Individual and population diversity of 20 representative olfactory receptor genes in pigs

Understanding the influence of genetic variations in olfactory receptor (OR) genes on the olfaction-influenced phenotypes such as behaviors, reproduction, and feeding is important in animal biology. However, our understanding of the complexity of the OR subgenome is limited. In this study, we analyzed 1120 typing results of 20 representative OR genes belonging to 13 OR families on 14 pig chromosomes from 56 individuals belonging to seven different breeds using a sequence-based OR typing method. We showed that the presence of copy number variations, conservation of locus-specific diversity, abundance of breed-specific alleles, presence of a loss-of-function allele, and low-level purifying selection in pig OR genes could be common characteristics of OR genes in mammals. The observed nucleotide sequence diversity of pig ORs was higher than that of dogs. To the best of our knowledge, this is the first report on the individual- or population-level characterization of a large number of OR family genes in livestock species.

Olfactory receptors (ORs) are members of the seven transmembrane G protein coupled receptor (GPCR) superfamily and are expressed mainly on olfactory neurons 1 .ORs bind to volatile odorants present in the air and transmit olfactory signals from bound odorants to the brain 2,3 .However, gene expression analysis of ORs in non-neuronal tissues, including testis, lung, intestine, skin, heart, and blood indicate their additional functions in modulating cell-cell recognition, cell migration, proliferation, pathfinding, and secretion [4][5][6][7][8][9][10] .
The sense of smell integrates a plurality of olfactory signals created by the binding of odorants to ORs; thus, a large degree of polymorphism is necessary to distinguish various olfactory cues by OR molecules 11,12 .OR genes constitute the largest gene family in mammalian genomes, and the size of the OR repertoire for intact OR genes in terrestrial tetrapods varies from < 400 in primates to 4267 in African elephants 13 .In addition, the percentage of OR pseudogenes among all OR genes in the genome have increased significantly in certain species favoring a reduction in the OR repertoire.For example, whales show a significant increase in OR pseudogenes compared to their terrestrial counterparts [14][15][16] .
OR genes are traditionally divided into two classes, class I and class II, based on amino acid sequence homology 17,18 .Class II OR genes have a much larger number of genes than class I genes and form clusters on multiple mammalian chromosomes.In contrast, class I OR genes commonly exist as a single cluster on chromosomes in mammalian species [19][20][21][22] .However, detailed characteristics of the functions and phylogenetic relationships of each class of ORs have been studied in a limited number of species [23][24][25] .
In addition, the allelic diversity of OR genes in an individual or population can influence their olfactory specificity and sensitivity by modulating the perception of odorants.A recent study showed that genetic variation in a single OR is frequently associated with changes in odorant perception in humans 26 .Another study showed that the allelic distribution of OR genes varies among different dog breeds [27][28][29] .However, the extent of allelic diversity of OR genes and the effects on phenotypic differences are not yet clearly understood due to the difficulty in identifying associations between genetic variations in a large number of OR genes and olfactionrelated phenotypes.In addition, the large genetic variations in OR genes, including pseudogenization, nucleotide substitutions, indels, and copy number variations (CNVs), make it challenging to evaluate the genome-scale allelic diversity of OR genes in an individual genome.
No studies on allelic diversity and population variation of OR genes in livestock animals have been reported to date.Understanding the genetic variation in the OR repertoires of livestock species may reveal interesting aspects of ORs in animal behavior, reproduction, and feeding.For example, the stereotypical mating posture of an estrus female pig exposed to a compound in the saliva of boars is also mediated by the olfactory system 30 .
The OR repertoire of pigs (Sus scrofa) consists of 1301 OR genes, including 1113 functional and 188 pseudogenes 22,31 , which is much higher than the OR repertoire of cattle (970), which is another artiodactyl, and dogs (811), who are known for their excellent olfactory abilities 15,32 .Pigs are attractive animal models for studying olfaction due to their agricultural importance and strong reliance on their sense of smell in various behavioral contexts 33 .Approximately 200 to 300 breeds of pigs and wild boars exist worldwide 34,35 .Studies on OR diversity in diverse pig breeds provide important insights into how OR gene diversity is maintained and differentiated in populations.
A previous study showed that OR genes in the pig genome exist in 46 clusters scattered across 16 chromosomes and can be classified to 17 gene families 22 .In this study, we developed primer sets allowing the specific amplification of 20 pig OR genes covering 13 OR families on 14 pig chromosomes which include a large part of pig OR gene families.We compared the allelic diversity of each OR gene across 56 animals from seven pig breeds.Despite a limited population size for each breed, our results showed a significant difference in allelic constitution and diversity among different breeds.

Successful development of a sequence-based typing (PCR-SBT) of 20 pig OR genes
To understand the genetic characteristics of OR genes in pigs, we selected 20 pig OR genes belonging to 13 OR families (1, 2, 5, 6, 8, 9, 10, 11, 13, 14, 51, 52, and 53) located on 14 chromosomes (chr1, 2, 4, 5, 6, 7, 9, 10, 12, 13, 14, 15, 18, and X) according to the OR classification in our previous study 22 and the available sequence information in the pig genome assembly Sscrofa11.1 (NCBI; S1 Table ).To construct a representative list of OR genes for the pig genome, we compared the previously identified OR genes based on the pig genome assembly Sscrofa10.2from a prior study with those from the current assembly, Sscorfa11.1.The majority of identified OR genes were consistent between the two assemblies, with some minor discrepancies.Notably, the comparison revealed the presence of sOR5P1 exclusively in Sscrofa10.2,while sOR8F2P, a pseudogene, initially located in the telomeric cluster of chr10 in Sscrofa10.2,had been relocated to unplaced scaffold NW_018085136.1 in Sscrofa11.1 (S1 Table ).
To unbiasedly assess the genetic diversity of pig OR genes, we aimed to encompass sequences from at least one OR gene within each OR gene family.However, we were unable to include sequences of four OR families, 3, 4, 7, and 12, due to failure in locus-specific amplification.A pseudogene, sOR8F2P, was incorporated to address the influence of loss of function on the genetic diversity of OR genes.Consequently, a total of 20 OR genes representing 13 OR gene families across 14 pig chromosomes were selected, forming the basis for evaluating the allelic diversity of pig OR genes.Furthermore, an assessment of the expression of these selected OR genes was conducted utilizing publicly accessible data (NCBI Gene Expression Omnibus repository, accession number GSE197184; available at https:// www.ncbi.nlm.nih.gov/ geo/).The study revealed the expression of 18 out of the 20 chosen OR genes in the pig olfactory epithelium.Notably, sOR5P1 and sOR8F2P were exceptions, as they were absent from the OR gene annotation of Sscrofa 11.1 (S2 Table ).
Locus-specific primers were designed for each of the 20 OR genes (S3 Table ), and PCR was conducted using the genomic DNA of 56 pigs from seven breeds (eight individuals each) including Berkshire (BER), Duroc (DUR), Korean Native (KNP), Landrace (LAN), National Institutes of Health miniature (NIH), Seoul National University miniature (SNU), and Yorkshire pigs (YOR) for each OR.PCR amplicons were obtained and sequenced to confirm target specificity.PCR primers were reiteratively designed to achieve locus-specific amplification of the target OR genes from all animals.Allelic determination of the 1120 typing results was conducted by blasting the sequencing results of amplicons to the OR reference sequences and pig reference genome.After cloning, a minimum of three colonies were independently analyzed to confirm a new allele for heterozygotes and to separate each allele.Finally, we obtained successful typing results for the 20 pig OR genes at the sequence level, and the sequences were submitted to GenBank (accession numbers OQ379512 -OQ379644) (S4 and S5 Tables).

Confirmation of copy number variation for OR53A1
Our typing results showed that OR sequences from 18 out of 20 OR genes were mapped to a single chromosomal position, but sOR6Y2 and sOR53A1 were mapped to two different positions: S. scrofa chromosome (SSC) 14 90982296-90983234 (expected) and 91028428-91029366 (additional) and SSC9 3549170-3550117 (expected) and 3858315-3859262 (additional) bp regions, respectively (S3 Table ,  S1 Fig).This finding indicates the possible presence of CNV or gene duplication in the two OR genes in the pig genome.To experimentally validate the results, we designed additional primers specific to each of the differently mapped sOR6Y2 and sOR53A1 sequences based on identified sequence variations between the OR sequences mapped to two different regions and analyzed differences in amplification specificity in our sample panel (29 for sOR53A1 and 20 for sOR6Y2; S2 Fig) .For sOR6Y2, only the SSC14 90982296-90983234 bp (expected) region-specific sequences were amplified for all samples, but none of the animals possessed the 91028428-91029366 bp (additional) region-specific sequence.In contrast, 18 out of 29 animals (38%) showed a 3858315-3859262 bp region-specific additional sequence for sOR53A1 (S2 Fig) .Therefore, we experimentally confirmed the presence of CNV in sOR53A1.Subsequent RT-qPCR specific to all sOR53A1-like sequences residing at both SSC9 3549170-3550117 and 3858315-3859262 in our sample panel showed the presence of CNVs in SNU and LAN individuals (S6 Table, S2 Fig) .The estimated copy number ranged from two to five.The increase in the sOR53A1 copy number was associated with specific alleles, including sOR53A1*02 and *05 in SNU and sOR53A1*05 in LAN (S6 Table ).We considered all sOR53A1like sequences as allelic forms because we were unable to differentiate locus-specific differences among the sOR53A1 amplicons. www.nature.com/scientificreports/

Allelic diversity of 20 pig OR genes and the presence of nonfunctional alleles
The typing results corresponding to the 20 pig OR coding sequences (CDS) ranging from 912 to 1038 bp (304-346 amino acids) in length from 56 animals belonging to seven pig breeds resulted in the identification of 351 single nucleotide polymorphisms (SNPs) (S7 Table ).The number of SNPs for each OR gene ranged between 1 and 75, with a mean of 17.55.We assigned putative allele names for each OR allele in which two-digit numbers were given as allele names and separated with the "*" symbol after the locus names, as in sOR10L1*07, shown above.The allele with the highest frequency was assigned to *01, and additional numbers were assigned in the decreasing order of allele frequencies.Among these variations, a nucleotide deletion in the transmembrane 3 region of sOR10L1*07 causes premature termination in LAN pigs (S5 Table , S3  A significant difference was observed in the number of alleles for the OR genes, ranging from two (sOR5J1 and sOR8F2P) to 15 (sOR6T3; Fig. 1).However, allelic diversity is not always proportional to the number of SNPs observed in each OR gene.For example, the number of alleles for OR53A1 and sOR6T3 was 10 and 15, respectively, whereas the corresponding number of SNPs were 75 and 28, respectively (S7 Table ).This could be attributed to the abundant presence of rare mutations in an allele from the OR genes with a large number of alleles, such as sOR6T3.
The mean allele numbers (Na) of 20 OR genes at the nucleotide and amino acid sequence level were 6.65 and 5.68, respectively (Fig. 1, S7 Table ), indicating a minimum level of reduction in allelic numbers in protein sequences.Most OR genes showed minimal differences (≤ 3) in allele numbers between nucleotide and protein sequences, except for sOR1O1, in which only three protein sequences were coded from eight alleles.In the case of sOR1O1, the ratio of nonsynonymous to synonymous SNPs (nd/sd) was low (0.167) compared to that of other ORs (S7 Table ).The nd/sd ratio of the ORs ranged from 0 to 2.500, with a total mean of 1.014.The mean value of the nucleotide diversity (π) of the 20 pig OR genes was 0.0052, ranging from 0.0001 to 0.0216 (S7 Table ).Positive correlations between π and the cluster size, and between π and the number of OR genes in the cluster were observed, with correlation coefficients I of 0.52 and 0.42, respectively (Table 1).Allelic diversity of 20 OR genes was compared among the seven pig breeds (Table 2).The average value of the Na of the 20 OR genes of each of the seven breeds was 2.607, ranging from a minimum of 1.750 in NIH to a maximum of 3.500 in LAN.The comparison among the breeds showed that the value was much higher in commercial breeds (mean of 2.810), including BER, DUR, LAN, YOR, and KNP, than in inbred miniature pigs (mean of 2.100), such as NIH and SNU.Moreover, we observed the presence of breed-specific alleles in 18 of the 20 OR loci, except for sOR13B10 and sOR5J1 (S10 Table ).The largest number of breed-specific alleles (n = 3) was identified in sOR6T3 of NIH.The mean value of breed level observed heterozygosity (Ho) from the typing results of 20 ORs for the seven pig breeds was estimated to be 0.331, ranging between 0.036 and 0.643 (Table 2 and S7 Table), showing significant differences in Ho among different breeds.LAN showed the highest values for Ho (0.469), with an Na of 3.500 and a mean breed-specific allele number (Np) of 0.700.NIH showed the lowest Ho (0.181) and was homozygous for 12 out of 20 OR genes (60%).For DUR, KNP, SNU, and YOR, allelic diversity was similar.
The extent of locus-specific OR diversity was consistent across different breeds.For example, the allelic diversity of sOR5J1 was the lowest across all seven breeds among functional ORs, and similarly, that of sOR53A1 was the highest among all breeds typed (Fig. 1, S8 Table ).sOR5J1 and sOR5P1 were homozygous in most animals across all seven breeds, and sOR5J1 was fixed to a single allele at BER, DUR, KNP, NIH, and SNU (S8 and S11 Tables).This finding could suggest the characteristics of each OR gene regarding functional restriction or diversity in the detection of various odorants.Contrary to expectations, sOR8F2P, a pseudogene, showed fixation of an allele in most breeds, except SNU.In addition, we observed nine alleles (sOR5P1*01, sOR5J1*01, sOR13A6*01, sOR1A7*01, sOR53A1*02, sOR8Q4*01, sOR8F2P*01, sOR5AF1*01, and sOR2R3*01) common across all seven breeds (S8 and S9 Tables).The frequency of these common alleles was relatively high in all analyzed breeds, with frequencies of 0.88 (sOR5P1*01) and 0.54 (sOR2R3*01) in the breed combined analysis (S8 Table ).In addition, sOR10L1*07, a non-functional allele identified in this study, was observed in a LAN individual in the form of a heterozygote (S3 and S8 Tables).
A principal component analysis (PCA) was conducted using the typing results of 20 OR genes from all seven breeds (Fig. 2).The results using principal components 1 and 2 showed the clustering of the European breeds, including BER, DUR, YOR, and LAN, and the formation of separate clusters for NIH and SNU, respectively.www.nature.com/scientificreports/Moreover, KNP was somewhat separated from the cluster of European breeds but was closer to the European breeds than to the miniature pigs.

Higher inter-locus differences of class I OR over class II OR genes
To understand the genetic relationships of the 20 OR genes analyzed in our study, a unweighted pair group method with arithmetic mean (UPGMA) tree was constructed from the typing results of the 20 OR genes using nucleotide and protein sequences (Fig. 3).The combined analysis of three class I genes (sOR53A1, sOR51B1, and sOR52M1) located in a cluster on SSC9 and 17 class II OR genes (sOR13B10, sOR5J1, sOR5P1, sOR14A3, sOR10L1, sOR10B1, sOR13A6, sOR2R3, sOR11A7, sOR8Q4, sOR8F2P, sOR1O1, sOR5AF1, sOR6Y2, sOR6T3, sOR9K3, and sOR13K1) located on 14 different chromosomes formed two separate clusters.The OR clusters formed subclusters within each cluster, indicating that the 20 OR genes selected in this study represent a large genetic diversity of pig OR genes.The higher bootstrapping values (> 90%) of class I OR subfamilies relative to those of class II OR subfamilies in clustering analysis indicate clear inter-locus differences among class I OR genes as compared to class II OR genes.The trees constructed using nucleotide and amino acid sequences showed consistent results, but the clustering of OR genes in the branches with low bootstrapping values was less conclusive.
In addition, the mean nucleotide diversity of class I OR genes (π mean = 0.0132) was significantly higher (paired sample t-test, P = 0.000366) than that of class II OR genes (π mean = 0.0038; S7 Table ).This finding may indicate an evolutionary effect against the duplication of class I OR genes in mammals to maintain a relatively lower number of class I OR genes than class II ORs by limiting the functional redundancy of class I OR repertoires.However, when comparing the mean Ka/Ks ratio between class I and II OR genes with more than 10 SNPs (sOR5P1, sOR14A3, sOR10L1, sOR10B1, sOR2R3, sOR1O1, sOR5AF1, sOR6Y2, sOR6T3, sOR9D3, sOR13K1, sOR53A1, sOR51B1, and sOR52M1), most of them were under weak purifying selection and no significant difference was observed in the ratio between the two (Ka/Ks class I = 0.307, Ka/Ks class II = 0.316, Student's t-test, P = 0.935; S7 Table ).This result could be attributed to the relatively recent gene expansion in Class II OR genes compared to Class I OR genes.

Extent of genetic variation among different structural domains of pig OR genes
Pig OR genes consist of 15 structural domains, including N-terminal (Nterm), C-terminal (Cterm), seven transmembrane (TM1-7), three extracellular (EC1-3), and three intracellular domains (IC1-3) 3,36 .The domain structures of 133 different sequences from our typing results of 1120 OR amplicons were determined based on conserved sequences for each motif, and intra-locus variations were analyzed for each domain.Values of the number of variable sites, pairwise differences, and Ka/Ks ratio were estimated for each domain of the 20 OR genes (Table 3 Although the number of variable sites were somewhat different from 0 to 18 for the 15 OR domains, the normalized values for the length (in bp) of each domain were not significantly different.The percentage of variable sites for each domain ranged from 1 to 3% of total sites, with an average of 1.93%, whereas TM5 and TM6 were slightly higher (2.85 and 2.62%, respectively) than IC1 and TM3 (1.17 and 1.39%, respectively).The Ka/Ks ratio of each structural domain showed that all OR domains were under purifying selection, with values ranging from 0.020 to 0.826 in pigs (Table 3).

Discussion
Olfaction is critical for the survival of several animal species.OR genes are highly expanded in mammals, who largely rely on smell for finding food, recognizing danger, and reproducing.The genetic variations in OR genes are likely to be associated with functional differences in olfaction, together with anatomical differences in the olfactory organs.Therefore, understanding the genetic differences in OR genes should allow for the prediction of the diversity and sensitivity of olfaction capacity in animal species.However, target-specific typing of OR genes in mammals is challenging because of the extremely large number of paralogs in their genomes and genetic variations 13,15,19,20,22 .Thus, systematic analyses of the genetic diversity of mammalian OR genes at the population level have been limited to a few studies 28,37,38 .
The number of OR genes varies greatly among species and reflects their dependency on the olfaction of each species.The analysis of OR subgenomes in mammalian species showed that pigs are animals with a large number (~ 1110) of functional OR genes 13,22,25 .Previous studies suggested that ORs with more than 60% identity in protein sequence may recognize odorants with related structures 11,39 .However, the presence of a large number of OR genes with > 60% of protein sequence similarity or subfamilies in the mammalian genomes indicate functional differences among closely related sequences.Nguyen et al. 22 classified 349 OR subfamilies belonging to three class I and 14 class II OR families in the pig genome.In addition, allelic diversity can contribute to the functional diversity of a gene.
In this study, we analyzed the genetic variations of 20 annotated pig OR genes belonging to 13 OR families using locus-specific PCR-SBT typing to maximize the diversity of selected OR genes and presented the allelic diversity of OR genes in pigs.Although the number of OR genes analyzed were limited, we covered OR genes from 13 of the 17 families in the pig genome.Our results showed that pig OR genes have similar characteristics in genetic diversity to those of humans and dogs 27,28,40 in the polymorphic spectrum of OR genes consisting of both weakly (mean π < 0.0052; i.e., sOR13B10, sOR5J1, sOR5P1, sOR10L1, sOR10B1, sOR13A6, sOR11A7, sOR8Q4, sOR8F2P, sOR1O1, sOR5AF1, sOR6T3) and highly polymorphic genes (mean π > 0.0052; i.e., sOR53A1, sOR51B1, sOR52M1, sOR14A3, sOR2R3, sOR6Y2, sOR9D3, sOR13K1; S7 Table ).Compared to humans and dogs, pigs have a higher allelic diversity in OR genes belonging to large clusters than to small (Table 1), presence of breed-specific alleles (S10 Table ), weak selection (S7 Table ), and higher allelic diversity in class I over class II OR genes (S7 Table ).Moreover, we observed significant differences in the allelic constitution of OR genes across different pig breeds, but the availability of data from other species is limited.
The nucleotide sequence diversity of ORs in pigs (mean π = 0.0052) observed in this study was significantly higher (Z-test, P = 0.00107) than that reported for dogs (π = 0.0018) 27,28 , consistent with the fact that the size of the pig OR repertoire is larger than that of dogs 21,22,32 , showing a positive correlation between the extent of gene expansion and nucleotide sequence variations in OR genes, which reflects inter-locus and intra-locus diversity, respectively.The number of OR genes in a given species can be positively correlated with the number of odorants that can be discriminated by the species.However, sensitivity to a specific odorant may be determined by the expression levels of OR genes and anatomical characteristics of the olfactory epithelium.Although speculative, the discrimination spectrum and specificity of the OR system to various odorants could be better in pigs than in dogs, considering the revealed genetic diversity of the OR genes of the two species.The positive correlation between the extent of gene expansion and nucleotide sequence diversity of OR genes is interesting, suggesting the importance of olfaction in animal survival although the dependency varies on species.Animals with larger sizes of OR repertoire may also maintain higher allelic diversity than those with smaller OR repertoire.The higher level of inbreeding for domestic dog breeds than for pigs may also contributed to the reduced overall diversity of OR genes in domestic dogs 41,42 .
OR molecules consist of 15 structural domains, and the extent of genetic variation and conservation of each domain may vary.In previous studies, multiple alignments of numerous mammalian ORs showed that the extent of conservation was relatively low between TM3 and TM6 43 .In this study, we observed higher degrees of genetic variation in TM5 (π = 0.0128), TM6 (π = 0.0103), and EC3 (π = 0.0104) domains than in others (mean π = 0.0075) for pig ORs (Table 3).However, the mean Ka/Ks ratio of EC3 (0.020) was lower than that of TM5 (0.449) and TM6 (0.219), indicating a difference in domain diversification depending on the selective advantages of the region in olfaction.
Major histocompatibility complex (MHC) and OR are similar in that they recognize and differentiate a large number of ligands and have evolved to create high genetic diversity through gene duplication and allelic diversity.Therefore, the genetic characteristics of the two protein families are similar.For swine leukocyte antigen (SLA) genes, which is the MHC genes of pig, it has been well documented that the presence of an extremely large number of alleles, a large difference in allelic constitution among different breeds, and breed-specific alleles [44][45][46][47][48][49] are most likely to result from independent breeding histories for different breeds.In addition, the characteristics of pig OR genes analyzed in this current study showed similarities to those of the SLA genes described above.Moreover, the observation of a larger number of OR alleles in LAN and YOR breeds than in other breeds is consistent with the results for SLA class I genes 45 .LAN and YOR are well-recognized as preferable maternal lines in pig breeding.
Conservation of the extent of the genetic diversity of individual OR genes across different pig breeds is interesting.This could suggest the characteristics of each OR gene regarding functional restriction or diversity in the detection of various odorants.The locus-specific diversity of individual OR genes could indicate either differences in the age of the locus after duplication or the permissibility of genetic changes among OR genes.A positive correlation between the genetic diversity of individual OR genes and the cluster size of OR genes suggests an association between gene duplication and the allelic diversity of OR genes, consistent with a previous study on dog ORs 27 .
Notably, only two alleles were observed for sOR8F2P, a pseudogene, across most breeds, despite the loss of selection pressure on pseudogenes [50][51][52] .In addition, we observed limited allelic diversity among a few OR genes, including sOR5J1, sOR5P1, sOR11A7, sOR13B10, and sOR8Q4 (S7 Table ).In contrast, larger genetic diversity was observed for sOR14A3, sOR10B1, sOR2R3, sOR53A1, sOR6Y2, and sOR6T3 than for other ORs.The allelic diversity of individual OR genes may be related to the characteristics of the target odorants binding to individual OR molecules.In addition, we observed the presence of conserved alleles with a high frequency (> 0.5) in a few OR genes across all seven pig breeds (S13 Table ), which may indicate the functional importance of these alleles.
The mean nucleotide diversity (π pig = 0.0052) calculated from the 20 pig OR genes in this study was much higher than that of other species, including humans (π human = 0.0008, 40 ) and dogs (π dog = 0.0018, 27,28 ), suggesting accumulation of a larger number of genetic changes in pigs than in the other two species (approximately three times higher than in dogs).Therefore, the OR system of pigs has evolved to have large genetic diversity, together with a large number of functional OR genes and less pseudogenization than in humans or dogs.
The large number of OR genes in the genome and the structural characteristics of OR genes as a single exon gene may expose OR genes to continuous genetic changes.Thus, a large number of functional OR genes, as many as 400-2000 in mammals, has been dictated by continuous change in the OR repertoire, varying with the dependence on olfaction by individual species.Moreover, the size of the functional OR repertoire can rapidly decrease once the dependence on olfaction is weakened, as in marine mammals 16,25,53 .
Analysis of evolutionary dynamics from the 20 pig OR genes showed the presence of weak purifying selection consistent with that of previous studies 13,25,54 .However, our results indicated no clear difference in selection pressure between class I and class II OR genes (S7 Table ), which differs from the results of a few previous studies 13,25,54 .This difference could be attributed to the differences in the characteristics of the datasets among the studies.Our analysis was based on the genetic variations of each OR locus separately, rather than using the combined genetic variations of OR orthologs in previous studies.
In conclusion, we presented the results of the population diversity of 20 representative pig OR genes belonging to 13 OR families.Genetic diversity varies among OR genes and depends on the breeding structure.The large number of OR genes in pigs indicates a large dependence on olfaction for their survival.Although this study was limited by the number of OR genes and individuals analyzed for each breed, we believe that the genetic characteristics of the remaining OR genes in the pig genome are likely to be consistent with the results presented here.Identification of the ligands for major OR alleles that are common across different breeds and for highly polymorphic OR genes may help clarify the characteristics of functionally critical ORs for pigs.Analyses of the genetic diversity of the complete OR repertoire of an individual and comparisons among different individuals, breeds, or other Suidae animals are challenging, considering the genetic complexities of OR genes described in this study.However, our study is the first step toward achieving these objectives.

Animals and tissues
Tissue samples of seven pig breeds, BER, DUR, LAN, YOR, KNP, NIH, and SNU, were collected and stored at -80 °C in previous studies 45,55 and used to prepare genomic DNA for this study.Eight individuals from each breed were randomly selected.All experiments were approved and performed in accordance with the guidelines and Fig).In addition, an SNP (T→G) at the first position of the stop codon of sOR13A6*03 resulted in disruption of the stop codon (S5 Table, S4 Fig).Notably, sOR13A6*03 was the major allele in LAN along with sOR13A6*01, with a frequency of 0.38 (S8 and S10 Tables).

Figure 1 .
Figure 1.Allelic diversity of 20 pig olfactory receptor genes analyzed from seven pig breeds.(a) The mean number of alleles and mean nucleotide diversity (π) of 20 OR genes from the typing results of 56 pigs from seven breeds were plotted.The scale bars are indicated on the left and right, separately.(b) The mean heterozygosity (observed and expected) was plotted.

Figure 2 .
Figure 2. The results of principle component analysis among different pig breeds from the genotype information of 20 OR genes.The results of principle component analysis using principal component (PC) 1 and 2 are shown.Each individual is represented by a dot, and different breeds are indicated by different colors, as shown on the upper left of the plot.

Figure 3 .
Figure 3. Phylogenetic relationships of 20 pig OR genes.UPGMA trees were plotted based on nucleotide (a) and protein sequences (b), respectively.Alleles were clustered together for all 20 OR genes and, therefore, alleles of the same OR genes were collapsed and indicated by triangles at the end of the branch.Branches of Class I and II OR genes are indicated in red and blue, respectively.Bootstrap values less than 50% are not shown in the plots.The genetic distance is scaled at the upper center.

Table 1 .
Correlation between OR gene related genetic characteristics.Parameter values for the correlation analysis are described in S7 Table.

Table 2 .
Comparison of the mean allelic diversity of 20 OR genes among seven pig breeds.
a number of alleles, b number of effective alleles, c Shannon's information index, d number of private alleles, e observed heterozygosity, f expected heterozygosity.

Table 3 .
Comparison of the genetic diversity of different OR structural domains from 20 pig OR genes.a Nterm, N-terminal; TM, transmembrane; IC, intracellular; EC, extracellular; Cterm, C-terminal; b the mean number of non-synonymous sites; c the mean number of synonymous sites; d Ka, number of nonsynonymous substitutions per nonsynonymous site; Ks, number of synonymous substitutions per synonymous site.