Bacteriocins Prediction and Growth evaluation as a Function of Temperature and pH using Mathematical Modelling Approach

Our skin is home to many microorganisms, including viruses and fungi that make up the skin microbiota. Like those in the gut, skin microorganisms play a fundamental role in protecting against attack pathogens [1]. As the largest organ in the human body and the first line of defense, the skin is colonized by advantageous microorganisms and acts as an effective physical barrier to block pathogen invasion. In a few conditions where the barrier is damaged or when the balance between commensals and pathogens is out, skin ailments or even infections can occur. Human skin may be sebaceous, dry, or oily. However, most of the skin areas are dry, thus developing an unfavorable environment for the growth of microorganisms, whereas the oily and sebaceous regions favor their growth. Contemplating the structure of the microbiome in various regions of the skin is valuable for explaining the aetiology of common skin ailments such as the high frequency of dermatitis inside the elbow [2] and psoriasis on the outside of the elbow [3]. The human skin microbiome includes Acinetobacter, Corynebacterium, Propionibacterium, Pseudomonas, Staphylococcus, and Streptococcus [4].

Bacteriocins are ribosomal peptides with antimicrobial activity. All types of eubacteria and some archaea produce bacteriocins [5]. Bacteriocins are produced by both Gram-positive and Gram-negative bacteria. Gram-negative microbes produce large bacteriocins ranging from 449 to 629 amino acids [6], whereas gram-positive bacteria produce peptides with less than 70 amino acids [7]. Bacteriocins are characterized based on their peptide structure into three classes: class 1, class 2 (subclasses a, b, c, d), and class 3. They are active against other microbes, either inside similar species or crosswise across genera [8]. With respect to human health, bacteriocins have a limited target range, stability, lower lethality, and strong action, positioning them as feasible options or supplements to existing small antibiotics. The increase in the number of antimicrobial- resistant pathogens and the development of awareness regarding the significance of the microbiome in human health makes the search for bacteriocins more significant [9]. There is an increasing understanding that bacteriocins are key representatives of probiotic strains used to avoid or constrain the growth of diarrheal pathogens. Probiotics are health-enhancing bacteria employed in various industrial settings to improve human health. Probiotic strains are also used to treat a number of diseases, including mucosal surface infections, as alternatives to antibiotics in several ailments, and as antimicrobial agents. Probiotic bacteria such as bifidobacteria, enteric, lactic acid bacteria, Saccharomyces, and streptococci produce bacteriocins as therapeutic agents [10,11]. Some bacteriocins, the most striking Nisin, have a long history of utilization as additives in the food industrial settings, Pyocins (bacteriocin) produced by Pseudomonas aeruginosa, are used to treat several infections, and Streptococcal bacteriocins are used as probiotics and in the resistance [12], these antimicrobials are additionally accepting wider consideration as a potential treatment instead of using antibiotics [13-20]. A few bacteriocins of the skin microbiota and their mechanisms of action are shown in Table 1.

Table 1: Key Bacteriocins reported to be produced in Skin Microbiome.

As only a limited number of bacteriocins have been identified and reported in the skin microbiota, there is great room for exploration regarding the prediction and characterization of novel bacteriocins from the skin microbiome. Bacteriocins can be identified using various conventional and Insilco methods [21]. To date, in silico bacteriocin screening procedures have prompted the distinguishing proof of numerous novel lantibiotics (class 1 bacteriocins, which act as antibiotics) [22], microcin (very small bacteriocin, removes enteric pathogens from the body) [23], and the sactibiotic (subclass of bacteriocins, plays a role in post-translational modification) [24]. In various examples, standard BLAST-based methodologies have been utilized to recognize bacteriocins, and the BAGEL bacteriocin mining tool has become an especially important asset [25]. BAGEL consolidates the direct mining of structural genes with aberrant mining for genes related to bacteriocin. This is especially helpful for distinguishing peptides that experience critical post-translational alterations, such as those in antibiotics.

Here, we present a new methodology for distinguishing bacteriocins with antimicrobial activity. Our approach is to predict novel bacteriocins and mine microbial genomes to identify genes and ORFs consisting of novel bacteriocins and modelling their production at constant pH and temperature. This modelling will also enable the prediction of the behaviour of microbes at a particular pH and temperature.

In this section, we propose a novel procedure for bacteriocin prediction. This method is based on the mining of microbial genomes to unravel new insights into the genomes and to predict novel bacteriocins can be used for several purposes. The details of this methodology are presented in Figure 1.

Information Retrieval and Screening of Bacterial Genomes

Information about the skin microbes was obtained from the literature [26], and their whole chromosomal and plasmid genomes were retrieved from the NCBI genome browser1 in Fasta format. It contains detailed genomic information, including annotations, assemblies, chromosomes, maps, and sequences.

Identification of Potential Genes and Operons

Potential genes and operons of each microbial strain were identified using the FGenesB tool. This is the most accurate ab initio tool for prokaryotic gene prediction. Its prediction calculation depends on the Markov chain models of coding regions, translation stops, and start sites. The program utilizes specific parameters for the genome learned by the FGENESB-prepare script [27].

Detection of Open reading frames and Promoter Regions

The Open reading frames (ORF) of the microbial species were detected using the ORF Finder tool of NCBI. The minimum ORF length was set to 75 nucleotides as an arbitrary threshold, and the ATG codon was set as the start of each ORF. The promoter regions were identified using the BTSS Finder tool. The promoter is a region of the chromosome that determines where and how transcription of a specific transcript is initiated. Promoter acknowledgment is vital for characterizing the transcription units responsible for gene regulation and particular pathways. The BTSS Finder is a unique tool for the prediction of putative promoter regions in prokaryotes, with an accuracy rate of 86 – 93 % [28].

Identification of Histidine Kinases and Response Regulators

Histidine kinase and response regulator genes in each strain of skin microbes were identified using the P2CS tool [29]. P2CS is a specified database for the two-component systems of prokaryotes, the Two-component system (TCS) is a signal transduction pathway found in each organism, enabling versatile reactions to changes in environmental conditions. However, they are mostly present in microorganisms, where they control various aspects of bacterial digestion, such as cell separation and morphogenesis, digestion, virulence, motility, and biofilm formation. These frameworks are traditionally depicted as the association of two proteins, histidine kinases and response regulators [30].

Identification and validation of Bacteriocins Genes

A typical issue in the annotation of ORFs is the distinguishing proof of genes that are functionally similar to each other but have no sequence homology. This is especially the case for bacteriocins, where ORFs covering bacteriocin genes are regular biosynthetic genes. Using this information, structural bacteriocins could be identified. For this purpose, the BAGEL3 [31] web server was used in this study to distinguish putative bacteriocin ORFs in DNA sequences using novel learning-based bacteriocin databases and motif databases.

Inference of Genome-Genome Distance

Genome sequence data can be utilized to compare similitude among various species, which is also helpful for the assurance of in ter-genomic distances between various species. Therefore, keeping this thing in mind, the Genome-Genome distance among all the skin microbial strains was calculated by the GGDC tool [32].

Modelling The Growth of Bacteria and Bacteriocins Expression

The bacteriocin production model for the aforementioned six bacteriocin-producing bacterial strains was developed in the Sim Biology application of MATLAB, and the growing and resting phases of bacteria were also designed to determine bacteriocin production, and the effects of temperature and pH on bacteriocin growth were also generated. It has been revealed that to improve the simulations of the model and to completely understand the expression of each model component, a particular set of parameters is required [33]. For this reason, parameters including half-life, molecular weight, isoelectric point, degradation and elimination rates of bacteria, required temperature, and pH for their growth were obtained from the Bactibase database and literature [34-37], and few parameters were assumed. These parameters are essential for exact simulations of the model.

Parameters Evaluation and Mathematical Formulation

The parameters were evaluated through simulations using the ordinary differential equation (ODE) solver tool of MATLAB. This tool works on the Runge-Kutta estimation strategy in view of unequivocal cycles for the discretization of ODE solutions. After parameter evaluation of the bacteriocin growth mode, differential equations were generated. The model was simulated using the fminsearch algorithm. After parameter evaluation, simulations were performed to determine the rate of expression of bacteriocins and the growth of bacteria in several pH and temperature ranges.

Procedure of Modelling

First, the maximum bacterial growth rate was determined through model simulations at different pH and temperature values. The exponential rate of bacteriocin production and resting phase of the bacteria were also determined. The resting phase was modelled to check for bacterial growth. When the resting phase (in hours) ends, the species starts their growing phase as a function of time, both the resting\ and growing phases are shown by Equations 1 and 2.

In equations 1 and 2 Bmax shows the maximum production of bacteriocins in both phases under normal pH In the absence of any inhibitory effect, krs shows the response rate of bacteria to a slight change in pH and temperature; kdeath indicates the death of bacteria due to the presence of any inhibitory object, whereas the change in growth rate due to the effect of pH and temperature is shown by EFFECT. The bacteriocin production rate is shown in terms of differential equations 3 and 4 as follows:

kdeg shows the degradation rate of bacteriocins after their maximum production, due to changes in temperature and pH, or due to the presence of inhibitory substances.

The microbial species that form the skin microbiota, retrieved from the literature, include Acinetobacter Johnsonii, Corynebacterium, Pseudomonas aeruginosa, Propionibacterium acnes, Staphylococcus aureus, Staphylococcus epidermidis, Staphylococcus warneri, Streptococcus mitis and Streptococcus pyogenes. Acinetobacter johnsonii consists of 1 chromosomal and eight plasmid genomes; Staphylococcus epidermidis has one chromosomal and two plasmid genomes, whereas other microbes have only a single chromosomal genome. The number of potential genes, transcription units, and operons predicted for each microbial genome are presented in Table 2.

Table 2: The number of genes, transcription units, and operons in skin microbial flora predicted through FgenesB server [FgenesB server takes genomic DNA as input and predicts the genes, operons, and their location in the genome.

The genes were predicted because it is the most important step in the understanding of the genome of a species; once the genes are known, the functions of the organisms can be determined, predicting the functions of the genes, and affirming that gene prediction is still challenging and the most demanding task in vivo experimentation [33]. Similarly, the operons were identified, as it is also important to consider operon structures as an essential family among the functional genes. In addition, these units regularly appear in different genomes and perform exceedingly compartmentalized actions in biochemical pathways [34]. Understanding the system that manages gene expression and distinguishing the key administrative components that guide gene expression are noteworthy challenges in molecular biology. Promoters control the expression of each transcriptional unit or gene in the genome [35]. In the present study, the ORFs of each species were also determined. In prokaryotes, genes act as an essential hierarchical unit at the genome level, and compared to eukaryotes, the coding density of bacterial genomes is extremely high [36]. The genome of a regular bacterium is usually 106– 107 base pairs (bp) long, containing approximately 103 to 104 genes. However, the aggregate number of conceivable ORFs is typically in the range 104–105.Although the number and average length of ORFs may differ, microorganisms share normal characteristics due to the distribution of their open reading frame length, which is related to their GC content. Most prokaryotes are considered firmly related based on the same number of ORFs and their lengths [37]. The identified promoter regions and ORFs for each microbial genome are listed in Table 3.

Table 3: The number of promoter regions and ORFs predicted in microbes through BTS finder and ORF finder.

Histidine kinases and response regulators were determined for each microbial strain in the skin microbiome. Both histidine kinase and a response regulator form a two-component system (TCS), which is a significant signal transduction system in microor ganisms [38]. TCS performs important functions in bacteria, such as bacteriocin production, cell-cell correspondence, adjustment to conditions, and pathogenesis on account of pathogens. Bacterial species usually respond to environmental changes via TCS. Because of their absence in mammals, TCS proteins are viewed as a potential target for developing new antibiotics [39]. Drugs that inhibit TCSs target upstream functions that control these basic proteins. Therefore, anti-TCS drugs work in a different manner than regular medications and are likely to be compelling against drug-resistant bacterial pathogens. Because TCS proteins are missing in humans, drugs that target TCSs may have fewer toxic effects [40]. Therefore, the purpose of identifying TCS was to determine the genes in microbes that produce TCS proteins that may produce bacteriocins. Histidine kinases and response regulators of skin microbial flora retrieved from the P2CS database are displayed in Table 4.

Table 4: The number of Histidine kinases and Response regulators of skin microbial strains retrieved through P2CS database.

From Table 4, it is clear that all six bacteriocin-producing strains produced TCS proteins. Bacteriocins are exceptionally strong bactericidal agents. As mentioned previously, some of them have a restricted range of action, focusing only on firmly related species, especially those produced by lactic acid bacteria, which are active against gram-positive bacteria [34]. Bacteriocins of S. epidermidis are accepted to have advantageous effects on human skin because they may confine the colonization of pathogenic life forms [41-46].

The ability of these S. epidermidis bacteriocins to eliminate human pathogens has sustained therapeutic interest for bacteriocins as a treatment choice [42,44]. Novel bacteriocins of only six bacterial species, Propionibacterium acnes, Pseudomonas auroginosa, Streptococcus pyogenes, Staphylococcus epidermidis, Staphylococcus Warneri, and Staphylococcus aureus are identified in this research study. The predicted bacteriocins are shown in Figure 2. Figure 2 demonstrates novel bacteriocins predicted in skin microbiome along with other genes that play important roles in immunity and regulation through Bagel 3. Only six microbes were found to encode novel bacteriocins (Figure 2). The advantage of Bagel 3 is that it predicts novel bacteriocins by genome mining and aligning [31]. The bacteriocin produced by P. acnes, S. aureus, S, epidermidis, S. pyogens, and P. aeuroginosa shows similarity to Colicin-10 -, which acts as a bacterial killer by forming ion channels. Pyocins-S2 - inhibits lipid synthesis and breaks down chromosomal DNA, S. warneri produces bacteriocins that act as a probiotic against resistant bacterial species. The genome-genome distance and phylogenetic distance among the skin microbial flora were also calculated because the prokaryotic scientific classification has been generally viewed as comprising three interrelated parts: nomenclature, classification, and characterization [43-47]. The organizing rule of microbial scientific categorization is to group related life forms together, which differs from several other groups. The genome-genome distance based on the DNA-DNA hybridization (DDH) method is presented in Table 5. From Table 5, it is clear that all the microbial species are related to each other, with a probable distance ranging from 0.9 – 1.0. Figures 3 shows the high similarity among several species such as Staphylococcus aureus and Staphylococcus warneri, Streptococcus pyogenes and Streptococcus mitis, similarly, Pseudomonas aeruginosa and Acinetobacter johnsonii are highly similar to each other. The remaining Staphylococcus epidermidis and Propionibacterium acnes were closely related up to 99 % to other species. The results of this scientific study have provided a novel dataset of predicted bacteriocins, along with phylogenetic distance and genomic information of the skin microbiome. For the modelling of the bacteriocins, their growth model was developed, as shown in Figure 4.

where some of the parameters were assumed, while others were evaluated by the ODE solver and Fminsearch with the formula used in Equation 5.

Table 5: The genome-genome distance calculation for species based on DDH technique.

[𝑡, 𝑦] = 𝑜𝑑𝑒45([𝑡, 𝑦] = 𝑜𝑑𝑒45(𝑜𝑑𝑒𝑓𝑢𝑛, 𝑡𝑠𝑝𝑎𝑛, 𝑦0, 𝑜𝑝𝑡𝑖𝑜𝑛𝑠) eq. 5

where @fun is the function vector showing the action of bacterial species, tspan symbolizes the time consumed in the simulation process, and y0 is the state variable for differential equations. The fminsearch function reduces the error functions and assigns the evaluated values to each parameter. In Figure 4, the production of bacteriocin is represented by the growth rate constant kgrowth, where the growing phase represents the time when bacteriocin production acquires its maximum development in a particular medium; the growth reaction is controlled by the kdeath+effect rule, according to which, if any bacterial species do not receive proper pH and temperature or if any substance causes an inhibitory effect on bacterial growth, resulting in the death of a species and the bacteriocins will not be produced; on the other hand, after production of bacteriocins, the bacterial species moves to the resting phase. Some of the parameters used for this model were assumed and the others were evaluated. The assumed parameters and their evaluated values for the modelling of bacteriocins are listed in Table 6.

Table 6: The parameters and their values used for the bacteriocins production model.

For each bacterial species, several variants were developed with different values of K-deg, K-e, pI, pH, and temperature. However, the other parameters were kept constant. A relationship between pH and temperature was established, and the parameters were fitted to the model for a particular pH and temperature. The production rate of bacteriocin was related to the temperature and pH, and the function f(pH), assessed at different pH values (pH 0–8) and temperature (0-50°C), was modelled for each bacteriocin-producing species. For modelling the growth of bacteriocins of Propionibacterium acnes the pH was set to 6.5, Temp at 400C, for Pseudomonas auroginosa the values are 7.0 pH and temperature 400C, Streptococcus pyogenes requires 370C and 7.6 pH, for Staphylococcus epidermidis the pH was set to 6.8 and temp 300C, 4500C and 6.5 pH for Staphylococcus Warneri, and for Staphylococcus aureus the parameter values were 370C and pH 7.0. The simulation results for bacteriocin growth are shown in Figure 5 and 6.

In Figure 5 and Figure 6, the X-axis represents the number of hours required for bacterial species to grow and produce their desired bacteriocins, and the Y-axis shows the rate of bacteriocin production. Yellow lines represent the growth rate of bacteriocins, blue lines show the growth rate of bacteria, and red lines indicate the resting phase of bacteria after the production of bacteriocins. From Figure 5 and Figure 6, it is clear that, at a particular pH and temperature, all the bacteria started to produce bacteriocins after 4 hours and the concentration reached 10 × 107 to 10 × 108. A slight reduction in pH or temperature results in a decrease in the production of bacteriocins. Propionibacterium and Psuodomonas produced small amounts of bacteriocins, whereas the bacteriocin production rates of staphelyococi and Streptococi were observed to be better. Therefore, for the appropriate production of bacteriocins, the aforementioned temperature and pH can be applied to specific bacteria.

In this study, the genomes of skin microbial flora were gathered and mined to unravel interesting pieces of information and bacteriocins. It has been reported that 30–99 % of microorganisms can produce bacteriocins. This in silico study identified only six microbial strains that produced more than one novel bacteriocin, along with several other genes important for immunity and regulation. Simulation analysis of the bacteriocin growth model provided the specific pH and temperature for the appropriate production of bacteriocins. These bacteriocins have diverse modes of action that may be helpful in developing a mixture of therapeutics with synergistic effects against target living organisms. The evolutionary development of bacteriocins from other microorganisms can also be predicted using the same approach.

AM: Assisted in the data collection, prepared and drafted the manuscript, contributed to the analysis and interpretation of data for the manuscript and revised the work for critically important intellectual content.

AJ: conceived the study, its design and methodology, contributed to the analysis, data interpretation, manuscript writing and revised the work for critically important intellectual content and acted as the corresponding author.

Acknowledgements: The author expresses sincere gratitude to LCUCK, located in London, United Kingdom, for providing the necessary platform to carry out this work. Additionally, I would like to thank Ms J. Kaur for meticulously proofreading the whole manuscript draft and offering valuable feedback.

The author reiterates that there are no conflicting interests that could potentially compromise the integrity of this research.

This work did not get any financial support.

This research does not include the participation of any human or animal participants.

The methods carried out in the studies involving human participants followed the moral standards set by the institutional and national research committee, the 1964 Helsinki Declaration and its subsequent revisions or similar ethical standards.

4,488 words, excluding references

href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7356928/#:~:text=It%20was%20demonstrated%2C%20in%20fact,based%20methods%20of%20human%20identification.">Tozzo P, D’Angiolella G, Brun P, Castagliuolo I, Gino S, Caenazzo L (2020) Skin Microbiome Analysis for Forensic Human Identification: What Do We Know So Far?. Microorganisms 8(6): 873.
1. Timm CM, Loomis K, Stone W, Mehoke T, Brensinger B, et al. (2020) Isolation and characterization of diverse microbial representatives from the human skin microbiome. Microbiome 8(1): 1-2.
2. Grice EA (2015) The intersection of microbiome and host at the skin interface: genomic- and metagenomic-based insights. Genome Res 25(10): 1514-1520.
3. Byrd AL, Belkaid Y, Segre JA (2018) The human skin microbiome. Nature Reviews Microbiology 16(3): 143-155.
4. Cotter PD, Ross RP, Hill C (2013) Bacteriocins-available alternative to antibiotics? Nat Rev Microbiol 11: 95-105.
5. Zheng J, Gänzle MG, Lin XB, Ruan L, Sun M (2015) Diversity and dynamics of bacteriocins from human microbiome. Environmental Microbiology17(6): 2133-2143.
6. Radaic A, de Jesus MB, Kapila YL (2020) Bacterial anti-microbial peptides and nano-sized drug delivery systems: The state of the art toward improved bacteriocins. Journal of Controlled Release 321: 100-118.
7. Simons A, Alhanout K, Duval RE (2020) Bacteriocins, Antimicrobial Peptides from Bacterial Origin: Overview of Their Biology and Their Impact against Multidrug-Resistant Bacteria. Microorganisms 8(5): 639.
8. Nes IF History, current knowledge, and future directions on bacteriocin research in lactic acid bacteria. InProkaryotic Antimicrobial Peptides 2011; (pp. 3-12). Springer, New York, NY.
9. Hegarty JW, Guinane CM, Ross RP, Hill C, Cotter PD (2016) Bacteriocin production: a relatively unharnessed probiotic trait?. F1000Research 5: 2587.
10. Gillor O, Etzion A, Riley MA (2008) The dual role of bacteriocins as anti-and probiotics. Applied microbiology and biotechnology 81(4): 591-606.
11. Deegan LH, Cotter PD, Hill C, Ross P (2006) Bacteriocins: Biological tools for bio-preservation and shelf-life extension. Int Dairy J 16(9): 1058-1071.
12. Ghequire MGK, Kemland L, Anoz-Carbonell E, Buchanan SK, De Mot R, et al. (2017) A Natural Chimeric Pseudomonas Bacteriocin with Novel Pore-Forming Activity Parasitizes the Ferrichrome Transporter Mbio 8(1): e01961-16.
13. Fujimura S, Nakamura T (1978) Purification and Properties of a Bacteriocin-Like Substance (Acnecin) of Oral Propionibacterium acnes. Antimicrobial Agents and Chemotherapy 14(6): 893-898.
14. Aung MS, San T, Urushibara N, San N, Oo WM, et al. (2020) Molecular Characterization of Methicillin-

Susceptible and-Resistant Staphylococcus aureus Harboring Panton-Valentine LeukocidinEncoding Bacteriophages in a Tertiary Care Hospital in Myanmar. Microbial Drug Resistance 26(4): pp.360-367.

15. Garcia-Gutierrez E, Walsh CJ, Sayavedra L, Diaz-Calvo T, Thapa D, et al. (2020) Genotypic and Phenotypic Characterization of Fecal Staphylococcus epidermidis Isolates Suggests Plasticity to Adapt to Different Human Body Sites. Frontiers in Microbiology 11: 688.
16. Prema P, Bharathy S, Palavesam A, Sivasubramanian M, Immanuel G (2006) Detection purification, and efficacy of warnerin produced by Staphylococcus warneri. World Journal of Microbiology and Biotechnology,2006; 22(8): 865-872.
17. Lux T, Nuhn M, Hakenbeck R, Reichmann P (2007) The diversity of Bacteriocins and Activity Spectrum in Streptococcus pneumoniae. Journal of Bacteriology 189(21): 774-7751.
18. Armstrong BD, Herfst CA, Tonial NC, Wakabayashi AT, Zeppa JJ, et al. (2016) Identification of a two-component Class IIb bacteriocin in Streptococcus pyogenes by recombinase-based in vivo expression technology. Scientific Reports 6(1).
19. Walsh CJ, Guinane CM, Hill C, Ross RP, O’Toole PW, et al. (2015) In silico identification of bacteriocin gene clusters in the gastrointestinal tract, based on the Human Microbiome Project’s reference genome database. BMC Microbiology 15.  
20. Marsh AJ, O’Sullivan O, Ross RP, Cotter PD, Hill C (2010) In silico analysis highlights the frequency and diversity of type 1 lantibiotic gene clusters in genome sequenced bacteria. BMC Genomics 11:679.
21. Scholz R, Molohon KJ, Nachtigall J, Vater J, Markley AL, et al. (2011) Plantazolicin, a novel microcin B17/streptolysin S-like natural product from Bacillus amyloliquefaciens FZB42. J Bacteriol 193(1): 215-224.
22. Murphy K, O’Sullivan O, Rea MC, Cotter PD, Ross RP, et al. (2011) Genome Mining for Radical SAM Protein Determinants Reveals Multiple Sactibiotic-Like Gene Clusters. PLoS One 6(7): e20852.
23. van Heel AJ, de Jong A, Montalban-Lopez M, Kok J, Kuipers OP (2013) BAGEL3: Automated identification of genes encoding bacteriocins and (non-) bactericidal posttranslationally modified peptides. Nucleic Acids Res 41(Web Server issue): W448-W453.
24. Chiller K, Selkin BA, Murakawa GJ (2001) Skin Microflora and Bacterial Infections of the Skin. Journal of Investigative Dermatology Symposium Proceedings 6: 170-174.
25. V Solovyev, A Salamov (2011) Automatic Annotation of Microbial Genomes and Metagenomic Sequences. In Metagenomics and its Applications in Agriculture, Biomedicine and Environmental Studies (Ed. R.W. Li), Nova Science Publishers p. 61-78.
26. Shahmuradov IA, Mohamad Razali R, Bougouffa S, Radovanovic A, Bajic VB (2016) bTSSfinder: a novel tool for the prediction of promoters in cyanobacteria and Escherichia coli Bioinformatics 33(3):334-340.
27. Barakat M, Ortet P, Whitworth DE, (2011) P2CS: a database of prokaryotic two-component systems. Nucleic Acids Research 39: D771-D776.
28. Behr S, Fried L, Jung K (2014) Identification of a Novel Nutrient-Sensing Histidine Kinase/Response Regulator Network in Escherichia coli. Journal of Bacteriology 196(11): 7.
29. de Jong A, van Hijum SAFT, Bijlsma JJE, Kok J, Kuipers OP et al. (2006) BAGEL: a webbased bacteriocin genome mining tool. Nucleic Acids Research 34: W273-W279.
30. Auch AF, Klenk HP, Göker M (2010) Standard operating procedure for calculating genometo-genome distances based on high-scoring segment pairs. Standards in Genomic Sciences 2(1): 142-148.
31. Tian T, Song J (2012) Mathematical modelling of the MAP kinase pathway using proteomic datasets. PLoS One. 2012;7(8): e42230.
32. Achermann Y, Goldstein EJ, Coenye T, Shirtliff ME (2014) Propionibacterium acnes: from commensal to opportunistic biofilm-associated implant pathogen. Clinical microbiology reviews 27(3): 419-440.
33. Pornsunthorntawee O, Wongpanit P, Chavadej S, Abe M, Rujiravanit R (2008) Structural and physicochemical characterization of a crude biosurfactant produced by Pseudomonas aeruginosa SP4 isolated from petroleum-contaminated soil. Bioresource Technology 99(6): 1589-1595.
34. Kloos WE, Schleifer KH (1975) Isolation and Characterization of Staphylococci from Human Skin II. Descriptions of Four New Species: Staphylococcus warneri, Staphylococcus capitis, Staphylococcus hominis, and Staphylococcus simulans1. International Journal of Systematic and Evolutionary Microbiology 25(1): 62-79.   
35. Alomar J, Lebert A, Montel MC (2008) Effect of temperature and pH on growth of Staphylococcus aureus in co-culture with Lactococcus garvieae. Current Microbiology 56(4): 408-412.
36. Roy D Sleator (2010) An overview of the current status of eukaryote gene prediction strategies". Gene 461: 1-4.
37. Zheng Y, Szustakowski JD, Fortnow L, Roberts RJ, Kasif S (n.d.). Computational Identification of Operons in Microbial Genomes 12(8): 1221-1230.
38. Rangannan V, Bansal M (2010) High-quality annotation of promoter regions for 913 bacterial genomes. Bioinformatics 26(24): 3043-3050.
39. Patthy L (1999) Genome evolution and the evolution of exon-shuffling–a review. Gene 238: 103-114.
40. Mir K, Neuhaus K, Scherer S, Bossert M, Schober S (2012) Predicting Statistical Properties of Open Reading Frames in Bacterial Genomes. PLoS ONE 7(9): e45103.
41. Jung K, Fried L, Behr S, Heermann R (2012) Histidine kinases and response regulators in networks. Current Opinion in Microbiology 15(2): 118-124.
42. Wang S (2012) Bacterial Two-Component Systems: Structures and Signalling Mechanisms. In C. Huang (Ed.), Protein Phosphorylation in Human Health. InTech.
43. Guarnieri MT, Zhang L, Shen J, Zhao R (2008) The Hsp90 inhibitor radicicol interacts with the ATPbinding pocket of bacterial sensor kinase PhoQ. J Mol Biol 379(1): 82-93.
44. Martínez B, Rodríguez A, Suárez E (2016) Antimicrobial Peptides Produced by Bacteria: The Bacteriocins. In T. G. Villa & M. Vinas (Eds.), New Weapons to Control Bacterial Growth (pp. 15–38). Cham: Springer International Publishing.
45. Christensen GJM, Brüggemann H (2014) Bacterial skin commensals and their role as host guardians. Beneficial Microbes 2014;5(2), 201-215.
46. Cotter PD, Ross RP and Hill C (2013) Bacteriocins-a viable alternative to antibiotics? Nature Reviews Microbiology 11: 95-105.