Effective breeding programmes down under.

Abstract

Australia is a very large country far away from the rest of the world. And while Australia may have had its fair share of well-known geneticists, what happens on the ground is what really matters. The breeding programmes in sheep and cattle are remarkably successful, even though breeding decisions are made by many independently operating individuals. Progressive and proactive breeders have been keen to embrace the technology and schemes delivered by the science of animal breeding. They have acted in their own interest; with a shared fate of contributing via the breeding programme to the same common benefits. How does that work? First, a bit of history. Australian animal production systems are mostly very extensive. They are based on low input costs and are directly competing in world markets. For much of the 20th century, the animal products, in particular wool from Merino sheep, helped Australia maintain one of the highest living standards in the world. By the 1950s, wool was synonymous with the Australian way of life along with the saying “Australia rides on the sheep's back.” Since the early 1970s income from wool has declined considerably due to competition from cotton and synthetic materials, both in absolute terms and also in relation to the price of the other major sheep product—meat. Consequently the Australian sheep population has reduced from 180 million to approximately 70 million over the last 25 years. Beef cattle numbers in Australia were long steady at 10 million, but with the introduction of Bos indicus breeds in the 1950s, the cattle population increased rapidly to approximately 30 million in the 1970s. It is still around that level, although the world market as well as severe droughts can have a large impact on the total herd. Regarding other species, there are around 2 million pigs, 1.5 million dairy cows and close to 100 million chickens in Australia. Much of the seedstock in poultry and dairy is imported, whereas Australia has a ban on importing genetic material for pigs. Australia is a competitive agricultural exporter, with two thirds of total production exported. Agricultural products account for 16 per cent of Australian merchandise exports, with nearly half of these products being from livestock. In general, when reproductive rates are high in the target species of selection, there are only a few nuclei or breeders playing a role. This is particularly the case in plant breeding. Poultry and pigs follow the same pattern. In dairy cattle, there is a very open world market of top genetics, drawing on approximately 10–15 breeding programmes across a number of developed countries, and at the end of the day, much of the genetic improvement in most countries is imported, reflecting global flows of genetic material. The global concentration of breeding programmes may not be desirable, especially when production environments differ or are subject to change. Australia has a unique environment with heat waves and particularly droughts, phenomena that are predicted to increase in intensity. The ability to survive heat and drought is already part of animal breeding programmes implicitly, but the specific form of genotype by environment interaction or adaptation demands explicit attention. Genomic information can provide scope for a much better handle on estimating the interaction. Challenges may lie in data recording in extreme harsh conditions. Beef cattle producers in the tropical Northern Australia endure highly difficult production environments, where survival is often higher on the agenda than data collection. Australia also has a unique breeding industry, especially in sheep and beef cattle. Due to the extensive nature of these industries, limited use of AI arising from the scale of enterprises and the low reproductive rates of the species, there are very many contributors to genetic improvement. Since the introduction of national genetic evaluation systems, such as BREEDPLAN for beef and LAMBPLAN for meat sheep in the 1980s, much genetic progress has been achieved. There are no large breeding or AI companies in sheep and beef. The breeding structures can be considered as large dispersed nucleus breeding programmes. What makes the activities of the many breeders coherent is the existence of national multitrait genetic evaluation. The ability to select the highest merit animals across herds (or flocks) seems so trivial that we hardly realize that this is at the core of dispersed breeding programmes. The national breeding programme is working via Adams Smith's invisible hand. The structure of a breeding industry is quite critical when it comes to implementing and modifying breeding programmes. In strictly multitiered structures with closed nuclei, large companies are in charge and there is usually limited debate about strategies while there is good access to funds for introducing new technologies. Further, breeding values are not a currency, and seedstock is protected. Breeding objectives are usually pretty clear and consistent as well. Therefore, these breeding structures face few hurdles in achieving substantial genetic gain. Even though dairy cattle breeding programmes are concentrated around (large) AI companies, bull proofs are an extremely important currency— perhaps overemphasized (Bichard, 2002; Livest. Prod. Sci. 75: 1-10)—and, interestingly, the seedstock of the nucleus is traded rather than protected. Beef cattle and sheep are at the minimal end of this structural continuum, and their breeding programmes are relatively flat—bull or ram studs derive most of their income from sale of herd bulls or flock rams that will be used for natural mating. There are many decision-makers, each with a small role relative to the total breeding population. Most importantly, breeders generally receive only a fraction of the value of genetic improvement that they are generating. These parameters are not conducive to the uptake of efficient improvement, and they are the reason why breeding programmes in the extensive industries often fail, for example in developing countries. However, the sheep and beef industries in Australia are making genetic gains of more than 1% per annum, which, relative to potential gain, is at par with many breeding programmes in the intensive industries. New technologies can change the structure of breeding programmes. AI caused the first technology wave and shaped dairy cattle breeding programmes towards progeny testing, as was mapped out by Robertson and Rendel as early as 1950 (J. Genet. 50, 21–31). The embryo technologies in the 1980s led to the formation of centralized MOET dairy breeding programmes, but did not topple the traditional progeny testing schemes, for reasons related to risk, as outlined by Bichard (2002; Livest. Prod. Sci. 75: 1–10). Genomic selection has given the young bull testing a final blow, and it will be interesting to see what breeding structure will emerge when this is combined with embryo and possibly gene editing of monogenic traits. The benefits of genomic selection in beef and cattle are less obvious than in dairy cattle. The predicted rates of gains are much lower, as target traits are not sex-limited, and selection takes place in young animals with much less scope to reduce generation interval. And rather than saving money from avoiding an expensive progeny test, the investment in the breeding programme might have to double or triple because of animal genotypings. For an individual breeder, maximum economic returns are the main objective, rather than maximum genetic gain, and although investing in genetic improvement may be cost effective at the national (population) level, it is often not so at the level of the breeder that has to make the investment (Smith, 1978; Anim. Prod. 26:101-110). Therefore, the rates of gain are often not as high as they could be, simply because there is market failure, resulting in suboptimal use of technologies and suboptimal recording of important traits. The invisible hand is therefore not working perfectly. In Australia, levy (and tax) payers’ funds have been used to address some of these issues. For example, “information nuclei” have been created to have reference populations for hard to measure traits. Genomic selection has been implemented in the Australian beef and sheep breeding programmes and is now used by the more progressive breeders who can now select on an index that improves meat eating quality. Interesting scientific challenges have emerged as well, notably those related to genomic prediction and selection in the cattle and sheep populations that have a multibreed and cross-bred nature. New applications and tools can be invented, related to selection, culling and sorting of animals into markets according to their genetic potential. An effective R&D programme is useful only if there is an effective industry that actually implements the new ideas. In Australia, this seems to work pretty well.