Correspondence between typing methods results: a web-based quantitative analysis tool

Abstract number: 1732_92

Carrico J.A., Pinto F.R., Melo-Cristino J., de Lencastre H., Almeida J.S., Ramirez M.

Objectives: Molecular epidemiology studies of clinically relevant microorganisms frequently focus on a comparison between the assigned types of different typing methods. Therefore, there is a critical need to understand the correspondence between types produced by different methods. This may be useful not only for the comparison of the genetic backgrounds of the particular set of isolates under study but also to produce a broader view of how the results of the different typing methods are related. To achieve this goal, a framework of measures for the quantitative comparison of typing methods results was developed (Carriço et al., Illustration of a common framework for relating multiple typing methods by application to macrolide-resistant Streptococcus pyogenes, J. Clin. Microbiol. 2006) and is now implemented in a free-access web-based interface, on which users can analyse their data anonymously and retrieve the results.

Methods: To facilitate the use of the framework developed, a web-based interface was implemented to provide researchers with a user-friendly interface to compare their typing methods results. On the website, we also provide a set of BionumericsTM scripts to allow the users to implement the proposed analysis on their Bionumerics databases while offline.

Results: The web-based interface and the Bionumerics scripts allow for the calculation of several measures: Simpson's index of diversity with confidence intervals for accessing the diversity of typing results for different methods; Adjusted Rand and Wallace coefficients for bidirectional and unidirectional (respectively) comparison between the results of two typing methods results. Based on these measures the users can quantitatively evaluate the discriminatory power of the typing methodologies used and their concordance.

Conclusion: The web-based interface and the Bionumerics scripts provide users with the ability for a quantitative assessment of correspondence between typing methods results. This can be used to evaluate the predictive power of a typing methodology when compared to another, which can guide the user in the choice of a `gold standard' for clone definition. Furthermore, as new microbial typing methods are proposed, this methodology allows for their comparison in terms of type assignments with established methodologies.

Session Details

Date: 31/03/2007
Time: 00:00-00:00
Session name: European Society of Clinical Microbiology and Infectious Diseases
Location: ICC, Munich, Germany
Presentation type:
Back to top