suitable for laboratory-housed pigs (Mench et al., 1998), appropriate methods for organ-source pigs require development and evaluation (Orlans, 2000).
Farm animals might be genetically engineered for human biomedical applications other than xenotransplantation or the production of pharmaceuticals. Research is underway, for example, to produce a porcine model of cystic fibrosis, and there already are farm animal models for retinal degeneration (Petters et al., 1997) and neurodegenerative disease (Theuring et al., 1997). As genetic engineering techniques for farm animals improve— particularly such that single base coding changes that are typical of many human genetic diseases can be introduced, and the production and use of farm animal models becomes more economically feasible—it is likely that more models for disease research and toxicity testing will be developed. Discussion of the potential issues raised by these biomedical uses of farm animals is outside the scope of this report. However, the welfare implications will depend upon specific features of the model under study, including any unalleviated pain and suffering associated with the disease process itself, as well as the need for specialized husbandry and veterinary care requirements (Dennis, 2002).
If genetic technology becomes more efficient and affordable, the primary farming applications of transgenesis and cloning likely will be to produce animals with increased growth, improved feed conversion, leaner meat, increased muscle mass, improved wool quality, improved disease resistance, and increased reproductive potential. The technology also can be used to produce food of improved nutritional quality (nutraceuticals) or appeal.
The primary difference between traditional breeding and genetic engineering is the speed at which change typically occurs (although naturally occurring mutations and recombination events also can cause rapid and dramatic change), and the single-gene nature of genetically engineered change. Traditional methods of selection are more likely to be subject to the checks and balances imposed by natural selection. Many related and apparently unrelated traits are correlated genetically; thus, selective breeding involves selecting for a whole phenotype rather than a single gene product. Because most production and behavioral traits in livestock are polygenic and our understanding of livestock genomes is poor, few traits can reliably and predictably be engineered or introduced by manipulating only one gene (Moore and Mepham, 1995). For this reason, the production of a line of transgenics will require generations of