Current transgenic livestock technology commonly relies on the random insertion of a gene construct into the genome. In this case the activity of the transgene can be strongly influenced by the specific chromosomal context of the integration site a phenomenon commonly referred to as position effect. As a consequence the level of expression of a randomly inserted transgene can vary greatly or even result in adverse effects to the animals due to disruption of or interference with an endogenous gene. This unpredictability of transgene activity requires the generation and characterization of multiple transgenic cell and animal lines which is a major disadvantage for applications involving large animals due to the long gestation times and high costs involved. The randomness of the process can be avoided by the application of homologous recombination (HR) technology for the introduction of site-directed modifications. Due to the unavailability of livestock ES cells HR approaches in livestock are presently restricted to the use of somatic cells which are hampered by very low efficiencies. The combination of the random chromosomal insertions of recognition sites for specific DNA recombinases (such as Cre or Flp) and the subsequent site-directed insertion of a transgene into this locus by a recombinase potentially offers much greater efficiencies. This recombinase mediated cassette exchange strategy provides a valuable system to efficiently produce transgenic livestock with predictable transgene expression levels. It also allows for the production of transgenic animals without antibiotic selection markers which are commonly used for the isolation of stably transfected cells but typically become part of the modified genome although they serve no useful function in the animal. An alternative approach is the use of episomal vectors which do not integrate into the genome but are maintained alongside the chromosomes as independent entities. They can thus ensure predictable expression and provide an elegant solution for potential problems associated with integration into the genome. However emerging episomal vector systems are commonly of viral origin and dependent on viral sequences and factors. This has major drawbacks because the viral elements may be recognised as invading DNA and become permanently silenced or may even trigger the immune system raising general biosafety concerns. The recent development of a novel self-replicating episomal vector system based on the presence of a scaffold/matrix attachment region hold much promise as it can function independent of any viral factors. Moreover the system has been engineered for the recombinase-mediated deletion of the bacterial vector backbone to limit the episomal vector that replicates in synchrony with the host cell chromosomes to its essential functional components. In the future it will be important to demonstrate that these new concepts can live up to their promise and deliver greater predictability efficiency and safety.