Comparison of Probe Hybridization Array Typing to Multilocus Sequence Typing for Pathogenic Escherichia coli

Abstract

Probe hybridization array typing (PHAT) is a previously validated, high-throughput, highly discriminatory binary typing method based on the presence or absence of genetic material. To increase the utility of PHAT, we identified a refined PHAT probe set using 24 known and potential Escherichia coli virulence genes, by which groups similar to multilocus sequence typing (MLST) clonal groups (CGs) could be determined. We PHAT typed 1,132 E. coli isolates, representing at least 62 MLST CGs and diverse disease states, using a “library-on-a-slide” microarray format. Using 24 PHAT probes, all 62 MLST CGs in the representative E. coli collection were distinguished. For major CGs, PHAT correctly classified all sequence types within CG7 and CG17 but misclassified between one and four sequence types for CG13, CG14, CG23, CG38, and CG58, giving an overall sensitivity and specificity of 80.4 and 98.7%, respectively. After application of the PHAT classification to the whole collection, MLST validation of the PHAT probe classification resulted in sensitivities from 0.0 to 100.0% and specificities from 75.0 to 100.0% for individual CGs and an overall sensitivity and specificity of 64.7 and 88.3%, respectively. The refined PHAT probe set is capable of classifying isolates into groups in a manner similar to major clonal complexes of MLST, indicating coevolution between the chromosomal background and the flexible gene pool. Further refinement is needed to distinguish between closely related groups. For analysis of large bacterial collections, PHAT is a relatively time- and cost-efficient method and is ideal for a first level of analysis.