Escherichia coli (E. coli) is mostly a commensal bacterium, part of the intestinal microbiota of a variety of animals, including humans (Bélanger et al., 2011; Vila et al., 2016). However, some E. coli strains can be pathogenic and depending on the spectrum of encoded virulence factors E. coli can cause either intestinal or extraintestinal infections (Kaper et al., 2004). It is known that the species E. coli has an extensive genetic substructure (Chaudhuri and Henderson, 2012) and that the substructure of E. coli populations differs among distinct geographical regions (Freitag et al., 2005; Walk et al., 2009) and bacterial hosts (Vadnov et al., 2017).

Initially, four different phylogenetic groups of E. coli were defined, A, B1, B2, and D (Chaudhuri and Henderson, 2012). Clermont et al. (2000) established a PCR-method, the so called triplex PCR, for assigning E. coli strains into these four phylogenetic groups, a method that was widely used to type and subtype commensal and pathogenic E. coli. Phylogenetic classification has been extensively used to compare with serogroup, virulence and resistance traits as well as distribution among various hosts. However, subsequently, on the basis of multi-locus sequence typing and complete genome data, additional E. coli phylogenetic groups were recognized (Walk et al., 2009; Luo et al., 2011). The number of defined phylogenetic groups thus rose to 8 (A, B1, B2, C, D, E, F that belong to E. coli sensu stricto, and the eighth—the Escherichia cryptic clade I). Thus, Clermont et al. (2013) proposed a revised, so called extended quadruplex method for assigning E. coli strains to phylogenetic groups that is now replacing the triplex method. The authors validated the extended quadruplex method on a set of 234 strains, which included the ECOR strains (Clermont et al., 2011) and 133 strains from Australia (Gordon et al., 2008). In addition, Clermont et al. (2013) used the new extended quadruplex method for phylogroup assignment of 293 human fecal E. coli strains from France and 373 human fecal E. coli strains from Australia (Clermont et al., 2013). The authors reported that 12.8% of the tested strains belonged to the new phylogroups C, E, F, and clade I and that strains previously assigned, with the triplex method, to the A and D group should be retested with the new extended quadruplex method. None of the investigated strains were not typeable (NT). Recently, Logue et al. (2017) performed a comparative analysis of phylogenetic assignment of human and avian extraintestinal pathogenic (ExPEC) and fecal commensal E. coli (FEC) strains and showed that in total 13.05% of studied human E. coli strains and 40.49% of avian E. coli strains had to be reclassified. The majority of reassignments among the human E. coli strains involved changes from phylogroup D to F (45 out of 139 reclassifications), A to C (29 out of 139 reclassifications) and D to B2 (26 out of 139 reclassifications), while among the avian E. coli strains, the majority were reclassified from phylogroup A to C (162 out of 377 reclassifications), D to F (139 out of 377 reclassifications), and D to E (26 out of 377 reclassifications) (Logue et al., 2017). Here, we compared phylogroup classification of our strain collections: E. coli from skin and soft tissue infections (Petkovšek et al., 2009), fecal E. coli strains from healthy humans (Starčič Erjavec et al., 2010), and avian fecal strains (Salmič and Stele, 2012), with both PCR methods and with the results presented in Logue et al. (2017) (Table 1). Compared to the latter study (Logue et al., 2017), among our strain collections, more human (27.60% of human) and less avian (23.33% of avian) strains had to be reclassified. Further, among our human strains, the majority involved reclassification from the D to E phylogroup (19 out of 53 reclassifications), and D to F (9 out of 53 reclassifications). On the other hand, only 4 out of 53 involved reclassification from A to C as well as from B1 to E, with the latter not reported by Logue et al. (2017). Further, among our avian fecal strains, the majority of reclassifications were from the D to NT (7 out of 21 reclassifications), D to E (5 out of 21 reclassifications) and A to NT (3 out of 21 reclassifications). Our results thus showed that among distinct E. coli populations, reclassifications to different groups occurred with different prevalences. This is also evident from our data obtained on our collection of 86 fecal E. coli strains from brown bears (Vadnov et al., 2017), where the most prevalent reclassification was from group D to Clade III/IV/V with 25 out of 61 reclassifications, followed by reclassification from D to E (10 out of 61 reclassifications) (Table 1). Further, the high number of reclassifications to NT observed among our avian fecal strains is striking, especially as Logue et al. (2017) reported only a small number for reclassifications to the NT, in total from A to NT only 5 out of 516 reclassifications, while from B1 to NT and from B2 to NT only 1 out of 516 reclassifications. A survey on published studies using the extended quadruplex Clermont method for assigning the phylogenetic group showed even higher numbers of NT E. coli strains: 16 E. coli strains (7.92%) from 202 E. coli strains isolated from in- and outpatients in Burkina Faso (Ouedraogo et al., 2016), 8 E. coli strains (24.24%) from 33 vaginal E. coli strains isolated from pregnant women in Mozambique (Sáez-López et al., 2016a), 9 E. coli strains (6.21%) from 145 vaginal and obstetric infection E. coli strains from women in Barcelona (Sáez-López et al., 2016b), 12 E. coli strains (3.05%) from 393 E. coli strains from mallard ducks in Germany (Rödiger et al., 2015), 38 E. coli strains (27.14%) from 140 uropathogenic E. coli strains in Iran (Iranpour et al., 2015).

To conclude, Logue et al. (2017) stated that the new extended quadruplex method had a significant impact on avian pathogenic E. coli classification and definition of some human uropathogenic strains, a statement we fully support. However, we would like to emphasize that the extent of reclassifications into different groups differ among distinct E. coli populations and that, as with the new extended quadruplex method many strains are NT, the question arises whether there is a need for a revised revised phylogenetic typing method? It is well known that E. coli is a highly diverse bacterial species therefore, it is not unexpected that all strains from novel environments were not phylotyped with the new extended quadruplex method. We believe that there is a clear need to search for NT E. coli strains from novel environments (new hosts in not yet explored geographic regions). Whole genome sequence analysis of such strains should be performed in the search for markers that can be incorporated into a new rapid PCR phylotyping method.