Limits...
In silico genomic analyses reveal three distinct lineages of Escherichia coli O157:H7, one of which is associated with hyper-virulence.

Laing CR, Buchanan C, Taboada EN, Zhang Y, Karmali MA, Thomas JE, Gannon VP - BMC Genomics (2009)

Bottom Line: Each lineage contains significant phenotypic differences, with lineage I strains being the most commonly associated with human infections.The results of this study highlight the similarities in relationships derived from multi-locus genome sampling methods and suggest a "common genotyping language" may be devised for population genetics and epidemiological studies.Future genotyping methods should provide data that can be stored centrally and accessed locally in an easily transferable, informative and extensible format based on comparative genomic analyses.

View Article: PubMed Central - HTML - PubMed

Affiliation: Laboratory for Foodborne Zoonoses, Public Health Agency of Canada, Lethbridge, AB, Canada. chad_r_laing@phac-aspc.gc.ca

ABSTRACT

Background: Many approaches have been used to study the evolution, population structure and genetic diversity of Escherichia coli O157:H7; however, observations made with different genotyping systems are not easily relatable to each other. Three genetic lineages of E. coli O157:H7 designated I, II and I/II have been identified using octamer-based genome scanning and microarray comparative genomic hybridization (mCGH). Each lineage contains significant phenotypic differences, with lineage I strains being the most commonly associated with human infections. Similarly, a clade of hyper-virulent O157:H7 strains implicated in the 2006 spinach and lettuce outbreaks has been defined using single-nucleotide polymorphism (SNP) typing. In this study an in silico comparison of six different genotyping approaches was performed on 19 E. coli genome sequences from 17 O157:H7 strains and single O145:NM and K12 MG1655 strains to provide an overall picture of diversity of the E. coli O157:H7 population, and to compare genotyping methods for O157:H7 strains.

Results: In silico determination of lineage, Shiga-toxin bacteriophage integration site, comparative genomic fingerprint, mCGH profile, novel region distribution profile, SNP type and multi-locus variable number tandem repeat analysis type was performed and a supernetwork based on the combination of these methods was produced. This supernetwork showed three distinct clusters of strains that were O157:H7 lineage-specific, with the SNP-based hyper-virulent clade 8 synonymous with O157:H7 lineage I/II. Lineage I/II/clade 8 strains clustered closest on the supernetwork to E. coli K12 and E. coli O55:H7, O145:NM and sorbitol-fermenting O157 strains.

Conclusion: The results of this study highlight the similarities in relationships derived from multi-locus genome sampling methods and suggest a "common genotyping language" may be devised for population genetics and epidemiological studies. Future genotyping methods should provide data that can be stored centrally and accessed locally in an easily transferable, informative and extensible format based on comparative genomic analyses.

Show MeSH

Related in: MedlinePlus

Supernetwork constructed from in silico data of six typing methods. The supernetwork created from the combination of each maximum parsimony tree from the following typing methods: Stx-phage insertion site typing, MLVA, CGF, SNP genotyping, mCGH, and GISSH-based novel region distribution typing. Maximum parsimony trees were combined using the un-weighted mean distance and Z-closure with 1000 iterations; the resulting supernetwork was displayed using the equal angle method.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC2719669&req=5

Figure 2: Supernetwork constructed from in silico data of six typing methods. The supernetwork created from the combination of each maximum parsimony tree from the following typing methods: Stx-phage insertion site typing, MLVA, CGF, SNP genotyping, mCGH, and GISSH-based novel region distribution typing. Maximum parsimony trees were combined using the un-weighted mean distance and Z-closure with 1000 iterations; the resulting supernetwork was displayed using the equal angle method.

Mentions: In order to obtain a more complete understanding of the relationships among E. coli O157:H7 strains, the maximum parsimony trees derived using in silico data from the six separate molecular typing methods included in this study were combined to form a supernetwork. We accomplished this by combing the six maximum parsimony trees generated using PAUP* instead of concatenating the data into a single matrix and generating a single most parsimonious tree. This ensured a method with many data points such as mCGH did not outweigh a method like MLVA that contains few data points. The supernetwork based on the six trees given equal weighting (Figure 2) shows competing signals, rather than an inferred absolute tree [16]. As can be seen, the 19 strains were distributed into four main groups. One contained strains K12 MG1655 and O145:NM EC33264 and another the single lineage II strain EC869. The remaining two groups of the supernetwork were lineage-specific, comprising one cluster of lineage I strains and another of lineage I/II strains. The distribution of the four lineage I strains was consistent with the clade breakdown of Manning et al. [12], as the clade 2 strains TW14588 and EC4501 were more closely related to each other than clade 3 strain EDL933 or clade 1 strain Sakai.


In silico genomic analyses reveal three distinct lineages of Escherichia coli O157:H7, one of which is associated with hyper-virulence.

Laing CR, Buchanan C, Taboada EN, Zhang Y, Karmali MA, Thomas JE, Gannon VP - BMC Genomics (2009)

Supernetwork constructed from in silico data of six typing methods. The supernetwork created from the combination of each maximum parsimony tree from the following typing methods: Stx-phage insertion site typing, MLVA, CGF, SNP genotyping, mCGH, and GISSH-based novel region distribution typing. Maximum parsimony trees were combined using the un-weighted mean distance and Z-closure with 1000 iterations; the resulting supernetwork was displayed using the equal angle method.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC2719669&req=5

Figure 2: Supernetwork constructed from in silico data of six typing methods. The supernetwork created from the combination of each maximum parsimony tree from the following typing methods: Stx-phage insertion site typing, MLVA, CGF, SNP genotyping, mCGH, and GISSH-based novel region distribution typing. Maximum parsimony trees were combined using the un-weighted mean distance and Z-closure with 1000 iterations; the resulting supernetwork was displayed using the equal angle method.
Mentions: In order to obtain a more complete understanding of the relationships among E. coli O157:H7 strains, the maximum parsimony trees derived using in silico data from the six separate molecular typing methods included in this study were combined to form a supernetwork. We accomplished this by combing the six maximum parsimony trees generated using PAUP* instead of concatenating the data into a single matrix and generating a single most parsimonious tree. This ensured a method with many data points such as mCGH did not outweigh a method like MLVA that contains few data points. The supernetwork based on the six trees given equal weighting (Figure 2) shows competing signals, rather than an inferred absolute tree [16]. As can be seen, the 19 strains were distributed into four main groups. One contained strains K12 MG1655 and O145:NM EC33264 and another the single lineage II strain EC869. The remaining two groups of the supernetwork were lineage-specific, comprising one cluster of lineage I strains and another of lineage I/II strains. The distribution of the four lineage I strains was consistent with the clade breakdown of Manning et al. [12], as the clade 2 strains TW14588 and EC4501 were more closely related to each other than clade 3 strain EDL933 or clade 1 strain Sakai.

Bottom Line: Each lineage contains significant phenotypic differences, with lineage I strains being the most commonly associated with human infections.The results of this study highlight the similarities in relationships derived from multi-locus genome sampling methods and suggest a "common genotyping language" may be devised for population genetics and epidemiological studies.Future genotyping methods should provide data that can be stored centrally and accessed locally in an easily transferable, informative and extensible format based on comparative genomic analyses.

View Article: PubMed Central - HTML - PubMed

Affiliation: Laboratory for Foodborne Zoonoses, Public Health Agency of Canada, Lethbridge, AB, Canada. chad_r_laing@phac-aspc.gc.ca

ABSTRACT

Background: Many approaches have been used to study the evolution, population structure and genetic diversity of Escherichia coli O157:H7; however, observations made with different genotyping systems are not easily relatable to each other. Three genetic lineages of E. coli O157:H7 designated I, II and I/II have been identified using octamer-based genome scanning and microarray comparative genomic hybridization (mCGH). Each lineage contains significant phenotypic differences, with lineage I strains being the most commonly associated with human infections. Similarly, a clade of hyper-virulent O157:H7 strains implicated in the 2006 spinach and lettuce outbreaks has been defined using single-nucleotide polymorphism (SNP) typing. In this study an in silico comparison of six different genotyping approaches was performed on 19 E. coli genome sequences from 17 O157:H7 strains and single O145:NM and K12 MG1655 strains to provide an overall picture of diversity of the E. coli O157:H7 population, and to compare genotyping methods for O157:H7 strains.

Results: In silico determination of lineage, Shiga-toxin bacteriophage integration site, comparative genomic fingerprint, mCGH profile, novel region distribution profile, SNP type and multi-locus variable number tandem repeat analysis type was performed and a supernetwork based on the combination of these methods was produced. This supernetwork showed three distinct clusters of strains that were O157:H7 lineage-specific, with the SNP-based hyper-virulent clade 8 synonymous with O157:H7 lineage I/II. Lineage I/II/clade 8 strains clustered closest on the supernetwork to E. coli K12 and E. coli O55:H7, O145:NM and sorbitol-fermenting O157 strains.

Conclusion: The results of this study highlight the similarities in relationships derived from multi-locus genome sampling methods and suggest a "common genotyping language" may be devised for population genetics and epidemiological studies. Future genotyping methods should provide data that can be stored centrally and accessed locally in an easily transferable, informative and extensible format based on comparative genomic analyses.

Show MeSH
Related in: MedlinePlus