic scale. Results showed that the gene was highly conserved within a broad range, especially in the 5′ terminal sequence. The molecular evolution of the gene was then analyzed using published re-sequencing data of 30 wild boars from Tibet, 3 wild boars from Sichuan, and 15 native pigs from other regions in China. Eight polymorphic sites were identified, and the nucleotide diversity (π) value within the H19 gene body was significantly higher (Z-test, P < 0.05) in domesticated pigs than in wild pigs. However, no significant divergence occurred between domesticated and wild pigs. Single nucleotide polymorphisms in the 3′ terminal sequence were surveyed in other Chinese local breeds and foreign pig breeds. We observed a consistently higher diversity in domesticated pigs than in wild pigs. The methylation pattern of the H19 gene in pigs was subsequently analyzed using published methylated DNA immunoprecipitation data and an unpublished single-base resolution liver methylome. Analysis results showed distinct methylation levels in some tissues. Among the samples surveyed, Landrace showed the lowest methylation level, followed by the Guizhou wild boar, whereas the Enshi pig exhibited the highest methylation level in the 2 kb upstream region of the H19 gene. Liver transcriptome data suggested that Landrace harbored the highest expression of the H19 gene, followed by the Guizhou wild boar, whereas the Enshi pig harbored the lowest expression of the gene. Differential methylation sites (DMSs) among the three breeds were mainly identified in the 2 kb upstream region of the H19 gene. In the Enshi pig, we detected allele-specific methylation (ASM) regions in the 2 kb upstream region of the H19 gene. Most of the DMSs in the upstream 2 kb region of the gene were also located in the ASM region in this breed. Conclusions Molecular analyses suggest that the H19 gene was highly conserved during large-scale evolution and exhibited genotype differentiation during domestication and breed differentiation. The drastic diversity pattern between domestic and wild pigs in the H19 gene body, which was highly conserved during large-scale evolution, suggests that this gene might have played roles in the breed differentiation of domestic pigs. Methylation analysis indicates an opposite epigenetic regulation direction between Chinese and European pig (EU) domestication, which resulted in opposite expression changes in this gene between the two domesticated groups. Our preliminary analyses on DMSs among different pig breeds and ASM imply that imprinting was associated with methylation differences. This study systematically demonstrates the genetic and epigenetic patterns of H19 during pig domestication and provide valuable cues and basis for further research on the function of H19 in pig domestication. Electronic supplementary material The online version of this article (doi:10.1186/s12862-016-0657-5) contains supplementary material, which is available to authorized users.


(Continued from previous page)
Conclusions: Molecular analyses suggest that the H19 gene was highly conserved during large-scale evolution and exhibited genotype differentiation during domestication and breed differentiation. The drastic diversity pattern between domestic and wild pigs in the H19 gene body, which was highly conserved during large-scale evolution, suggests that this gene might have played roles in the breed differentiation of domestic pigs. Methylation analysis indicates an opposite epigenetic regulation direction between Chinese and European pig (EU) domestication, which resulted in opposite expression changes in this gene between the two domesticated groups. Our preliminary analyses on DMSs among different pig breeds and ASM imply that imprinting was associated with methylation differences. This study systematically demonstrates the genetic and epigenetic patterns of H19 during pig domestication and provide valuable cues and basis for further research on the function of H19 in pig domestication.
Keywords: Pig, Long non-coding gene, H19, Molecular evolution, Methylation, Domestication, Breed divergence Background Pigs are important domesticated animals that provide an important source of meat in China and the whole world. Pigs in East Asia originated from the East Asian wild boar and were domesticated approximately 10,000 years ago [1]. Natural and artificial selection have promoted the development of significant differences in morphology, growth, fertility, and local fitness between domesticated and wild pigs [2]. Various Chinese local pig breeds have been produced through domestication and breeding. Statistical data from the United Nations Food and Agricultural Organization (FAO) indicated that 118 native pig breeds, which accounted for 24 % of the global count, were available in China in 2000 [3]; these breeds may be classified into six types: North China (NC), South China (SC), Central China (CC), Jianghai (JH), Southwest China (SW), and Plateau. Diverse breeds of local pigs present valuable genetic resources [4]. Pigs from different regions show substantial variations in physiological features. For instance, Taihu pigs from JH exhibit high fecundity; CC-and SC-type pigs show high adipose contents, which possibly contribute to the good taste of meat from these pigs; and Plateau-and NC-type pigs contain more lean meat than fat. Clearly, pigs from different regions in China have developed different characteristics that have enabled them to adapt to particular environments and conditions during microevolution. However, the molecular evolutionary mechanisms underlying the diversity of the same traits among different pig breeds during domestication and improvement remain unclear. Several genes plays important roles in species domestication and breed divergence [5][6][7][8]. As key factors in pig domestication, non-coding RNAs have been speculated to serve important functions in pig domestication and breeding [9].
Non-coding genes are DNA sequences that produce RNA with no translational ability [10]. Most non-coding RNAs can regulate gene expression [9]. The long non-coding gene H19 is located in the H19/IGF2 imprinted gene cluster. Studies on mice and human have shown that the H19 gene contributes to the growth, development, and generation of muscle fibers [11][12][13][14].
Aside from genetic mutations, epigenetic mechanisms also influence gene expression and explain how geneenvironment interactions yield particular phenotypes during development [15]. In particular, some complex genetic phenomena cannot be explained by DNA variations alone, but integrated with epigenetic systems, such as DNA methylation with strong reversibility, microRNA regulation or histone modification, which can be passed to subsequent generations [16]. Recent studies have suggested that the H19 gene influences embryonic growth and development as well as skeletal muscle differentiation and regeneration via epigenetic mechanisms, such as DNA methylation. Abnormal methylation of the H19 gene may be associated with cancer formation [17,18]. In addition, research on the epigenetic mechanisms underlying Beckwith-Wiedemann Syndrome (BWS) [19] and Silver-Russell Syndrome (SRS) [20] has shown that methylation of the H19 gene exerts a pronounced effect on disease and muscle development. Pidsley [21] analyzed DNA methylation and genotypes in three differential methylated regions (DMRs) and investigated the H19 promoter in the H19/IGF2 cluster in humans; results showed that DNA methylation is strongly associated with cerebellum weight [21]. These findings indicate that H19 methylation exerts a pronounced effect on disease and muscle development.
Pigs are highly domesticated animals that drastically vary in meat features depending on the breed and geographic location. The H19 gene contributes to the skeletal muscle growth and development of animals through both genetic variation and epigenetic modification. Therefore, H19 was speculated to play a vital role in porcine domestication and breed divergence via genetic mutation and/or epigenetic changes.
To understand the molecular evolution pattern and methylation pattern of the H19 gene in relation to pig domestication and breed divergence, the H19 gene and its flanking region were analyzed using the homologous sequence of 18 species by Phast software [22] and Mega software [23]. Results revealed that the H19 gene is relatively conservative in a large-scale evolution. Published re-sequencing data [24] of 30 wild boars from Tibet, 3 wild boars from Sichuan, and 15 native pig breeds from other regions in China were initially used to detect single nucleotide polymorphisms (SNPs) in the H19 gene. Most SNP sites were distributed in the 3′-terminal of the H19 gene. Candidate SNPs from 14 Chinese native pig breeds and 3 foreign pig breeds (a total of 234 individuals) were also subjected to clone sequencing to survey the SNPs obtained using polymerase chain reaction (PCR). The nucleotide diversity and F-statistics (F st ) along the upstream, gene body, and downstream regions of the H19 gene were analyzed using the re-sequencing data of 30 Tibetan wild boars, 3 wild boars from Sichuan, and 15 Chinese native pig breeds. The H19 gene was conserved during domestication and had a high divergence in the domestication pig population, indicating the functional role of this gene. Separate methylated DNA immunoprecipitation sequencing (MeDIP-Seq) data were obtained from eight adipose tissues and two muscle tissues of three specimens of Tibetan wild boars, Rongchang pigs, and Landrace pigs. Single baseresolution bisulfate methylation sequencing data of liver tissue from one Chinese Enshi pig, one Landrace pig, and one Guizhou wild boar were also obtained (unpublished data). Data from MeDIP-Seq and bisulfite sequencing were then used to compare the methylation patterns of the H19 gene between domesticated pigs and boars. The possible role of the H19 gene during pig domestication and breed divergence was systematically described on the basis of molecular evolution analysis results, epigenetics differences, and gene expression data.

Evolution of H19 among mammals
The evolution of the H19 gene was first analyzed on a large phylogenetic scale to assess its evolutionary status. Basing from the H19 gene transcript sequence and mitochondrial sequence of 18 mammals, we constructed a gene tree and a phylogenetic tree by using MEGA6. The gene tree (Fig. 1a) was largely consistent with the phylogenetic tree (Fig. 1b). Results showed that primates and rodents cluster into independent clades and exhibited a close relationship than Sus scrofa.
Conservation scores were calculated using PhastCons [22] to prove that the H19 gene was conserved during large-scale evolution and to identify the conserved and accelerated evolutionary regions in the gene among the 18 species. Results showed that the upstream region of the H19 gene was lowly conserved whereas the gene body, specifically exon1, was highly conserved (Fig. 2a, Additional file 1: Figures S1A and S2A). PhyloP was used to calculate the p-value of conservation or acceleration, and results indicated that the H19 gene was conserved during evolution within a broad range (Fig. 2b, Additional file 1: Figures S1B and S2B). The evolution of H19 on a large phylogenetic scale indicates that the gene was highly conserved within abroad range, including the 3' region of H19.  [43]. The bootstrap consensus tree inferred from 1000 replicates [43,44,54,55] is taken to represent the evolutionary history of the taxa analyzed. Branches corresponding to partitions reproduced in less than 50 % bootstrap replicates are collapsed. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown next to the branches

Screening for selective signal
We analyzed the published re-sequencing data [24] on H19 to detect a selective sweep of the H19 gene during domestication on a genomic scale. We identified eight alleles among the domestic pigs, wild boars, and Tibetan boars. Although all of the eight alleles were detected in the domestic pigs, a large portion of these alleles were homozygous in pig individuals (Additional file 1: Tables  S2 and S3). This high degree of homozygosity implies that the gene is associated with pig breeding [25]. We further calculated the π and pairwise F st of the linked upstream, gene body, and downstream genomic regions of the H19 gene in 30 Tibetan wild boars, 15 domesticated pigs, and 3 wild boars from Sichuan with 1 kb sliding windows and 300 bp steps ( Fig. 3 and Additional file 1: Figure S3). The π value within the H19 gene body was significantly higher in domestic pigs than in wild boars ( Fig. 3a and c). This result suggests that the H19 gene sequence conservation within wild boars is sharply contrasted with the sequence diversity among the H19 alleles found within domestic pigs. Within the domesticated pig population, the upstream and downstream regions of the H19 gene showed a low diversity, which is similar to that in wild populations (Fig. 3b). By contrast, the H19 gene body exhibited a high diversity (Student's t-test, P < < 0.05, Fig. 3a). This phenomenon may be caused by selection relaxation. However, considering the abundant occurrence of breed differentiation within domesticated pigs, we suspected that H19 might have undergone genetic differentiation during breed differentiation and pig domestication. We also detected that the F st of the H19 gene body between domesticated pigs and wild boars was higher than that of H19 downstream (Student's t-test, P < < 0.05, Fig. 3c).
To detect whether or not H19 is under selection, we obtained the whole-genome nucleotide divergence (π dome_G = 0.0025, π wild_G = 0.0030) of domesticated and wild pigs, as well as the population difference (Fst d_w_G = 0.501) between domesticated and wild pigs, by using Tibetan re-sequencing data with GATK software [26]. Comparison of the sliding windows of π value and F st between domesticated and wild pigs ( Fig. 3c and Additional file 1: Figure S3), showed that the domesticated pigs have a significantly higher diversity than wild pigs in the H19 gene (Z-test, P = 0.0128). This results suggests that H19 in domestic pig breeds might have experienced a more rapid evolution than that in wild ones. The F st between these two populations in H19 (F st.H19 = 0.393) had no significant difference with the average F st value of the whole genome (F st.whole = 0.501) (Fig. 3a) [27,28]. However, the drastic diversity pattern between domestic and wild pigs in the H19 gene body, which is a highly conserved region during large-scale evolution, suggests that this gene might have played roles in the breed differentiation of domestic pigs.

Nucleotide diversity
We obtained all 2608 bp sequences in the H19 gene, including five exons and four introns from the published re-sequencing data of Tibetan wild boars [24], to identify the polymorphism sites in the gene. Results showed that the 3'-region of the H19 gene has abundant polymorphisms. Partial sequences of the H19 gene from 1660 bp to 2460 bp (the transcription start site is 1 bp) were obtained from 234 pig individuals consisting of China native pigs and three foreign breeds to detect the roles  Table 1, Additional file 1: Tables S2, S3 and Figure S4. The SNP statistics table revealed eight SNPs in the region; of these SNPs, seven were observed only in domesticated pigs and one was present in both domesticated and wild pigs. However, clear differentiation of the genotype of this SNP site was observed. The proportion of the T allele was higher than or equal to that of the C allele in wild pigs, but nearly all T alleles were observed in the Tibetan wild boars. This result indicates that this site underwent differentiation in the wild pigs.
Among the Chinese breeds, a number of populations were C-alleles homozygous, and other populations showed a particular frequency of T-alleles. At the population level, the SNPs were present in the heterozygous state, indicating that these sites have not been completely fixed. Interestingly, the SNP at 1818 bp exhibited a high mutation proportion in the CC and SC populations but was nearly completely homozygous in the other populations.
Estimates of nucleotide diversity (π) were consistently higher in pigs from CC (π = 0.00127) and SC (π = 0.00203) than in pigs from other locations, especially 1664  Patterns of nucleotide diversity and pair-wise population differentiation of H19 locus in domesticated pigs and wild boars. Sliding window analysis of nucleotide diversity and genetic differentiation coefficient between groups at H19 for domesticated pigs (Dome) and wild boars (Wild). π (right) and F st (left) was calculated for segments of 1000 bp with 300 bp intervals (c). The distribution of nucleotide diversity (π) and F st within H19 gene (a) and linked upstream and downstream 6 kb (b) was shown in the boxplot, the middle thick black line represent median.θ π ratio was calculated by π .dome /π .wild , the red dash line and green dash line respectively represent the average θ π ratio and the average Fst of the whole genome of Susscrofa Tibetan wild boars (π = 0.00071) (Additional file 1: Table  S4). The same findings were observed for θ and θπ (Additional file 1: Table S4). Such results suggest that genetic diversity increased during Chinese pig differentiation, although no significant selection signal was detected through Tajima's D test [29], Fu and Li's H and D test [30] or Fay and Wu's H test [31] (Additional file 1: Table S4). The results of HKA test were also not significant (P Dome = 0.8796). We constructed a neighborjoining tree by using the net distance between populations and found that CC and SC clustered together (Fig. 4). A considerable degree of mixture possibly decreased the selection signal.
Methylation analysis of the H19 gene between wild and domesticated pigs We selected the DNA immunoprecipitation (MeDIP) data of one tissue (longissimus dorsi muscle) of every pig sample and conducted Pearson correlation test to test whether or not a significant difference exists between the mapping results of the whole genome and the H19 gene. Pearson correlation coefficient ranged from 0.949 to 1. This result indicates that the mapping results are highly correlated on the basis of the Pearson correlation coefficient. These results suggest that the result of whole genome mapping was consistent with that of H19 gene mapping after validation of the selected representative data. The methylation levels of the H19 gene body and linked upstream and downstream genomic regions were analyzed using the MeDIP data of eight fat tissues and two muscle tissues of three Tibetan pigs (the Plateau pig breed from the Tibetan wild boar lineage), Landrace (foreign pig breed), and Rongchang pig (Chinese native pig breed) to understand the methylation pattern of the H19 gene between wild and domesticated pigs ( Fig. 5 and Additional file 1: Figure S5). Results showed that the methylation levels within the H19 gene among the three above mentioned pig breeds were lower than 6 kb upstream (Student's t-test, P < <0.01) and 6 kb downstream (Student's t-test, P < <0.01), whereas the methylation level was the highest in the 6 kb upstream region. Several tissues, including the greater omentum (GOM) and mesentery adipose (MAD), from the Rongchang pig showed significantly higher gene body methylation levels than the corresponding tissues in Landrace (Student's t-test, P < <0.01) and Tibetan pigs (Student's t-test, P < <0.01). Several tissues showed distinct levels of methylation. In Landrace, the methylation levels at both 6 kb upstream and 6 kb downstream regions in the abdominal subcutaneous adipose tissue (ASA), psoas major muscle (PMM), and retroperitoneal adipose tissue (RAD) were lower than those in other tissues, whereas the methylation levels in the upper lipid of back (ULB) were higher than those in other tissues (two-way ANOVA, P tissue = 4.48 × 10 -6 ). In the Chinese native pig (Rongchang pig), the methylation levels were obviously higher in the GOM and MAD but lower in the ASA, longissimus dorsi muscle (LDM), and PMM in all three regions (two-way ANOVA, P tissue = 0.0308). Interestingly, the methylation levels at the upstream and downstream regions were clearly lower in the inner lipid of back (ILB) than in the gene body and other tissues of Tibetan pigs (two-way ANOVA, P genebodyupanddown = 2.02 × 10 -12 , P ILB-others = 1.36 × 10 -7 ). The methylation patterns in ASA and PMM differed between Chinese native pigs and Tibetan pigs (Student's t-test, P = 0.067 to 1.962 × 10 -5 ) (Fig. 5).
The methylation levels of the H19 gene and linked 2 kb upstream and downstream regions, as well as the expression levels in the liver, were determined and compared on the basis of our unpublished single-base liver methylome and transcriptome data of the wild boar (Guizhou wild pig), the Chinese native pig (Enshi black pig), and a foreign breed (Landrace) (Fig. 6, Additional file 2: S7). The methylation levels of the gene were the lowest in the foreign pig (Landrace), followed by that in the Chinese wild pig (Guizhou wild boar). The methylation levels in the Chinese native pig (Enshi pig) were the highest at the 2 kb upstream region of the H19 gene (Fig. 6a-c and f ). By contrast, the methylation levels in the gene body of Landrace were higher than those in the two other samples (pairwise comparisons Fig. 4 The Neighbor-joining tree of H19 in different pig breeds. The neighbor-joining tree of H19 base on the net distance between breeds was constructed according to catalogs (a) and pig breeds (b) by Mega 6.0 using t tests, **P < 0.001 with the adjusted Bonferroni method). The expression levels of the H19 gene were the highest in the liver of foreign pigs (Landrace), followed by those in the Chinese wild pig and Chinese native pig (Enshi pig) (one-way ANOVA, P < <0.01) (Fig. 6g). These results are consistent with the theory that methylation in the promoter region decreases expression levels and represses transcription noise in the gene body [32].
We further explored the relation between imprinting and methylation difference. In this step, we identified the DMSs among pig breeds and the ASM region by using our unpublished single-base resolution methylome data of the liver from the three pig breeds. MeDIP data were not in single base resolution; thus, we were unable to use the data to address tissue-specific methylated sites. We previously surveyed the methylation status of CpGs of the H19 gene in the three breeds. In the present study, we first identified the DMSs among the three pig breeds by using Chi-square test. We found a total of 6, 9, and 12 DMSs between the Enshi pig and the Guizhou wild pig, the Guizhou wild pig and Landrace, and the Enshi pig and Landrace, respectively, in which 3, 6, and 9 sites were in the upstream 2 kb region; 2, 2, and 3 sites were in the gene body region; and 1, 1 and 0 sites were in the downstream 2 kb region, respectively. This result indicates that DMSs were mainly located in the upstream 2 kb region (Additional file 1: Table S5). We also identified regions with ASM in each breed. We identified 8 ASM regions, all of which were located in the and LB (Landrace, c) and the expression level of H19 gene (g). The methylation level of every site was calculated by number of methylated reads divide number of total reads which was mapping to this site. The expression level was estimated by FPKM value which was calculated by Cuffdiff software using transcriptome data. The grey horizontal lines below the a, b, c were the ASM region, the vertical bars in a, b, c were DMS, the different color represent DMS between this pig breed and the other referred by the related color. (***P < 0.001;**P < 0.02;*P < 0.05) upstream 2 kb region of the H19 gene. Interestingly, all of these ASM regions were in the Enshi pig. In the 8 ASM regions, no SNP was found in the two other pig breeds; hence, we were unable to distinguish the allelic methylation status in these two breeds. We then used the Enshi pig as a model to explore the DMSs and ASM of H19. Most of the DMSs between the Enshi pig and the other pigs were located in the ASM region; specifically, 2 of 3 DMSs between the Enshi pig and the Guizhou wild pig and 7 of 9 DMSs between the Enshi pig and Landrace were situated in the ASM region. Our preliminary results imply that imprinting may be associated with methylation differences among different breeds.

Discussion
Increasing amounts of genomic sequence data and highthroughput annotation of genes have enabled the bioinformatics analysis of various genes of interest; this analysis has consequently provided important insights into the evolutionary link between genes of interest and particular phenotypic traits associated with several diseases [33]. As one of the first long non-coding imprinting genes to be described, H19 reportedly promotes skeletal muscle development and regeneration [14]. Epigenetic studies have reported that abnormal methylation of the H19 gene may induce BWS syndrome [34] and influence fetal growth [35]. Thus, the H19 gene is hypothesized to serve an important function in the evolution and breed differentiation of pigs.
To test this hypothesis, molecular evolution was first analyzed on a large scale by using the homologous H19 sequence of 18 species. The H19 gene was highly conserved in mammals. Relative to the upstream and downstream regions, the H19 gene body was highly conserved, especially in the 5'-region of the gene body. The functional importance of the H19 gene was confirmed by the relatively large-scale evolutionary perspective observed. Molecular evolution analyses indicated a drastic divergence of diversity pattern between domestic and wild pigs. This finding is interesting because this region was highly conserved during large-scale evolution, suggesting that this region may have been under a strong evolutionary constraint. Therefore, we suspected that the obviously increased diversity within the domestic pig population might be associated with breeding differentiation. Further analyses within Chinese domestic breeds showed that several SNP genotypes within the H19 gene clearly differentiated among the different pig breeds, with the SC and CC types showing distinct genotypes compared with the others (Additional file 1: Table S3, Fig. 4). Although we were unable to associate their genotypes with their specific meat characters, our results provide a basis for further functional efforts on the alleles we identified. Over all, the present results demonstrate that the H19 gene serve a critical function in pig breed differentiation and local adaptation. Further functional significances on the different genotypes of this gene are needed to confirm.
H19 is an imprinted gene. Only its maternal allele is expressed, and paternal alleles are imprinted in normal individuals. Abnormal methylation of the H19 DMR located upstream of the H19 transcript start site induces abnormality and diseases [36][37][38]. Therefore, the potential epigenetic impact of H19 on pig domestication is an interesting issue. Maximizing both the published MEDIP data and the single-base bisulfite methylation data, we can obtain a general view of the epigenetic pattern on pigs. We found that the methylation levels within the gene body were relatively low and that tissue-specific methylation differences existed among different samples. Furthermore, we obtained a precise pattern of H19 and successfully detected differential methylation associated with the differential expression of H19 on the basis of our unpublished single-base resolution methylome and transcriptome data of livers from three representative pig samples. Chinese local breeds (ES) showed the highest upstream methylation but the lowest downstream methylation. Accordingly, H19 expression in ES breeds drastically decreased compared with that in the wild boar and European breeds (LB). Promoter methylation in mammals and plants reportedly represses gene expression. The function of gene body methylation is intriguing, especially because gene body methylation may contribute to alternative gene expression [39,40]. During Chinese pig domestication, a hyper-methylation of H19 occurred in the liver from its wild ancestor, which affected the repression of this gene. Borensztein [13] found that Myod as a muscle source adjunction factor interacts with the H19/IGF2 cluster to regulate myofiber differentiation and inferred that H19 participates in growth and muscle development [13]. A recent study has also concluded that H19 plays a crucial role in regulating skeletal muscle differentiation and generation mediated by miR-675-3p and miR-675-5p, which are encoded within H19 by knockdown of H19 in myoblast cells and knockout of H19 in mouse satellite cells, respectively [14]. Thus, the inhibition of H19 expression can influence skeletal muscle differentiation. Interestingly, epigenetic regulation on H19 in European pig breeds is another case. Obvious de-methylation or hypomethylation occurred in the promoter region in this breed compared with the two other breeds. Specifically, DMSs in the breed differential methylation were also identified. The methylation status of the H19 gene in its wild ancestor, the European wild boar, remains unknown; nevertheless, the present results clearly indicated a distinctively high expression of this gene in European pig breeds and thus suggested an opposite epigenetic regulation direction between Chinese and European pig domestication. ASM plays a crucial role in epigenetic phenomenon and parental imprinting [41]. In this study, consistent to the differential methylation sites between the Enshi pig and the other breeds, all of these ASM regions were identified in the 2 kb upstream region of the H19 gene in the Enshi pig. Our preliminary results implied that imprinting may be associated with methylation differences of H19 genes between pig breeds. Both the differential methylation sites and ASM in the upstream region may have important impact on gene expression and further functions. Whether or not imprinting participated in epigenetic evolution during pig domestication and breed differentiation remains to be discusses in further research.

Conclusion
The evolution patterns of the H19 gene during pig domestication and breed differentiation were systematically analyzed at the genetic to epigenetic levels. The H19 gene was highly conserved during large-scale evolution and exhibited genotype differentiation during domestication and breed differentiation, especially in the CC and SC populations. Epigenetic investigation showed that the H19 gene exhibited a specific methylation pattern and suggested an opposite epigenetic regulation direction between Chinese and European pig domestication. This opposite epigenetic regulation direction resulted in opposite expression changes of this gene between the two domesticated groups. Most DMSs in the upstream 2 kb region were also located in the ASM region, which implied a great relevance between imprinting and differential methylation. Our results demonstrated the genetic and epigenetic patterns of H19 during pig domestication and provided valuable cues and basis for further research on the function of H19 in pig domestication.

Ethics
Animals care and all the experimentation in this study were carried out in accordance with the pre-approved guidelines from Regulation Proclamation No.5 of the Standing Committee of Hubei People's Congress. All experimental procedures and sample collection methods were approved by the Institutional Animal Care and Use Committee of Huazhong Agricultural University, Wuhan, China (permit HZAUSW2015-0003). We obtained a written informed consent from the animal owners to take blood and tissue samples.
Identification of conserved region of H19 between 18 species musculus) were downloaded from NCBI. The gene tree was constructed by using the transcripts of 18 animals and the phylogenetic tree was constructed by using mitochondrial sequences of 18 animals. All the sequences was aligned by MEGA6.0 and constructed gene tree and phylogenetic tree using maximum likelihood method by MEGA6.0 respectively. PHAST software was used to calculate the conservative score of the H19 gene among the 18 species based on the theory that mtDNA assumes neutral selection between species. Initially, phyloFit was used to fit tree models (constructed using maximum likelihood by MEGA6.0 based on the mtDNA sequences) to achieve multiple alignment of the H19 sequences via maximum likelihood. Then, phastCons was used to generate conservation scores of the H19 gene sequences between the 18 species listed above. Finally, phyloP was used to compute conservation or acceleration p-value based on the alignment of the H19 gene sequences and a model of neutral evolution (http://compgen.bscb.cornell.edu/ phast/) [22]. Results were analyzed using R studio.

Sample collection
A list of samples used is shown in Additional file 1: Table S2. All domesticated pig breeds were collected from the local laboratory. All blood samples from domesticated pigs were collected from local farmers to represent local species. The total genomic DNA of Duroc, Landrace, and Large White pigs were extracted from tissue samples according to a standard phenol-chloroform protocol. The extracted DNA was then used to analyze the genetic diversity of H19 of the different pig breeds.

Cloning of the H19 gene
Five primer pairs, shown in Additional file 1: Table S1, were designed to amplify the H19 gene from the first exon to the last exon based on the sequence published in NCBI (GenBank: AY044827) using Primer Premier 5.0 software. Each PCR contained a final volume of 25 μL and consisted of the following components: 125 ng genomic DNA, 0.1 μM each primer, 5 μL 5 × Pri-meSTAR GXL buffer (5 mM Mg 2+ plus), 4 μL dNTP Mixture (2.5 mM each) and 0.5 U GXL Taq DNA polymerase (Takara, China). PCR conditions was set as follows: initial denaturation at 94°C for 5 min; 38 cycles of 10 s of denaturation at 98°C, 15 s of annealing at the appropriate temperatures listed in Additional file 1: Table S1, and 1 min extension at 72°C; and a final 5 min extension at 68°C. Amplified products were detected via gel electrophoresis, and the positive products were sent to BGI Tech for sequencing. All sequences were assembled using the SeqMan program in the DNASTAR software suite.

Molecular evolution analysis
First, DNA sequences were aligned using MUSCLE by MEGA6 [42] and then adjusted manually. MEGA6 was also used to infer phylogenetic relationships based on maximum likelihood method (ML). The ML tree was constructed by maximum likelihood method with Kimura 2-parameter model [43]. Reconstruction of ML tree are generated by the software MEGA 6, with bootstrap values set at 1000 times.
The SeqMan program was used to calculate the genetic distance between populations and count the SNP information of the H19 gene.
The nucleotide diversity of the H19 gene was analyzed by calculating the number of segregating sites (S), nucleotide diversity per site (π) calculated for whole sequences (π total ), Watterson's theta estimator (θw), and haplotype diversity (Hd) using DnaSP version 5.0 excluding insertion/deletions (indels).
To examine deviations from the neutral evolution model and reveal the history of evolution and divergence between wild boar and domesticated pigs, Tajima's D test, Fu and Li's D* test, Fu and Li's F* test, Fay and Wu's H* and HKA test were performed using DnaSP version 5.0 [44].
To observe variance in polymorphisms along the 14.6 kb H19 gene region, including the 6 kb upstream and 6 kb downstream regions of the H19 gene, the π and F st values of H19 gene were analyzed via sliding window with segments of 1 kb and 300 bp intervals. Besides, in order to test whether there have selection signals, the π and F st of whole genome was calculated by sliding window with segments of 100 kb and 10 kb intervals by re-sequencing data. Then use the average π and F st of whole genome as the sample mean to test the potential selection region using Z-test.

Re-sequencing data
To obtain additional samples for the Tibetan wild boar population and improve evolutionary analysis, resequencing data of Tibetan wild boars (SRA065461) were collected according to the literature published in NCBI (ftp://ftp-trace.ncbi.nlm.nih.gov/sra/sra-instant/ reads/ByStudy/sra/SRP/SRP018/SRP018123/) which included 30 Tibetan wild boars, 3 wild pigs, 15 Chinese local pigs and used to analyze nucleotide diversity and F st statistics along the upstream and downstream regions of the H19 gene.

Single-base resolution methylome data
Three unpublished sets of methylome data from three pig livers were used to detect the methylation pattern of the H19 gene (the fastq file used in this article have uploaded in NCBI, the SRA accession is SRP070531).

RNA-seq data
Unpublished transcriptome data of three pig livers were used to detect the expression level of the H19 gene in liver (the fastq file used in this article have uploaded in NCBI, the SRA accession is SRP070533).

Re-sequencing data analysis
Using SRA toolkit software (version 2.3.4), *.sra was transferred to *.fastq with -I -split-files parameters. Clean data were obtained by removing low quality bases, repeats, and other impurity sequences using a Perl script (Additional file 3: S9). Processed sequences were mapped to the 2.6 kb-long H19 gene via SOAPaligner using default parameters, and SNPs were called via SOAPsnp using the -M parameter with other parameters held at their default setting. Results were used to analyze the evolution and genetic diversity of the H19 gene by combining the sequence data of native pig and European breeds previously sequenced [45].
Clean paired-end re-sequencing data were mapped to the upstream and downstream regions of the H19 gene, which is 56 kb-long, using Bowtie2 with default parameters (with -U parameters, other parameters all are default parameters) [46]. SNP calling was then performed using SAMtools and bcftools [47] with default parameters to obtain the SNPs of the H19 gene in Tibetan wild boars, wild boars, and domesticated boars. Results were used to determine whether significant divergence in the H19 gene occurred during the course of evolution and differentiation.
Clean paired-end re-sequencing data were mapped to pig genome (Sscrofa10.2) using bwa software with default parameters and called SNP using GATK software with default parameters, then based on the SNP information to calculate the average π and F st of whole genome.

MeDIP data analysis
Similar to process the re-sequencing data, we obtain the cleaned data and then mapping the MeDIP data to the H19 gene including the upstream and downstream of it by soapaligner using the default parameters. The number of reads was mapped to the CpG site and designated as the methylation level. Besides, in order to excluding the repeats and duplications, Susscrofa genome version 10.2 and H19 was merged as the reference to be mapped by MeDIP data using soapaligner and then extract the results which mapped to H19.

Single-base resolution methylome data analysis
Low quality and repeat sequences were removed using a Perl Script to obtain clean single-base resolution methylome data. The transfer efficiency of C to T was then detected using the mitochondrial genome as reference genome by Bismark Bisulfite Mapper-v0.10.1 [48]. The clean data were mapped to the 56 kb-long H19 gene (GenBank: AY044827) by Bismark Bisulfite Mapper-v0.10.1 using the default parameters and methylation statues of every CpG site was generated in Additional file 4: S1, Additional file 5: S2 and Additional file 6: S3 (the sam files which were produced in the process of mapping can be found in the Additional file 7: S8).

RNA-seq data analysis
Low-quality and repeat sequences were removed using a Perl script to obtain clean single-base resolution methylome data. Tophat software [49] was used to map the RNA-seq data to the H19 gene (GenBank:AY044827), which was used as the reference sequence, with default parameters (the sam files which were produced in the process of mapping can be found in the Additional file 7: S8) [50]. Cufflinks software was used to remove redundancy with default parameters [51] and transcripts of the H19 gene were then obtained, the results was shown in Additional file 8: S4, Additional file 9: S5 and Additional file 10: S6. FPKM (fragments per kilobase of exon model per million mapped reads) values were calculated as expression levels [52].

Comparison of methylation level
Methylation levels between wild boars and domesticated pigs were analyzed using the single-base resolution liver methylome of Landrace, Enshi black pig, and a wild pig from our laboratory (unpublished). Initially, the complete H19 gene sequence (downloaded from NCBI, AY044827.1) was indexed as the reference sequence and converted using Bismark Genome Preparation in Bismark version 0.10.1 software to allow alignment. Bowtie version2 combined with Bismark was then used to align the single baseresolution methylome sequences to the converted and indexed genome and subsequently generate the SAM format file using default parameters. Finally, methylation levels and information of each CpG site were extracted using the Bismark methylation extractor with default parameters. The methylation level of each site was calculated using the number of methylation reads of a particular site divided by the number of total reads mapped to this site. Differential methylation site between each two different pig breeds were identified by Chi-square test. Identification of Allelic specific methylation (ASM) region to identified allelic methylation status, we firstly determined SNPs with a Bayesian algorithm using SOAPsnp by aligned files which were generated by Bismark and Bowtie. All heterozygous SNPs with only two genotypes were collected and BS-seq read uniquely covering a heterozygous SNP were divided into two groups based on their assigned genotypes. We selected regions exhibiting allele-specific methylation (ASM) using the following criteria: the windows which fold change reached twice and the absolute difference methylation level reached 0.3 using five CG sites as sliding window, then overlapped the selected windows and performed t-test to select the allelic specific methylation (ASM) region [53]. MeDIP data [GenBank: SRP007462] from 10 tissues of Rongchang, Landrace, and Tibetan wild boars published in NCBI were downloaded. After cleaning, SOAPaligner and SOAPsnp were used to align the data to the complete H19 reference gene sequence and extract the mapping reads of every site. The number of reads was defined with respect to the methylation level of each site. In order to identify the difference of methylation analysis, ANOVA or pairwise comparisons using t tests was employed for comparisons of more than two groups, Student's t-test was used for two group comparisons. Resulting P values of pairwise comparisons using t tests were corrected with adjusted Bonferroni method.