Identification of Novel Genomic Islands in Liverpool Epidemic Strain of Pseudomonas aeruginosa Using Segmentation and Clustering

Pseudomonas aeruginosa is an opportunistic pathogen implicated in a myriad of infections and a leading pathogen responsible for mortality in patients with cystic fibrosis (CF). Horizontal transfers of genes among the microorganisms living within CF patients have led to highly virulent and multi-drug resistant strains such as the Liverpool epidemic strain of P. aeruginosa, namely the LESB58 strain that has the propensity to acquire virulence and antibiotic resistance genes. Often these genes are acquired in large clusters, referred to as "genomic islands (GIs)." To decipher GIs and understand their contributions to the evolution of virulence and antibiotic resistance in P. aeruginosa LESB58, we utilized a recursive segmentation and clustering procedure, presented here as a genome-mining tool, "GEMINI." GEMINI was validated on experimentally verified islands in the LESB58 strain before examining its potential to decipher novel islands. Of the 6062 genes in P. aeruginosa LESB58, 596 genes were identified to be resident on 20 GIs of which 12 have not been previously reported. Comparative genomics provided evidence in support of our novel predictions. Furthermore, GEMINI unraveled the mosaic structure of islands that are composed of segments of likely different evolutionary origins, and demonstrated its ability to identify potential strain biomarkers. These newly found islands likely have contributed to the hyper-virulence and multidrug resistance of the Liverpool epidemic strain of P. aeruginosa.

A hypervariable genomic island identified in clinical and environmental Mycobacterium avium subsp. hominissuis isolates from Germany

Mycobacterium avium subsp. hominissuis (MAH) is an opportunistic human pathogen widespread in the environment. Genomic islands (GI)s represent a part of the accessory genome of bacteria and influence virulence, drug-resistance or fitness and trigger bacterial evolution. We previously identified a novel GI in four MAH genomes. Here, we further explored this GI in a larger collection of MAH isolates from Germany (n=41), including 20 clinical and 21 environmental isolates. Based on comparative whole genome analysis, we detected this GI in 39/41 (95.1%) isolates. Although all these GIs integrated in the same insertion hotspot, there is high variability in the genetic structure of this GI: eight different types of GI have been identified, designated A-H (sized 6.2-73.3kb). These GIs were arranged as single GI (23/41, 56.1%), combination of two different GIs (14/41, 34.1%) or combination of three different GIs (2/41, 4.9%) in the insertion hotspot. Moreover, two GI types shared more than 80% sequence identity with sequences of M. canettii, responsible for Tuberculosis. A total of 253 different genes were identified in all GIs, among which the previously documented virulence-related genes mmpL10 and mce. The diversity of the GI and the sequence similarity with other mycobacteria suggests cross-species transfer, involving also highly pathogenic species. Shuffling of potential virulence genes such as mmpL10 via this GI may create new pathogens that can cause future outbreaks.

Pangenome Evolution in the Marine Bacterium Alteromonas

We have examined a collection of the free-living marine bacterium Alteromonas genomes with cores diverging in average nucleotide identities ranging from 99.98% to 73.35%, i.e., from microbes that can be considered members of a natural clone (like in a clinical epidemiological outbreak) to borderline genus level. The genomes were largely syntenic allowing a precise delimitation of the core and flexible regions in each. The core was 1.4 Mb (ca. 30% of the typical strain genome size). Recombination rates along the core were high among strains belonging to the same species (37.7-83.7% of all nucleotide polymorphisms) but they decreased sharply between species (18.9-5.1%). Regarding the flexible genome, its main expansion occurred within the boundaries of the species, i.e., strains of the same species already have a large and diverse flexible genome. Flexible regions occupy mostly fixed genomic locations. Four large genomic islands are involved in the synthesis of strain-specific glycosydic receptors that we have called glycotypes. These genomic regions are exchanged by homologous recombination within and between species and there is evidence for their import from distant taxonomic units (other genera within the family). In addition, several hotspots for integration of gene cassettes by illegitimate recombination are distributed throughout the genome. They code for features that give each clone specific properties to interact with their ecological niche and must flow fast throughout the whole genus as they are found, with nearly identical sequences, in different species. Models for the generation of this genomic diversity involving phage predation are discussed.

The Dynamics of Genetic Interactions between Vibrio metoecus and Vibrio cholerae, Two Close Relatives Co-Occurring in the Environment

Vibrio metoecus is the closest relative of Vibrio cholerae, the causative agent of the potent diarrheal disease cholera. Although the pathogenic potential of this new species is yet to be studied in depth, it has been co-isolated with V. cholerae in coastal waters and found in clinical specimens in the United States. We used these two organisms to investigate the genetic interaction between closely related species in their natural environment. The genomes of 20 V. cholerae and 4 V. metoecus strains isolated from a brackish coastal pond on the US east coast, as well as 4 clinical V. metoecus strains were sequenced and compared with reference strains. Whole genome comparison shows 86-87% average nucleotide identity (ANI) in their core genes between the two species. On the other hand, the chromosomal integron, which occupies approximately 3% of their genomes, shows higher conservation in ANI between species than any other region of their genomes. The ANI of 93-94% observed in this region is not significantly greater within than between species, meaning that it does not follow species boundaries. Vibrio metoecus does not encode toxigenic V. cholerae major virulence factors, the cholera toxin and toxin-coregulated pilus. However, some of the pathogenicity islands found in pandemic V. cholerae were either present in the common ancestor it shares with V. metoecus, or acquired by clinical and environmental V. metoecus in partial fragments. The virulence factors of V. cholerae are therefore both more ancient and more widespread than previously believed. There is high interspecies recombination in the core genome, which has been detected in 24% of the single-copy core genes, including genes involved in pathogenicity. Vibrio metoecus was six times more often the recipient of DNA from V. cholerae as it was the donor, indicating a strong bias in the direction of gene transfer in the environment.

Evolutionary Divergence of Aggregatibacter actinomycetemcomitans

Gram-negative facultative Aggregatibacter actinomycetemcomitans is an oral pathogen associated with periodontitis. The genetic heterogeneity among A. actinomycetemcomitans strains has been long recognized. This study provides a comprehensive genomic analysis of A. actinomycetemcomitans and the closely related nonpathogenic Aggregatibacter aphrophilus. Whole genome sequencing by Illumina MiSeq platform was performed for 31 A. actinomycetemcomitans and 2 A. aphrophilus strains. Sequence similarity analysis shows a total of 3,220 unique genes across the 2 species, where 1,550 are core genes present in all genomes and 1,670 are variable genes (accessory genes) missing in at least 1 genome. Phylogenetic analysis based on 397 concatenated core genes distinguished A. aphrophilus and A. actinomycetemcomitans. The latter was in turn divided into 5 clades: clade b (serotype b), clade c (serotype c), clade e/f (serotypes e and f), clade a/d (serotypes a and d), and clade e' (serotype e strains). Accessory genes accounted for 14.1% to 23.2% of the A. actinomycetemcomitans genomes, with a majority belonging to the category of poorly characterized by Cluster of Orthologous Groups classification. These accessory genes were often organized into genomic islands (n = 387) with base composition biases, suggesting their acquisitions via horizontal gene transfer. There was a greater degree of similarity in gene content and genomic islands among strains within clades than between clades. Strains of clade e' isolated from human were found to be missing the genomic island that carries genes encoding cytolethal distending toxins. Taken together, the results suggest a pattern of sequential divergence, starting from the separation of A. aphrophilus and A. actinomycetemcomitans through gain and loss of genes and ending with the divergence of the latter species into distinct clades and serotypes. With differing constellations of genes, the A. actinomycetemcomitans clades may have evolved distinct adaptation strategies to the human oral cavity.

The extended regulatory networks of SXT/R391 integrative and conjugative elements and IncA/C conjugative plasmids

Nowadays, healthcare systems are challenged by a major worldwide drug resistance crisis caused by the massive and rapid dissemination of antibiotic resistance genes and associated emergence of multidrug resistant pathogenic bacteria, in both clinical and environmental settings. Conjugation is the main driving force of gene transfer among microorganisms. This mechanism of horizontal gene transfer mediates the translocation of large DNA fragments between two bacterial cells in direct contact. Integrative and conjugative elements (ICEs) of the SXT/R391 family (SRIs) and IncA/C conjugative plasmids (ACPs) are responsible for the dissemination of a broad spectrum of antibiotic resistance genes among diverse species of Enterobacteriaceae and Vibrionaceae. The biology, diversity, prevalence and distribution of these two families of conjugative elements have been the subject of extensive studies for the past 15 years. Recently, the transcriptional regulators that govern their dissemination through the expression of ICE- or plasmid-encoded transfer genes have been described. Unrelated repressors control the activation of conjugation by preventing the expression of two related master activator complexes in both types of elements, i.e., SetCD in SXT/R391 ICEs and AcaCD in IncA/C plasmids. Finally, in addition to activating ICE- or plasmid-borne genes, these master activators have been shown to specifically activate phylogenetically unrelated mobilizable genomic islands (MGIs) that also disseminate antibiotic resistance genes and other adaptive traits among a plethora of pathogens such as Vibrio cholerae and Salmonella enterica.

The mosaic genome structure and phylogeny of Shiga toxin-producing Escherichia coli O104:H4 is driven by short-term adaptation

Shiga toxin-producing Escherichia coli (STEC) O104:H4 emerged as an important pathogen when it caused a large outbreak in Germany in 2011. Little is known about the evolutionary history and genomic diversity of the bacterium. The current communication describes a comprehensive analysis of STEC O104:H4 genomes from the 2011 outbreak and other non-outbreak-related isolates. Outbreak-related isolates formed a tight cluster that shared a monophyletic relation with two non-outbreak clusters, suggesting that all three clusters originated from a common ancestor. Eight single nucleotide polymorphisms, seven of which were non-synonymous, distinguished outbreak from non-outbreak isolates. Lineage-specific markers indicated that recent partitions were driven by selective pressures associated with niche adaptation. Based on the results, an evolutionary model for STEC O104:H4 is proposed. Our analysis provides the evolutionary context at population level and describes the emergence of clones with novel properties, which is necessary for developing comprehensive approaches to early warning and control.

Complete genome sequence of producer of the glycopeptide antibiotic Aculeximycin Kutzneria albida DSM 43870T, a representative of minor genus of Pseudonocardiaceae

BACKGROUND

Kutzneria is a representative of a rarely observed genus of the family Pseudonocardiaceae. Kutzneria species were initially placed in the Streptosporangiaceae genus and later reconsidered to be an independent genus of the Pseudonocardiaceae. Kutzneria albida is one of the eight known members of the genus. This strain is a unique producer of the glycosylated polyole macrolide aculeximycin which is active against both bacteria and fungi. Kutzneria albida genome sequencing and analysis allow a deeper understanding of evolution of this genus of Pseudonocardiaceae, provide new insight in the phylogeny of the genus, as well as decipher the hidden secondary metabolic potential of these rare actinobacteria.

RESULTS

To explore the biosynthetic potential of Kutzneria albida to its full extent, the complete genome was sequenced. With a size of 9,874,926 bp, coding for 8,822 genes, it stands alongside other Pseudonocardiaceae with large circular genomes. Genome analysis revealed 46 gene clusters potentially encoding secondary metabolite biosynthesis pathways. Two large genomic islands were identified, containing regions most enriched with secondary metabolism gene clusters. Large parts of this secondary metabolism "clustome" are dedicated to siderophores production.

CONCLUSIONS

Kutzneria albida is the first species of the genus Kutzneria with a completely sequenced genome. Genome sequencing allowed identifying the gene cluster responsible for the biosynthesis of aculeximycin, one of the largest known oligosaccharide-macrolide antibiotics. Moreover, the genome revealed 45 additional putative secondary metabolite gene clusters, suggesting a huge biosynthetic potential, which makes Kutzneria albida a very rich source of natural products. Comparison of the Kutzneria albida genome to genomes of other actinobacteria clearly shows its close relations with Pseudonocardiaceae in line with the taxonomic position of the genus.

Genetic characteristics and pathogenic mechanisms of periodontal pathogens

Periodontal disease is caused by a group of bacteria that utilize a variety of strategies and molecular mechanisms to evade or overcome host defenses. Recent research has uncovered new evidence illuminating interesting aspects of the virulence of these bacteria and their genomic variability. This paper summarizes some of the strategies utilized by the major species - Aggregatibacter actinomycetemcomitans, Tannerella forsythia, Treponema denticola, and Porphyromonas gingivalis - implicated in the pathogenesis of periodontal disease. Whole-genome sequencing of 14 diverse A. actinomycetemcomitans strains has revealed variations in their genetic content (ranging between 0.4% and 19.5%) and organization. Strikingly, isolates from human periodontal sites showed no genomic changes during persistent colonization. T. forsythia manipulates the cytokine responses of macrophages and monocytes through its surface glycosylation. Studies have revealed that bacterial surface-expressed O-linked glycans modulate T-cell responses during periodontal inflammation. Periodontal pathogens belonging to the "red complex" consortium express neuraminidases, which enables them to scavenge sialic acid from host glycoconjugates. Analysis of recent data has demonstrated that the cleaved sialic acid acts as an important nutrient for bacterial growth and a molecule for the decoration of bacteria surfaces to help evade the host immune attack. In addition, bacterial entry into host cells is also an important prerequisite for the lifestyle of periodontal pathogens such as P. gingivalis. Studies have shown that, after its entry into the cell, this bacterium uses multiple sorting pathways destined for autophagy, lysosomes, or recycling pathways. In addition, P. gingivalis releases outer membrane vesicles which enter cells via endocytosis and cause cellular functional impairment.

The deletion of TonB-dependent receptor genes is part of the genome reduction process that occurs during adaptation of Pseudomonas aeruginosa to the cystic fibrosis lung

Chronic Pseudomonas aeruginosa infections are the main cause of morbidity among patients with cystic fibrosis (CF) due to persistent lung inflammation caused by interaction between this bacterium and the immune system. Longitudinal studies of clonally related isolates of a dominant CF clone have indicated that genome reduction frequently occurs during adaptation of P. aeruginosa in the CF lung. In this study, we have evaluated the P. aeruginosa population structure of patients attending the Universitair Ziekenhuis Brussel (UZ Brussel) CF reference center using a combination of genotyping methods. Although the UZ Brussel P. aeruginosa CF population is characterized by the absence of a dominant CF clone, some potential interpatient transmissions could be detected. Interestingly, one of these clones showed deletion of the alternative type I ferripyoverdine receptor gene fpvB. Furthermore, we found that several other TonB-dependent receptors are deleted as well. The genome of one potentially transmissible CF clone was sequenced, revealing large deleted regions including all type III secretion system genes and several virulence genes. Remarkably, a large number of deleted genes are shared between the P. aeruginosa CF clone described in this study and isolates belonging to the dominant Copenhagen CF DK2 clone, suggesting parallel evolution.

Genomic analysis of Pseudomonas aeruginosa PA96, the host of carbapenem resistance plasmid pOZ176

Pseudomonas aeruginosa PA96 is a clinical isolate from Guangzhou, China, that is multiresistant to antibiotics. We previously described the 500-kb IncP-2 plasmid, pOZ176 that encodes many resistance genes including the IMP-9 carbapenemase. Whole-genome sequencing of PA96 enabled characterization of its genomic islands, virulence factors, and chromosomal resistance genes. We filled gaps using PCR and used optical mapping to confirm the correct contig order. We automatically annotated the core genome and manually annotated the genomic islands. The genome is 6 444 091 bp and encodes 5853 ORFs. From the whole-genome sequence, we constructed a physical map and constructed a phylogenetic tree for comparison with sequenced P. aeruginosa strains. Analysis of known core genome virulence factors and resistance genes revealed few differences with other strains, but the major virulence island is closer to that of DK2 than to PA14. PA96 most closely resembles the environmental strain M18, and notably shares a common serotype, pyoverdin type, flagellar operon, type IV pilin, and several genomic islands with M18.

A markerless protocol for genetic analysis of Aggregatibacter actinomycetemcomitans

BACKGROUND/PURPOSE

The genomes of different Aggregatibacter actinomycetemcomitans (A actinomycetemcomitans) strains contain many strain-specific genes and genomic islands (defined as DNA found in some but not all strains) of unknown functions. Genetic analysis for the functions of these islands will be constrained by the limited availability of genetic markers and vectors for A actinomycetemcomitans. In this study, we tested a novel genetic approach of gene deletion and restoration in a naturally competent A actinomycetemcomitans strain D7S-1.

METHODS

Specific genes' deletion mutants and mutants restored with the deleted genes were constructed by a markerless loxP/Cre system. In mutants with sequential deletion of multiple genes loxP with different spacer regions were used to avoid unwanted recombinations between loxP sites.

RESULTS

Eight single-gene deletion mutants, four multiple-gene deletion mutants, and two mutants with restored genes were constructed. No unintended non-specific deletion mutants were generated by this protocol. The protocol did not negatively affect the growth and biofilm formation of A actinomycetemcomitans.

CONCLUSION

The protocol described in this study is efficient and specific for genetic manipulation of A actinomycetemcomitans, and will be amenable for functional analysis of multiple genes in A actinomycetemcomitans.

Vibrio cholerae O1 epidemic variants in Angola: a retrospective study between 1992 and 2006

Cholera is still a major public health concern in many African countries. In Angola, after a decade of absence, cholera reemerged in 1987, spreading throughout the country until 1996, with outbreaks recurring in a seasonal pattern. In 2006 Angola was hit by one of the most severe outbreaks of the last decade, with ca. 240,000 cases reported. We analyzed 21 clinical strains isolated between 1992 and 2006 from several provinces throughout the country: Benguela, Bengo, Luanda, Cuando Cubango, and Cabinda. We used two multiplex PCR assays to investigate discriminatory mobile genetic elements (MGE) [Integrative Conjugative Elements (ICEs), VSP-II, GI12, GI14, GI15, K, and TLC phages] and we compared the profiles obtained with those of different reference V. cholerae O1 variants (prototypical, altered, and hybrid), responsible for the ongoing 7th pandemic. We also tested the strains for the presence of specific VSP-II variants and for the presence of a genomic island (GI) (WASA-1), correlated with the transmission of seventh pandemic cholera from Africa to South America. Based on the presence/absence of the analyzed genetic elements, five novel profiles were detected in the epidemic strains circulating in the 1990s. The most frequent profiles, F and G, were characterized by the absence of ICEs and the three GIs tested, and the presence of GI WASA-1 and the WASA variant of the VSP-II island. Our results identified unexpected variability within the 1990s epidemic, showing different rearrangements in a dynamic part of the genome not present in the prototypical V. cholerae O1 N16961. Moreover the 2006 strains differed from the current pandemic V. cholerae O1 strain. Taken together, our results highlight the role of horizontal gene transfer (HGT) in diversifying the genetic background of V. cholerae within a single epidemic.

Chromosomal targeting by CRISPR-Cas systems can contribute to genome plasticity in bacteria

The clustered regularly interspaced short palindromic repeats (CRISPR) and their associated (Cas) proteins form adaptive immune systems in bacteria to combat phage and other foreign genetic elements. Typically, short spacer sequences are acquired from the invader DNA and incorporated into CRISPR arrays in the bacterial genome. Small RNAs are generated that contain these spacer sequences and enable sequence-specific destruction of the foreign nucleic acids. Occasionally, spacers are acquired from the chromosome, which instead leads to targeting of the host genome. Chromosomal targeting is highly toxic to the bacterium, providing a strong selective pressure for a variety of evolutionary routes that enable host cell survival. Mutations that inactivate the CRISPR-Cas functionality, such as within the cas genes, CRISPR repeat, protospacer adjacent motifs (PAM), and target sequence, mediate escape from toxicity. This self-targeting might provide some explanation for the incomplete distribution of CRISPR-Cas systems in less than half of sequenced bacterial genomes. More importantly, self-genome targeting can cause large-scale genomic alterations, including remodeling or deletion of pathogenicity islands and other non-mobile chromosomal regions. While control of horizontal gene transfer is perceived as their main function, our recent work illuminates an alternative role of CRISPR-Cas systems in causing host genomic changes and influencing bacterial evolution.

Mobile elements, zoonotic pathogens and commensal bacteria: conduits for the delivery of resistance genes into humans, production animals and soil microbiota

Multiple antibiotic resistant pathogens represent a major clinical challenge in both human and veterinary context. It is now well-understood that the genes that encode resistance are context independent. That is, the same gene is commonly present in otherwise very disparate pathogens in both humans and production and companion animals, and among bacteria that proliferate in an agricultural context. This can be true even for pathogenic species or clonal types that are otherwise confined to a single host or ecological niche. It therefore follows that mechanisms of gene flow must exist to move genes from one part of the microbial biosphere to another. It is widely accepted that lateral (or horizontal) gene transfer (L(H)GT) drives this gene flow. LGT is relatively well-understood mechanistically but much of this knowledge is derived from a reductionist perspective. We believe that this is impeding our ability to deal with the medical ramifications of LGT. Resistance genes and the genetic scaffolds that mobilize them in multiply drug resistant bacteria of clinical significance are likely to have their origins in completely unrelated parts of the microbial biosphere. Resistance genes are increasingly polluting the microbial biosphere by contaminating environmental niches where previously they were not detected. More attention needs to be paid to the way that humans have, through the widespread application of antibiotics, selected for combinations of mobile elements that enhance the flow of resistance genes between remotely linked parts of the microbial biosphere. Attention also needs to be paid to those bacteria that link human and animal ecosystems. We argue that multiply antibiotic resistant commensal bacteria are especially important in this regard. More generally, the post genomics era offers the opportunity for understanding how resistance genes are mobilized from a one health perspective. In the long term, this holistic approach offers the best opportunity to better manage what is an enormous problem to humans both in terms of health and food security.

Multiple mechanisms contribute to lateral transfer of an organophosphate degradation (opd) island in Sphingobium fuliginis ATCC 27551

The complete sequence of pPDL2 (37,317 bp), an indigenous plasmid of Sphingobium fuliginis ATCC 27551 that encodes genes for organophosphate degradation (opd), revealed the existence of a site-specific integrase (int) gene with an attachment site attP, typically seen in integrative mobilizable elements (IME). In agreement with this sequence information, site-specific recombination was observed between pPDL2 and an artificial plasmid having a temperature-sensitive replicon and a cloned attB site at the 3′ end of the seryl tRNA gene of Sphingobium japonicum. The opd gene cluster on pPDL2 was found to be part of an active catabolic transposon with mobile elements y4qE and Tn3 at its flanking ends. Besides the previously reported opd cluster, this transposon contains genes coding for protocatechuate dioxygenase and for two transport proteins from the major facilitator family that are predicted to be involved in transport and metabolism of aromatic compounds. A pPDL2 derivative, pPDL2-K, was horizontally transferred into Escherichia coli and Acinetobacter strains, suggesting that the oriT identified in pPDL2 is functional. A well-defined replicative origin (oriV), repA was identified along with a plasmid addiction module relB/relE that would support stable maintenance of pPDL2 in Sphingobium fuliginis ATCC 27551. However, if pPDL2 is laterally transferred into hosts that do not support its replication, the opd cluster appears to integrate into the host chromosome, either through transposition or through site-specific integration. The data presented in this study help to explain the existence of identical opd genes among soil bacteria.

Prediction of genomic islands in three bacterial pathogens of pneumonia

Pneumonia is one kind of common infectious disease, which is usually caused by bacteria, viruses, or fungi. In this paper, we predicted genomic islands in three bacterial pathogens of pneumonia. They are Chlamydophila pneumoniae, Mycoplasma pneumoniae and Streptococcus pneumoniae, respectively. For each pathogen, one clinical strain is involved. After implementing the cumulative GC profile combined with h and BCN index, eight genomic islands are found in three pathogens. Among them, six genomic islands are found to have mobility elements, which constitute a kind of conserved character of genomic islands, and this introduces the possibility that they are genuine genomic islands. The present results show that the cumulative GC profile when combined with h and BCN indexes is a good method for predicting genomic islands in bacteria and it has lower false positive rate than the SIGI method. Specially, three genomic islands are found to contain clusters of genes coding for production of virulence factors and this is useful for research into the pathogenicity of these pathogens and helpful for the treatment of diseases caused by them.

IGIPT - Integrated genomic island prediction tool

IGIPT is a web-based integrated platform for the identification of genomic islands (GIs). It incorporates thirteen parametric measures based on anomalous nucleotide composition on a single platform, thus improving the predictive power of a horizontally acquired region, since it is known that no single measure can absolutely predict a horizontally transferred region. The tool filters putative GIs based on standard deviation from genomic average and also provides raw output in MS excel format for further analysis. To facilitate the identification of various structural features, viz., tRNA integration sites, repeats, etc. in the vicinity of GIs, the tool provides option to extract the predicted regions and its flanking regions.

AVAILABILITY

The database is available for free at http://bioinf.iiit.ac.in/IGIPT/

n-Gram characterization of genomic islands in bacterial genomes

The paper presents a novel, n-gram-based method for analysis of bacterial genome segments known as genomic islands (GIs). Identification of GIs in bacterial genomes is an important task since many of them represent inserts that may contribute to bacterial evolution and pathogenesis. In order to characterize and distinguish GIs from rest of the genome, binary classification of islands based on n-gram frequency distribution have been performed. It consists of testing the agreement of islands n-gram frequency distributions with the complete genome and backbone sequence. In addition, a statistic based on the maximal order Markov model is used to identify significantly overrepresented and underrepresented n-grams in islands. The results may be used as a basis for Zipf-like analysis suggesting that some of the n-grams are overrepresented in a subset of islands and underrepresented in the backbone, or vice versa, thus complementing the binary classification. The method is applied to strain-specific regions in the Escherichia coli O157:H7 EDL933 genome (O-islands), resulting in two groups of O-islands with different n-gram characteristics. It refines a characterization based on other compositional features such as G+C content and codon usage, and may help in identification of GIs, and also in research and development of adequate drugs targeting virulence genes in them.