Imported from Food Pathogen Team Kwan Lab CUHK: Proposed analytical Workflow in a bacterial pathogen outbreak investigation

H.S. Kwan

A. Introduction

The German Escherichia coli outbreak causing serious diseases, notably, Haemoragghic Uremic Syndrome started in early May. The conventional typing methods including serotypinc, MLST sequence typing and pulse-field gel electrophoresis seemed to yield only limited information about the strain. The situation was a little chaotic. The strain could not be typed definitively as EAEC or EHEC. It was suspected to be a hybrid of EAEC and EHEC. More detail analysis was needed.

Nick Loman started a very interesting “Crowdsourcing” genomic analysis of the German Escherichia coli outbreak strain.  The first set of analyses was based on the BGI genome sequencing reads.  The first 2 sets of reads were provided by BGI and Life Technologies, both used the Ion Torrent Personal Genome Machine. Ion Torrent PGM was the machine of choice for its quick operation time. It seemed that The draft sequences were completed in only 3 days albeit the coverage was low. The quick turn around time allowed “Crowdsourcing” to start very early after the strains were given to BGI and Life Technologies to sequence. The speed of sequencing was important to give an initial quick analysis of the outbreak strain to show its major characteristics. Treatment could be chosen with this information. For deadly diseases, timing is always very important. Certainly, a few days later, the more powerful machines, Illumina HiSeq and Roche 454 Titanium, provided higher coverage for better assembly which was important for high resolution analyses.

My laboratory, Kwan Lab, CUHK, has noted the progress very early in the outbreak. For the last 2 and a half years, we have been constructing a Foodborne Pathogen Genome Sequence Database and sequencing the genomes of Hong Kong isolates of Salmonella Typhimurium and Vibrio parahaemolyticus. We have been asking ourselves how we can use the database for outbreak investigation.  The “Crowdsourcing” activity has provided us first-hand experience of how genome sequences are used in a real outbreak investigation.

B. Questions and approaches to answer them

My understanding from the “Crowdsourcing” experience is that we have been asking several important questions about the strain responsible for the outbreak and using genome analysis to find the answers.

The questions are:

1.     Is the outbreak strain isolated previously? That is, is it a known strain?

2.     If not known, is it related to known strains?

3.     If answer is no to above, then it is novel?

4.     What are the pathogenicity features of the strain?

5.     Is a typing scheme available to follow further cases?

6.     If not, can we develop a typing scheme?

Answering these questions would allow us to understand the strain better and help us to

1.     trace the outbreak strain,

2.     find appropriate treatment for the disease,

3.     understand the pathogenicity of the outbreak strain,

4.     develop preventive measures for further and future outbreaks.

The genome sequencing combined with “Crowdsourcing” seemed to work well. We could answer the questions and further use the answers for treatment and control of the outbreak.  Amazed by the efficiency and effectiveness of the process, I analyzed the genome analyses reported by the laboratories contributing to “Crowdsourcing” to find out which analysis is for answering each question.

C. Synthesis: Analytical workflow in a pathogenic bacteria outbreak

My synthesis from the genome sequencing-crowdsourcing experience is shown in Figure 1.  The advantages of Genome sequencing approach are apparent comparing to conventional methods:

1.     Genomic approach is definitive with little uncertainty

2.     Genomic approach reveals whether the outbreak strain is known, related, or novel

3.     Genomic analysis provides comprehensive information on pathogenicity of the outbreak strain

4.     Comparative Genomics provide information to develop typing schemes for diagnosis of the outbreak strain in patients and food, important for management of the outbreak

5.     Genomics provide lots of data for further comprehensive understanding of the outbreak strain, important for prevention of further and future outbreaks

 

                                   Figure 1.  Putative analytical workflow in a bacterial outbreak.  

1. Top group, conventional workflow: after typing and analysis, there is still great uncertainty about its relationship to previously isolated and analysed strains. It is uncertain whether it is novel. Not much is known about the pathogenicity feature of the strain. The data cannot be used to generate a typing scheme.
2. Bottom group, genome sequencing and genomic analysis: after the analysis, the strain can be shown to be a known strain, or related to a known strain, or a novel strain. The pathogenicity is well studied at the genomic/gene level, providing knowledge for furhter analysis.
3. The right columns of boxes show the questions and whether they can be answered. The ? marks indicate much uncertainties. The ✓ marks indicate answers can be provided.  The  ✓ mark in the box means some information can be obtained for further analysis.

D. Concluding remarks

Genome sequencing and analyses with an appropriate workflow would be the new paradigm in bacterial outbreak investigation. It is important to have a fast genome sequencing turn-around time and later a high coverage sequencing for further detail analyses. This approach still costs dearly and can be prohibitive. When we can perform inexpensive, fast, and high-throughput genome sequencing, the genomic analysis paradigm will be established. 

Contributors to this proposal: Billy CHENG, Lei LI, Qianli HUANG, Tommy AU, Jackie WONG, Wenyan NONG and Simon CHEUNG.