The swine draft genome sequences (Sscrofa10.2) were retrieved from the National Center for Biotechnology Information (NCBI). A translated basic local alignment search tool (TBLASTN) search was performed to identify regions containing OR related sequences that had at least two of the following conserved motifs: MAYDRYVAIC (TMIII), KAFSTCASH (TMVI), and PMLNPFIY (TMVII), or their variants with less than 50% of sequence difference from the conserved motifs. From the identified regions, we selected the sequences in the region one kilobase (kb) upstream and downstream of the BLAST matches. From the analysis, we identified 1,644 OR candidate sequences that were 2 kb in length and translated to amino acid sequences in all six frames. Then, we retrieved 24,809 OR protein sequences from 222 species from NCBI and performed a protein BLAST (BLASTP) analysis against the translated OR candidate sequences to determine the positions of the start and stop codons of the open reading frames (ORFs) on the basis of structural similarity to known OR proteins. For sequences that deviated from the sequences of reported OR proteins, the methionine and stop codon most similar in sequence context to those of the coding sequences of known OR proteins were selected as the start and end of the coding regions. We again performed TBLASTN analysis against the 1,644 sequences to evaluate the presence of all four conserved motifs [GN, MAYDRYVAIC (TMIII), KAFSTCASH (TMVI), and PMLNPFIY (TMVII)]. The candidate sequences were considered “functional ORs” if they were at least 300-amino acid long without any interrupting stop codons and/or frameshifts within the ORFs, “OR pseudogenes” if they were at least 300-amino acid long but contained stop codons or frameshifts within the ORFs, and “partial ORs” if they were shorter than 300 amino acids in length but matched the sequences of the known OR genes. Sequences similar to non-OR G-protein-coupled receptors or partial sequences were removed from our analyses, leaving 1,301 OR genes (including pseudogenes).

The nucleotide sequences of 3,511 OR genes from human (457), mouse (908), dog (845), and pig (1,301, 1644 putative ORs minus 343 partial ORs) were combined and aligned together using CLUSTALW 22 . An unrooted phylogenetic tree was constructed after 1,000 rounds of bootstrapping. The tree was used for classifying OR gene families and subfamilies. Pig OR sequences that did not form a cluster with any reference ORs from the other three species were additionally classified using a sequence similarity matrix (data not shown) in which 40% and 60% amino acid similarity were used as the thresholds to distinguish between families and subfamilies, respectively, as previously described 23 .

For naming pig OR genes, we followed the OR gene classification system described by Glusman et al. 23 . Functional pig OR genes were named “sORmXn” whereas pseudogenes were named “sORmXnP”, where “s” stands for S. scrofa, “OR” is the root name indicating an olfactory receptor, “m” is an integer representing the family that the gene belongs to, “X” is a single letter denoting the subfamily of the gene, and “n” is an integer representing an individual family member. The names of the pig OR sequences were devised on the basis of on their phylogenetic relationships. For example, sOR1A1 is an OR gene of family 1, subfamily A, and is the first member of this subfamily. In the case of pseudogenes, a name such as sOR7E12P indicates an OR pseudogene of family 7, subfamily E, that is the twelfth member of this subfamily. Duplicated genes with the exact same coding sequences were indicated by adding the suffix A, B, or C at the end of their names, i.e., sOR51N3A and sOR51N3B.

Multispecies OR gene clustering analysis was performed with OR protein sequences from humans, dogs, mice, and pigs using the OrthoMCL 3 software 24 , in order to group them on the basis of their sequence similarity and divergence. In total, 706 clusters were formed from 3,511 sequences. The cutoff value for a cluster was 60% similarity at the level of the protein sequence, resulting in sequences with greater than 60% similarity being clustered together regardless of the species of origin.

To detect conserved motifs in predicted OR protein sequences, sequence logos were generated from an alignment of functional OR gene sequences using the WebLogo program 25 . The PRATT 26 program from the Pattern Discovery Platform 27 was used to define pig OR-specific patterns with the criteria listed in Additional file 1.

The four motif sequences, GN, MAYDRYVAIC, KAFSTCASH and PMLNPFIY, which are common to mammalian OR genes were used to search the full repertoire of ORs in the pig genome (Figure 1A). We identified 1,301 OR gene-related sequences with lengths of 900–1,000 base pairs (bp). We also analyzed their ORFs and grouped them into the following two categories: functional and pseudo genes. In total, 1,113 OR sequences were identified as functional and 188 were identified as pseudogenes. Among the identified functional genes and pseudogenes, 91.19% of the sequences contained all three OR domains and the rest were missing one of the conserved motifs (Figure 1B). For the GN motif, the presence of the motif was difficult to evaluate because the motif was defined by only two amino acids and may also have sequence variations. Therefore we did not include the result.

The locations of the OR genes were analyzed on the basis of their relative positions in the pig genome by grouping them into positional regions according to their positional proximity. If the coding sequences of the OR genes were more than one megabase (Mb) apart, they were considered to be present on different regions. Of the 1,301 functional genes and pseudogenes, 1,290 were mapped to 46 different chromosomal regions across 16 pig chromosomes and the remaining 11 were located on chromosome U, which contains unmapped sequences (Figure 2). Except for chromosomes 11, 16, 17, and Y, which were devoid of OR genes, all the other chromosomes contained one to 406 OR genes (Table 1). Chromosome 2 had the largest number of OR genes (341), followed by chromosomes 7, 9, and 1. Accordingly, chromosome 2 contained the largest number of OR subfamilies with 121 subfamilies, while only a single subfamily was present on both chromosomes 8 and 10 (Table 1).

Note: In the case of the absence of both OR functional genes and pseudogenes, the pseudogene % was not indicated.

We observed extensive variations in the number of OR genes at individual OR gene clusters from one to 123 OR genes per locus/cluster (Table 2). Due to the presence of a large number of OR genes in the genome, the number of pseudogenes was also high (n = 188). The percentage of pseudogenes varied among clusters and ranged from 0 to 100% (Table 1). Of the 46 OR gene clusters, the locus “10-78” was the only OR gene locus that had only one pseudogene, while the other 45 clusters had at least one functional gene (Table 2). In the current swine genome assembly Sscrofa10.2, 11 OR genes (nine functional genes and two pseudogenes) were located on unmapped contigs without any chromosome information. Complete information on the distribution of all OR functional genes and pseudogenes in the pig genome is detailed in Additional file 2.

¹Number of OR gene clusters with 0 to 123 functional OR genes.

²Number of subfamilies whose members are encoded at one to five clusters.

³Number of clusters that encode members of one to 52 subfamilies.

Understanding the diversity of OR genes is important for elucidating the differences in their functional responses to various odorants. ORs with more than 60% identity in protein sequence are suggested to recognize odorants with related structures 29 30 . To evaluate the diversity in the OR gene repertoire of pigs, the identified pig OR genes were classified into families and subfamilies according to the results of phylogenetic analyses (data available upon requested) and their sequence similarity. Then, the results obtained after the classification were compared with those previously obtained for from humans, dogs, mice, and rats 9 13 16 . Our analysis showed that the pig OR repertoire comprises 17 families and 349 subfamilies; this repertoire is largest among the known repertoires of mammals (Additional file 3). This suggests that compared to other species, pigs may have a more sophisticated system to sense smell and may be able to distinguish more diverse odorants. Although humans and dogs have relatively large number of OR subfamilies (300 each), humans have a higher pseudogene frequency (52%) than pigs, and dogs have a lower number of functional genes (n = 872) than pigs (n = 1,113). This supports the idea that the functional complexity of the pig olfactory system could be attributed, in part, to genetic complexity. Similar to the OR genes of other mammals, pig OR genes could also be classified into two classes, with three Class I families and 14 Class II families (Additional file 3).

The number of OR genes belonging to each subfamily may represent the importance of the specific subfamilies for the species, as the OR gene subfamilies that are important for the survival of the species are likely to expand in the genome through evolution. Therefore, we counted the number of ORs in each subfamily (Additional file 4). The size of pig OR subfamilies was extremely variable with one to 52 OR genes per subfamily. While most subfamilies had one to six members, six subfamilies had more than 20 genes each. The most common type of subfamily comprised only a single OR gene, accounting for 146 subfamilies. In contrast, subfamily sOR6A consisted of 52 genes (data not shown).

To study the possible associations between the subfamily structure and the chromosomal organization of OR genes in pigs, the chromosomal locations of all OR gene members of the 349 pig OR subfamilies were analyzed (Table 1). The largest OR cluster in the pig genome was the cluster “9-2” on chromosome 9, which contained 123 OR genes making up 52 subfamilies. We observed that 275 (78.8%) subfamilies were encoded by genes at a single chromosomal cluster, suggesting possible functional similarities among OR genes within a cluster. When we determined the subfamily composition of individual OR gene clusters, the number of subfamilies within a cluster ranged from one to 52 (Table 2). About 26% (12/46) of the OR clusters encoded only one OR subfamily, while 74% of clusters (34/46) encoded OR genes of more than two subfamilies. The general characteristics of the OR subgenome including the number of functional OR genes within a cluster, the number of clusters within a subfamily, and the number of subfamilies within a cluster in the pig (Table 2) were consistent with those reported for other species such as mouse and human 9 13 .

Gene duplication plays an important role in establishing the biological characteristics or diversity of organisms during evolution 31 . From our analysis to identify OR genes in the pig genome, we found 100% identical coding sequences of OR genes that mapped to different regions in the pig genome. Further analysis showed that the sizes of duplications ranged from 1.1 to 120 kb (data not shown). Duplicated OR genes were found for both functional genes (n = 166) and pseudogenes (n = 22) (Additional file 5), although most of the duplications were of functional genes. There are 80 functional and 11 pseudo genes that have one identical copy each, making 160 and 22 OR genes in total, and two OR genes sOR7A6[ABC] and sOR5AT1[ABC] were found three times each in the pig genome assembly Sscrofa10.2 (Additional file 6). In total, 93 duplication events consisting of 87 intra- and six inter-chromosomal duplications (data not shown) were observed at 11 chromosomes with duplication of two to 41 genes depending on the chromosome (Additional file 5). The most frequent duplication pattern was the presence of two identical OR coding sequences in the genome (Additional file 6). However, we also were not able to entirely exclude the possibility that some of these duplications might result from the errors in the genome assembly. Although we reexamined the partial or duplicated OR genes with respect to assembly issues such as locations in the contigs and relationship between individual members of identical duplicates, we did not find any logical evidences to support that a part of partial or duplicated OR genes were caused by assembly errors.

Using the criteria in Additional file 1, we performed a pattern discovery analysis for pig OR genes. Table 3 shows five motif patterns identified from four conserved transmembrane domains of pig OR genes, TMII, TMIII, TMVI and TMVII, which are similar to those reported from other species including dogs 16 , rats 16 , and humans 13 except for minor differences at variable amino acid sites. Analysis of the similarities and differences in conserved OR transmembrane motifs among different species could elucidate the functional importance of each site within the motifs.

Note: The pattern for dogs and rats was taken from Quignon et al. 16 . [XYZ] means X or Y or Z. The lower case letter “X” can be used as a pattern element to denote any amino acid. X(m) is equivalent to the repetition of X exactly m times. X(m,n) is equivalent to the repetition of X exactly k times for any integer k satisfying: m ≤ k ≤ n.

To identify potential target specificity of pig OR subfamilies in odor perception, we compared the amino acid sequences of the 1,113 translated pig OR genes to those of other species with previously described information on odorant specificity, including two human ORs 32 33 and 20 mouse ORs 29 30 34 35 36 37 38 . From the analysis, we found that 18 pig ORs matched ORs from other species with known specificity with at least 60% sequence identity, suggesting that these ORs may share similar olfactory specificities (Table 4). There were three mouse ORs, Olfr672, Olfr586 and Olfr545, showing less than 60% sequence similarity to pig ORs and they are known to sense n-aliphatic acids, n-aliphatic alcohols, n-aliphatic dicarboxylic acids, and (−) citronellal. In addition, our analysis also showed that no pig OR has sequence similarity to OR3A1; this human OR is known to perceive helional, which has sweet and hay-like smell.

Note: A dash (−) indicates the absence of corresponding pig ORs. The order of pig OR clusters was based on the amino acid sequence identity (4th column).