I read with much interest this manuscript that describes Yqi as a new adhesin involved in the early steps of colibacillosis.
When analyzing in more detail the yqiGHI region of strain APEC O1, it turns out that the proteins encoded by these genes in strains APEC O1 and UTI89 are very similar to the proteins encoded by the genes that are present at the same place in strains MG1655 and W3110.
Indeed, when performing a blast search among E. coli proteins using the YqiI protein as bait, a set of proteins clustered in three groups comes out. One group contains 5 Yqi proteins including that from APEC O1 and UTI89, a second group that contains 16 Yqi proteins including that from several O157 strains, strain 536 and strain W3110 and a last group that contains 9 Yqi proteins including that from strain MG1655.
This indicates that Yqi proteins from APEC O1, IMT5155 and UTI89 are actually members of a larger family.
This does not change the conclusions from the authors showing a specificity of the yqi gene for STC95 strains. Indeed, similarities at the nucleotide level are not sufficient to allow positive PCR detection of the yqi gene in non-ExPEC strains.
Rather, this highlights the fact that there may be more than a simple matter of presence of yqi in STC95 strains and absence in non ExPEC strains: one should consider allelic variation among yqi genes. Interestingly, strain CFT073 contains two members of the yqi family that code for proteins belonging to two different clusters.
Additionally, one should also consider the expression of these yqi clusters: indeed, in K-12 strains MG1655 and W3110, an IS2 element is present just upstream of the yqiG gene. It is very likely that this IS2 changes the promoter controlling expression of yqi genes and therefore modifies in some way or another the expression of Yqi proteins.

Whether these variations, in terms of allele or promoter, have any consequences on adhesion properties remains to be tested.