Transgenesis may allow the generation of farm animals with altered phenotype, animal models for research and animal bioreactors. Although such animals have been produced, the time and expense involved in generating transgenic livestock and then evaluating the transgene expression pattern is very restrictive. If questions about the ability and efficiency of expression could be asked solely in vitro rapid progress could be achieved. Unfortunately, experiments addressing transcriptional control in vitro have proved unreliable in their ability to indicate whether a transgene will be transcribed or not. However, initial studies suggest that cell culture may be able to predict in vivo post-transcriptional events. We review these issues and propose that strategies which engineer the transgene integration site could enhance the probability for efficient expression. This approach has now become feasible with the development of techniques allowing animals to be generated from somatic cells by nuclear transfer. The important step in this procedure is the use of cells grown in culture as the source of genetic information, allowing the selection of specific transgene integration events. This technology which has dramatically increased the potential use of transgenic livestock for both agricultural and biotechnological applications, is based on standard cell culture methodology. We are now at the start of a new era in large animal transgenics.