Study of a classification algorithm for AIEC identification in geographically distinct E. coli strains

Abstract

Adherent-invasive Escherichia coli (AIEC) have been extensively implicated in Crohn's disease pathogenesis. Currently, AIEC is identified phenotypically, since no molecular marker specific for AIEC exists. An algorithm based on single nucleotide polymorphisms was previously presented as a potential molecular tool to classify AIEC/non-AIEC, with 84% accuracy on a collection of 50 strains isolated in Girona (Spain). Herein, our aim was to determine the accuracy of the tool using AIEC/non-AIEC isolates from different geographical origins and extraintestinal pathogenic E. coli (ExPEC) strains. The accuracy of the tool was significantly reduced (61%) when external AIEC/non-AIEC strains from France, Chile, Mallorca (Spain) and Australia (82 AIEC, 57 non-AIEC and 45 ExPEC strains in total) were included. However, the inclusion of only the ExPEC strains showed that the tool was fairly accurate at differentiating these two close pathotypes (84.6% sensitivity; 79% accuracy). Moreover, the accuracy was still high (81%) for those AIEC/non-AIEC strains isolated from Girona and Mallorca (N = 63); two collections obtained from independent studies but geographically close. Our findings indicate that the presented tool is not universal since it would be only applicable for strains from similar geographic origin and demonstrates the need to include strains from different origins to validate such tools.

Figure 1

Geographical distribution of the isolates assessed in four groups of analysis and the percentage of strains that are correctly (green) or incorrectly (red) predicted by the SNP algorithm in comparison with their previous phenotypic characterisation. (A) AIEC/non-AIEC strains from Girona (Spain), including LF82 as a reference strain; (B) ExPEC (Spain and USA)^,, and AIEC/non-AIEC strains from Girona (Spain); (C) AIEC/non-AIEC strains from Girona (Spain) and AIEC/non-AIEC from France, Chile, Spain (Mallorca), Australia and ExPEC-Spain^, and ExPEC-America; (D) AIEC/non-AIEC strains from Girona and Mallorca (Spain).

Figure 2

Classification algorithm for AIEC identification. Assessed in our collection and external strain collections (France, Chile, Spain (Mallorca), Australia and ExPEC-Spain^, and ExPEC-America). Percentages represent the proportion of strains that are correctly predicted as AIEC or non-AIEC based on the result for each SNP combination. The number of total strains corresponding to each condition is indicated. (−): no amplification; other: a nucleotide different from guanine (G) or overlapping peaks.