Model meat systems were produced using both aseptically procured and irradiated raw minced beef to initially compare the growth characteristics of a three-strain mixture of Escherichia coli O157:H7 in each medium. A multiplex PCR assay (detecting VT1, VT2 and eae genes) was used to determine the proportion of individual strains recovered at each sampling time by virtue of the different combinations of these virulence factors encoded by each strain, and to investigate their genetic stability. No differences in the growth characteristics of the pathogen (P>0.05) were recorded in the meat matrices, irrespective of the preparation method, thus validating the use of irradiation to sterilise (42 kGy, in vacuo at <-5 o C) minced meat in the production of control meat matrices for application to further research investigations. A novel plating method, incorporating a period of catalase-induced resuscitation, was found to give significantly higher recovery of the pathogen (P<0.05) from these meat matrices when compared to conventional spread-plating on Trypticase Soy Agar (TSA). A proportion of the recovered populations, however, was found to be failing to produce an amplicon relating to a Verotoxin (VT) gene. A further investigation demonstrated that O157:H7 strain E 90197 (VT2 and eae positive) was displaying genetic instability during growth in and recovery from meat matrices with the apparent loss of the VT2 gene. An alternative VT2 and eae positive strain (ATCC 43889) demonstrated no such genetic instability, demonstrating inter-strain variation for this phenomenon. Thus, it should be considered that an essential pre-requisite to studies relying upon the maintenance of the pathogenic potential of E. coli O157:H7 should be an evaluation of the genetic stability of proposed strains of this serotype.