Previous PageTable of ContentsNext Page

CHAPTER 20
SELECTION AMD CROSSBREEDING STRATEGIES IN ADVERSE ENVIRONMENTS

by

E.P. Cunningham
The Agricultural Institute, Dublin, Ireland

Summary

The methods used for the improvement of livestock populations in the developed parts of the world have evolved in recent decades to a high degree of organization, and often of efficiency also.

In contrast, livestock improvement programmes are now being undertaken in many under­developed countries in which the environment is adverse in two respects:

  1. There may be severe limitations to animal productivity imposed by climate, nutrition, health, management practices, or more usually, a combination of these.
  2. The infrastructure which supports the livestock improvement programme in more developed economies may be partially or entirely lacking.

Where nutritional, climatic, health and managerial stresses limit animal production, the models used as a basis for breeding programmes in favourable environments may no longer apply. Genetic variances and co-variances for production traits may be quite different. In addition, the traits connected with stress tolerance may be very important. Generally, large inter-population differences can be observed for these traits. Depending on whether these inter-population differences are additive or non-additive in nature, different strategies for exploiting them may be called for. furthermore, the extent of additive or non-additive variation may depend on the degree of stress in the environment. The literature in this area is explored and some possible breeding strategies are described.

The absence of intrastructural support means that improvement programmes which are standard in developed countries may not be operable. However, because the populations con­cerned are at a different stage of development, it may be possible to operate highly effect­ive but much less sophisticated programmes. Such an approach is described.

20.1 Introduction

The methods used for the improvement of livestock populations in the developed parts of the world have evolved in recent decades to a high degree of organization, and often of efficiency also. In poultry, it is common for 100% of the breeding population to be involved in intensive recording and selection programmes. In dairy cattle, over 50% of the population is frequently involved, while in beef cattle and sheep, the percentage is usually lower. However, in all cases, sophisticated techniques of evaluation and of data analysis are now part of the working programme. These systems are usually developed in parallel with an intensification of the husbandry conditions, and the limitations to their effectiveness for genetic improvement are set by the rate of generation turnover, and sometimes by the exhaus­tion of genetic variance. Occasionally, population structure inhibits the rate of improve­ment, or the investment in testing and selection may be insufficient. In general, however, the constraints are simply those inherent in the reproductive potential of the species.

This rapid development has been made possible first by the solid economic base provided by modern animal husbandry techniques. This justifies the expenditure on recording, artificial insemination, technical advice, and other elements of the infrastructure which make a planned breeding programme possible. Secondly, these programmes have developed on the basis of well tested models of the genetic background for the traits and species concerned. Thirdly, since many of the programmes require the rapid assembly and evaluation of large amounts of information, their full" development has depended on the evolution of data processing machinery and analytical methods.

In contrast, livestock improvement programmes are now being undertaken in many developing countries in which the environment is adverse in two respects:

  1. There may be severe limitations to animal productivity imposed by climate, nutrition, health, management practices, or more usually a combination of these. The genetic models which have been used in benign environments may no longer be appropriate in these circumstances.
  2. The infrastructure which supports the livestock improvement programmes in more developed economies may be partially or entirely lacking. Difficulties of cost, literacy, communication, and even economic security may greatly inhibit the consistent operation of a livestock improvement programme.

In this paper, I wish to consider some aspects of the genetic models which underlie planned breeding sohemes, and to examine some limited evidence on their applicability in stressful environments. I also want to discuss some aspects of the operation of breeding schemes in situations where the infrastructural support is minimal.

20.2 Selection vs crossbreeding

Selection and crossbreeding are often regarded as mutually exclusive avenues for the improvement of the livestock population. Particularly in countries with poor infrastructure and with relatively unselected local strains of livestock, there are often strong temptations to regard crossbreeding as the only necessary step to be taken.

However, as I have pointed out in a contribution to a previous FAO consultation (Cunningham, 1979a) any crossbreeding strategy which is adopted for a population requires, at some point in the programme, an indigenous selection operation. Some of the broad strategic options are:

  1. Improving the local population from within. This naturally requires a domestic selection programme.
  2. Population replacement by top crossing with an exotic strain. An internal selection programme will be required after the first few generations in order to make further progress, and in order to mould the introduced genotype to the local environment.
  3. A gene pool or synthetic population formed by planned proportions of domestic and exotic genes. Selection will be required once the planned proportions have been achieved, though it may be complicated by heterotic and maternal effects.
  4. Rotational crossbreeding, designed to maintain high levels of heterozygosity and at the same time to achieve specific proportions of the domestic and exotic strains. Variants of this could be reciprocal backcrossing between a native and an exotic strain, or between a native strain and an F1 between native and exotic. In either case, a selection programme will be required in the domestic population from which the crossing sires are derived.
  5. Grading up to F1 males. This option, already suggested by Hickman (1978) seems particularly attractive for populations which it is desired to stabilize as a 50% mixture of local and exotic strains. However, it requires the development and selection of a nucleus within the domestic population.

The common feature of all these programmes is that, while they may rely on crossbreeding to exploit heterosis, to achieve breed replacement, or to establish and maintain a combination of two or more strains, they all require an indigenous selection programme as part of the operation. This situation is shown schematically in Figure 20.1.

20.3 Genetic models for crossbreeding and selection programmes.

The genetic models underlying selection programmes within most developed livestock populations are relatively simple and straightforward. For quantitative traits, a knowledge of heritabilities, genetic and phenotypic correlations, phenotypic variances and. appropriate economic weights, permits the development of selection programmes which can then be adjusted to give optimum gains in either genetic or economic terms. While full optimization is seldom achieved, results of such programmes, where measurement of genetic change has been possible, have generally been found to be reasonably in line with expectation.

Within-population selection programmes are much more difficult to implement in developing countries, for the obvious reasons that technical performance per animal if often a secondary concern, and that the organizational infrastructure may be very weak. Further, information on the nature of genetic variation in a population may be scanty. Nevertheless, as I have indicated above, almost any attempt to improve the population will require some domestic selection exercise. In these circumstanoes, it may be possible to ignore the absence of infrastructure, and to combine intensive recording and selection in a central herd or flock with recruitment of individual animals from a much wider population on the basis of some very simple screening procedures. This, in effect, is a kind of open nucleus breeding structure with a different kind of recording system operating in the nucleus and in the supporting population. I have described, in somewhat more detail, how such a system might operate for a dairy cattle selection scheme (Cunningham, 1979a).

A key element in the implementation of selection (and crossbreeding) schemes in cattle, and to a lesser extent in other species, is artificial insemination (A.I.). The easy transport of frozen semen, and now frozen embryos, makes possible the rapid and widespread introduction of new genotypes into a population. While, in general, the effects of this transfer of genetic material are probably beneficial, at least in the short term, this easy access to exotic genotypes is in some instances leading to the obliteration of local populations. Where population replacement is the deliberate and planned intention, then this may be reasonable. However, I fear there must be many instances where population replacement is taking place simply because exotic semen is the only option available through A.I. I believe therefore that there is a very good case for promoting the idea of a counterpart selection and semen producing scheme in the domestic population wherever a major exotic semen importation programme is planned. This would serve several useful purposes. It would enlarge the strategic options open to those involved in planning breed improvement. It would provide a medium for the propagation of improved strains of the local population, and therefore stimulate the interest in developing a local improvement programme. By so doing, it could help to generate improved versions of the local population which could be the local partner in any attempt to develop a synthetic strain, or to use in rotational crossing with the exotic. Finally, it would ensure that the chances of survival of the local strain in the face of competition from the exotic strain would be enhanced. The converse of this last point is that the surest guarantee of the eventual extinction of the local strain is to abandon all attempts at internal improvement, particularly in the face of extensive crossbreeding with a selected exotic strain.

Crossbreeding systems, other than those aimed at breed replacement, are designed to exploit complementarity between the strains involved, or heterosis, or both. Often, it is not possible to distinguish between these two. However, it is necessary to have some idea of their relative importance in order to make rational decisions about the choice of breeding structure. The simple, cumulative dominance model used as the usual explanation for heterosis, implies that the degree of heterosis is a linear function of heterozygosity. There is a good deal of evidence to support this model, e.g. McGloughlin, (1980). However, there are also instances which appear to contradict it. In particular, it has been a fair general experience that, in dairy cattle, the extensive heterosis observed in F1 crosses between temperate and tropical breeds is almost completely lost in the F2 and subsequent generations, e.g. Buvanendran, (1977).

Is it possible to reconoile these conflicts of evidence? I believe it is, but it requires a rather complicated genetio model which involves heterosis, additive differences, threshold effects, and interactions of all of these with environmental levels. Such a model is shown in Figure 20.2.

This version of the model is related to the experience of crosses of temperate and trop­ical cattle in stressful and benign environments. A low-producing local breed and a high-potential exotic breed may not differ greatly in production (however this is defined) in very stressful conditions. However, the F1 frequently displays remarkable performance, indicating heterosis, sometimes of the order of 100% or more. By contrast, in the very favourable environment, the unimproved local strain does not significantly improve in performance while the high potential exotic strain performs exceptionally well. The F1 is not substantially different from the mid-parent indicating very modest heterosis. In most of the trials con­ducted to-date, the exotic has not been included in a pure form, so the only measures avail­able are the performance of the local strain and the F1 with, occasionally contemporary F2 's. On this model, the substantial difference between the F1 and the local strain is largely due to heterosis in the poor environment, and is largely due to the additive differ­ence between the strains in the benign environment. This could perhaps be explained if the physiological limitations in the poor environment were those of stress tolerance, while the same difference in the benign environment were largely due to the difference in additive genetic merit for production, perhaps mediated largely through appetite. The nation of a threshold can be invoked to mark the point on the scale of environment at which stress tolerance ceases to be the dominant factor, and additive differences in production become more important. Of course, there may not be a sharp dividing point of this kind, but it is conceivable that there are physiological thresholds of, e.g. temperature maintenance, below which an additive genetic potential for production is largely inhibited. A somewhat similar explanation of the differences in performance between Bos taurus and Bos indicus strains in tropical conditions have been put forward by Frisch and Vercoe (1978).

Such a model can also be phrased in another way: as a double genotype by environment interaction, in which both additive and non-additive genetic differences are dependent on the environment. To explore the plausibility of the model we should therefore look for evidence that such interactions do occur. Examples of additive genetic x environmental interactions have been reviewed by Pani and Lasley (1972). A systematic search for inter­actions of heterosis with environment has not, to my knowledge, been undertaken before. The literature in this area is currently being reviewed by Barlow (1980), and such inform­ation as there is produces a fairly complex picture.

Following the general theory that heterozygotes are in some way better buffered against the environment than homozygotes, one might expect heterosis to be more strongly expressed in stressful environments than in favourable ones. Broadly speaking, where stress is a function of temperature, this hypothesis is supported by the experimental data in both plants and animals. In fungi, Connolly (1977) has shown the opposite to be the case, with percent heterosis increasing linearly with performance over a temperature scale from 15° to 30° C. Where stress is nutritional in origin, conflicting experimental results exist, in part, no doubt, because the nature of the stress is much more variable and less well defined. In five studies in pigs in which heterosis was measured under varying nutrit­ional levels, the general indication was that it was greater under less intensive feeding regimes (Sellier, 1976). In six studies in beef cattle, heterosis was measured, for the most part, under intensive feedlot systems, and also under more extensive pasture finishing. In five of the six studies, heterosis appeared to be greater on the more intensive feeding system. In seven studies in mice, the balance of evidence indicated higher percentage heterosis under conditions of temperature stress than at more favourable temperatures. There was no evidence of differential heterosis with varying nutritional level.

The picture therefore is far from clear. This is perhaps not surprising, because, for the most part, these observations are drawn from experiments not specifically set up to test this point, because the range of environments involved may not have extended to a point which could be called stressful, because we know nothing about the selection history of the strains involved, and because the traits and species are so diverse.

However, the question posed has a fairly fundamental bearing on the immediate genetic future of many livestock populations. The basic problem is this: livestock populations in many developing countries have now begun massive top-crossing programmes with exotic strains; the F1 progeny of these crosses are generally outstanding; where they have been tested, F2 's ana subsequent generations have often been well below expectation; the decision on whether to proceed to the development of a synthetic, to continue upgrading, or to aim for some systematic crossbreeding system which maintains heterozygosity depends on a knowledge of how important a part heterosis plays in the superiority of the crosses. Until this question is resolved, it is impossible to plan either the defence or the development of these populations on a sound basis. I would therefore consider that the two problems in the genetic evolution of local livestock populations which most orgently need scientific attention are:

  1. Clarification of the role of heterosis, additive differences between populations and the interaction of both of them with variations in the environment, leading to adequate genetic models on which breed planning can be based.
  2. Development of open nucleus selection schemes within local populations, which require minimal infrastructure, and which can be used as either a complement to planned cross­breeding programmes, or as a major contribution to ensuring the competitiveness of the local population in the future.

20.4 References

Barlow, R. 1980. Personal communication.

Buvanendran, V. 1977. Production characteristics of Jersey-Sindi grades in Sri Lanka. Aust. J. Agric. Res. 28, 747-53.

Connolly, V. 1977 • Unpublished data.

Cunningham, E.P. 1979a. The importance of continuous genetic progress in adapted breeds. Report of the FAO Expert Consultation on Dairy Cattle Breeding in the Humid Tropics, PP. 35-41 FAO, Rome.

Cunningham, E.P. 1979b. The state of quantitative genetic theory. Proceedings of the Inaugural Conference, Australian Association for Animal Breeding and Genetics, Armidale, Australia, pp. 18.26.

Frisch, J.E. and Vercoe, J.E. 1978. Utilizing breed differences in growth of cattle in the tropics. World Animal Review 25, 8-12.

Hickman, C.G. 1978. The estimation and use of non-additive genetic variability in cattle and buffalo. Mimeo, 10 pp. FAO, Rome.

McGloughlin, Patricia 1980. The relationship between heterozygosity and heterosis in reproductive traits in mice. Animal Production 30, 69-77.

Sellier, P. 1976. The basis of crossbreeding in pigs: a review. Livestock Production Science 3: 203-226.

Figure 20.1 Strategic options involving crossbreeding and selection

animal genetic resources conservation and management

Figure 20.2 A model for differential heterosis and additive effects of two strains in good and poor environments.

animal genetic resources conservation and management

Résumé

Les méthodes utilieseés pour l'amélioration des populations animales dans les régions déiveloppées du monde sont devenues, au cours des dernières décennies, très systématiques et souvent très efficaces. Dans l'aviculture, il est commun qu'une population reproductrice tout entière fasse l'objet de programmes intensifs d'enregistrement des performances et de sélection. Dans le cas du bétail laitier, de tels programmes s'appliquent souvent à plus de 50 pour cent de la population, tandis que pour les bovins de boucherie et les ovins, le pourcentage est généralement inférieur. Toutefois, dans tous les cas, les programmes de travail comportent désormais des techniques raffinées d'evaluation et d'analyse des données. En général, l'application de ces systèmes va de pair avec 1'amélioration des conditions d'élevage, et leur efficacité, en matière d'amélioration génétique, n'a pour limite que le taux de re-nouvellement des générations et parfois l'épuisement d'une variance génétique. II arrive que la structure de la population freine le rythme de l'amélioration, ou que les investissements consacrés aux essais et à la sélection soient insuffisants. Cependant, d'une manière generale, il n'existe d'autres contraintes que celles qui soient inhérentes au pouvoir de reproduction de l'espèce.

Cette rapide évolution a été rendue possible, en premier lieu, par la solide base économique qu'assuraient les techniques modernes d'élevage. C'est ce qui justifie les dépenses consacréesà l'enregistrement des performances, à l'insémination artificielle, aux services techniques et aux autres éléments d'une infrastructure qui permet l'application d'un programme d'amelioration génétique. En second lieu, ces programmes ont été établis en partant de modèles dont on avait soigneusement controlé les antécédents génétiques relatifs aux caractéristiques et à l'espèce. Troisièmement, étant donné que beaucoup de programmes exigent le rassemblement et l'évaluation rapide d'un grand nombre d'informations, leur plein développement a ' également été conditionné par l'évolution des procédés de traitement mécanique des données et des méthodes d'analyse qui les accompagnent.

En revanche, des programmes d'amélioration du bétail, sont à l'heure actuelle entrepris dans de nombreux pays en développement où l'environnement leur est défavorable sous deux rappor

  1. Il peut exister, dans ces pays, de sérieuses limitations à la productività animale en raison du climat, de la nutrition, de la santé, des techniques gestion, ou plus souvent d'une combinaison de ces divers éléments.
  2. L'infrastructure sur laquelle s'appuie, dans les economies plus avancées, le programme d'amélioration du bétail peut faire partiellement ou entièrement défaut.

Dans ces conditions les difficultés nées du coût de l' intervention du niveau d'alphabétisation; de la communication et même de la sécurité éconoraique peuvent faire largement obstacle au fonctionnement cohérent d'un programme d'amélioration du bé.tail.

La présents étude ports sur énvolution des strategies d'amélioration du bétail face à ces deux sortes d'obstacles.

Lé où des stress d'ordre nutritionnel, climatique, sanitaire et relevant de la gestion, limitent la production animale, il se peut que l'on ne puisse pas utiliser les modeles sur lesquels se fondant les programmes d'amélioration génétique dans un environnement favorable. Les variances et les co-variances génétiques qui sont à 1'origins des caractàristiques de la production peuvent être tout à fait differentes. De plus les caractéristiques relatives à la tolerance au stress peuvent être très importantes. A cet égard on observe généralement de grandes differences entre les populations. Selon que ces differences ont un caractère additif ou non on pourra recourir à differéntes strategies pour les exploiter. De plus, l'ampleur des variations à caractère additif ou non additif peut dépendre de l'intensité du stress dans 1'environnement. La littérature qui porte sur ce problème est étudiée ici et un certain nombre de strategies possibles d'amélioration génétique sont décrites.

L'absence d'appui infrastructurel signifie que lee programmes d' amélioration auxquels on a normalement recours dans les pays développés peuvent être inapplicables. Toutefois, étant donné que les populations envisagées se trouvent à différents stades de développement, il pourrait être possible de mettre en oeuvre des programmes extrêmement efficaces, mais beaucoup moins complexes. Ce genre d'approche est décrit ici.

Resumen

Los métodos empleados gara mejorar las poblaciones ganaderas en los paises desarrollados han alcanzado en los últimos años un alto grado de organización, y frecuentemente, también de rendimiento. Es habitual que el 100 por ciento de las poblaciones de eria de ganado aviar estén sometidas a programas intensivos de registro y selección, asi como más del 50 por ciento del ganado vacuno de leche, mientras para el ganado vacuno de came, y para los ovinos, el porcentaje suele ser inferior. Pero actualmente siempre forman parte del programa de trabajo complejas técnicas de evaluacidn y de análisis de datos. Estos métodos suelen coincidir con una intensificación de las condiciones de eria. Los resultados del mejoramiento genético están limitados por el intervalo entre las generaciones y, a veces, por el agotamiento de la varianza genética. La estructura de la población inhibe, a veces, la taza de mejoramiento o bien pueden ser insuficientes las inversiones en pruebas y selecciones. Pero, en general, laB mayores dificultades son simplemente aquellae inherentes al potencial reproductive de las especies.

Está rapida evolución ha sido posible, en primer lugar, por la sólida base económica que proporciona la zootecnia moderna. Lo cual justifica los gastos en registros, inseminación artificial, asesoramiento técnico y otros elementos de la infraestructura que hace posible un programa de selección genética planificada. En segundo lugar, estos programae se han establecido tomando como base, para los caracteres y especies en cuestión, modelos genéticos bien establecidos y perfectamente verificados. En tercer lugar, como muchos de los programas requieren la rapida copilación y evaluacióh de numerosos datos, su pleno desarrollo ha dependido también de la evolución de los proceiimientos de tratamiento de datos y de los métodos analiticos que dichos mecanismos requieren.

Se están, en cambio, llevando a cabo actualmente programas de mejoramiento pecuario en muchos paises en desarrollo, donde el medio es adverso, debido a:

  1. Las graves limltaciones a la productividad animal que pueden imponer el clima, la nutrición, la salud, las prácticas de explotación o, lo que es más normal, combinación de estos factores.
  2. La falta total o parcial de una infraestructura para los programas de mejoramiento pecuario en los paises con economias más desarrolladas.

Las dificultades para financiar los gastos, los problemas de analfabetismo, las comunicaciones, incluso la inseguridad econtómica pueden obstaculizar considerablemente el funcionamiento de los programas de mejoramiento pecuario.

El presente estudio trata sobre los métodos de mejoramiento pecurario que, ante las condiciones adversas, arriba citados, habria que adoptar.

Cuando la fatiga debida a las malas condiciones climáticas, nutricionales, sanitarias y de manejo limitan la producción animal, los modelos empleados como base para programas de cria en ambientes favorables pueden dejar de ser aplicables. Las varianzas y covarianzas genéticas para los caracteres de produeción pueden ser completamente diferentes. Además, los caracteres relacionados con la tolerancia a la fatiga pueden ser muy importantes. En general, se pueden observar grandes diferencias intra-población para estos caracteres. Se requerirán diferentes métodos de explotación según esas diferencias intra-poblacidh sean de naturaleza aditiva o no aditiva. Además, la amplitud de los variaciones aditivas o no aditivas puede dependerdel grado de fatiga producida por el medio. Se estudian las publicaciones sobre esta materia y se describen algunos métodos de crianza.

La falta de una infraestructura puede significar la imposibilidad de aplicar los programas de mejoramiento que son usuales en los paises desarrollados. Sin embargo, como las poblaciones afeotadas se encuentran en una fase de desarrollo diferente, se podrian aplicar programas muy eficaces pero mucho menos complejos. Este enfoque se describe en el presente reporte.

Previous PageTop of PageNext Page