In studies where no genotyping method was used, it was assumed that each isolate represented a strain. Results and discussion Comparative performance

of the five molecular methods The percentage of correctly identified strains obtained using the five identification methods, and the number of misidentified non-targeted species greatly depended upon the method used (Tables 1 and 2). The percentage of misidentified strains ranged from 16.8% to 67.4% (Table 2). The m-PCR method of Kabeya et al. [15] had the worst performance, and produced unreliable results for all three of its targeted species (Tables 1 Akt inhibitor and 2). Although all strains of A. cryaerophilus and A. skirrowii were correctly identified, a further eight and six non-targeted species, respectively, were mistakenly identified as one of these two species (Table 1). Furthermore, only 4.8% of the A. butzleri strains were correctly identified, with six non-targeted species being confused with this species (Tables 1 and 2). Globally, the Kabeya et XAV-939 cost al. m-PCR method correctly identified just 32.6% (31/95) of the studied strains. Although this method

was also designed to differentiate subgroups 1A and 1B of A. cryaerophilus, not all strains of these subgroups were correctly identified (Table 2). This correlates with the in silico observations of Douidah et al. [9] who reported that the primer used [15] were not specific enough to provide correct identification of A. cryaerophilus at the subgroup level. Further to this, Debruyne et al.[21] have suggested, that based on results of AFLP and hsp60 analyses, the subgroup nomenclatures 1A and 1B should be abandoned. The second least reliable method analysed was the m-PCR technique described by Houf et al.[14]. This correctly

identified 55.8% (53/95) of the strains (Table 2), including all those belonging Ixazomib price to its targeted species (A. butzleri, A. cryaerophilus, and A. skirrowii; Table 1). This method was 100% reliable for the identification of A. butzleri, and there was no confusion with other species. However, nine of the fourteen non-targeted species generated the typical amplicon of A. cryaerophilus; two that of A. skirrowii; and two simultaneously generated both amplicons (Tables 1 and 2). Only A. cibarius produced no amplification when using this method (Table 2). These results agree with previous studies that showed the existence of misidentifications when using this method [1, 5–7]. A similar number of correctly identified strains (83.2%) were obtained when using the other three evaluated methods (Pentimalli et al.[16]; the combined method of Douidah et al. [9] and De Smet et al.[17]; and Figueras et al.[18]). However, the number of misidentified non-targeted species differed depending upon the method used (Tables 1 and 2). Most misidentification occurred when using the method of Pentimalli et al.[16]. In this case, four non-targeted species were confused with A. butzleri, one with A.