Introduction
The development, selection and spread of antimicrobial-resistant bacteria in human and veterinary
medicine is a major concern worldwide. AR phenotypes are the in vitro resistant characteristics of
bacterial isolates against the action of one or more antimicrobial agents. These characteristics often
are assessed by measuring the MIC using broth microdilution methods or by disk diffusion
methods. These values are usually further interpreted into susceptible or resistant phenotypes
according to breakpoints that are determined by a variety of committees (e.g., NCCLS). AR
resistance profiles for a bacterial isolate can range from pansusceptible to resistant to multiple
antibiotics.
Finding AR phenotypic groups with similar properties in hundreds of bacterial isolates is of great
significance to understand resistance patterns. Visualizing and consequently analyzing large
multinomial AR phenotypic data sets with high variability is difficult. Therefore, being better able
to classify AR phenotypes would help our understanding of the relationship among the resistance
clusters and the risk factors in a study, and also would aid in the interpretation of seemingly unique
versus more common ‘garden variety’ AR phenotypic patterns.