Escherichia coli have been associated with IBD in people, dogs and mice. The purpose of this study was to determine if E. coli associated with IBD across species share common features that are pathoadaptive for IBD. E. coli was isolated from the inflamed intestinal mucosa of people with Crohn's ileitis (ICD, 25 strains), dogs with granulomatous colitis (GC, 35 strains) or lymphoplasmacytic colitis (LP, 7strains), and mice with ileitis (2 strains). E. coli from the non-inflamed ileum of people (39 strains: 14 non-IBD (NIBD), 9 colonic Crohn's (CCD), 16 ulcerative colitis, UC), and the colon of healthy dogs (34 strains) were included as controls. CD-associated E. coli LF82, murine IBD inducing E. coli NC101 and DH5-a were used as reference strains. E. coli were examined for serotype (O:H), phylogroup (A, B1, B2, D), multi locus sequence type (MLST-7), virulence genes (ratA, PmtI, colV, hcp, lpfA-141, lpfA-154 kpsMII, fyuA, iss, malX, typeII, typeIV, afaBC, focG, ibeA, papC, sfaDE, cnf1, Stx1 and eae), and their ability to invade Caco-2 cells. A wide variety of serotypes were present in E. coli isolated from inflamed and uninflamed mucosa. O - was the most common among human (19%) and dog (24%) strains. O25 was the second most common human O type (9%), whereas O4 and O83 were detected in 16% and 9% of dog strains respectively. E. coli from ICD, GC and NIBD were distributed among all four phylogroups. B1 was most frequent in ICD (36%), GC (34%) and murine ileitis (100%), whereas A was most common in NIBD (36%), and B2 in healthy dogs (44%), but these differences were not significant. In contrast, E. coli isolated from the non-inflamed ileum of UC were predominantly B2 and D (63% B2 and 19% D, P= 0.001). Phylogenetic analysis of MLST-7 sequences revealed E. coli strains clustered by phylogroup independent of disease association or species of origin. It is noteworthy that a well-defined cluster of B2 strains (bootstrap value of 100) isolated from GC, ICD, UC and CCD all shared serotype O25:H4. Virulence genes segregated by phylogroup rather than disease association or species of origin: ratA, colV, kpsMII, fyuA, malX, focG, ibeA, papC, sfa DE and cnf1 were more prevalent among B2 strains, while presence of lpfA was associated with phylogroup B1 (p<0.001). There was no significant effect of disease association, species of origin or phylogroup on the ability of E. coli to invade Caco-2 cells. However, the presence of lpfA was associated with Caco-2 cell invasion by human E. coli strains (P<0.05). We conclude that no single cluster of E. coli is associated with IBD in people, dogs and mice. We found no evidence to link B2 and D strains to ileal CD, or GC in dogs. In contrast, we found that E. coli of phylogroup B1 were frequently isolated from ICD, GC and murine ileitis, and the B1 associated gene encoding long polar fimbriae (lpfA-154) was associated with increased invasion of caco-2 cells.