https://orcid.org/0000-0001-8889-5504
Figures
Abstract
Bacillus subtilis, a valuable industrial microorganism used in starter cultures in soybean fermentation, is a species of bacteria with interspecies diversity. Here, four multilocus sequence typing (MLST) schemes developed to assess the diversity of B. subtilis or Bacillus spp. were applied and compared to confirm the interspecies diversity of B. subtilis. In addition, we analyzed correlations between amino acid biosynthesis genes and sequence types (STs); this is important because amino acids are key taste components in fermented foods. On applying the four MLST methods to 38 strains and the type strain of B. subtilis, 30 to 32 STs were identified. The discriminatory power was 0.362–0.964 for the genes used in the MLST methods; the larger the gene, the greater the number of alleles and polymorphic sites. All four MLST methods showed a correlation between STs and strains that do not possess the hutHUIG operon (which contains genes required for the production of glutamate from histidine). This correlation was verified using 168 further genome-sequence strains.
Citation: Lee G, Heo S, Kim T, Na H-E, Lee J-H, Jeong D-W (2023) Comparison of four multilocus sequence typing schemes and amino acid biosynthesis based on genomic analysis of Bacillus subtilis. PLoS ONE 18(2): e0282092. https://doi.org/10.1371/journal.pone.0282092
Editor: Chih-Chieh Chen, National Sun Yat-sen University, TAIWAN
Received: September 14, 2022; Accepted: February 6, 2023; Published: February 21, 2023
Copyright: © 2023 Lee et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files. In particular, the genome sequences registered in NCBI database (http://ncbi.nlm.nih.gov/genomes) were used, and the access numbers for genome was provided in S1 Table. The allele sequence of the gene identified in the P1 scheme can be identified in pubMLST (https://pubmlst.org/organisms/bacillus-subtilis), and the allele sequence of S2, S3, and L1 can be identified in Figshare (https://doi.org/10.6084/m9.figshare.21818052.v1).
Funding: This work was supported by the National Research Foundation of Korea (NRF) [NRF-2019R1A2C1003639 to Do-Won Jeong and NRF-2022R1F1A1070021 to Jong-Hoon Lee].
Competing interests: The authors have declared that no competing interests exist.
Introduction
Bacillus subtilis is isolated from several fermented soybean products of East Asia, including meju and doenjang from Korea, douchi from China, and natto and miso from Japan [1–5]. The long history of consumption of fermented foods is considered to be the history of consumption of the microorganisms involved in the food fermentation, so B. subtilis is a “generally recognized as safe” microorganism. During fermentation, B. subtilis produces amino acids, fatty acids, and volatile compounds from macromolecules such as starch and proteins via the activities of extracellular amylases and proteases [6–10]. In particular, peptides, amino acids, and their derivatives contribute to the taste of fermented foods [11] and such compounds produced by B. subtilis contribute to the quality and sensory properties of fermented soybean [8, 12]. For this reason, protease activities for amino acid production have been used as criteria for selection of fermentation starter strains [13, 14].
Amino acids play an important role in the sensory properties of foods, providing sweet, bitter, and umami tastes. Each amino acid has its own taste, and their combination contributes to the overall sensory properties of foods [15]. In particular, glutamic acid confers umami (savory) taste [16]. Although glutamic acid in proteins does not contribute umami taste, free amino acid produced from protein by proteolytic activity during fermentation does [17]. Microbial cultures that are involved in proteolytic processes during fermentation also contribute to organoleptic properties of the product foods, such as texture and flavor, and provide essential micronutrients such as minerals and vitamins [18, 19]. Using glutamic acid-producing strains as starter candidates might contribute to the enhancement of umami tastes and product food quality.
Multilocus sequence typing (MLST) is a highly discriminatory typing method appropriate for genetically coherent organisms that can provide a measure of genetic relatedness among strains [20]. MLST data have often been used to distinguish between strains of species that exhibit close phylogenetic relatedness. MLST based on concatenated DNA sequences of internal fragments of multiple housekeeping genes shows greater discriminatory power and reproducibility than phenotypic typing [21]. Five MLST methods have been developed for Bacillus spp., including B. subtilis, to help understand the correlations between sequence type (ST) and antibacterial, antifungal, bio-degradation, antibiotic resistance, and biofilm forming activities, respectively [22–26]. To define STs in an MLST scheme, multiple housekeeping genes (usually 7–9) are partially and accurately sequenced. For each gene used, the distinct sequences found within a species are considered different alleles. Each allele (i.e., each observed unique gene sequence) is assigned an “allele number.” The combination of allele numbers at the multiple loci used in the particular typing system defines the ST of the given strain. Bacillus spp. show much variation within their housekeeping genes, which means that individual strains can be typed by this method. However, there are no comparison results between the different MLST methods for Bacillus. Nor is there any data on correlations between STs determined by MLST and amino acid biosynthesis by Bacillus strains. The aims of this study were to compare the discriminatory ability of the previously developed MLST methods, and to investigate correlations between genetic differences and amino acid biosynthesis ability of B. subtilis strains.
Materials and methods
Bacterial strains and culture conditions
Thirty-eight B. subtilis isolates with previously published genome sequences in the GenBank database were kindly provided by the Microbial Institute for Fermentation Industry (S1 Table). All these strains were isolated from food in Korea. The taxonomic identities of all isolates were confirmed by sequence analysis of near-complete 16S rRNA gene regions [27], and all showed >99.9% identity to the type strain, B. subtilis NCIB 3610T. All B. subtilis isolates were cultured in tryptic soy agar (Difco, USA) and/or tryptic soy broth (Difco) at 37°C for 24 h.
MLST schemes and data analysis
MLST scheme P1 for B. subtilis uses seven housekeeping genes: glpF, ilvD, pta, purH, pycA, rpoD, and tpiA [22] (Table 1). All isolates were assigned allelic profiles and STs according to the previously developed P1 scheme [22] (Table 1); partial gene sequences of housekeeping genes were retrieved from the genome sequences and the sequence at each location was compared with the known alleles from previous MLST results to specify the allele number. Then, the ST was derived from the combination of the allele numbers for all the genes used in the typing scheme. In schemes S2 and S3, nine and seven housekeeping genes are used, respectively [24, 25]. Scheme P1 provides information on allele numbers, nucleotide sequences, and ST profiles on the internet, but schemes S2 and S3 are not open-source, so we could not obtain exact information about alleles and STs for them. Thus, we newly designated allele numbers and STs for these schemes. In the case of scheme L1, a method previously developed for B. licheniformis and B. paralicheniformis [26], the same sequence as the allele identified at this time could not find the same nucleotide sequence as the alleles identified in the previous experiment, so ST was derived by newly assigning alleles following the previous allele number.
Phylogenetic analysis based on STs was performed using the maximum likelihood method in MEGA 7.0 software. STs were analyzed to identify clonal complexes (CCs), which are related clusters of STs, using the eBURST algorithm to identify single-locus variants, double-locus variants, and singletons [28]. Singletons were defined as STs with at least two allelic mismatches with all other STs [28]. Typing efficiency is defined as the number of alleles per polymorphic site for each housekeeping gene [29]. Discriminatory power is the likelihood that two strains differentiate when randomly selected from a population of unrelated strains [29]. The non-synonymous (dN) and synonymous (dS) nucleotide substitutions per site were estimated using MEGA 7.0 software [30]. Using phyloviz software, minimum spanning trees were drawn from the alleles and STs determined by each scheme. The minimum spanning trees were constructed using the goeBURST Full MST algorithm, and the level is assigned the number of loci in the dataset [31].
Comparison of genomes to assess amino acid biosynthesis
Genome sequence data for 39 B. subtilis strains was obtained from the NCBI database (http://ncbi.nlm.nih.gov/genomes; S1 Table). For metabolic pathway analysis based on protein coding genes, genome sequences of 39 strains were uploaded to the Rapid Annotations using Subsystems Technology (RAST) server for SEED-based automated annotation, and subjected to whole-genome sequence-based comparative analysis and Kyoto Encyclopedia of Genes and Genomes pathway analysis [32]. The generated metabolic pathways of the strains were verified using the iPath module (ver. 3) [33]. The Efficient Database framework for comparative Genome Analyses using BLAST score Ratios (EDGAR) was used for core genome, pan-genome, and singleton analyses [34]. Further comparative analyses were performed for specific regions and genes of interest using the BLASTN, BLASTX, and BLASTP tools.
Results and discussion
Comparison of four MLST schemes
Four MLST schemes have been developed to evaluate the genetic diversity of B. subtilis: schemes P1 and S1–S3 (Table 1) [22–25]. The MLST schemes are built on partial sequences of 7–9 housekeeping genes (Table 1). Among them, the types and fragments of the housekeeping genes used are the same in schemes S1 and P1; only the order of concatenation is different. Thus, scheme S1 was not used in this study. Also, we previously developed an MLST method for B. licheniformis (scheme L1), which was successfully applied to genotype analysis of B. subtilis [26, 35]. This scheme was also applied in the present work.
Genome sequences of 39 B. subtilis strains were retrieved from the GenBank database and then analyzed for alleles and ST according to each of the MLST schemes (S2 Table). Except for scheme P1, updated information on allele and ST was not known, so a new number was assigned. Scheme S2 gave the greatest diversity of genetic classification. The size of the housekeeping genes used in scheme S2 is more than double that used in the other schemes (Fig 1). After dividing the number of alleles by DNA size, simple calculating the strain per base, scheme P1 seemed to indicate the greatest genetic diversity.
Comparison of housekeeping gene lengths and variability in terms of number of alleles for 39 strains of B. subtilis.
The allelic variation was analyzed for each gene sequence. The number of polymorphic sites within each gene ranged from 4 (in rpoD, which is used in scheme P1) to 117 (in purH, used in scheme S2), and the number of alleles ranged from 5 (for rpoD) to 23 (for pycA and pta, used in scheme S2) (S3 Table). Although the number of polymorphic sites varied, the dN/dS ratio for each housekeeping gene showed no significant difference. The average dN/dS ratio and discriminatory power across all MLST genes was 0.4503 and 0.86, respectively (Fig 2 and S3 Table). The discriminatory power was 0.362–0.964 for the genes used in the MLST methods; the larger the gene, the greater the number of alleles and polymorphic sites. When comparing MLST schemes, scheme S3 showed that the highest nucleotide diversity, and this scheme generated 30 STs. However, two schemes (P1 and L1) and scheme S2 generated 31 and 32 STs from the 39 strains, respectively. Based on the above results, scheme S2 gave the greatest diversity of alleles and polymorphic sites. However, we determined that scheme P1 was better because the difference in diversity between S2 and P1 was not significant, and scheme P1 is already publicly accessible. If other housekeeping genes with higher discriminatory power were used for a modified scheme P1 in place of rpoD and/or tpiA (which have low discriminatory power), the scheme might identify greater genetic diversity.
Genomic analysis of amino acid biosynthesis in B. subtilis
To identify correlations between STs and amino acid production, first, in silico prediction of the amino acid biosynthesis pathways in B. subtilis was performed (Fig 3) using the complete genome sequences of the 38 B. subtilis strains and the type strain that were analyzed above. Thirty-one of the B. subtilis strains and the type strain appeared to contain loci coding for all the enzymes required for the biosynthesis of all amino acids. However, seven of the strains did not encode two or more of the biosynthesis genes hutH, hutI, and hutG that are required for conversion of histidine to glutamate. Glutamate is a precursor of glutamine, arginine, and proline; thus, these strains could not synthesize these amino acids de novo. However, the genes for biosynthesis of arginine, glutamine, and arginine from glutamate were retained, so those amino acids could be synthesized if glutamate were provided from outside the cell (Fig 3).
Red crosses indicate strains that do not possess hutH, hutI, and/or hutG. Genes and amino acids are depicted in green italics and blue boxes, respectively. The black arrows correspond to potential enzymatic reactions catalyzed by gene products encoded in the genome of B. subtilis.
Amino acid biosynthesis profiles of B. subtilis isolates grouped by ST
To examine correlations between the amino acid production-gene patterns and STs, we compared the STs and amino acid biosynthesis components using phylogenetic trees (S1 Fig). All the schemes were able to classify strains that did not possess genes for production of glutamate from histidine into STs and there was no significant difference between them, so scheme P1 was applied going forward. Using scheme P1, the seven strains lacking glutamate biosynthesis genes were identified into six STs: ST208 (strains SRCM101441 and SRCM103886), ST210 (SRCM103576), ST211 (SRCM103581), ST213 (SRCM100333), ST236 (SRCM103612), and ST240 (SRCM103697).
Fig 4 shows minimum spanning trees illustrating predicted genetic relationships among the 39 strains based on the four MLST schemes. Strains that do not possess the hutHUIG operon were distinguished from those that do, and the strains lacking these genes clustered together in the trees. It can be seen that S3 and L1 are bound from S3-10 and L1-22.
Numbers in the trees indicate STs. Circle sizes are proportional to the number of isolates belonging to that ST.
To confirm the correlation between apparent ability to synthesize glutamate from histidine and ST, we analyzed the presence of four genes, hutH, hutU, hutI, and hutG, involved in glutamic acid synthesis in B. subtilis genomes in public databases as of July 2022 (Table 2). By that time, 207 B. subtilis genomes had been registered (the 39 we analyzed above, plus another 168). Based on analysis of the genomes, 26 of the 168 strains did not possess hutG and hutI, and 28 of the strains did not possess hutH. Six strains among the 168 B. subtilis strains corresponded to ST208, and none of them had hutH, hutI, and hutG genes. Seventeen strains corresponded to ST4, which also lacked these three genes. The remaining six strains corresponded to different STs, among which new STs were included (strains DKU_NT_02, BS38, MH-1, PJ-7, JAAA, and FX-1 were assigned to ST207, ST209, ST212, ST214, ST245, and ST256, respectively). The data for ST208 and ST4 showed that there is a correlation between the absence of the genes for glutamate synthesis and ST. In the case of other STs, it will only be possible to draw conclusions when further data (i.e., more sequences) are accumulated.
General discussion
MLST is a useful tool for identifying genetic diversity within species [21]. Studies have investigated correlations between MLST and the presence or absence of virulence factors, differences between regions, and the presence or absence of antibiotic resistance [36–40]. Five MLST methods have been developed so far for B. subtilis and genus Bacillus. In this study, comparative analysis was performed between four of these MLST methods (omitting scheme S1) [22–26]. In addition, relationships between the amino acid biosynthesis genes of B. subtilis and the STs derived from MLST were analyzed.
MLST scheme P1 has the advantage of being available anywhere in the world because it uses a public database [22]. It is method of identifying genetic diversity based on seven housekeeping genes. In this study, MLST using scheme P1 derived 31 STs from 39 B. subtilis strains; this result was similar to that obtained using the other three MLST methods (S1 Fig). It was shown that six STs did not possess the hutHUIG operon, which encodes enzymes that produce glutamate from histidine; thus, there appears to be a correlation between amino acid biosynthesis and ST in B. subtilis. The rpoD gene used in scheme P1 is highly conserved (having few polymorphic sites), resulting in a low discriminatory power. If that gene were to be replaced with a gene that exhibits higher discriminatory power, more effective diversity results could be achieved using the modified scheme P1.
Scheme S2 was developed to compare the genotype with the secondary metabolites of B. subtilis [24]. MLST by scheme S2 uses nine housekeeping genes, and the entire housekeeping gene sequences are used; thus, the average gene size is about 1.6 kb, compared with about 500 bp in the other schemes. As a result, the genetic diversity appears greater than that indicated by the other MLST schemes. Nevertheless, the number of STs produced by scheme S2 was 32 (compared with 30 or 31 by the other methods), which indicated that the number and size of the genes used for typing did not contribute much to the genotyping. The discriminatory power of scheme S2 was higher than that of the other methods.
Scheme S3, which includes seven genes for typing, was developed to identify the sfp gene, which is responsible for the production of a lipopeptide biosurfactant, and genetic diversity in Bacillus species [25]. Using this scheme, 30 STs were derived from the 39 strains. Although this scheme showed the correlation between hutHUIG operon-free strains and ST, it did not show superiority over other methods.
MLST scheme L1 is a method built to distinguish between B. licheniformis and B. paralicheniformis, which have high genetic homogeneity [26]. It was expanded to re-identify B. velezensis [41], or B. subtilis from other Bacillus species [35]. Here, it was applied to identify genetic diversity within B. subtilis, and results similar to those from the other MLST methods were obtained. This scheme discriminated strains according to the presence or absence of the hutHUIG operon. However, the scheme uses genes gmk and tpi for typing—these genes show low polymorphism, so we suggest that more effective results could be obtained if genes with greater discriminatory power were used instead.
In conclusion, comparing four MLST schemes for B. subtilis shows that no one method is obviously superior to the others. However, scheme P1 has a significant advantage because it is already a well-established and publicly accessible system. We determined that there are correlations between ST and amino acid biosynthesis genes in B. subtilis. If further research is conducted, MLST can be used to secure desired amino acid-producing strains.
Supporting information
S1 Fig. Phylogenetic trees for the four MLST schemes.
The data were compared using simple matching coefficients and were clustered by the maximum likelihood method. Branches with bootstrap values of <50% have been collapsed. The scale the diagram is the pairwise distance expressed as the percentage dissimilarity.
https://doi.org/10.1371/journal.pone.0282092.s001
(DOCX)
S1 Table. Bacterial strains and genome accession numbers.
https://doi.org/10.1371/journal.pone.0282092.s002
(DOCX)
S2 Table. Sequence types of Bacillus subtilis using four MLST schemes, as well as the numbers of alleles.
https://doi.org/10.1371/journal.pone.0282092.s003
(DOCX)
S3 Table. Compositional characteristics of genes used in the four MLST schemes for B. subtilis.
https://doi.org/10.1371/journal.pone.0282092.s004
(DOCX)
References
- 1. Lee JH, Kim TW, Lee H, Chang HC, Kim HY. Determination of microbial diversity in meju, fermented cooked soya beans, using nested PCR-denaturing gradient gel electrophoresis. Lett Appl Microbiol. 2010;51(4):388–94. pmid:20704651.
- 2. Kim TW, Lee JH, Kim SE, Park MH, Chang HC, Kim HY. Analysis of microbial communities in doenjang, a Korean fermented soybean paste, using nested PCR-denaturing gradient gel electrophoresis. Int J Food Microbiol. 2009;131(2–3):265–71. pmid:19324443.
- 3. Chen T, Wang M, Li S, Wu Q, Wei H. Molecular identification of microbial community in surface and undersurface douchi during postfermentation. J Food Sci. 2014;79(4):M653–8. pmid:24621312.
- 4. Homma K, Wakana N, Suzuki Y, Nukui M, Daimatsu T, Tanaka E, et al. Treatment of natto, a fermented soybean preparation, to prevent excessive plasma vitamin K concentrations in patients taking warfarin. J Nutr Sci Vitaminol. 2006;52(5):297–301. pmid:17190098.
- 5. Onda T, Yanagida F, Tsuji M, Shinohara T, Yokotsuka K. Time series analysis of aerobic bacterial flora during miso fermentation. Lett Appl Microbiol. 2003;37(2):162–8. pmid:12859661.
- 6. Dash BK, Rahman MM, Sarker PK. Molecular identification of a newly isolated Bacillus subtilis BI19 and optimization of production conditions for enhanced production of extracellular amylase. Biomed Res Int. 2015;2015:859805. pmid:26180814.
- 7. Barbieri G, Albertini AM, Ferrari E, Sonenshein AL, Belitsky BR. Interplay of CodY and ScoC in the regulation of major extracellular protease genes of Bacillus subtilis. J Bacteriol. 2016;198(6):907–20. pmid:26728191.
- 8. Yi SH, Hong SP. Characteristics of bacterial strains with desirable flavor compounds from Korean traditional fermented soybean paste (doenjang). Molecules. 2021;26(16):5657. pmid:34443655.
- 9. Nguyen T, Nguyen CH. Determination of factors affecting the protease content generated in fermented soybean by Bacillus subtilis 1423. Energy Rep. 2020;6:831–6. https://doi.org/10.1016/j.egyr.2019.11.011.
- 10. Shukla S, Lee JS, Park HK, Yoo JA, Hong SY, Kim JK, et al. Effect of novel starter culture on reduction of biogenic amines, quality improvement, and sensory properties of doenjang, a traditional Korean soybean fermented sauce variety. J Food Sci. 2015;80(8):M1794–803. pmid:26147854.
- 11. Zhao CJ, Schieber A, Ganzle MG. Formation of taste-active amino acids, amino acid derivatives and peptides in food fermentations—A review. Food Res Int. 2016;89(Pt 1):39–47. pmid:28460929.
- 12. Dajanta K, Chukeatirote E, Apichartsrangkoon A. Improvement of thua nao production using protein-rich soybean and Bacillus subtilis TN51 starter culture. Ann Microbiol. 2012;62(2):785–95. https://doi.org/10.1007/s13213-011-0319-1.
- 13. Jeon HH, Jung JY, Chun BH, Kim MD, Baek SY, Moon JY, et al. Screening and characterization of potential Bacillus starter cultures for fermenting low-salt soybean paste (doenjang). J Microbiol Biotechnol. 2016;26(4):666–74. pmid:26718466.
- 14. Kim T, Heo S, Na HE, Lee G, Kim JH, Kwak MS, et al. Bacterial community of galchi-baechu kimchi based on culture-dependent and -independent investigation and selection of starter candidates. J Microbiol Biotechnol. 2022;32(3):341–7. pmid:35001009.
- 15. Zhang Q, Liu M, Ruan J. Integrated transcriptome and metabolic analyses reveals novel insights into free amino acid metabolism in Huangjinya tea cultivar. Front Plant Sci. 2017;8:291. pmid:28321230.
- 16. Frerot E, Chen T. Identification and quantitation of new glutamic acid derivatives in soy sauce by UPLC/MS/MS. Chem Biodivers. 2013;10(10):1842–50. pmid:24130027.
- 17. Rafiq S, Huma N, Pasha I, Sameen A, Mukhtar O, Khan MI. Chemical composition, nitrogen fractions and amino acids profile of milk from different animal species. Asian-Australas J Anim Sci. 2016;29(7):1022–8. pmid:26954163.
- 18. Amani E, Eskandari MH, Shekarforoush S. The effect of proteolytic activity of starter cultures on technologically important properties of yogurt. Food Sci Nutr. 2016;5(3):525–37. pmid:28572938.
- 19. Berdague JL, Monteil P, Montel MC, Talon R. Effects of starter cultures on the formation of flavour compounds in dry sausage. Meat Sci. 1993;35(3):275–87. pmid:22061221.
- 20. Maiden MCJ. Multilocus sequence typing of bacteria. Annu Rev Microbiol. 2006;60:561–88. pmid:16774461.
- 21. Boers SA, Reijden WA, Jansen R. High-throughput multilocus sequence typing: bringing molecular typing to the next level. PLoS One. 2012;7(7):e39630. pmid:22815712.
- 22. Kamada M, Hase S, Fujii K, Miyake M, Sato K, Kimura K, et al. Whole-genome sequencing and comparative genome analysis of Bacillus subtilis strains isolated from non-salted fermented soybean foods. PLoS One. 2015;10(10):e0141369. pmid:26505996.
- 23. Le XT, Pham DT, Pham TA, Tran TT, Khuat TH, Le HQ, et al. Exploration of genetic diversity of Bacillus spp. from industrial shrimp ponds in Vietnam by multi-locus sequence typing. Fish Aquat Sci. 2019;22(1):1–9. https://doi.org/10.1186/s41240-019-0132-5.
- 24. Boka B, Manczinger L, Kocsube S, Shine K, Alharbi NS, Khaled JM, et al. Genome analysis of a Bacillus subtilis strain reveals genetic mutations determining biocontrol properties. World J Microbiol Biotechnol. 2019;35(3):52. pmid:30868269.
- 25. Kabore D, Gagnon M, Roy D, Sawadogo-Lingani H, Diawara B, LaPointe G. Rapid screening of starter cultures for maari based on antifungal properties. Microbiol Res. 2018;207:66–74. pmid:29458870.
- 26. Jeong DW, Lee B, Lee H, Jeong K, Jang M, Lee JH. Urease characteristics and phylogenetic status of Bacillus paralicheniformis. J Microbiol Biotechnol. 2018;28(12):1992–8. pmid:30394048.
- 27. Jeong DW, Kim HR, Jung G, Han S, Kim CT, Lee JH. Bacterial community migration in the ripening of doenjang, a traditional Korean fermented soybean food. J Microbiol Biotechnol. 2014;24(5):648–60. pmid:24548930.
- 28. Feil EJ, Li BC, Aanensen DM, Hanage WP, Spratt . eBURST: inferring patterns of evolutionary descent among clusters of related bacterial genotypes from multilocus sequence typing data. J Bacteriol. 2004;186(5):1518–30. pmid:14973027.
- 29. Roman F, Iniguez AM, Yeo M, Jansen AM. Multilocus sequence typing: genetic diversity in Trypanosoma cruzi I (TcI) isolates from Brazilian didelphids. Parasites Vectors. 2018;11(1):107. pmid:29471851.
- 30. Agarwal A, Gupta S, Yadav AK, Nema RK, Ansari K, Biswas D. Molecular and phylogenetic analysis of Chikungunya virus in Central India during 2016 and 2017 outbreaks reveal high similarity with recent New Delhi and Bangladesh strains. Infect Genet Evol. 2019;75:103940. pmid:31247338.
- 31. Zolfo M, Tett A, Jousson O, Donati C, Segata N. MetaMLST: multi-locus strain-level bacterial typing from metagenomic samples. Nucleic Acids Res. 2017;45(2):e7. pmid:27651451.
- 32. Aziz RK, Bartels D, Best AA, DeJongh M, Disz T, Edwards RA, et al. The RAST server: rapid annotations using subsystems technology. BMC Genom. 2008;9:75. pmid:18261238.
- 33. Darzi Y, Letunic I, Bork P, Yamada T. iPath3.0: interactive pathways explorer v3. Nucleic Acids Res. 2018;46(W1):W510–W3. pmid:29718427.
- 34. Blom J, Albaum SP, Doppmeier D, Puhler A, Vorholter FJ, Zakrzewski M, et al. EDGAR: a software framework for the comparative analysis of prokaryotic genomes. BMC Bioinform. 2009;10:154. pmid:19457249.
- 35. Lee G, Heo S, Kim T, Na HE, Park J, Lee E, et al. Discrimination of Bacillus subtilis from other Bacillus species using specific oligonucleotide primers for the pyruvate carboxylase and shikimate dehydrogenase genes. J Microbiol Biotechnol. 2022;32(8):1011–6. https://doi.org/10.4014/jmb.2205.05014.
- 36. Martino ME, Fasolato L, Montemurro F, Rosteghin M, Manfrin A, Patarnello T, et al. Determination of microbial diversity of Aeromonas strains on the basis of multilocus sequence typing, phenotype, and presence of putative virulence genes. Appl Environ Microbiol. 2011;77(14):4986–5000. pmid:21642403.
- 37. Jeong DW, Kim HR, Lee JH. Genetic diversity of Staphylococcus equorum isolates from saeu-jeotgal evaluated by multilocus sequence typing. Antonie Van Leeuwenhoek. 2014;106(4):795–808. pmid:25103946.
- 38. Willems RJL, Hanage WP, Bessen DE, Feil EJ. Population biology of Gram-positive pathogens: high-risk clones for dissemination of antibiotic resistance. FEMS Microbiol Rev. 2011;35(5):872–900. pmid:21658083.
- 39. Miragaia M, Carrico JA, Thomas JC, Couto I, Enright MC, de Lencastre H. Comparison of molecular typing methods for characterization of Staphylococcus epidermidis: proposal for clone definition. J Clin Microbiol. 2008;46(1):118–29. pmid:17989196.
- 40. Lee JH, Shin D, Lee B, Lee H, Lee I, Jeong DW. Genetic diversity and antibiotic resistance of Enterococcus faecalis isolates from traditional Korean fermented soybean foods. J Microbiol Biotechnol. 2017;27(5):916–24. pmid:28237994.
- 41. Heo S, Kim JH, Kwak MS, Sung MH, Jeong DW. Functional annotation genome unravels potential probiotic Bacillus velezensis strain KMU01 from traditional Korean fermented kimchi. Foods. 2021;10(3):563. pmid:33803098.