Gene transfer by pronuclear injection has been accomplished in farm animals by a number of research groups. Applications of this technology for improving milk composition, producing pharmaceutical proteins and manipulating physiology are described. Recent developments in our understanding of gene expression at the molecular level will increase the precision with which genetic changes can be made by gene transfer.

The generation of transgenic pigs is an entirely new way of breeding. In contrast to classical breeding techniques the objects of manipulation in this case are individual genes rather than the entire genome of an organism. In pigs DNA-microinjection into the pronuclei of zygotes is the only available technique of transferring genetic material developed so far. The process involves collection, manipulation, microinjection, cultivation, and transfer of early embryos and also molecular-biological techniques allowing cloning of gene constructs, preparation of suitable injection solutions, and techniques allowing detection of integrated and expressed transgenes in transgenic animals. Gene transfer in pigs usually yields less than 1% transgenic piglets per injected zygote. Our own experiments have shown that simultaneous transfer of untreated control embryos increases yields from 0.5% to 1%.

Gene transfer in pigs can be employed in particular to increase growth performance and carcass composition by using genes encoding hormones of the growth hormone cascade (GHRH, GH, IGF-I). So far, the effects already known from experiments in mice have not been reproduced in pigs.

We are currently investigating whether the transfer of the influenza resistance gene Mx+ of mice will yield disease-resistant pigs.

Breeding with transgenic animals must take into account that approximately 30% of the primary transgenic animals will be mosaics which will not pass on the transgene to their offspring. Unwanted side effects may also occur during gene transfer. Most important examples are instability of integrated transgenes and variability of gene expression over many generations.

In about 5% of all primary transgenic animals integration of the transgene can be assumed to lead to the generation of insertion mutations. Animals carrying these mutations should not be used for breeding. Furthermore severe health problems may be caused by uncontrolled over-expression of the transgene.

Much more work will be necessary in future before we will be able to employ gene transfer techniques in practical breeding programmes.

Gene insertion into the avian germ line is reviewed. Useful pathogen-derived and animal-derived resistance genes that could be inserted into the avian germ line to convey resistance to pathogens are discussed. Methods of germ line insertion that have been successful in mice are summarized and their possible application in future avian germ line research are described. The successful transfer of retroviral genes into the chicken germ line using replication-competent recombinant retroviruses is summarized and several potentially useful transgenic chicken lines are described. The attempts to use one-round replication-defective retroviral vectors have been complicated by the detection of replication-competent retrovirus in chicks from injected eggs.

Because fish lay large numbers of eggs, in which both fertilisation and development is external, they are rather suitable organisms for transgenic induction. To date fish of many different species have been used in such work - rainbow trout, atlantic salmon, channel cat fish, tilapia, medaka, zebrafish, goldfish, carp and loach. Normally the induction of transgenics has involved injection of multiple copies of the chosen novel gene into the egg cytoplasm soon after fertilisation, followed by assays for subsequent gene integration, expression and transmission to progeny.

Genes which have been utilised for transgenic induction in fish include somatotrophin (growth hormone) of both mammalian and piscine origin, metallothionein, 'antifreeze' protein, and crytallin coding sequences. A number of laboratories have demonstrated by Southern blotting of restricted DNA fragments that genomic incorporation has been achieved, although in others there appears to be a continuing problem of long term persistance of extrachromosomal DNA material. Expression of the injected sequences tends to be at a low level and in some cases not to be detectable. This is probably due to the inefficiency of mammalian gene promoters in fish cells and will hopefully be overcome by the use of genes of piscine origin. Germ line integration and transmission to progeny of the novel sequence has been established in a few laboratories.

Transgenic fish may also offer some potential as expression systems for valuable pharmaceutical products. Problems of containment and planned release of transgenic fish are also briefly discussed.

Conventional animal breeding is in a healthy state, with good rates of genetic change possible, and economic efficiency being continually improved in the different livestock species. New reproductive techniques offer faster rates of change, as with the use of embryo transfer in sheep and cattle breeding. It is now possible to introduce foreign DNA, transgenes, into the germ line of farm animals and this offers new opportunities in the genetic improvement of economic merit. The transgenes should be directed at traits of high economic value, and be made in elite breeding stocks to benefit from past genetic improvements. Much still needs to be learned about engineering transgenes and about the basic biochemical-physiological processes they are designed to affect. However many transgenes will be produced over the next decade, and some may be useful in animal improvement. Founder transgene individuals are unique for incorporation site, copy number and expression, and will require screening, multiplication, testing and evaluation, in both hemizygous (TO) and homozygous (TT) form, for all economic traits. For a transgene with a large useful net effect on economic merit, the most rapid genetic response will come from fixing it in the elite breeding stock. An effect of 5-10 percent in economic merit will be needed to offset the normal genetic improvement displaced. Current experience with transgenes is of very large effects on the target trait, but with deleterious effects on fitness and economic merit. Control of the level, tissue and time of expression are crucial to the practical use of transgenes. With uncertainty in the results of a transgenic program, conventional programs will have to be maintained, so extra investment will be required. Patent rights may bring extra benefits to breeders, but there are also risks if the stock prove defective. Breeding strategies and organization may also change. Despite initial problems and uncertainties, the power of the transgenic methods is so great that the development of transgenic stocks is likely to become a very important tool in genetic improvement of livestock in the future.

In an effective animal breeding programme the objectives are appropriate, the population structure is good, suitable records are collected, and these records are used in an optimal way so that animals likely to have the best offspring are chosen.

Estimates of genetic parameters such as heritabilities and correlations influence the design of the programme: for example, the optimal number of daughters to progeny test per dairy sire or, more fundamentally, whether to adopt performance or progeny testing, whether to record food conversion efficiency or simply growth rate, and whether or not to collect records on some indicator trait such as blood urea nitrogen level. The estimates of parameters also influence the weightings given to different traits and to individuals and relatives' performance in breeding value estimation and selection. Methods for estimating parameters have become much more sophisticated as a result of advances in theory, particularly, restricted maximum likelihood (REML), in computer power, and in development of special computer programs.

The beef cattle industry in the United States has undergone dramatic changes over the past decade with the adoption of genetic evaluation programs. The method of choice has been Henderson's mixed model methodology for best linear unbiased prediction (BLUP). The most prevalently used model is the animal model (Henderson and Quaas, 1976) computed by the equivalent reduced animal model (Quaas and Pollak, 1980).

Neither the methodology or the models being used are particularly new. What is new in this industry is the widespread application of these techniques to the analysis of the data banks maintained by the breed organizations. Today many breed associations publish a national sire evaluation, and most of these have published their first in the last three years. This rapid proliferation of published evaluations has coincided with an attitude in the industry of promoting specification beef and predictable performance. Genetic evaluations provide information not only to achieve goals in selection but as well for merchandizing cattle based on quantifiable potential. The enthusiasm for genetic evaluations right now in the U.S. beef industry is high.

Past improvement programmes in pigs have concentrated on lean growth, and have relied on direct comparisons of contemporaries in a common environment. The new mixed model technology now allows comparisons across environments and generations, and, by utilizing all possible genetic relationships, more accurate prediction of genetic merit for traits of low heritability. In addition, breeding programme design will benefit from more precise estimation of genetic parameters. Principal benefits of the new technology are greater flexibility of structure of breeding programmes, and faster improvement of litter size. The consequence will be larger populations under selection with lower costs per unit of genetic improvement. For national improvement schemes using a high proportion of AI matings, central testing stations could become unnecessary. Main research priorities are the optimum family structure to balance selection and inbreeding, more efficient computing strategies for large numbers of traits, and the incorporation of single genes into predictions of merit. The new technology will be important for the exploitation of biological advances in manipulation of both reproduction and the genome.

Methods of evaluating dairy cattle using mixed models with Best Linear Unbiased Prediction properties have progressed from the sire model to the animal model. Definitions of effects in models need refinement, particularly for contemporary groups. Pedigree selection and progeny testing is the standard for producing sires used in artificial insemination, but multiple ovulation and embryo transfer schemes are being tried. Efficient production is necessary under conditions of surplus. Efficiency can be achieved by higher production per cow and reducing costs by improved reproduction, increased herd life, reduced health costs, and reduced dystocia. Preferential treatment is a major problem. New biotechnological developments such as bovine somatotropin, mitochondrial genetics, sexing semen, embryo transfers, cloning, transgenic animals, and markers are considered as potential new technologies that may be useful for dairy cattle improvement.

1. It is shown how the growth of an animal in non-limiting conditions can be predicted from the values of a small set of animal variables. The functions used are derived from a small set of simple, and reasonable, assumptions.

2. Growth may be limited by intakes of nutrients and energy which are less than those needed. The prediction of growth where such limitation occurs calls for the prediction of the intakes. There may be genetic variation in the abilities of animals to deal with the relevant constraints which are, commonly, feed bulk or hot environments.

3. Breeders need to pay more attention to giving sufficient descriptions of their stocks.

There is an increasing possibility that animal performance may be manipulated by new techniques incorporated into efficient breeding programmes. In order to fully exploit these techniques to increase the efficiency of growth, gene products involved in the control of growth, development and also reproduction must be identified. Correct expression or manipulation of these genes to provide optimum effect on a trait requires complete knowledge of their biological action. The transgenic approach has provided some success using growth hormone and growth hormone releasing hormone to increase growth in mice. Further possibilities exist, since recent work suggests that peptide growth factors such as the insulin-like growth factors, fibroblast growth factor and transforming growth factor-β may be primary regulators of normal growth. This brief overview introduces some of these peptide growth factors, provides basic information on them and offers them for discussion as potential targets for manipulation.

Data from 10 years of German random sample tests (1978-87) were analyzed to document recent changes in 5 white-egg and 5 brown-egg strains which had participated fairly regularly in the tests. In the absence of a control population, it is only possible to compare rates of changes between strains and the overall trends of white-egg vs. brown-egg strains. The data show that improvement continues at average levels of performance which must have seemed utopian 20 years ago, when it was popular to question whether any genetic progress was being made.

The methods used by different breeders to achieve the apparent improvements are generally not disclosed, but it seems reasonable to assume that all programs more or less follow the same pattern: combined pure-line and cross-line testing, approximations to classical index selection using individual and family information, with emphasis on different traits depending on economic importance and pressures to keep pace with competitor stocks.

The five brown-egg strains on average have made more progress than the five white-egg strains, but there are also obvious differences in rate of change among strains with the same shell colour. In some instances, the changes may involve the replacement of one or more lines making up the strain cross, which is not reflected by a change in product name.

Overfatness of lamb and beef carcasses is a problem in many countries. Genetic improvement of carcass composition, particularly in specialised terminal sire (meat) breeds, should lead to permanent, cumulative, and cost-effective improvements. In the past, selection in these breeds has been mainly on liveweight. However, the use of in vivo estimates of carcass composition is expected to lead to higher rates of progress. In ruminants index selection is likely to be most useful, and early results from selection experiments are encouraging. Although schemes for recording performance and estimating breeding values are in operation in many countries, structural problems such as small herd or flock sizes, and lack of comparisons between herds or flocks impede progress. To some extent, these problems could be overcome by cooperation amongst breeders and wider use of new statistical techniques. Genetic improvement of meat production could be further enhanced by improved techniques for in vivo estimation of carcass composition, and by use of embryo transfer to increase “annual selection intensity” of females.

Genetic selection for litter size received little attention either by scientists or commercial pig breeders until recently. Low heritability of litter size, its sex-limited expression and the fact that it cannot be measured until reproductive maturity are often given as reasons. But the major reason may be that the reproductive rate of pigs is already quite high, and the economic value of genetic changes in litter size have been small relative to those of efficiency and composition of growth.

The economics of pig production have changed and technological and biological advances have caused us to re-examine the economic value of traits. Reducing fat in the carcass and more efficient conversion of feed to lean meat have always been, and will continue to be, high priority selection objectives for pig breeders. But expensive, environmentally controlled gestation, farrowing and nursery houses are the norm of modern pig production and increasing litter size will result in a significant reduction in the cost of producing lean meat (Tess et al., 1983).

Selection for meat production in poultry includes commercial development of meat-type chickens, turkeys and waterfowl and laboratory studies that involve these types of poultry. Heritabilities for growth traits are moderate to high, enabling response to selection for these traits. Changes in growth traits result in correlated responses in feed intake and have a negative relationship with reproductive traits. That is, deviations from intermediate growth rates are deleterious to reproduction. In this paper we discuss genetic aspects of poultry meat production and long-term laboratory selection experiments for growth. Emphasis is given to the species with the most extensive research base, the chicken.