Support vector machine applied to predict the zoonotic potential of E. coli O157 cattle isolates

Abstract

Sequence analyses of pathogen genomes facilitate the tracking of disease outbreaks and allow relationships between strains to be reconstructed and virulence factors to be identified. However, these methods are generally used after an outbreak has happened. Here, we show that support vector machine analysis of bovine E. coli O157 isolate sequences can be applied to predict their zoonotic potential, identifying cattle strains more likely to be a serious threat to human health. Notably, only a minor subset (less than 10%) of bovine E. coli O157 isolates analyzed in our datasets were predicted to have the potential to cause human disease; this is despite the fact that the majority are within previously defined pathogenic lineages I or I/II and encode key virulence factors. The predictive capacity was retained when tested across datasets. The major differences between human and bovine E. coli O157 isolates were due to the relative abundances of hundreds of predicted prophage proteins. This finding has profound implications for public health management of disease because interventions in cattle, such a vaccination, can be targeted at herds carrying strains of high zoonotic potential. Machine-learning approaches should be applied broadly to further our understanding of pathogen biology. machine learning | zoonosis | Shiga toxin | E. coli | cattle

Whole genome sequencing and metagenomics for outbreak investigation, source attribution and risk assessment of food‐borne microorganisms

“…Typhimurium ST313, often responsible for invasive disease, to the human host (Okoro et al., ). Another example showed variable zoonotic potential among bovine E. coli O157 isolates, with only a minority predicted to be associated with human disease, contrary to preliminary assumptions (Lupolova et al., ).…”

Multi‐Exposure Pathway Model to Compare Escherichia coli O157 Risks and Interventions

“…Recent research has demonstrated variability in E. coli O157 pathogenicity with respect to strain. However, given the paucity of information on occurrence of specific strains in Canadian cattle, and the lack of available dose–response information, the impact of bacterial strain on the outcome of disease was not assessed herein.…”

Enterohaemorrhagic and other Shiga toxin-producingEscherichia coli(STEC): Where are we now regarding diagnostics and control strategies?

“…To date, phenotypic and molecular epidemiological techniques have had limited success at addressing these questions. However, the recent adoption of whole genome sequencing (WGS) accompanied by developments in pan‐genomic pipelines and, in particular, the use of machine learning algorithms, for example, the Support Vector Machine (SVM) (Lupolova et al., ) are beginning to provide novel information about the proportion of strains, which can cause human disease (see section Direct next generation sequencing (NGS) approaches). Unfortunately, such approaches are currently highly dependent on expert bioinformaticists/mathematicians/statisticians, who use dedicated and jargon‐ladened language largely uninterpretable by non‐expert clinicians and clinical microbiologists.…”

“…In addition, efforts have been undertaken to improve vaccine efficiency, such as by including vaccine candidates that induce a cellular response (Corbishley et al., ). The ongoing mining of WGS data (Lupolova et al., ) may allow vaccine development to focus on vaccine candidates expressed only by virulent STEC.…”