Comprehensive genomic and computational insights into Brucella suis: pan-genome analysis, evolutionary perspectives, and in-silico vaccine design Authors: Ghulam Muhammad Sami Ullah Asif, Amna Chudhary, Yu Feng, Pengtao Jiao, Munkhduuren Shatar, Gerelmaa Ulziibat, Silvana Beutinger Marchioro, Yan li, Jianxin Ye, Yu Huan, Hongjin Li, Hui Jiang, Jiabo Ding, Guangzhi Zhang Published: 2026 ABSTRACT AbstractBackgroundBrucella suis is a zoonotic intracellular pathogen responsible for brucellosis, mainly in swine and humans. Although numerous genome sequences are publicly available, an integrative genomic analysis combining pan-genome architecture, structural organization, evolutionary relationships, and vaccine-associated targets remains limited.ResultsIn this study, we analyzed 91 publicly available B.suis genomes to characterize their pan-genome composition and genomic structure. The pan-genome exhibited an open configuration, indicating continued genomic diversification. A total of 2,146 core genes were identified, representing conserved functions essential for species maintenance, while the accessory genome reflected strain-level variability. Phylogenetic reconstruction based on single-copy orthologs revealed distinct evolutionary clades among the strains. A complementary phylogenetic analysis of pan-genome gene presence–absence patterns further supported clade differentiation and highlighted variation in accessory gene repertoires. Comparative synteny and genome structural analyses demonstrated largely conserved chromosomal organization with localized rearrangements across strains. Screening of the core proteome identified 64 putative antigenic proteins with predicted surface localization and immunogenic properties. Additionally, resistance-associated determinants related to tetracycline and doxycycline were detected in one genome within the dataset.ConclusionsThis comprehensive genomic analysis defines the pan-genome structure, evolutionary relationships, and genome organization of B.suis. The integration of core and pan-genome-based phylogenies provides complementary insights into strain diversification, while the identified conserved antigenic candidates offer a foundation for future experimental validation and rational vaccine development strategies.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12864-026-13105-y.Keywords: Pan-genome, Zoonotic bacteria, Antibiotic resistance, Vaccine development, Phylogenomics Introduction Brucella suis (B.suis) is a Gram-negative, facultative intracellular bacterium and a primary causative agent of brucellosis in swine, with established zoonotic transmission to humans [1–3]. Human infection typically occurs through direct contact with infected animals or consumption of contaminated animal products and may result in chronic febrile illness, arthritis, and reproductive complications [4, 5]. The ability of B.suis to survive and replicate within host macrophages contributes to immune evasion and persistent infection, presenting ongoing challenges for disease control and prevention [6–8]. Whole-genome sequencing has substantially advanced the understanding of genetic diversity within the genus Brucella. Although Brucella species are generally characterized by high genomic conservation, strain-level differences contribute to variability in host preference, pathogenic potential, and ecological adaptation [9, 10]. Comprehensive comparative genomic analyses are therefore essential to distinguish conserved genetic components from strain-specific variations and to clarify patterns of evolutionary divergence [11, 12]. Pan-genome analysis provides a systematic framework to evaluate species-wide genetic diversity by partitioning genes into core and accessory components. The core genome represents conserved genes shared across all strains, reflecting essential biological functions, whereas the accessory genome encompasses variable genes that may contribute to niche adaptation and functional diversification. Characterizing the pan-genome structure of B.suis across a large number of isolates can provide insights into genomic conservation, diversification patterns, and functional variability within the species [11, 13, 14]. In addition to variation in gene content, genome organization and structural conservation are important determinants of bacterial evolution. Comparative synteny and genome structural analyses enable the identification of chromosomal rearrangements, inversions, insertions, and other structural variations that may influence genome stability and evolutionary trajectories. Integrating pan-genome characterization with structural genomic analysis allows a more comprehensive assessment of species-level diversity. Phylogenetic reconstruction further clarifies evolutionary relationships among strains. Phylogenies derived from single-copy orthologs reflect conserved evolutionary signals, while gene presence–absence–based phylogenetic approaches capture variability within accessory genomes. The combined application of these complementary strategies can provide a more robust framework for understanding strain diversification and population structure [15, 16]. In parallel with genomic characterization, identifying conserved antigenic proteins remains a key objective in vaccine research [16, 17]. Although live attenuated vaccines against B.suis have demonstrated efficacy in livestock, concerns regarding residual virulence and zoonotic risk necessitate alternative approaches. Reverse vaccinology enables genome-wide screening of conserved proteins to identify potential subunit vaccine candidates with broad strain coverage [18, 19]. In this study, we conducted a comprehensive integrative genomic analysis of 91 publicly available Brucella suis genomes utilizing well-established bioinformatics pipelines. We characterized the pan-genome architecture, reconstructed phylogenetic relationships using both single-copy ortholog- and gene presence–absence–based approaches. We performed comparative synteny and genome structural analyses to assess chromosomal conservation and variation. Furthermore, we screened the core proteome to identify putative antigenic candidates and evaluated the distribution of resistance-associated determinants across the dataset. This comprehensive genomic framework provides a resource to support future functional studies and translational research on B.suis. Materials and methods Data collection and pangenome construction Collection of genome assemblies Ninety-five publicly available B.suis genome assemblies were retrieved from the NCBI Genome Assembly database (https://www.ncbi.nlm.nih.gov/datasets/genome/?taxon=29461). Only high-quality draft assemblies were considered for inclusion. Assembly quality was evaluated using QUAST (v5.2) [20]. Based on assembly metrics, including contig number and N50 values, 91 high-quality genomes were retained for downstream analyses. The accession numbers of all genomes included in this study are provided in Supplementary File F1 to ensure reproducibility. Genome annotation All retained genome assemblies were reannotated using Prokka (v1.14.6) [21], a widely used prokaryotic genome annotation pipeline. Prokka generated standardized annotations, including coding sequences (CDSs), transfer RNAs (tRNAs), ribosomal RNAs (rRNAs), and hypothetical proteins. The use of a uniform annotation pipeline ensured consistency across datasets and provided a reliable basis for subsequent comparative and pangenome analyses. Pangenome construction The pangenome of B.suis was constructed using the Roary pipeline (v3.13.0) [22], which clusters protein-coding sequences into orthologous groups across multiple genomes. A minimum sequence identity threshold of 95% was applied. Genes were categorized into core (≥ 99% of genomes), soft-core, shell, cloud, and unique groups based on their distribution across strains. Phylogenetic analysis Two phylogenetic reconstructions were performed to evaluate evolutionary relationships from distinct genomic perspectives. The OrthoFinder-based tree [23], derived from concatenated single-copy orthologs, primarily reflects vertical inheritance, whereas the core genome tree (Sect. 2.2.1), derived from the Roary-generated [22] core gene alignment and inferred using FastTree [24] captures variation within conserved genomic regions that may be influenced by recombination or gene gain/loss. Comparison of these trees enabled assessment of phylogenetic congruence and potential genome plasticity within B.suis. Core genome–based phylogeny from pan-genome analysis Phylogenetic relationships based on shared core gene content were inferred using the core genome alignment generated during pangenome construction (Sect. 2.1.3). The concatenated alignment of core genes (present in all strains) produced by Roary was used for tree reconstruction. An approximately maximum-likelihood phylogeny was inferred using FastTree [24], under the generalized time-reversible (GTR) model, as implemented within the Roary workflow. The resulting Newick tree was visualized and annotated using ChiPlot [25]. Single-copy ortholog–based phylogeny Orthologous gene clusters among the 91 Brucella suis genomes and Ochrobactrum anthropi as an outgroup genome were inferred using OrthoFinder v2.5.4 under default settings [23]. Core single-copy orthologous genes shared by all genomes were extracted and individually aligned using MAFFT. Poorly aligned regions were trimmed with trimAl v1.4 using automated settings to remove ambiguously aligned positions. The filtered alignments were concatenated into a supermatrix for phylogenetic inference. A maximum-likelihood (ML) phylogeny was reconstructed using IQ-TREE v2.0.3 [26]with ModelFinder enabled (-m MFP) to determine the best-fitting substitution model. Branch support was assessed using 1,000 bootstrap replicates with ultrafast bootstrap approximation and branch optimization (-B 1000 --bnni --seed 123). The resulting tree was visualized and annotated using iTOL [27]. Functional annotation of pan-genome Functional annotation of the pan-genome was performed using EggNOG-mapper (v2) based on the EggNOG 5.0 database [28]. EggNOG-mapper assigns orthology-based functional annotations, including Gene Ontology (GO) terms, KEGG pathways, and Clusters of Orthologous Groups (COG) functional categories. This approach provides evolutionary context and enables consistent functional classification of genes across the core and accessory genomes. Annotated functions were further analyzed to identify genes associated with metabolic pathways, virulence-related features, and antimicrobial resistance. Insertion element analysis Insertion sequences (IS elements), a class of mobile genetic elements (MGEs), were identified in the 91 B.suis genomes using ISEScan v1.7.2.3 [29]. Genome sequences in FASTA format were used as input, and default parameters were applied for detection. Predicted IS elements were classified according to IS families (e.g., IS3, IS5, IS6, IS66, IS481) based on ISEScan output. To facilitate visualization and summary statistics, TE-hub was used to organize IS annotations, including abundance per strain and genomic distribution. It should be noted that this analysis is limited to IS elements; other MGEs such as transposons, integrons, plasmids, and prophages were not assessed in this workflow. Antibiotic resistance and biosynthetic gene clusters profiling Antibiotic resistance genes were identified using STARAMR (v0.7.2) [30], which integrates ResFinder [31] and other resistance detection tools. The tool identified point mutations, acquired resistance genes, and other genomic features associated with antimicrobial resistance. Results were cross-verified with functional annotations to explore the relationship between resistance determinants and the accessory genome. Biosynthetic gene clusters were identified using antiSMASH [32]. Comparative synteny and genome structural analysis A total of 10 representative genomes were selected for comparative structural analysis. Genome selection was based on geographical origin to capture phylogeographic variability, as well as on strain-specific or accessory gene content identified during pan-genome analysis, to assess structural variation associated with accessory genome diversity. Both chromosomes of B.suis were concatenated in silico before alignment to enable a comprehensive comparison of genome structure across strains. Chromosomal coordinates were maintained in the concatenated sequences to allow accurate mapping of Locally Collinear Blocks (LCBs) to their original chromosomal locations. Whole-genome multiple alignment was performed using progressive Mauve to assess the conservation of gene order and overall genome architecture [33]. One genome, GCA_000007505.1, was designated as the reference. Locally Collinear Blocks (LCBs) were identified as homologous genomic regions conserved in sequence similarity and gene order across the compared genomes. The resulting alignment was visualized as a synteny map, where each genome was represented as a linear track with genomic coordinates along the x-axis. LCBs were displayed as coloured blocks, with consistent colouring across genomes indicating homologous conserved regions. Structural variations, including inversions, translocations, insertions, and deletions, were inferred based on deviations from collinearity and changes in block orientation. Synteny relationships and genomic context were further examined using MCScanX to evaluate conservation and disruption of collinear gene clusters. Patterns of block fragmentation, rearrangement, and absence of LCBs were interpreted as indicators of structural divergence. However, the absence of collinear regions was interpreted cautiously, as such patterns may result from lineage-specific gene loss, long-term divergence, insertion sequence (IS)-mediated rearrangements, or assembly fragmentation rather than recent horizontal gene acquisition. The analysis was designed to characterize genome structural variability and conserved genomic architecture without presupposing ongoing horizontal gene transfer. Data visualization and comparative analysis The distribution of core and accessory genes across genomes was visualized using ggplot2 in R. Functional enrichment analysis was performed to identify overrepresented pathways and gene families in the accessory genome. In silico prediction and analysis of an antigenic protein for vaccine design Antigen selection Antigenic tool was run on the core genome, and potential genes were selected on the basis of antigenic score [34]. Functional annotation of these genes having a top score was performed using eggNOG-mapper. These potential proteins were selected on the basis of their functions. Conserved outer membrane proteins in B.suis were used for potential vaccine prediction [35]. The selected protein was subjected to antigenicity prediction using the VaxiJen v2.0 server [36], applying a default threshold of 0.5 to determine its potential as an antigenic target. MHC class I T-cell epitopes were predicted using the NetCTL server [37], selecting high-scoring epitopes based on a default threshold of 0.75. The immunogenic potential of the identified epitopes was further evaluated using the IEDB server [38], where only those with positive antigenicity values were shortlisted. To ensure the reliability of the selected epitopes, additional analyses were performed. Safety testing Antigenicity was reconfirmed using VaxiJen v2.0 [36], conservancy analysis was conducted using the IEDB Conservancy Analysis Tool [38], and toxicity was assessed using the ToxinPred tool [39] to ensure host-specific immune targeting. MHC class II epitopes were identified through the IEDB-AR server, which utilizes a consensus-based approach integrating the average relative binding matrix method and stabilization matrix alignment method. Furthermore, MHC-restricted allele clustering was analysed using MHCcluster v2.0 [40], and the results were visualized with heat maps and phylogenetic trees to determine functional relationships between peptides and HLAs. B-cell epitopes were predicted using BCPred [41] and FBCPred [42]servers with a stringent cut-off score of 0.8. The biochemical properties of these epitopes, including hydrophobicity, surface accessibility, and secondary structure, were assessed using the IEDB B-cell epitope prediction server. The selected T-cell and B-cell epitopes were assembled to construct a multi-epitope vaccine candidate with low toxicity, low allergenicity, and high immunogenicity [43]. The construct incorporated four epitope sequences linked by GGGS linkers, and a Pan HLA-DR reactive epitope (PADRE) was included to enhance CD4 + T-cell activation. Additionally, four adjuvants, HBHA, HBHA-conserved, ribosomal, and beta-defensin, were integrated to boost immune responses via toll-like receptor (TLR) activation. The Allergenicity of the predicted vaccine construct was evaluated using the AlgPred tool [44], which integrates IgE epitope analysis, MAST, ARPs BLAST, and SVM-based predictions, applying a threshold of -0.40. Antigenicity was further validated using the ANTIGENpro program [45], while solubility was predicted using SOLpro [46]. The physicochemical properties of the vaccine construct were analysed using the Expasy ProtParam server [47], assessing molecular weight, amino acid composition, isoelectric point (pI), GRAVY values, aliphatic index, instability index, and estimated half-life. Secondary structure modelling was performed using the SWISS-MODEL homology modelling pipeline, where homology models were constructed based on experimentally validated protein structures in the PDB database and UniProtKB, mapped using SIFTS. Structural validation was carried out using PSIPRED [48] and PROCHECK [49], and sequence similarity was assessed via PSI-BLAST [50], with models showing > 80.6% similarity selected for further analysis. Secondary structure modelling, molecular docking, and in silico cloning The secondary structure modelling of the predicted vaccine construct was performed using the online tool SWISS-MODEL. The homology models were generated through the automated SWISS-MODEL homology modelling server pipeline, where experimental crystal structures of proteins were mapped between the PDB database and UniProtKB using SIFTS. Model structure evaluation was conducted using PSIPRED and PROCHECK, while Psi-BLAST selected related sequences with similarity > 80.6% to reported proteins. The best model for each protein was chosen for further analysis. For molecular docking, the interaction of the predicted vaccine construct (V1) with six different human leukocyte antigen (HLA) alleles was analysed using the PatchDock online tool [51]. The HLA alleles were retrieved from the PDB database with the following: HLA-A02:01 (PDB: 6J1W), (B) HLA-B40:01 (PDB: 3C9N), and (C) HLA-A*24:02 (PDB: 5JLZ). Further refinement of interactions was performed using FireDock (Fast Interaction Refinement in Molecular Docking) [52]. Similarly, the predicted vaccine construct with TLR4 was docked using GRAMMX [53]. The molecular dynamic simulation of the vaccine-TLR4 complex was conducted using GROMACS [54], where solvation was performed with an SPC water model in a cubic box, followed by energy minimization using the steepest descent algorithm. The system was equilibrated using the NVT ensemble, followed by the NPT ensemble, and the molecular dynamic simulation was executed for 10 ns. Additionally, the flexibility and structural stability of the docked vaccine-TLR4 complex were assessed using the iMODs server [55], which provided insights into B-factors, deformability, covariance, and eigenvalues. For optimization and in silico cloning, the amino acid sequence of the vaccine was reverse-translated into a DNA sequence to enhance expression in the E. coli system using the Java Codon Adaptation Tool (JCAT) [56]. The JCAT tool predicted the GC content and Codon Adaptation Index (CAI) while avoiding restriction enzyme cleavage sites, Rho-independent transcription termination, and prokaryotic ribosome binding sites. Finally, the optimized amino acid sequence of the vaccine construct was inserted into the E. coli pET-28a (+) expression system using SnapGene. Expression analysis The solubility of the predicted vaccine construct was evaluated using Protein-Sol, while extracellular release potential was predicted using SignalP-5.0 to determine secretion efficiency and expression feasibility. Results Data processing and pangenome construction Genome assembly and reannotation In this study, we analysed 91 genomes of B.suis collected from 11 different countries, including the USA, South Africa, China, Mongolia, Portugal, Croatia, Malaysia, Switzerland, India, France, and Germany. The genome assemblies varied in size, ranging from 3.1 to 3.4 MB, with the number of contigs per genome spanning 249 to 2. On average, each genome contained 3099 protein-coding genes. The genomic GC content remained consistent across all isolates, averaging 57.1%. Pangenome composition Pan-genome analysis of 91 B.suis genomes identified 4424 genes cluster. Of these, 2146 genes (48.5%) were classified as core genes (present in 99–100% of strains), representing the conserved backbone of the species. An additional 693 genes (15.7%) were assigned to the soft-core genome (present in 95–99% of strains). The accessory genome consisted of 177 shell genes (4.0%), detected in 15–95% of isolates, and 1,408 cloud genes (31.8%), present in fewer than 15% of strains. Overall, the core and soft-core fractions together accounted for 64.2% of the total pan-genome, whereas accessory genes (shell + cloud) comprised 35.8%, indicating moderate genomic variability across the 91 analysed genomes. Phylogenetic analysis Phylogenetic analysis based on the core genome A phylogenetic tree was constructed based on the core genome alignment obtained from Roary (Fig. 1). The tree illustrates the evolutionary relationships among the analysed bacterial strains, with branch lengths representing their genetic divergence. The phylogenetic analysis of all strains based on the pangenome. The left panel shows a maximum-likelihood phylogenetic tree based on core genome alignments, illustrating evolutionary relationships among strains. The middle panel shows the distribution of gene categories per strain: core genes (purple), soft-core genes (light blue), shell genes (orange), cloud genes (red), and unique genes (green). The upper-right pie chart summarizes the distribution of gene categories, while the lower-right heat map illustrates core (purple) and variable genes (yellow), highlighting conservation and diversity among strains Most strains clustered into distinct clades, indicating high similarity within each group. Closely related strains were identified based on shorter branch lengths and shared recent common ancestors. Notably, GCA_014884445.1, GCA_014884465.1, and GCA_014884485.1 exhibited minimal divergence, forming a tight cluster indicative of high genetic similarity. Similarly, GCA_904066045.1, GCA_904066055.1, GCA_904066065.1, and GCA_904066075.1 grouped closely together, reflecting strong evolutionary relatedness. In contrast, GCA_000157775.1 displayed the longest branch length, indicating it is the most divergent genome in the dataset. Additional genomes, including GCA_014884525.1, GCA_001908355.1, and GCA_000331635.1, also showed relatively longer branches, suggesting increased genetic divergence within specific lineages. The core genome analysis categorized genes into five major groups: core genes, soft-core genes, shell genes, cloud genes, and unique genes. Core gene count was 2146, representing 48.5% of the total gene content, and was highly conserved across all strains. Soft-core genes accounted for 15.7%, while shell and cloud genes comprised 4.0% and 31.8%, respectively. A significant proportion of cloud genes suggests substantial genomic variability among the strains. A heat map analysis of gene presence-absence data revealed a clear separation between conserved and variable gene content across the strains (Fig. 1). Most core genes were ubiquitously distributed across all strains, whereas cloud genes were present only sporadically, further supporting the observed genetic diversity. Unique genes were primarily associated with GCA_009218085, GCA_0019083355, and GCA_000331635, which may reflect possible strain-specific adaptations. (Fig. 1). The pan-genome analysis indicated that all strains contained a similar total number of genes, as determined from genome annotation outputs. However, differences were observed in gene composition: strains with an increased number of unique genes exhibited a corresponding reduction in core genes, suggesting a redistribution of gene categories while maintaining a relatively constant total gene count. Genome phylogeny based on single-copy orthologs A maximum likelihood phylogenetic analysis of Brucella suis was performed using concatenated single-copy orthologs and inferred with IQ-TREE2 under the LG + F+G4 substitution model. The resulting phylogeny demonstrates clear intra-species structuring, resolving the analyzed genomes into nine distinct clades (I–IX) (Fig. 2). The phylogenetic tree of 91 Brucella suis strains based on single-copy orthologs shows nine distinct clades (I–IX with bootstrap support values. Strains are color-coded by clade, with geographic origins indicated. The right panel displays genomic features, including CDS, rRNAs, tRNAs, and tmRNAs. The scale bar represents genetic distance The overall topology is well resolved, with most internal nodes supported by moderate to high bootstrap values, indicating that the inferred evolutionary relationships are statistically strong. Within individual clades, isolates cluster tightly and are separated by short branch lengths, consistent with low levels of sequence divergence and suggesting recent common ancestry among members of each lineage. In contrast, the deeper nodes separating the major clades exhibit longer branch lengths, reflecting progressive accumulation of genetic variation within the B.suis population. Although a few basal relationships show comparatively lower bootstrap support, this pattern is typical for closely related bacterial genomes and does not affect the stability of the primary clade structure. The phylogeny reveals substantial genetic diversity within B.suis while maintaining overall species cohesion. Clades II, V, and VI appear particularly compact, suggesting relatively recent expansion or limited diversification within these groups, whereas other clades display slightly greater internal branch lengths, indicative of more heterogeneous lineage histories. Geographically, isolates originating from multiple regions, including the United States, China, Croatia, Portugal, France, Switzerland, India, Malaysia, and South Africa, are distributed across several clades rather than forming strictly region-specific clusters. This pattern suggests widespread global dissemination of B.suis lineages and implies that geographic origin alone does not fully explain the observed phylogenetic structure. Nevertheless, partial regional enrichment within certain clades may reflect localized transmission dynamics or host-associated expansion events. Comparative genome feature profiles shown alongside the phylogeny (CDS, rRNA, tRNA, and tmRNA counts) are highly conserved across the dataset, supporting the taxonomic consistency of the analyzed strains. Minor variation in CDS numbers likely represents strain-level genome plasticity, such as small insertions, deletions, or pseudogene formation, rather than major functional divergence. Taken together, the single-copy ortholog–based maximum likelihood phylogeny indicates that B.suis comprises multiple well-defined yet closely related evolutionary lineages within a genomically cohesive species framework. The combination of strong bootstrap support, short intra-clade divergence, and conserved genome architecture supports a model of gradual diversification accompanied by global dissemination of major lineages. Functional classification of pangenome The functional classification of genes in the B.suis pangenome was assessed using the Clusters of Orthologous Groups (COG) database, providing insights into the distribution of gene functions across the core, shell, soft-core, and cloud gene categories (Fig. 3). Core genes, which are conserved across all analysed B.suis genomes, predominantly belong to essential functional categories such as amino acid transport and metabolism (E), transcription (K), replication, recombination, and repair (L), and carbohydrate transport and metabolism (G). The high proportion of core genes assigned to these categories suggests their fundamental roles in maintaining cellular processes and genomic integrity in B.suis. Additionally, a significant fraction of core genes falls under the category of unknown function (S), indicating the presence of conserved yet functionally uncharacterized genes that may contribute to essential biological processes. In contrast, shell genes present in a subset of B.suis genomes exhibit a more variable functional distribution. The majority of these genes are associated with signal transduction mechanisms (T), secondary metabolite biosynthesis and transport (Q), and inorganic ion transport and metabolism (P). These functional classes are likely linked to strain-specific adaptations, including regulatory responses to environmental changes and host interactions. Functional annotation of the Core genome. Functional annotation is based on core, cloud, shell, and soft-core genes. The legend describes the function of all the annotation categories These categories suggest that cloud genes may contribute to niche specialization, virulence, or antimicrobial resistance in specific strains. Cloud genes, which are found in only a few B.suis strains, represent the most functionally diverse group and exhibit a high proportion of genes related to defence mechanisms (V), post-translational modifications (O), and cell motility (N). These categories suggest that cloud genes may contribute to niche specialization, virulence, or antimicrobial resistance in specific strains. Similarly, soft-core genes present in most but not all B.suis genomes show an intermediate functional profile, with contributions from core cellular processes and adaptive functions. This analysis reveals that the B.suis pangenome consists of a highly conserved core genome essential for cellular maintenance, alongside a variable accessory genome that likely contributes to environmental adaptation and strain-specific traits. A substantial number of genes with unknown functions across different gene categories highlight the need for further functional characterization to understand the genomic diversity of B.suis fully. These findings provide valuable insights into the evolutionary dynamics of B.suis and offer a foundation for future studies on host-pathogen interactions, antimicrobial resistance, and strain differentiation. Mobile genetic elements Mobile genetic elements (MGEs) are important contributors to genome variation, plasticity, and structural variation in bacteria. In this study, insertion sequence (IS) elements were identified across 91 B.suis genomes, revealing clear differences in their distribution and abundance. IS5 was the most prevalent and variable element, with copy numbers ranging from 0 to 16 per genome, and many genomes exhibited high copy numbers (≥ 9), indicating substantial expansion. In contrast, IS3 and IS6 were relatively conserved, typically present at 2–3 and 1–2 copies, respectively, suggesting stable genomic maintenance. The elements IS66 and IS481 were rare, occurring sporadically and at low copy numbers. Genomes with the highest IS content consistently exhibited a characteristic pattern of IS3 (3 copies), IS6 (2 copies), and elevated IS5 (15–16 copies), suggesting a coordinated expansion of these elements. Overall, variation in total IS content was primarily driven by IS5, whereas IS3 and IS6 remained stable, and IS66 and IS481 contributed minimally to overall IS diversity (Fig. 4). Notably, GCA_000740235.1 harbored the highest total number of IS elements (n = 22), largely due to IS5 expansion. At the same time, genomes such as GCA_000292125.1 and GCA_949837105.1 showed the highest occurrence of minor IS elements (IS66 and IS481), though at low absolute counts. This analysis was limited to insertion sequences; other MGEs, including plasmids, integrons, transposons, and prophages, were not assessed and may also contribute to genome plasticity. Circular presentation of Brucella suis strains showing genomic feature distribution. This circular phylogeny illustrates the relationships among Brucella suis strains based on core genome alignments. Concentric rings indicate the presence of key genomic elements: antibiotic resistance genes (ARGs), biosynthetic gene clusters (BGCs), and insertion elements (IS). The distribution patterns highlight genomic diversity across strains Antibiotic resistance and biosynthetic gene clusters profiling A total of 91 B.suis genomes were screened for antimicrobial resistance (AMR) genes using STARMR. (Fig. 4). The analysis revealed a limited presence of known AMR determinants across the dataset. STARAMR analysis of genome GCF_000157755.1 detected a partial tet(C) tetracycline resistance gene with 100% nucleotide identity and 60.46% sequence coverage relative to the reference plasmid sequence (GenBank accession AY046276). The identified fragment (711 bp) was located on contig NZ_DS999726.1. While tet(C) is known to confer resistance to tetracycline and doxycycline via efflux, the incomplete coverage suggests that functional confirmation would require phenotypic validation. In parallel, the presence and distribution of secondary metabolite biosynthetic gene clusters (BGCs) were analysed across the B.suis genomes (Fig. 4). The screening focused on four major classes: β-lactones, arylpolyenes, terpenes, and acylamino acids, revealing strain-specific variations in metabolite biosynthetic potential. Three genomes displayed notable differences in their secondary metabolite profiles. GCA_000480155.1 exhibited β-lactones and acyl amino acids, while arylpolyenes and terpenes were absent. GCA_000292125.1 contained β-lactones and acyl amino acids, with no evidence of arylpolyenes or terpenes. GCA_001908355.1 harboured β-lactones, arylpolyenes, and terpenes, but lacked acyl amino acids. These differences may reflect strain-specific metabolic variation, gene loss, or regulatory mechanisms influencing the presence of BGCs. The absence of certain secondary metabolites in specific strains may be linked to ecological niche specialization or selective pressures within B.suis. Further comparative genomic and functional analyses are required to elucidate the evolutionary significance of these variations and their potential impact on bacterial physiology, host interactions, and antimicrobial properties. Comparative synteny and genome structural analysis Whole-genome multiple alignment revealed a mosaic pattern of conserved and structurally variable regions across the analysed Brucella suis genomes (Fig. 5). Several large locally collinear blocks (LCBs) remained collinear across strains, forming parallel tracks with preserved order and orientation, indicative of a conserved core genomic framework. However, multiple LCBs exhibited orientation changes, providing clear evidence of inversion events and underlying structural variation among strains. The most extensive rearrangements were observed in strain GCA_949836965.1, which harbored several large inversions spanning approximately 0.36–0.98 Mb, 1.29–1.70 Mb, 1.82–2.06 Mb, 2.16–2.27 Mb, 2.37–2.47 Mb, 2.65–2.80 Mb, and 2.87–3.20 Mb. Additionally, less extensive inversion events were detected in GCA_040406615.1 and GCA_019998625.1, whereas the remaining strains retained a largely collinear genomic organization. Synteny map showing genome alignment across multiple species. Each row represents a genome, with coloured blocks indicating conserved regions (LCBs). Matching colours across genomes shows synteny, while connecting lines reveal structural variations such as inversions and translocations. Genomes near the bottom show greater fragmentation and rearrangement, indicating higher structural divergence Missing or truncated LCBs were also observed in certain genomes, which may reflect lineage-specific gene loss, historical genomic rearrangements, IS element-mediated reshuffling, or assembly-related artefacts. Consequently, the absence of syntenic regions was not interpreted solely as evidence of recent gene gain or loss. Overall, these findings indicate that B.suis genomes share a conserved core architecture, while strain-level structural variability, primarily driven by inversion events, reflects evolutionary diversification through internal genome rearrangements rather than extensive horizontal gene acquisition. In-silico Vaccine Prediction Functional annotation of selected core genes All antigenic scores for all 2146 core genes obtained from the Antigenic tool were assessed, and those with a score ≥ 1.362 were selected. To refine this selection, functional annotation was performed using EggNOG-mapper, which categorized genes according to their biological roles and relevance to immune responses. The Clusters of Orthologous Groups (COG) classification revealed that the most represented functional categories were amino acid transport and metabolism (E), function unknown (S), and inorganic ion transport and metabolism (P), while chromatin structure and dynamics (B) and defence mechanisms (V) were the least represented (Fig. 6). Through this process, 64 genes were shortlisted as they were predicted to play critical roles in host-pathogen interactions and immune system activation. These findings provide insights into the functional composition of the selected genes and their potential roles in vaccine target identification. In the end, we used the MarC family of integral membrane proteins for potential vaccine prediction. Clusters of Orthologous Groups (COG) classification of selected antigenic genes. The bar chart shows the distribution of 200 antigenic genes across COG categories. The most represented categories include amino acid transport and metabolism (E), function unknown (S), and inorganic ion transport and metabolism (P), while chromatin structure and dynamics (B) and defence mechanisms (V) were the least represented. This classification highlights key biological functions relevant to vaccine target selection Clustering analysis of MHC-I and MHC-II epitopes Following gene selection, epitope prediction was conducted using the Immune Epitope Database (IEDB) to identify highly immunogenic regions within the selected genes. Three categories of epitopes were predicted: MHC Class I epitopes, which are crucial for cytotoxic T lymphocyte (CTL) activation; MHC Class II epitopes, which stimulate helper T cells (Th cells); and B-cell epitopes, which are essential for antibody-mediated immunity. The top 20 epitopes (Table 1) from each category were selected based on their high immunogenicity, low toxicity, and potential to trigger a strong immune response. Table 1Predicted non-toxic, non-allergenic, and immunogenic epitopes for MHC-I, MHC-II, & B cellsB-cellsMHC-IMHC-IIMSRHKRALRASIISFAFRVAGGLLLFWIAFFEKRTERKEKSAEVAITRDHIRFIVMALFAIAGGLLLFWIAFEMIFFPSVGSFIVMALFAIAAITPGMSRHKRGQVASHPPSRRHPGSTCRAVRRRWYFPSVGSRAALALRASIISFIVMALFQDTHHWLERFRSKSIDFGLAPLFLAIAPLFLAITPGMSRHKGLLLFWIAFFIVMALFAIAGAQILIISFIVMALFFLAITPGMSRHKRGQISFIVMALFGAFRVAGGLLLFWIAIVMALFAIAGAFRVAGGLLLFWIALFWIAFEMIFGGLLLFWIAFEMIFELGMFGITIGAGLAPLFLAITPGMSRLLFWIAFEMGLLLFWIAFEMIFEKLLFWIAFEMIGMFGITIGAFRVAGGMSRHKRGQVGQVALRASIISFIVMRVAGGLLLFWHKRGQVALRASIISFSTCRAVRRRWHPGSTCRAVRRRWYQVAGGLLLFWIGAFRVAGGLLLFWIVALRASIISFIVMALFAIAGAQILGLAITPGMSRHKRGQVLAPLFLAITPGMSRHLLFWIAFEMIFEKRTMFGITIGAFRVAGGLMSRHKRGQVALRASIPGLAPLFLAITPGMSPLFLAITPGMSRHKRQVALRASIISFIVMARHKRGQVALRASIISSFIVMALFAIAGAQISRHKRGQVALRASIIVAGGLLLFWIAFEMIVALRASIISFIVMAL Predicted non-toxic, non-allergenic, and immunogenic epitopes for MHC-I, MHC-II, & B cells Toxicity, allergy, and immunogenicity evaluation The allergy assessment identified non-allergenic epitopes for MHC-I, MHC-II, and B-cells, suggesting their potential safety for immunological applications. Similarly, toxicity evaluation confirmed that all predicted epitopes for MHC-I, MHC-II, and B-cells were non-toxic, indicating their suitability for further development. Immunogenicity analysis identified highly immunogenic epitopes that were both non-toxic and non-allergenic. A detailed list of these epitopes and their immunogenic properties is provided in Table 1. The findings support the potential of these epitopes for vaccine design and therapeutic applications. Final in-silico vaccine design prediction A final vaccine construct was designed by selecting the most promising epitopes, incorporating 10 MHC Class II, 6 MHC Class I, and 4 B-cell epitopes. To enhance the construct’s immunogenic potential and stability, additional elements such as the Pan HLA-DR Epitope (PADRE) sequence were included, along with suitable linkers to facilitate epitope presentation in (Fig. 7). These modifications were intended to improve the vaccine’s ability to stimulate cellular and humoral immunity, thereby ensuring a broad and effective immune response. The 3D model of the vaccine protein was generated and evaluated for its conformational properties. Ramachandran plot analysis revealed a high percentage of residues in the most favourable regions, indicating a well-folded and stable protein structure. Favourable energy values further supported the structural reliability of the vaccine candidate, suggesting that it would remain stable under physiological conditions. Structural analysis of the vaccine construct was performed to assess its stability and structural integrity. The docking of the vaccine protein with MHC Class I and Class II molecules demonstrated strong binding affinities, suggesting that the epitopes within the construct could efficiently interact with immune receptors (Fig. 8) Structural representations of various HLA class I and class II molecules. Ribbon models illustrate the 3D conformations of six different HLA alleles. Panels (A–C) show HLA class I molecules—(A) HLA-A02:01 (PDB: 6J1W), (B) HLA-B40:01 (PDB: 3C9N), and (C) HLA-A*24:02 (PDB: 5JLZ)—each displaying a peptide-binding groove formed by α-helices and β-sheets. Panels (D–F) show HLA class II molecules—(D) HLA-DQ8 (PDB: 3UPR), (E) HLA-DR1 (PDB: 3I6L), and (F) HLA-DR3 (PDB: 1A6A)—which consist of α and β chains forming a peptide-binding cleft. Protein backbones are colored in green, while bound peptides or structural helices important for peptide interaction are highlighted in cyan. This comparative visualization highlights the structural diversity among HLA alleles relevant to antigen presentation and immune response High docking scores confirmed the robustness of these interactions, reinforcing the vaccine’s potential to elicit a strong immune response. These findings provide valuable insights into the rational design of a multi-epitope vaccine against B.suis. The in-silico analyses strongly suggest that the vaccine construct is structurally stable, immunogenic, and capable of inducing cellular and humoral immune responses. Further experimental validation will be necessary to confirm its efficacy and safety in preclinical and clinical settings. Discussion Understanding B.suis genetic diversity and adaptive mechanisms is critical for elucidating its role in disease transmission, host–pathogen interactions, and potential antimicrobial resistance dynamics [2, 3, 6]. In this study, we performed a pan-genome analysis using publicly available genomic data, revealing a dynamic genome structure [15]. Our results indicate that the pan-genome of B.suis follows an open model, suggesting ongoing gene gain and loss [15, 16]. This observation is consistent with bacterial species that occupy diverse ecological niches and host environments [2, 3]. The open nature of the genome reflects genetic plasticity, which may be influenced by multiple evolutionary forces, including gene acquisition and environmental selection pressures [6–8]. We identified approximately 2,146 core genes conserved across all strains, representing essential cellular functions such as replication, transcription, and translation [20, 21]. In contrast, the accessory genome included strain-specific genes that may contribute to variation in virulence and host adaptation [57, 58]. It is important to note that although insertion sequences were identified in this study, other mobile genetic elements, such as transposons and prophages, were not directly analyzed. Therefore, any contribution of broader mobile elements or horizontal gene transfer (HGT) mechanisms remains speculative and requires further investigation. Although intracellular bacteria are often associated with more conserved genomes, this pattern is primarily observed in obligate intracellular organisms such as Rickettsia and Chlamydia [59]. In contrast, Brucella suis is a facultative intracellular pathogen capable of surviving in both intracellular and extracellular environments. This ecological flexibility may contribute to its genomic variability and the observed open pan-genome structure. The accessory genome harbours genes potentially involved in virulence and host adaptation. These include proteins related to adhesion, secretion systems, and stress response, which may facilitate host colonization and survival [6, 7]. The presence of secretion system components suggests potential mechanisms for host interaction; however, their functional roles were not experimentally validated in this study. Additionally, variation in metabolic genes, including those involved in carbohydrate and amino acid metabolism, may reflect strain-specific adaptations to different host environments [60]. With respect to antimicrobial resistance, we identified a tetracycline-associated resistance gene in a single strain, likely associated with an efflux mechanism [3, 10]. While this finding indicates the presence of resistance determinants within the dataset, it does not suggest widespread resistance across B.suis populations. Therefore, these results should be interpreted cautiously. The potential role of B.suis as a reservoir for antimicrobial resistance requires further investigation using broader datasets and functional validation. Phylogenetic analysis revealed distinct evolutionary lineages, with clustering patterns that may reflect host-associated genetic diversification and geographic origin [18, 19]. However, the influence of recombination or horizontal gene transfer on these relationships was not directly assessed in this study and warrants exploration in future work. Furthermore, we identified several putative antigenic proteins through in silico prediction approaches. These proteins, particularly those conserved across multiple strains, may represent potential vaccine candidates [22–29]. However, these findings are preliminary and should be considered hypothesis-generating. Experimental validation, including immunogenicity and protective efficacy studies, is necessary to confirm their suitability for vaccine development. Conclusions and future directions This study provides a comprehensive genomic overview of B.suis, including its genetic diversity, pan-genome structure, and phylogenetic relationships. The observed open pan-genome highlights the genomic variability within this species, which may contribute to its adaptability across different hosts and environments. While resistance genes and virulence-associated features were identified, their distribution was limited, and their functional significance remains to be experimentally validated. Similarly, the predicted antigenic proteins identified in this study represent promising candidates for future investigation but should be interpreted as preliminary findings derived from computational analyses. Overall, our results emphasize the importance of continued genomic surveillance and integrative analyses. Future studies should focus on functional validation of key genomic features, broader investigation of mobile genetic elements beyond insertion sequences, and experimental evaluation of candidate vaccine targets. Supplementary Information Supplementary Material 1. Acknowledgements We gratefully acknowledge the use of publicly available genomic data from the National Centre for Biotechnology Information (NCBI) database, which contributed to this study. Authors’ contributions GMSUA, AC, and PJ contributed equally to this work. They designed the study, performed the experiments or analyses, and drafted the manuscript. MS, GU, and SBM assisted with data interpretation and critically revised the manuscript for intellectual content. JY, YH, HL, and YF provided technical support and contributed to data collection. HJ, GZ and JD supervised the project and provided critical revisions to the final manuscript. All authors read and approved the final manuscript. Funding The work was supported by National Key R&D Program of China(2025YFE0123500) and Beijing Rural Revitalization Agricultural Science and Technology Project༈NY2502170100, NY2502170200). Data availability The corresponding author can be contacted to request the data by submitting a reasonable inquiry. Accession numbers of all the genomes used in this study have been provided in (Supplementary File F1) and can be downloaded from NCBI using these accession numbers. Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Ghulam Muhammad Sami Ullah Asif, Amna Chudhary, Yu Feng and Pengtao Jiao contributed equally to this work. Contributor Information Hui Jiang, Email: jianghui01@caas.cn. Jiabo Ding, Email: dingjiabo@caas.cn. Guangzhi Zhang, Email: zhangguangzhi@caas.cn. References Associated Data Supplementary Materials Supplementary Material 1. Data Availability Statement The corresponding author can be contacted to request the data by submitting a reasonable inquiry. Accession numbers of all the genomes used in this study have been provided in (Supplementary File F1) and can be downloaded from NCBI using these accession numbers.