- news, features, articles and disease information for the fish industry

This book brings together information often difficult to find, joining the worlds of aquaculture science and industry production, through the leveraging of their synergies and dependencies. It contains a wealth of scientific knowledge which can only be traced in a fragmented manner in specialised journals.
Chapter 6

Selective Breeding of Shrimp
James Cock, Thomas Gitterle, Marcela Salazar, Ben Hayes, Morten Rye

Man has been selecting plants and animals better suited to his needs since the moment he began to domesticate organisms and reproduce them himself rather than collect seed or capture animals in the wild. Initially the grower or herder simply collected and planted any seed or nurtured the offspring of the herds he cared for. Even in these cases as pointed out by Darwin (Darwin, 1893) “……….man, without any intention or thought of improving the breed, by preserving in each successive generation the individuals which he prizes most, and by destroying the worthless individuals, slowly, though surely, induces great changes. As the will of man thus comes into play, we can understand how it is that domesticated breeds show adaptation to his wants and pleasures.” Although in the initial stages of domestication of most species there was no conscious selection of parents with particular characteristics, later selection at the phenotypic level was frequently enhanced by information on relationships among individuals. This occurred long before there was an understanding of genetics and inheritance. Growers were effectively selecting for desired traits and applying the principles of genetics without understanding the underlying mechanisms (Siegel et al., 2006). Aquaculture genetics has its origins over 2000 years ago in China and the Roman empire when reproductive cycles were closed and species such as the common carp (Cyrpinus carpio) were domesticated and genetically enhanced to be more suitable for fish farming (Vandeputte, 2003; Dunham, 2004). Domestication in itself will frequently lead to improvement in growth rate under aquaculture conditions as indicated by the high growth rates of the oldest (89 years) domesticated stocks of channel catfish (Dunham at al., 2001).

Chinese references to Koi, coloured common carp, go back to the Western Chin Dynasty (AD 265-316), with Japanese breeding of Koi reported from the seventeenth century (Blasiola, 2005). However, it was only with the rediscovery of Mendelian genetics and Darwin and Wallace’s demonstration of the cumulative effect of many small changes in the evolution of populations, followed by a much fuller understanding of genetics, that selective breeding was supported by a sound theoretical base. This sound base has been the corner stone of modern selective breeding programmes that began to be established at the beginning of the twentieth century, principally with various farm animals and crop plants.

The earliest modern genetic studies on fish were related to Mendelian inheritance of colour (Purdom, 1993). By 1925 Embody and Hayford selected brook trout (Salvelinus fontinalis Mitchill) for three generations for resistance to furunculosis and increased survival from 2 to 69% under natural outbreaks (Embody and Hayford, 1925). Donaldson began breeding rainbow trout (Oncorhynchus mykiss Walbaum) in 1932 and increased growth and fecundity. After more than 35 years of directed individual selection this strain was widely appreciated in the USA and distributed throughout the world (Hulata, 2001). By the early 1960s breeding carp was underway in Israel and in Hungary (Jeney et al., 2008; Hollebecq and Haffrey, 1999; Bakos et al., 2006). In these programmes many of the techniques for selection of aquaculture species were developed. Amongst these the use of a common environment for testing was one of the most important (Moav et al., 1978). Mass selection in carp provided only limited genetic gain for growth and the Israeli (Moav and Wohlfarth, 1976) and Hungarian (Bakos et al., 2006) programmes focussed on crossbreeding of different lines, which provides a one off improvement, but does not offer the possibility of continuous long term improvement. Nevertheless at the beginning of the new millenium, it was estimated that 80% of common carp production was from the Hungarian “Szarvas” crossbreeds (Dunham et al., 2001). It was only with major efforts to improve trout and Atlantic salmon populations in the 1970s, and later tilapia, that modern selective breeding programmes based on genetic selection began for cultivated aquaculture species (Gjedrem, 1996; Dunham et al., 2001; Ponzoni, 2006). The genetic variance of fish and shellfish was found to be high compared to terrestrial livestock. This, coupled with the high prolificacy of these species and the consequent possibility of high selection intensity, led to rapid genetic improvement for traits such as growth (Gjedrem, 1997; Dunham et al., 2001.)

It is now generally accepted that systematic genetic improvement programmes are key for the development of sustainable aquaculture production. This viewpoint is clearly stated in the FAO’s Code of Conduct of Responsible Fisheries (FAO, 1995). Well-designed breeding programmes are crucial for the development of high performance stocks that optimise the use of limited feed, water and land resources. Large selection programmes are now established for all major cultured fish species. Similar programmes have recently been set up for major shrimps species such as Penaeus vannamei in USA (Carr et al., 1997; Argue et al., 2002), Colombia (Gitterle et al., 2005a; 2005b), Brazil (Rocha et al., 2008), and Mexico (Jiang et al., 2005a; Castillo-Juarez et al., 2005); for P. monodon in Australia (Kenway et al., 2006) and in India (Krishna et al., unpublished); for P. japonicus in Australia (Hetzel et al., 2000; Preston et al., 2002), for P. stylirostris in New Caledonia (Goyard et al., 2002); and for P. chinensis in China (Wang et al., 2006). By 2005, about 60 selective breeding programmes utilizing sib information were in operation for fish and shellfish species world wide (Gjerde, 2006) and the number is steadily increasing. In addition there are numerous mass or individual selection programmes in which broodstock are selected by taking the larger survivors from commercial ponds. Nevertheless, the production of many species used in aquaculture is still based on populations that have not been genetically improved in a systematic manner.

Much of the theoretical basis and practical experience for animal breeding is based on experience with warm-blooded farm animals that have long been domesticated and that are often either housed or kept in relatively stable environments. Shrimps are poikilothermic, and are grown under highly variable conditions, with large fluctuations in such parameters as salinity, temperature, oxygen content and ammonia concentration. Furthermore, shrimps are highly prolific with a large proportion of the population normally dying soon after hatching. Later those that survive face stressful conditions in the highly intensive production systems that are becoming the norm. These features differentiate shrimps from most animals that have been the subject of long term selective breeding programmes. Nevertheless, the high fecundity, short generation intervals, and ease of control of the reproductive cycle make some shrimp species excellent target candidates for genetic improvement through selection. Due to the high reproductive rate and the large number of offspring, breeders can apply high selection intensities, and substantial additive genetic variation is reported for key traits such as growth rate. Artificial insemination is now feasible for several species and facilitates a wide range of breeding designs and population structures. These characteristics of shrimps, that differentiate them from many other animal production systems, indicate that optimum breeding strategies and methodologies may diverge from those used in those species which have formed the basis for the development of selective breeding methods for animals. Furthermore, it has been suggested that in certain instances methods and experiences from plant breeding may be usefully included in the design of shrimp breeding programmes (Cock et al., 2009).

In this chapter breeding schemes for shrimp are described with emphasis on family based selective breeding schemes involving sib testing, which have become the industry standard for major aquaculture species including shrimp. In depth coverage of the theoretical basis for improvement strategies is outside the scope of this chapter; for a comprehensive introduction to the field of quantitative genetics and selection theory, the reader is referred to such sources as Falconer and Mackay (1996) or Lynch and Walsh (1998) and for applications to aquaculture species Gjedrem (2005). Due to the recent domestication of shrimps and establishment of selective breeding programmes, breeding schemes and methodologies are still in a dynamic state of development with many unknowns. Under these circumstances the reader will find that many of the observations and ideas presented are somewhat speculative: we make no excuses for this, as we believe that it is only by speculating and trying new ideas that shrimp breeding will progress.

Identification of potential traits and their relation to production practices

Shrimps have only recently been cultivated by man and it is only since about 1980 that some shrimp populations have been reproduced in captivity, domesticated and isolated from wild populations. Almost all cultivated P. vannamei is now reproduced in captivity with populations totally isolated from the wild populations, whereas in the case of P. monodon the broodstock is still largely from broodstock caught in the wild. Populations reproduced in captivity have, in evolutionary terms, had a relatively short period to evolve and adapt to the environment in which farmed shrimp are produced, although in some cases the selection pressure may have been extremely intense and hence the populations may have changed substantially. Furthermore, the shrimp production systems have evolved over the past thirty years as management practices have, in general, intensified. When shrimp were first cultivated the densities were low and most of the feed came from the natural productivity of the ponds. Nowadays, the tendency is towards higher densities, a reliance on balance feed rations and aerated systems. The more intensive systems have been associated with a series of devastating disease epidemics and these have highlighted the need for shrimp populations that are highly productive even in the presence of these epidemics (Cock et al., 2009). Genetically improved stock is an attractive option for producers. However, genetic improvement is not a panacea: improved production technology will depend not only on the provision of improved stock, but also on good management.

In any breeding programme in species that can readily be reproduced in captivity one of the most difficult tasks is the choice of the traits to be improved or incorporated into the stock. Shrimp producers may indicate that they would like to have shrimp populations with a certain trait: in fact producers normally have a long list of desired traits. However, the genetic gain in each individual trait is reduced as other traits are added, even in the absence of unfavourable genetic correlations between the traits. The overall selection pressure for any given breeding scheme is a limited resource and inclusion of additional traits reduces the gains obtained for the other traits already included. In the simplest situation with mass or individual selection for genetically uncorrelated traits with equal heritabilities and phenotypic variances, each additional trait reduces the selection response for current traits by a factor of 1/n, with n indicating the total number of traits. In the presence of unfavourable genetic correlations among traits, the rate of progress for the individual traits may be considerably less than that obtained when selecting for individual traits.

Consequently, breeders have to weigh up the trade offs when selecting several traits for improvement. Techniques and models do exist to identify the ponderation given to various traits (Hazel, 1943; Harris and Newman, 1994). However, to be effectively deployed, they require a large stock of knowledge on the economic importance, genetic variances, heritabilities and breeding values of each trait. This information is rarely available in the case of shrimps, at least when setting up a new breeding programme, and hence it is generally difficult to use these models to define the traits to be included in a shrimp breeding programme. A more subjective approach, based on suppositions and field observations (see the “population size and selection” of this chapter), and the felt needs of the commercial growers, appears to be the only viable alternative. In this subjective approach there are several basic guiding principles that should be applied.

First and foremost breeding should only be considered as a solution to serious problems or to open up major new opportunities. If a breeding programme tries to include many traits of minor importance it will likely achieve little or nothing. Secondly, breeding requires high up front costs and improved stock should be compared with other means of resolving the perceived problem. For example, if low partial pressure of oxygen is a problem in an intensive production system it may be simpler and more cost effective to aerate the ponds. Thirdly, useful genetic variation must exist for the desired trait. It is important to note the qualification of genetic variation as useful. Simply showing that there is genetic variation for a trait does not suffice, the variation must be sufficient to eventually allow a useful increment in the desired trait in commercial populations. Furthermore, as performance testing schemes for aquaculture species often require individual records from a large number of breeding candidates and/or their close relatives, low cost and ease of recording information for the trait in question is important. Obviously it is preferable to have good data on heritabilities of traits and estimates of potential genetic gain and these data can be developed in genetic improvement programmes, nonetheless even relatively simple observations can be useful in deciding to further explore the possibility of improving populations for a particular trait. At the start of a breeding programme the evidence for genetic variation may be from quite simple observations (see the “population size and selection” section of this chapter). In practice breeders weigh up the possible economic benefits of the various traits and use their best estimates of the heritabilities of those traits to decide, often in a fairly subjective exercise, which traits they will emphasise in the selection programme and the weighting to be given to each trait.

Shrimp producers base their production on the four pillars of: good pond management; healthy shrimps; nutritious feed; and shrimp populations that are capable of growing rapidly and surviving under the intense management systems that are becoming the norm. These four pillars cannot be considered independently: thus, for example, the health of the shrimps will depend on their nutrition, the genetic make up of the population, and the pond management. Similarly the optimal genetic make up of the population will depend on the pond management, the sanitary status of the farm and the feed sources. Consequently, the desired characteristics of a selected population are inextricably linked to the conditions and management practices that define the environment in which they will be produced. Furthermore, the characteristics of the shrimp population itself may well directly influence the environment. For example, if a shrimp population is highly resistant to a certain pathogen, the prevalence of the pathogen in the environment may be markedly reduced. This apparently occurred in the Atlantic Coast of Colombia in the early years of the 21st century with the introduction of Taura Syndrome resistant populations of P. vannamei: for several years it became extremely difficult to find infected animals as a source of inoculum for use in challenge tests aimed at maintaining high levels of resistance in the commercial populations. The interactions between all the four pillars of production are such that any selection programme must take them into account and understand how they interact in order to define the desired traits that breeders will attempt to incorporate or enhance in the selected populations that are the result of any selective breeding programme.

Improvement of populations, particularly in traits controlled by several genes and when several traits have to be combined, takes time as several generations are normally required to obtain a marked improvement in all the desired traits. Consequently, breeders have to envision what the production systems of the future will likely be so as to ensure that they produce animals apt for those conditions.

Breeders need to justify their activities indicating the economic benefits that are likely to accrue from their efforts. Research in agricultural production systems provides extremely high rates of return (Pardey and Beintema, 2001). Breeding for aquaculture is no exception with high benefit to cost ratios. In the particular case of Nile tilapia (Oreochromis niloticus) the cost benefit ratios were estimated at between 1:8.5 and 1:60 (Ponzoni et al., 2007). One of the problems of financing breeding programmes is that the improved populations may be difficult to protect and become a public good (Ponzoni et al., 2007; Henson-Apollonio, 2006). Hence support from public agencies is a common means of financing breeding programmes.

Growth, survival and feed conversion

Shrimp producers, in general have focused on three principle desirable traits or characteristics of populations for the grow out phase of production. These are (i) high growth rates (ii) high survival and (iii) efficient feed conversion. The populations must express these characteristics or traits under the management systems that are used by the producers: it is of no use to develop stock that grow extremely rapidly at low stocking densities when fed on fresh feed if management in the grow out phase is based on high stocking densities and processed feed. Furthermore, growers require stocks that can readily be multiplied and distributed by hatcheries and that have no obvious defects such as deformities. The producers are not interested in such traits as disease resistance per se: they are interested in how these traits affect growth rates, survival and feed conversion efficiency under their management conditions.

Overall production of a shrimp pond is determined by the number of individuals that are seeded into the pond, their individual growth rates and their survival. The number of individuals seeded is a management decision. The growth rate of the individuals and the survival depends on the genotype of the individuals that make up the initial population seeded and how that population is fed and managed. Rapid growth shortens the production cycle, thus maximizing the use of infrastructure associated with ponds, and also reduces the biological risks of disease epidemics or other catastrophes such as loss of aeration due to power failure. However, fast growth on its own is not sufficient to guarantee high productivity: pond survival is also a key factor. Pond survival is highly dependent on the management of the ponds and the health status of the population and the interactions between them.

In intensive production systems diseases frequently cause devastating losses. The losses are principally related to high mortalities of animals, although there may be cases in which the presence of a disease, like IHHNV, reduces growth rates. Selection for disease resistance is directly related to the disease’s effect on growth and survival: the objective is not disease resistance per se but rather the impact that disease resistance will have on the desired performance characteristics of the selected stock (Cock et al., 2009). Currently most selection for disease resistance in shrimp is directed to improved survival in the presence of the disease, rather than the ability to both grow and survive under such conditions. The rationale behind this approach is twofold. First it is a great deal easier to select for survival than to select for the ability to grow well in the presence of a disease. Secondly, if a disease is endemic in an area, selection for growth in that area will also select for growth in the presence of that disease if the grow tests are carried out under commercial or at least similar conditions. It should, however, be noted that this approach becomes complicated when dealing with specific pathogen free (SPF) stocks when the pathogen that prejudices production is also on the SPF list for that breeding programme or nucleus.

Selecting for survival when there is one major disease or cause of low survival can normally best be carried out by selecting for resistance to the specific cause of low survival. Thus, when low survival rates are due to Taura Syndrome Virus (TSV), selection for TSV resistance rapidly increases survival rates with a high genetic correlation between survival and TSV resistance. On the other hand, when there is no one clear predominant cause of low survival selection is more difficult and the heritabilities of pond survival may be very low.

Feed constitutes up to 50% of the total cost of producing shrimps (Amaya et al., 2007) and hence feed conversion efficiency is an important parameter in determining profitability: growers unambiguously express their desire to have access to shrimp populations with high feed conversion efficiency. Currently no simple direct methodologies exist for evaluating feed conversion efficiency of individuals or families and hence it is not feasible to directly select for feed conversion efficiency. Selection for rapid growth would be expected, a priori, to increase feed conversion efficiency: a shorter time to reach market weight would likely lead to reduced metabolic maintenance requirements as these would be spread over a shorter period in faster growing animals. In the poultry industry in the United States of America improved stocks coupled with improved management and nutrition have increased potential growth rates (estimated as time to reach a given weight) and feed conversion efficiency approximately threefold over a 44 year period. The effects are principally due to genetic improvement (85-90%), with less improvement due to improvements in nutrition. The increased growth rates are associated with increased feed conversion efficiency (Havenstein et al., 2003). In teleosts several studies indicate that selection for growth will improve feed conversion efficiency as there is a positive genetic correlation between feed conversion efficiency and growth (Thodesen et al., 1999, 2001; Ogata et al., 2002; Silverstein et al., 2005). No data on genetic correlations for growth and feed conversion efficiency are published for shrimp species. Difference in the growth performance between animals raised under the same management and environmental conditions can be due to (i) a better capacity of the animals to convert the ingested food or (ii) to a greater voracity of the bigger animals (competition). The former will of course have a positive effect on the feed conversion efficiency, but the effects of more voracious animals on field conversion efficiency is not clear. Nevertheless, the experience in fish and poultry, both grown in intensively managed systems, would suggest that selection for growth will most probably also lead to concomitant selection for feed conversion efficiency. Thus, it is likely that those programmes that are selecting for growth are also indirectly selecting for improved feed conversion efficiency.


Shrimp improvement programmes have not directed their efforts, up to the present, to shrimp quality. Currently deformed animals are routinely discarded in selection programmes, but none of them are selecting specifically for particular traits related to quality. In many fish programmes selection for traits associated with quality is now routine. Selection for traits associated with the quality of the end product constitute an important part of the breeding objective in many breeding programmes for fish, and may also play a role in selection programmes for shrimp in the future. Information about the genetic background of product quality traits in shrimps is scarce (Jiang et al., 2005b). Traits such as improved head to tail ratio and reduced weight loss during cooking would increase edible and processed yields; and improved texture and colour of the processed product might affect the consumer’s perception of the product.

Current status of selected traits

Currently most selection schemes are directed to high growth rates and high survival in the grow-out period under a defined set of management conditions. Selection for disease resistance is seen as a means to improve survival, and to a lesser extent growth, and is also a high priority in several programmes. Although none of the current programmes are selecting directly for feed conversion efficiency, it is likely that they indirectly select for this trait when high growth rate animals are selected.

Evaluation of genetic variation and control in traits and selection protocols

Gjedrem (1996) in his introduction to selection and breeding programmes in aquaculture indicates that before a breeding programme can be established, breeding goal must be defined, estimates of genetic variance, heritability, phenotypic and genetic correlations among traits must be available. We suggest that although it is undoubtedly advantageous to have all this information, in practice it is rarely possible to be so well informed when the programme is established. The initial phases of a breeding programme are the source of much of the information on genetic variance and the genetic control of traits. In fact an important aspect of the design of the breeding programme is precisely the capacity of the elected design to provide essential information about the genetic make up of the target population.

Selection programmes are only appropriate when genetic variation not only exists for a given trait but also when breeders are able to access that variation. The variation available to breeders comes from three main sources: the base or founder population; mutations or recombinants that occur in the base population or in populations developed from the breeding nucleus that are readily accessible to breeders; and introductions of genotypes with specific traits from outside the breeders’ normal populations. In selective breeding programmes most of the initial progress is attributable to the variation in the initial base population, but over the longer term the role of mutations and new recombinants increase in importance (Hill and Bunger, 2004; Weber, 2004). The relative importance and use of these three sources of genetic variation, the base population, mutants and introductions varies according to the specific traits and their genetic control. In general mutants and introductions are of more importance when dealing with traits controlled by one or a small number of genes that are of low frequency in the overall population, whereas the base population is of fundamental importance for characteristics such as growth which are generally under polygenic control.

Sources of Genetic Variation: base populations and mutations

Base population

Several breeding programmes of various species from both the plant and animal kingdoms have progressed unsatisfactorily towards their goals due to low genetic variation in their base populations (Teichert-Coddington and Smitherman, 1988; Huang and Liao, 1990 and Cock et al., 2009.) There is little doubt about the importance of establishing a base population with a broad genetic variability and guidelines for this have been established for aquaculture species (Holtsmark et al., 2006; Holtsmark et al., 2008).

The primary objective of establishing a base population is to maximise the genetic variability that will be readily available to breeders. When establishing a base population variation in the traits that are likely to be useful is often stressed. Nevertheless, when a base population is established it is unlikely that those involved in setting it up will know precisely what traits will be desired in the future: new diseases may appear or novel quality traits may become important in the coming years and there is no way of predicting all the variation that may be required in the future. Consequently, the base population should be established with as much genetic variation as possible, including variation in traits of unknown importance and also unknown traits such as resistance to an unreported disease. Furthermore, new sources of genetic variability from outside the initial base population can be added as breeding programmes mature.

In many animal-breeding programmes the commercial stock have been developed to a stage where they are very different from the wild stock. In these cases breeders simply cannot introduce wild stocks to enhance their already improved stocks in the breeding nucleus as the wild populations are so far behind in performance that they are not competitive (Hill and Bunger, 2004; Cock et al., 2009). Most commercial shrimp populations are not far removed from wild populations and hence both commercial stock and wild populations can be used to establish base populations.

Collection of animals from a wide range of geographical locations or sites would, a priori, appear to be a logical starting point. Nevertheless, in Atlantic salmon (Gjerde, 1986) and in Rohu carp (Reddy et al., 2002) it has been demonstrated that the genetic variation for a trait like growth rate between rivers is much less than that within rivers suggesting that this strategy may not be as effective as one might expect. Molecular or DNA markers can be used to estimate genetic diversity both within and between populations (see the “marker assisted management of genetic variation in wild and domesticated populations” section of this chapter). The usefulness of molecular markers to estimate genetic variability in or between populations depends on the degree to which measures of molecular marker variation are correlated with traits which may be useful to breeders in meeting their objective. There is some controversy about the strength of this relationship (Reed and Frankham, 2001) and it varies widely between traits (Nielsen, 1998). A further alternative is to evaluate individuals from wild or domesticated populations and select those displaying a wide range of variation as part of the founder stock. This option also has serious difficulties. Firstly many traits are not readily evaluated in wild or commercial populations and secondly breeders often are not fully aware of the traits that will be of most importance as the breeding programme develops. In the absence of better guidelines and the difficulties of implementing sophisticated or onerous techniques, the optimum strategy is probably to collect as many populations as possible from the wild and from commercial populations to set up the base population. Later, if deficiencies are encountered in the base population they can be made up by returning to wild populations or commercial operations, at least in the early years of a breeding programme when the breeding nucleus is not so distinct from wild populations that their introduction would be a major setback in terms of genetic gain for desired traits.

Diseases and sanitary restrictions to manage them are a major restriction and may complicate the process of setting up base populations with a broad range of variability. These restrictions are particularly severe when specific pathogen free (SPF) breeding nuclei are being established and in areas in which certain diseases are absent.

Once the initial base population has been established, selection pressure in the early stages of the programme should be low and care should be taken to mate high ranking animals from different origins in order to minimise the initial reduction in genetic variation. Rapid and intense selection for one or two traits in the early stages of selection can potentially eliminate much of the genetic variation painstakingly accumulated in the initial establishment of the founder or base population.

Population size and selection

There are two principle means of managing selection protocols in aquaculture. The first is to grow separate populations in separate tanks or ponds. This system facilitates the maintenance of the identity of the animals from different populations with no need for marking or tagging of individuals from separate populations. However, there are severe disadvantages. Unless costly replication of tanks or ponds is employed the effects of differences between tank environments are confounded with the population effects: this severely limits the use of testing in separate tanks or ponds. The second alternative is to place all the animals in a common environment, however, this also has problems. The different populations will compete with each other and, unless animals are tagged, their identity is lost. Moav and Wohlfarth (1974) in common carp showed that competition effects were not serious and did not affect the rankings of different populations. Resolution of the problem of tagging or marking so as to indentify the origin of individuals is in a state of flux at present. Traditionally in shrimp breeding animals have been reared separately until they are large enough to be tagged and placed in common environment. In practice this separate rearing means that, for a substantial period, the genetic effects and the tank effects are confounded. However, as will be discussed in the “parental assignment and control of inbreeding” section of this chapter, new DNA fingerprinting technologies are likely to make separate rearing before tagging unnecessary. In consequence testing of mixed progeny from distinct families or populations in a common environment, which is already the norm for shrimp breeding, will be improved and continue to be the standard procedure.

In shrimp production large populations are the norm with hatcheries and farms routinely handling millions of individuals. Simple observation of these large populations can potentially provide breeders with valuable information on genetic variability. Inspection of large populations opens up the possibility of very intense selection and also the opportunity to select rare mutants or recombinants and incorporate them in the breeding nucleus.

Whilst recognizing that much of the information on genetic variance is not formally published or otherwise available when a breeding programme is established or when a new trait becomes a potential breeding goal, simple field observations can often provide a useful guide to the nature of the genetic variation in the population and the possibilities of selecting for a desirable trait. Some examples of both how this approach has been used in the past, and also how it has been neglected and the consequences of that neglect, are described in the following sections with further examples of current use of simple observations that may provide the basis for increasing genetic gain.

In the mid 1990s shrimp growers began to suffer from high shrimp mortalities due to some unknown cause in many shrimp farms in South America. Simple field observations indicated that about 70% of the population died. The other way of looking at this simple observation was that 30% of the animals survived and it was possible that they survived as they were resistant to whatever was causing the mortality. One of the major shrimp producers in Colombia, Oceanos S.A., after discussions with various researchers decided that it would be worthwhile to select survivors from infected ponds, and use them as broodstock for restocking of commercial ponds. Field observations suggested a significant (from the commercial point of view rather than statistical conventions) increase in survival and so the survivors were once again selected and used to produce broodstock. Within two generations the survival rates in commercial ponds reached 70% (Cock et al., 2009). This example clearly indicates how simple observations, in the absence of even a knowledge of the cause of the problem, can be used to provide the basis for establishing a selective breeding programme that can rapidly provide commercially useful stock.

In a somewhat similar manner in the late 1990s WSSV began to cause severe losses on the Pacific coast of the Americas. Mortalities were extremely high, being very close to one hundred percent. These simple observations should have suggested a very low frequency or the possibility of a total lack of useful genetic variation. Nevertheless, various breeders (including most of us) attempted to breed for resistance using both individual selection and family selection. The family selection was based on the assumption of polygenic resistance, in spite of early observations that suggested that useful levels of resistance were only found in two or three families which were much more resistant than all the others, suggesting that these families might have a small number of genes that conferred resistance. The breeding efforts for WSSV have undoubtedly increased the level of resistance in the populations, but the level has not been raised sufficiently to be useful in areas where epidemics are severe.

We have observed the occasional occurrence of extraordinarily large post larvae in hatcheries and others have made similar observations (Klimpel, 2001). Simple observations suggest that these super shrimps fall outside the normal bell curve distribution and that the large size is under genetic control. Whilst it is still too early to say how these super shrimps might perform under commercial conditions, it would appear worthwhile including them in breeding programmes, even before their usefulness is proven, precisely to determine whether they do indeed possess a genetically controlled trait which is commercially advantageous.

The examples given above highlight the importance of selecting from a large population size so as to be able to impose a high selection pressure, and also to detect the occasional superior mutant or recombinant. In the case of certain traits, such as resistance disease to those diseases that can be detected in the larval stage, large populations can readily be managed at the experimental level. On the other hand with such traits as extraordinarily fast growth in the grow-out phase large populations cannot easily be maintained under experimental conditions. Nevertheless, commercial growers routinely manage extremely large populations, with individual ponds frequently stocked with more than a million animals. We are currently exploring methods to select from commercial ponds the truly exceptional animals in terms of growth so that these animals can be included in the breeding nucleus and hopefully increase the rate of genetic gain for growth.

There are several major advantages of selecting individuals from large commercial populations. These include selection of individuals well adapted to commercial grow out conditions and the use of very intense selection pressure. However, currently a major difficulty has been the difficulties involved in determining the pedigree of the animals in commercial ponds. Currently most of the family based breeding programmes that involve sib selection rear a limited number of animals per family in separate environments until they can be tagged and then these animals are placed into commercial ponds. This scheme has various deficiencies. The initial environment for separate rearing differs from commercial conditions and hence there is little selection pressure for adaptation to commercial conditions in the early stages of the cycle. The number of animals that are tagged is only a small proportion of the total population in the ponds and it is difficult to recover superior tagged live animals. Due to the relatively small number of tagged animals and the difficulty of recovery, estimates of survival are poor and selection pressure is low. The situation is now changing rapidly with the possibility of DNA fingerprinting. If a commercial pond were stocked with progeny from a series of known crosses and the largest survivors were selected at the commercial harvesting time, then with current techniques using the DNA of both the parents of all the crosses that make up the population and the selected individual, then their pedigree could be determined with a high degree of certainty (for more details see the “parental assignment and control of inbreeding” section of this chapter). Thus, these new techniques, when their cost comes down to a reasonable level, open the possibility of applying high selection pressure in commercial ponds to obtain exceptional individuals of known pedigree that are well adapted to the commercial conditions under which they were evaluated.

In disease testing a specific trait is normally evaluated on its own with all other factors maintained as uniform as possible. However, for certain traits such as growth and general pond survival, the populations should be evaluated in conditions as close to commercial conditions as possible so as to ensure that the selected animals are well adapted to and perform well under commercial conditions. These field conditions are not as well controlled as those in challenge tests. The phenotypic variance, the additive genetic variance and consequently the heritability, which all affect the expected genetic gain, will all vary depending on the evaluation conditions. Hence, the expected genetic gain will also depend on the conditions under which the individuals or families are evaluated. Unfortunately, under less well controlled conditions the phenotypic variance tends to increases largely due to an increase in the environmental variance and hence the genetic gain is also reduced as compared to that which can potentially be obtained in more controlled conditions. Breeders have to constantly evaluate the potential gains from closely controlled conditions, like those normally obtained in challenge tests, to obtain a high degree of accuracy of selection (the correlation between true and estimated breeding values) against the possibility that the resulting populations may not be well adapted to the varying environment found in the real world of commercial production.

Challenge tests

In shrimps technology to prevent or treat most infectious diseases is still largely unavailable. Under these circumstances genetic selection for resistance to a given pathogen is a valuable tool for increasing pond survival after exposure to a particular pathogen. Three principal strategies have been applied for selection: (i) selection of survivors from commercial farm ponds, (ii) selection of survivors in artificial challenge test experiments and (iii) challenge tests in which in which populations of full-sib families are exposed to the pathogen in question and information on survival is used to select breeders from animals that have not been exposed to the pathogen, but that come from those families that have high survival. The latter scheme has the major disadvantage that it only capitalizes on 50% of the additive genetic variance.

Selection of survivors of a farm disease outbreak is an inexpensive, practical approach if shrimp are cultured under non-biosecure conditions (Gitterle, 1999) but it is not possible in a specific pathogen free (SPF) programme (Lotz, 1997). Similarly, selection of survivors in controlled challenge tests is not a practical option for SPF programmes. On the other hand, selection based on information from relatives that have been exposed to pathogens (e.g. family selection) is appropriate for an SPF programme because infected or exposed shrimp are not re-introduced into the breeding programme. The two options of selection of survivors from farm outbreaks and from controlled challenge tests can be used not only for mass selection but also for combined selection if information is available on the pedigree of the survivors.

For any strategy, challenge tests must be carefully designed to mimic natural infection conditions maximizing the animals’ ability to display their true level of innate resistance. All animals should be exposed equally to the risk of infection at the same time. This can only be achieved with well-designed inoculation techniques. Thus, for example, feeding the animals with contaminated tissue mimics natural infection, but does not control the time and dose of exposure. To solve this problem the inoculum may be placed directly into the oral cavity of the shrimps (Vidal et al., 2001) or infected material can be force fed by reverse gavage through the anus (Aranguren and Lightner, 2009).

For effective challenge tests, procedures to identify susceptible, infected, and recovered animals are needed. Ideally mortality should be recorded throughout the experiment, not only at the point of termination, to be able to identify differences in time until death. And finally, the tests should be designed to evaluate mechanisms of resistance that are likely to be useful under commercial conditions.

However, challenge tests under controlled conditions are not able to emulate pond conditions. For example, in some TSV challenge tests experiments shrimps are infected with TSV under controlled conditions, while in the pond they also encounter multiple pathogens (filamentous bacteria, protozoa, metazoa and even other viruses) as well as physical and chemical stressors (salinity, temperature, dissolved oxygen) that affect their capacity to defend themselves against the given pathogen.

Indirect selection

Lack of recording technology or high recording costs may prevent direct selection for a given trait of high economic importance. In this situation a secondary trait which is genetically correlated to the primary target trait may be used to make indirect selection for the secondary trait. The efficiency of indirect selection depends on the heritability ratio for the primary and secondary traits and their genetic correlation, as well as the selection intensity and phenotypic variance of the secondary trait. If the heritability of the secondary trait is substantially higher than for the primary target trait and the genetic correlation between the traits is sufficiently high, indirect selection may give larger response for the primary trait than direct selection.

Modern tools for selective breeding programmes

Molecular biology and information technology provide many tools that can be used to facilitate and improve selective breeding programmes. Today it is possible to compile and analyse massive databases with the computers and programmes that are readily available even in quite small research units. Whilst the use of modern information technology is currently being incorporated into many breeding programmes, the tools of molecular biology are only now beginning to be used on a routine basis and several extremely promising techniques are still in the development phase. The major opportunities for use of molecular biology in improving breeding programmes are in the areas of (i) identification of genetic variation in populations (ii) parental assignment or identification and (iii) marker assisted selection. All these uses depend to a greater or lesser extent on the use of molecular markers.

Marker assisted management of genetic variation in wild and domesticated populations

When choosing a strategy to maximise genetic variance of useful traits, it is important to consider the way in which traits are measured, particularly in wild populations when they are the source of genetic variance in base populations. Growth is relatively easy to measure. Disease traits are usually estimated by survival after a challenge test, and these are difficult to organise for wild animals. Traits such as many organoleptic quality traits are also difficult to measure on live wild shrimp. Therefore, marker based strategies should be considered when evaluating genetic variation for such traits as disease resistance and organoleptic quality traits, while phenotypic information may be more appropriate for traits like growth.

Microsatellites have also been used to evaluate maintenance of variability over several generations in the nucleus of breeding programmes. The diversity of wild P. monodon stocks estimated by 12 microsatellites was greater than in domesticated stocks, with diversity shifts observed in both allele numbers per locus and heterozygosis (Dixon et al., 2008). Similarly there was a significant loss of genetic variation, evaluated by RAPDs, in 15 lines from different hatcheries and bottleneck effects in their reproduction (Freitas et al., 2007).

Nevertheless a question remains as to whether levels of variability of neutral DNA markers, such as microsatellites, are related to genetic variability in quantitative traits and the effect this has on the fitness of the individual or the population. Although several studies report that individual heterozygosity at apparently neutral microsatellite markers is correlated with fitness components such as survival (Coulson et al., 1999), disease-resistance (Coltman et al., 1999), fecundity (Rodhouse et al., 1986) and reproductive success (Slate et al., 2000) other studies have failed to find strong associations. A meta-analysis of the association between neutral marker heterozygosity and fitness traits or components reported only weak associations (Coltman and Slate, 2003). Furthermore, heterozygosity at a set of allozyme loci, but not heterozygosity at an equal number of microsatellite loci, explained variation in quantitative traits in rainbow trout (Thelen and Allendorf, 2001) deep-sea scallop (Pogson and Zouros, 1994). Overall, in aquatic species, genetic diversity, as measured by neutral genetic markers, may be correlated with genetic variance of some traits, but not others.

In order to maximise genetic variance of potentially useful traits it is important to be able to select the molecular markers used. In immune related genes such as MHC (Major Histocompatibility Complex) loci, it is clear that heterozygosis is beneficial for resistance against diseases. For example in rainbow trout, Johnson et al., (2008) identified and typed microsatellites located within the MHC region and found an association of the MH-IB markers with survival in animals infected with Flavobacterium psychrophilum (Fp), the causal agent of bacterial cold-water disease. Although shrimp do not posses molecules of the MHC it could be valuable to explore polymorphisms at other immune-related genes.

Three hypotheses have been proposed to explain the potential heterozygote-fitness advantage reported. Firstly, a direct effect, in which heterozygote advantage is a direct result of the over-dominance of the loci studied. This could be potentially important in allozyme studies but not for microsatellite markers as shown by the studies previously described. Secondly, a local effect where the heterozygote advantage results from closed linked-disequilibrium between the neutral loci and a fitness locus (associative over dominance). Thirdly, there may be a general effect in which the apparent heterozygote advantage at the markers is a result of identity disequilibria at genome-wide distributed fitness loci.

In breeding programmes it is important to maintain overall genetic variance for sustained genetic progress and in order that traits that are currently unimportant, such as disease resistance to a disease that is currently not endemic in the production zone, but that may become important in the future. There are several, largely anecdotal, cases of loss of genetic variance for disease resistance in populations developed in areas where the disease was not endemic resulting in disastrous production losses when a disease epidemic occurred (Cock et al., 2009). We suggest that a combination of the available phenotypic information with marker based estimates of genetic variation may be used to assist in the establishment of base populations, to monitor genetic variance in the breeding populations and also to reduce the number of crosses between genetically similar individuals.

Parental assignment and control of inbreeding

Just as molecular markers can be used as a source of information when attempting to establish a founder population of maximum genetic variance; molecular markers can provide a means to ensure genetic variation is maintained over the multiple generations of the breeding programme. This is achieved by minimising short-term and long-term inbreeding. One way of managing inbreeding is to run a family based breeding programme, where enough families are maintained each year or each generation to keep inbreeding at or below an acceptable level.

Family breeding programmes in the past have required that families are kept separately until the animals are large enough to be physically tagged. This is costly and difficult to manage, as many separate tanks must be maintained. Additionally, estimating accurate breeding values for selection decisions can be problematic, due to potential confounding between additive genetic effects and environmental effects common to full-sibs raised for an extensive period in separate units. Parental assignment by use of molecular markers has the potential to overcome these problems. Using this technology, shrimps from different families could be reared together in the same tank even from the egg stage. As the need to keep each family in a separate tank is circumvented, using molecular markers would allow a larger number of families to be tested, without increasing investment in the number of tanks, and thus facilitating the use of higher selection intensities without rapid accumulation of inbreeding (Estoup et al., 1998).

Microsatellites have successfully been used to empirically reconstruct pedigrees with high accuracy using highly informative microsatellites. Wang et al., (2006) showed a success rate of 88% in assigning P. vannamei progeny to parents using only 6 microsatellite markers. Similarly, Jerry et al., (2006) were able to correctly assign all P. monodon progeny to their parents. Dong et al., (2006) were able to correctly exclusively assign 90.7% of the progeny to their parental pairs in mixed families groups of Chinese shrimp (Fenneropenaeus chinensis). However, technical problems such as the presence of null alleles, poor quality DNA and/or band stutter can decrease the accuracy of the test. One study employing simulations for parental assignment of Kuruma shrimp (Penaeus japonicus) showed that assignment success of progeny to their true mother was only 47% (Jerry et al., 2004). In other species, pedigree errors in microsatellite parentage assignment are 8.7%–15.5% in sheep (Barnett et al., 1999) and 2%–22% in cattle (Visscher et al., 2002). Whilst parental assignment using molecular tools offers many potential advantages, the techniques are still at the experimental stage and breeders will need to establish the reliability and the cost of the methods before employing them.

In order to reduce genotyping costs, Sonesson (2005) proposed a selection scheme combining two selection strategies: walk-back and optimum contribution selection. Walk-back selects only the best performing animals whilst also maximizing the number of families (Doyle and Herbinger, 1994) and optimum contribution maximises genetic gain at a desired level of inbreeding (Meuwissen, 1997). The proposed scheme indicates that with only 100 animals genotyped the inbreeding can be kept to an acceptable level and furthermore the contribution of the genotyped animals is acceptable. However, when EBVs are calculated from a selected subset of animals it will result in a selection bias. For more details of these selection strategies see the “designs based on long term response and optimal contribution theory” section of this chapter.

Marker Assisted Selection (MAS)

Understanding the genetic basis of important breeding traits, such as traits that increase an organism’s ability to growth, survive and reproduce under commercial conditions, would be a valuable tool in selective breeding programmes. Most of the traits such as growth and disease resistance are multifactorial or polygenic in nature and it is difficult to identify a specific molecular marker responsible for the phenotypic characteristics. Quantitative Trait Loci or QTLs, regions of DNA associated with a particular phenotypic trait are used as molecular markers for assisted selection. QTL mapping is primarily based on identifying segregation of a single gene marker and estimating the effect of its linkage to the polygenic trait. Statistical techniques, such as analysis of variance, are used to check if trait means for one marker genotype at any given locus differ significantly from the trait means for alternative marker genotypes.

The accuracy of MAS is improved if flanking markers are used to predict the QTL effect. This requires knowledge of the order of the markers along the chromosomes. Such genetic linkage maps at a low resolution are available for P. Monodon (Wilson et al., 2002), P. vannamei (Zhang et al., 2007) and P. chinensis (Li et al., 2006). However, greater numbers of markers need to be mapped in order to increase the probability of accurately identifying the QTLs.

Marker assisted selection (MAS) can be based on DNA in linkage equilibrium with a quantitative trait locus (QTL) (LE-MAS), molecular markers in linkage disequilibrium with a QTL (LD-MAS), or based on selection of the actual mutation causing the QTL effect (Gene-MAS) or Type I markers. With LE-MAS markers, the linkage between the markers and QTL is not sufficiently close to ensure that marker-QTL allele relationships persist across the population. Consequently the markers have to be identified for each family and can only be used for within family selection. Hence they are not likely to be widely used. On the other hand linkage disequilibrium marker associations QTL (LD-MAS) persist across the whole population and hence can readily be used for mass and combined selection.

At the time of writing there is only one report of a QTL in crustaceans. Lyons et al., (2007) using Amplified Fragment Length Polymorphism (AFLP) markers in M. japonicus identified a major QTL region, contributing 16% to phenotypic variation for growth rate. Allele variants associated with low and high growth are under evaluation in larger populations of M. japonicus.

One problem with MAS is that only a limited proportion of the total genetic variance is captured by the markers. An alternative to tracing a limited number of QTL with markers is to trace all the QTL. This can be done by dividing the entire genome up into chromosome segments, for example defined by adjacent markers, and then tracing all the chromosome segments. This method was termed genomic selection by Meuwissen et al. (2001). Genomic selection exploits linkage disequilibrium; the assumption is that the effects of the chromosome segments will be the same across the population because the markers are in linkage disequilibrium with the QTL that they bracket. Hence, the marker density must be high enough to ensure that all QTL are in linkage disequilibrium with a marker or haplotype of markers.

Single Nucleotide Polymorphisms (SNPs) are an alternative that could be used in MAS. Single Nucleotide Polymorphisms are nucleotide variations or single base changes. SNPs are the most frequent form of genetic changes. In shrimp SNPs have been identified and tested for its use as molecular markers for breeding traits. SNPs were associated with variation in resistance to TSV in P. vannamei populations in China (Zeng et al., 2008). On the other hand SNPs in alpha-amylase (AMY2) and cathepsin were not associated with growth in P. monodon, but there was a tendency for higher weight to be associated with allele G of CTSL SNP C681G (Glenn et al., 2005). Polymorphisms in moulting-related genes in P. monodon and P. vannamei produced no association with growth (Yu et al., 2006).

The proliferation of high throughput technologies and new protocols for identification of thousands of SNPs will make it possible to identify and genotype at a fraction of the current cost. This will facilitate analysis of large populations and we believe that SNPs will become a highly valuable tool for MAS selection in shrimp breeding programmes. Overall, modern high throughput genotyping technology which can identify tens of thousands of markers in livestock species and thus genomic selection is likely to become an important tool in breeding programmes.

Rather than looking for a marker closely linked to a gene that controls a particular trait, it would be preferable to identify the gene itself. Expressed Sequenced Tags (ESTs) are valuable tools for identifying genes associated to particular traits. ESTs are short (100-800 bps) sequences that indicate which genes are being expressed in particular tissues at a specific moment in time and space. Although ESTs are easy and inexpensive to generate their use is problematic for several reasons. In all ESTs databases there is the problem of redundancy, under-representation and overrepresentation of ESTs that may be caused, inter alia, by tissue differences or differences in isolation protocols. No less important is that ESTs do not provide information on changes in gene expression caused by mutations in regions that do not codify proteins.

If ESTs are to be used for selection of particular genes, rather than to be used as markers, then the effect of the genes that are expressed for particular traits must be known. Although the number of ESTs for P. vannamei is growing exponentially (Gross et al., 2001; O’Leary et al., 2006; Clavero-Salas et al., 2007), a high percentage of the sequences obtained do not have homologues in Gene Bank or other databases and therefore there is no information on their putative role or function. This suggests that the role of ESTs will be limited unless genomic programmes are developed for shrimp species (Benzie, 2005).

Differential gene expression may be particularly effective for identifying disease resistance. If a gene expresses itself in resistant, but not susceptible, animals when they are challenged by the causal agent of the disease, then these genes are likely to be associated with resistance. In aquaculture, differential gene expression in Atlantic salmon (Salmo salar) infected with the salmon anaemia virus (ISAV) was associated with disease-resistance (Jørgensen et al., 2008)

In shrimp Garcia et al., (2009) compared gene expression in hemocytes of WSSV-infected shrimp previously challenged (pre-challenged) and naïve. Genes related to shrimp immune response, such as the antimicrobial peptide penaeidin, crustin, C-type lectin, protease inhibitor and chitin binding domain-containing protein were up regulated in the pre-challenged shrimp. On the other hand, genes coding for viral structural and functional proteins were up regulated in the naïve animals. de Lorgeril et al., (2008) found differential expression of antimicrobial peptides genes in resistant and susceptible P. stylirostris to Vibrio penaeicida infection. Interestingly differential expression of lysozyme and penaeidin 3 genes was also found in non-challenged animals indicating differences in basal expression that could be used as markers for breeder selection.

Microarray analysis of gene expression of thousands of genes may become a useful tool for breeding programmes to study genotype x environment interaction. Gene expression levels in varied environments (such as heat or water stressed) could be screened to identify genes that show significant changes in expression over environments. Also, and potentially of great interest, these techniques might target those genes that show reduced levels of variance in gene expression across environments. This latter may identify loci that are well buffered across environments, leading to more stable phenotypes over a range of environments.

The effect of double stranded RNA in shrimp immunity and micro RNA silencing have also been studied and applied to shrimp viral infections. Initially Robalino et al., (2004) injected double stranded RNA (dsRNA) in P. vannamei infected with WSSV and TSV virus and decreased cumulative mortality by 50 to 75%. Interestingly this effect was independent of the sequence of the dsRNA but was limited to low doses of viral agents. In later experiments dsRNA sequences homologous to viral genes triggered a potent antiviral response implying the existence of RNA interference (RNAi-like) mechanisms in the antiviral response in shrimps. Furthermore, a dsRNA-induced gene-silencing pathway was identified in shrimps by targeting two housekeeping genes with injection of homologous dsRNA into the abdominal muscle (Robalino et al., 2005). Reverse genomics was then used to elucidate the role of crustins in shrimp immunity. A decrease in mRNA and protein expression of one crustin isoform was achieved by injection of dsRNA complementary to the gene into P. vannamei. Subsequent challenge of the injected animals with Vibrio penaeicida showed a significant increased in mortality as compared to infected animals without the dsRNA treatment (Shockey et al., 2009). Similarly, an increase in shrimp mortality in animals challenged with Vibrio harveyi was observed after silencing the Propheniloxidase activating enzyme in P. monodon (Charoensapsri et al., 2009).

Even though functional genomic studies are only beginning in shrimp, they have become an important tool for understanding shrimp immune function. There are many ESTs, which are associated with susceptibility/resistance or defence against infection, with no known homologues in databases. Consequently, their function is unknown but RNAi could be used to elucidate their role in shrimp immunity, providing shrimp breeders with more tools for selecting shrimp with increased resistance to pathogens.

Genetic gain

Breeders constantly strive to maximise genetic gain as it is the yardstick by which they are measured. The genetic gain is the improvement of the performance of a population due to genetic effects in response to selection. The genetic gain is defined as the average superiority of the offspring of selected parents compared with the generation before selection.

The genetic gain per generation is proportional to the selection intensity, the heritability and the phenotypic variance. Thus increasing both the heritability and the phenotypic variance will increase the genetic gain. However, the phenotypic variance and the heritability are not independent variables and increases in phenotypic variance due to random or environmental effects will tend to decrease heritability. The genetic gain over time is influenced by the generation interval. The shorter the interval the more rapid the genetic gain over a given time period.

Generation interval

One of the simplest means to increase the rate of genetic gain is through shortening the generation interval. In shrimp, mating and spawning can be induced by unilateral eyestalk ablation in some species (e.g. P. vannamei, P. styliostris and P. monodon) and breeders routinely use this technique to shorten the generation interval in females. Recently we have observed four month old males of P. vannamei that are sexually mature and can be used for artificial insemination. The use of these young males could greatly reduce the generation interval of males, with the caveat that this will only effectively increase genetic gain if early maturity is not associated with other undesirable traits.

In family breeding programmes, with current tagging technology, the rearing in separate tanks to 1 g weight at high densities slows initial growth and development. In the case of P. vannamei the separate rearing currently increases the generation interval from about twelve months in mass selection programmes that do not tag animals to approximately 14-15 months when the grow out period is equal to the commercial grow out time of about 100 days. In the Colombian P. vannamei breeding programme generation interval has been reduced to approximately one year by used of younger breeders. The use of molecular techniques to assign parents to progeny could reduce the generation interval by eliminating the need for rearing in separate tanks and hence reduce the generation interval and increase the genetic gain by close to 14%, which is not an insignificant amount.

Additive genetic and environmental variance

Genetic gain is increased when the additive genetic variance increases and when the environmental variation is reduced. From a breeders point of view this indicates that evaluation procedures should be designed to maximise the expression of additive genetic variance whilst at the same time minimizing the component of the phenotypic variance due to non-additive genetic effects and environmental variation. The reduction of the phenotypic variance due to environmental variation implies rearing and evaluating all animals in such a manner that they all encounter similar conditions and opportunities to express their genetic potential.

Controlled disease testing (see the “challenge tests” section of this chapter) are frequently used to evaluate individual traits under carefully controlled conditions in which the individuals or families can express their genetic variance for the trait in question, whilst minimizing variation due to all other factors. This effectively increases the accuracy of selection, with the estimated breeding values coming close to the true breeding values.

The phenotypic variance, the additive genetic variance and consequently the heritability, which all affect the expected genetic gain, vary according to the evaluation conditions. Hence, the expected genetic gain will also depend on the conditions under which the individuals or families are evaluated.

In combined selection schemes (see “combined selection” section of this chapter) breeding candidates are ranked according to both the relative performance of the family to which it belongs (mainly based on information of a number of close relatives) and also its performance as compared to other members of its own family. While its siblings may be tested under commercial conditions, the breeding candidate itself may not, due to restrictions on entering animals into breeding nucleus (as may be the case for SPF breeding nuclei). This may introduce loss of selection accuracy for the commercial production environment if that environment differs substantially from the environment where the individuals that are to become breeding candidates are selected and raised. This problem can be partially obviated by selecting breeders only on the basis of the performance of its siblings tested in the commercial environment; however this scheme only uses 50% of the additive genetic variance as it ignores the within -family variance.

Selection intensity

The prolificacy of shrimp coupled with the routine management of large numbers of shrimps opens up the possibilities for very high selection intensities as a small number of animals can be selected from populations in which large numbers of animals are evaluated. Some mass selection breeding programmes which evaluate large populations and heavy selection pressure have been very successful in obtaining animals well adapted to specific conditions, at least temporarily, through this approach. In Colombia, we have observed that mass selection of larger survivors over several generations lead to increased pond survival. However, high selection intensity in mass selection programmes, without tracking the genetic relationships between the selected animals, may quickly lead to genetic bottlenecks, rapid accumulation of inbreeding and loss of genetic variation. Hence the resulting populations are likely to respond less favourably to changing condition or threats, such as new diseases, that they have not encountered during the selection process. Furthermore, in mass selection the accuracy of selection may be low for traits of low heritability. Combined selection will increase the accuracy of selection as both family and individual data can be used to determine breeding values, thus increasing the accuracy of selection. Furthermore, as pedigree information is available in combined selection, inbreeding can be better monitored and controlled. However, as combined schemes work with a limited number of families and individually tagged animals the total number of animals performance tested is usually substantially lower than is the case for mass selection programmes and hence the selection intensity tends to be lower than in mass selection programmes. Furthermore, in combined selection using the Best Linear Unbiased Prediction (BLUP) methodology, where there is a high correlation of estimated breeding values within families (Wray and Thompson, 1990), as selection pressure is increased there is a danger that inbreeding and loss of genetic variance may become a serious problem, especially when selecting for a single or limited number of traits. Nevertheless, as discussed in the “parental assignment and control of inbreeding” section of this chapter, the use of molecular markers to assign parentage could increase the number of animals per family and the total number of animals per batch to such a level that selection intensities could be greatly increased.

Methods to measure Genetic Gain

Measuring the genetic gain is a key point in any breeding programme. Genetic gain can be estimated by several methods as discussed by Rye and Gjedrem (2005), but none of them are well suited to shrimps. The most common method of estimating genetic gain in shrimp is to compare unselected control groups or lines with the selected ones (Hetzel et al., 2000; Goyard et al., 2008; Argue et al., 2002). This method provides information on the response from one particular generation, from which unselected control groups have been maintained, to another later generation, but does not estimate the genetic trends over several generations. Shrimps have a relatively short period when they are reproductive and it is not currently possible to conserve sperm, eggs or embryos for long periods. For breeding programmes it is both costly and difficult to maintain unselected control groups from different generations and then to compare them (Rye and Gjedrem, 2005). Although it might appear simple to maintain unselected groups by simply mating random animals from a population there will inevitably be selection pressure applied to the progeny and even in the absence of selection pressure the frequency of alleles in the successive unselected populations will change due to random genetic drift.

Genetic trend analysis of multigenerational data based on mixed model methodology (Henderson, 1975) provides a good proxy to estimate genetic progress across generations of selection without the need to maintain unselected control populations (Gall and Bakar, 2002). Appropriate genetic ties between generations can be obtained by using a limited number of animals as parents in two successive generations. For shrimp species, repeated use of the males can be used for this purpose (Gitterle et al., 2007). Genetic ties provide connected data across generations and/or subpopulations and thus the means to compare the performance of animals in different generations or lines without physically evaluating them simultaneously. The genetic links are obtained by using the pedigree information across generations or by establishing pedigree relationships between lines or batches. Connected data can be obtained either from parent-offspring relationships (genetic ties) or from parents having offspring with records (direct ties) across different levels of fixed effects (e.g. within and across generations). The closer the relationship the stronger are the genetic ties. Thus, for example, using the same male and female breeders to produce progenies for two separate lines or for two separate generations provides a strong genetic link. Unbiased estimates of genetic gain can only be obtained if environmental and genetic differences across levels of fixed effects are accounted for (Sorensen and Kennedy, 1984a; 1984b).

Genetic trend analysis results and conclusions are valid for the breeding nucleus, but may not be appropriate to estimate genetic gain under commercial conditions. The genetic gain under commercial conditions can be estimated from well recorded field data from commercial operations (see the “dissemination schemes” section of this chapter). Commercial data sets are often not balanced and are frequently incomplete. Furthermore, the commercial results are influenced by multiple variables including the genetic stock, a multitude of management practices, weather and seasonal effects and the characteristics of the ponds. Nevertheless modern analytical tools such as mixed models combined with Best Linear Unbiased Prediction (BLUP) models and neural networks can be used to evaluate not only the performance of the improved genetic stock, and hence genetic gain, under commercial conditions, but also to evaluate the effects of management practices and the weather conditions on commercial production (Gitterle et al., 2009).

Breeding methods, design and selection methodologies

There are various methodologies that are open to breeders, each with its advantages and disadvantages. In the following sections the most important methodologies are discussed in terms of their relevance as appropriate means to improve shrimp populations.

Breeding methods


Inbreeding is the mating of close relatives and results in reduced genetic variation due to inevitable loss of allelic variability, as clearly demonstrated for P. styliostris (LeMoullac et al., 2003). Inbreeding is also known to be strongly associated with the fundamental biological phenomenon called inbreeding depression, which is expressed as loss of performance. Inbreeding depression is typically first seen for traits related to reproduction and general viability, but also affects regular production traits in terrestrial and aquatic animals. For penaeid shrimp, the number of studies on inbreeding depression is still limited. Bierne et al., (2000) report on an indication of inbreeding depression for growth rate in P. styliostris, and Keys et al. (2004) demonstrated substantial inbreeding depression for growth and survival in P. japonicus when comparing the performance of inbred and outbred families with known inbreeding levels. Closed populations of P. vannamei in Venezuela that have been mass selected for harvest weight (consciously) and survival for 11 generations show no obvious indications of severe inbreeding depression (De Donato et al., 2005). It should be noted, however, that since that paper was published Venezuela has faced severe outbreaks of TSV causing high mortalities. In Colombia mass selected material that has now passed through more than ten generations still performs well under commercial conditions (Ceniacua unpublished data). On the other hand, Moss et al., (2004) reported that inbreeding in P. vannamei negatively influenced hatch rate and hatchery survival, although growth was not affected for the inbreeding levels tested.

Moss et al., (2007) reviewing work at the Oceanic Institute conclude that under favourable conditions inbreeding effects are small for growth, and are minimal for grow-out survival, in the absence of viral pathogens. Furthermore, the deleterious effects of inbreeding on survival and growth are small under favourable conditions, but they increase when animals grow in poor environments (Doyle et al., 2006; Moss et al., 2007).

Moss et al., (2007) suggest that control of inbreeding is particularly important when improving such traits such as disease resistance; nevertheless, it is probably prudent for breeding programmes to manage inbreeding irrespective of the traits under selection. Due to its potential negative effects, the accumulation of inbreeding must be carefully controlled in any breeding programme. Selection in finite populations will always lead to increased inbreeding levels over time, accumulating at a rate proportional to the selection intensity. When the inbreeding level in a given population reaches an unacceptable level, unrelated or more remotely related animals from other populations may be introduced as parents. Deliberate inbreeding is only of interest when inbred lines are produced for crossing purposes to exploit non-additive genetic effects, as may be the case when double recessives confer a desirable trait.

Whilst there is much concern about inbreeding it has been suggested that build up of deleterious genes in shrimp populations mass selected for pond survival and growth under commercial conditions is unlikely; those animals that carry deleterious genes will be eliminated naturally as animals that carry them neither survive nor grow well (Cock et al., 2009). Hoffmann and Merilä, (1999) indicate that deleterious mutations expressed in environments commonly encountered by organisms will be rapidly removed by selection. Similarly it would be expected that deleterious genes existing in shrimp populations would also be rapidly removed. This type of self-elimination of poor types is readily achieved in highly prolific species grown under commercial conditions where a heavy loss of unfit animals is the norm (Cock et al., 2009).


The breeding method or strategy for genetic improvement within a population is known as purebreeding and is commonly chosen for continuous genetic improvement over a long period of time. In pure breeding mating of close relatives should be avoided to keep the rate of inbreeding at an acceptable level. Individuals that possess positive (desirable) genes have a high breeding value and are selected as parents for the next generation. Pure breeding is particularly associated with breeding of certain races or “breeds” of animals such as dogs and cattle. In many cases only those individuals with a known pedigree, indicating that they are purebred from an original founder stock, are accepted as being of that particular breed. In shrimps purebred lines from an original founder stock may exist, however, there seems to be no reason why a breeder should not introduce new stock into the breeding nucleus to increase variability or to improve a particular trait.

As any particular populations diverges from others, accumulating a series of desired traits that differentiate it from other populations, breeders will tend to maintain the population as a pure breed so as not to lose the desired traits by crossing with populations that lack them. At the same time, due to the loss of variability inherent in any purebreeding programme, whilst pure breeding may produce populations that are adapted to specific environmental conditions and management, these populations may not readily adapt to changes in the environment or the production systems.


Crossbreeding is mating of different species, breeds, strains or inbred lines with the objective to exploit non-addi­tive genetic variance (heterosis or hybrid vigour). Successful hybridization is reported for several penaeid shrimp species (Lawrence et al., 1984; Lin et al., 1988; Bray et al., 1990; Misamore and Browdy, 1997; Benzie et al. 1995, 2001), but due to lack of valid comparison of performance of hybrids and the parental species few studies assess the magnitude of non-additive genetic effects. Benzie et al. (2001) found no indication of heterosis for growth rate in hybrids produced by crossing two species of tiger shrimp, Penaeus monodon and P. esculentus. Recently crosses between inbred lines of Penaeus (Litopenaeus) stylirostris from Hawaii and New Caledonia grew 37% faster than the original inbred lines (Tian et al., 2006). Furthermore, survival was greater in the hybrid lines and they outperformed the inbred lines by a greater margin when conditions were adverse (as indicated by reduced survival). Similarly in the Chinese Shrimp (Fennropenaeus chinensis) crosses between Rushany and Korean populations generally outperformed the parent lines (Tian et al., 2008). However, systematic crossbreeding per se does not accumulate genetic gains over time, and should therefore be looked upon as a supp­lement to a programme for additive genetic im­prove­ment. Thus systematic selection programme based on additive genetic performance may be more effective than crossbreeding (Tian et al., 2008).

Developments in the techniques of gynogenesis and sex reversal (Thorgaard, 1986) could facilitate fast production of inbred lines. However, unless the inbred lines are derived from distinct base populations, inbreeding and crossing without selection for additive gene effects will not produce accumulating genetic gains. In addition, to evaluate crossbred performance, the lines must be developed and then crossed. The cost and time involved in developing and test-crossing inbred lines would only be justified if the heterotic effects were large when compared to the expected selection response over the same period of time using other methods. Currently in many producer countries imported seed from other countries is restricted to SPF animals. This restricts the genetic migration between populations developed in different countries and opens up the possibility of cross breeding of inbred lines from different countries or regions.

Other strategies

Studies on Atlantic salmon (Rye and Mao, 1998) and rainbow trout (Pante et al., 2002) breeding populations indicated that non-additive genetic effect constitute a significant proportion of the total variance for growth in these species. Extended models for genetic evaluation including non-additive genetic effects are expected to give improved accuracy for additive genetic merits and thus increased selection responses. Non-additive genetic effects may also be exploited through specific mate allocation. However, the level of improvement in progeny merit which can be obtained through simultaneous selection for additive and non-additive genetics effects in applied breeding programmes still remains to be investigated.

Strategies like ploidy and sex manipulation (to produce triploid or monosex populations) may be used to further increase the productivity of the commercial shrimp seed. Production of polyploid shrimp has been successfully achieved using chemical methods in Sicyonia ingentis (Xiang et al, 1991) and P. chinensis (Li et al. 2003, Bao et al., 1994; Xiang et al., 1993, 1998, 2001), and P. monodon (Norris et al. 2005). Temperature shock has also been used to produce polyploid animals in P. vannamei (Dumas and Ramos-Campos, 1999; Fast and Wyban, 2001), P. monodon (Fast and Wyban, 2001) P. japonicus (N. Preston, pers. comm. in Dunham et al. 2000), Sicyonia ingentis (Xiang et al, 1991), and P. chinensis (Li et al 2003, Li et al., 1999; Dai et al., 1993). At multiplier levels these strategies are less restricted than those that can be used for breeding within the nucleus.

Design of selection programme

The design of a selection programme defines the size and structure of the breeding nucleus and its development from the base population. The design determines the rate of genetic gain of individual traits. In order to optimise the design a balance must be reached between what is logistically feasible and what is desirable from a genetic point of view. Logistical limitations include such factors as the capacity to individually tag animals, infrastructure to evaluate a large number of families or individuals for various traits and to maintain a large number of selected breeders over time, the capacity to shorten the cycle between generations and the manner in which crosses can be made. Gjerde et al., (1996) provide an example of the trade offs with the observation that reducing the restrictions on inbreeding from 0.25% to 2% reduced the rate of genetic gain substantially more (40%) in small designs than in larger ones (10%).

The design of a selection programme is also directly influenced by the genetic control of the traits in question, the acceptable rate of inbreeding and the biological characteristics of the species which determines the types of crosses that can be achieved and the number of progeny obtained. In addition, in order to be able to estimate genetic gain over time, in species, where unchanged genetic stocks cannot be maintained and then compared with newer stocks as is the case with plants, it is necessary to connect or link populations over time.

As noted above the selection programme design is influenced by the genetic control of the traits to be selected, but the design itself determines how much information can be obtained on the genetic control (Hoffman and Merila, 1999).

Mating design

The choice of the mating design depends on the reproductive biology of the species, the ability to establish the parentage of populations or individuals and the genetic control of the traits to be selected. The mating design itself determines how much information can be obtained on the genetic control of individual traits and also on the rate of genetic gain. Thus, for example, in a mass selection programme in which breeders copulate naturally with no controlled crossing and no individual or family identification of evaluated progeny, little or no information is generated on the genetic control of traits. Hence, the mating design itself influences the ability to obtain information on the genetic control of traits which in turn is an important factor in the definition of the optimum mating design.

In order to obtain information on the genetic control of traits, controlled crossing is required with maintenance of the identity of the progeny from specific crosses. The manner in which these are produced depends on the reproductive biology of the species involved. In those shrimp species in which artificial insemination can routinely be carried out, common practice is to use each of the two spermatophores of a single male to fertilise an individual female (see nested design below). This produces both full and half-sib families in which the progeny are separately reared until they can be tagged and so later ascribed to a particular cross.

To a large extent the success of a shrimp breeding programme depends on the reproductive strategies of the species. Species where females store the sperm on the outside of the thelyca (exoeskeletal modifications of the thoracic sternities) are denoted as open thelycum, while those that store sperm within exoeskeletal invaginations of the thelycum are termed as closed thelycum (McVey, 1993). In species with open thelycum like P. vannamei and P. stylirostris females that are about to spawn can be artificially inseminated with a high probability of success (Gitterle et al 2005) while species with closed thelyca, like P. monodon and P. japonicus, have to be artificially inseminated as post-moult females irrespective of whether they are gravid or not. In both open and closed thelycum species sperm from the same male can use to inseminate more than one female (Gitterle et al., 2005; Kenway et al., 2006) to produce full and half sib families. However in the closed thelycum species, even when eye ablation is performed after artificial insemination to synchronise spawning, the success rate is highly variable and only a few females may mature and spawn. Hence, only a small number of families may be produced with only a small proportion of halfsibs.

In shrimp breeding programmes a trade-off has to be made between the optimal theoretical programme of crosses and the limited availability of animals that are ready to be crossed within a particular time period. In order to provide similar conditions to all individuals that are to be tested in a single batch the mating period for all the crosses should be as short as possible. The CENIACUA breeding programme aims to produce all the families of a particular batch of P. vannamei within a period of one week or less. During this period not all the selected females spawn as on any particular day approximately 5% of the females spawn. In other breeding programmes up to 15% of the females have been reported to spawn on any one day. Consequently, in practice mating is normally carried out between selected males and those pre-selected females that happen to be ready for the target period for family production. Hence more females than males need to be available in order to obtain the same selection intensity for females than for males.

Inbreeding is a major concern in breeding programmes, particularly with shrimps which are highly prolific. In mating designs in which the progeny cannot be assigned to a particular family, a very small number of prolific crosses may contribute a disproportionate number of animals to the populations to be evaluated, potentially contributing to a high rate of inbreeding and a loss of overall genetic variance in the population. It is only through mating designs which provide a means to ascribe individuals to particular crosses or population that inbreeding can be assessed and managed.

For a more detailed description of the general properties of various mating designs and when they are most advantageous see Gjerde (2005). In the following section various mating designs for shrimps are described.

Single pair

In single pair designs the spermatophores from one male are used to inseminate a single female. This scheme only produces full-sib families and consequently additive genetic and other effects common to full-sibs (non-additive genetic and environmental effects) are confounded and cannot be assessed independently. Hence, single pair mating should not be used unless the non-additive genetic effects and environmental effects common to full-sibs (normally due to separate rearing until the individuals can be tagged) are low.


This is the most commonly used mating design in today’s shrimp breeding programmes with open thelycum based on family selection or combined individual and family selection. Two mature female breeders are each artificially inseminated by covering their thelycum with the spermatic mass stripped from the two spermatophores of a male breeder. The females are then placed in individual spawning tanks and the progeny are separately reared until they can be tagged. Although it is possible to fertilise more than two females by further dividing the spermatic mass, in practice fertility is normally reduced and operationally it is more difficult than simple fertilizing each female with one of the two spermatophores produced by each male. However, males frequently regenerate mature spermatophores within about 14 days and these can be used to inseminate additional females. This scheme, and the other alternative of producing maternal half-sib families by fertilizing a single female twice both have the disadvantage that it prolongs the period in which crosses are made and leads to evaluation of families and individuals of different ages when they are placed in a common environment and a longer period of rearing in separate environments.


In a full factorial design, semen from a male is used to fertilise at least two different females, while a female is inseminated twice with at least two different males. Thus both full-sibs as well as paternal and maternal half-sibs are produced. This is a good design to obtain reliable estimates of both additive and non-additive genetic variance in the population. However, there are two serious drawbacks. Firstly, for a given number of rearing tanks available, a lower number of breeders will be tested than in a nested design and secondly the delay between the first and second spawning of the females leads to the problems of different ages and longer periods in separate environments described above.

Mass spawning

For natural mass spawning sexually mature male and female breeders are kept in a tank or pond where they mate freely. In mass spawning the relative contribution of each breeder to the total number of offspring is unknown and in current mass spawning programmes parents cannot be assigned to individual progeny. The progeny from one copulation event in shrimps is highly variable and can be as high as one million individuals in the case of P. monodon and populations derived from mass spawning may be dominated by a small number of highly prolific crosses. Consequently, as there is no way of ascertaining the parentage of individuals in current mass spawning programmes, individuals selected as breeders for the succeeding generation may be from a limited number of crosses. As a result, the effective population size may become low and the rate of inbreeding high (Brown et al., 2005). Collection of DNA from all potential parents and from selected individuals offers the potential to assign progeny to specific matings and hence to obviate this problem.

Designs based on long term response and optimal contribution theory

Parents are commonly selected on the criterion of highest estimated breeding values alone. This criterion maximises the immediate response in the following generation but does not necessarily result in maximum genetic response over a longer time horizon. Decisions on the selection of parents not only affect the following generation but also they influence both the rate of inbreeding of future generations and the overall genetic variance in the population (Wray and Goddard 1994). A direct consequence of this is that the selection of parents for the nucleus breeding population may differ substantially from that for a dissemination programme (see section 8). Estimated breeding values (EBV) of close relatives are highly correlated, and hence selection of parents based solely on EBVs may actually lead to even higher rates of inbreeding than with mass selection.

Caballero and Toro, (2000) suggest that in small captive populations the best means of maintaining genetic diversity is to minimise the co-ancestry of breeding individuals in the population. Algorithms have been developed to determine the number of offspring (or mates when assuming an equal number of offspring per mating) a given animal should get to maximise the genetic gain at a predefined rate of inbreeding (e.g., Meuwissen, 1997, Hinrichs et al., 2006). The use of optimum contribution selection implies that no specific pre-determined mating design will be used. The use of optimal contribution theory is difficult in shrimps due to the small proportion of females that spawn on any one day and the lack of methods to synchronise or control time of spawning.

Sonesson, (2005) simulated a methodology for selecting breeders obtained from breeding schemes in which animals from different families are mixed in a common environment without tagging and parents of individual animals are later identified using genetic DNA fingerprinting. Such breeding procedures would be costly and tedious if all individuals tested had to be fingerprinted. She suggests that a walk back scheme can be used to reduce the number of animals that need to be genotyped: first the individual with the highest phenotypic value is selected and genotyped, then the next highest phenotypic value is selected and genotyped and so on until the appropriate numbers of males and females needed for mating are obtained. The breeding values of the genotyped animals are then determined by BLUP (Best linear unbiased prediction) and then the optimum contribution selection method is used to determine which are the optimal crosses to maximise genetic gain and minimise inbreeding and loss of genetic variance in the overall population. This procedure is attractive for shrimp breeding if the DNA fingerprinting schemes are used. Modifications to the scheme will however be required to account for the fact that not all females spawn during the designated mating period. Hence it will be necessary to have a larger number of candidate females than those actually used to make the crosses. Furthermore, instead of determining just one female for each male several options should be provided for each male so that a male with a high breeding value is not excluded from the scheme due to the fact that the selected female does not spawn.


Genetic links are needed to establish the genetic gain over time as populations diverge from their original state by providing connected data. Genetic links have a further important role to play in breeding. For logistic and security reasons breeders may maintain separate populations or lines that they produce at different times or in different locations. If these lines are kept separate they will diverge over time with each line having particular characteristics. From a commercial point of view this may not be desirable as growers may begin to favour one particular line or population over another. By using genetic ties the separate populations or lines can be compared. Furthermore, the use of common parents in separate populations will tend to reduce their divergence. In addition, if a breeder is managing, let us say, two populations displaced in time, but with genetic links or ties, the breeder has the opportunity to select the individuals with the highest breeding values from the two populations and thus to increase the selection intensity.

Selection method

The selection method influences the probability of correctly ranking breeding candidates with respect to their breeding values: the higher the probability the greater the correlation between the true and the predicted breeding values and hence the accuracy of selection.

The most appropriate method depends on various factors amongst which the most salient are: the heritability of the trait; the reproductive capacity and the generation interval of the species; and the ability to use the animals that are evaluated as parents for the next generation. In shrimps the most commonly use selection methods are mass selection, family selection or a combination of the two (combined selection).

Mass selection

Mass selection selects breeding candidates solely according to their own phenotypic performance. Mass selection can be used to select for various traits and always selects for survival. Individual selection cannot be used for traits which require evaluation protocols that kill the animals. Mass selection is generally inefficient for binary traits like survival and sexual maturity when these occur at very high or low frequencies. Mass selection, which is often used on populations from mass spawning, can cause high levels of inbreeding due to a large number of animals per cross and hence the possibility that the selected individuals come from a limited number of families. This effect can be reduced by pooling a restricted number of individuals from each family at fertilization or shortly thereafter (Gjerde et al., 1996; Bentsen and Olesen, 2002) but this is not possible when mass selection is combined with mass spawning. In spite of the disadvantages of mass spawning combined with mass selection, it should not be discarded as an option. It has been used successfully to rapidly develop commercial populations with high growth rate and resistance to Taura Syndrome Virus (Cock et al., 2009) and can be used to provide selected broodstock to supply commercial hatcheries.

Within-family selection

Within-family selection selects individuals from within (full-sib) families and as is the case with individual selection, is restricted to trait(s) that can be recorded on the live breeding candidates. Furthermore it automatically selects for survival. It is usually applied when families are reared in separate tanks. When individuals from the families are pooled and reared in the same test unit, a within-family selection protocol called ‘walk-back’ selection (Doyle and Herbinger, 1994) has been proposed where DNA markers are used to identify the pedigree of the potential breeders (see section 5.3.5). A major disadvantage of within-family selection is it utilises only 50% of the additive genetic variance.

Sib selection

Certain traits, such as carcass quality, cannot be recorded on live breeding candidates. Other traits, such as disease resistance, may render potential breeding candidates as unfit for the breeding programme. For example, in an SPF breeding nucleus it will not be possible to use infected survivors of a challenge test in the breeding nucleus. In these cases full and half-sib selection can be used. Part of a full or half-sib population is evaluated, and the information on the evaluated animals is used to select the families that will provide the breeders for the next generation.

Sib selection is particularly effective when large families can be produced and when the heritability is low. The ability to produce large families allows a large number of individuals to be evaluated and hence an accurate estimation of the family mean, particularly when the heritability is low and environmental effects tend to mask the genetic effects. The high fecundity of shrimps makes sib selection appropriate for these species. To obtain high sustained long term selection response and low rate of inbreeding accumulation, the number of family groups tested when applying sib selection are usually high, usually greater than 100. The number of families becomes increasingly important when several traits are included in the breeding objective or when new traits may later be added to the breeding objectives.

Accurate sib selection normally requires evaluating individuals from known crosses in a common environment. The common environment is obtained by placing all the individuals in the same tanks or ponds. Hence marking or tagging of the animals is necessary. The current technology for physical tagging of shrimps requires a minimum tagging size of approx. 1 gram, which in the Colombian production system is reached 60 days after spawning. The families are reared in separate tanks until tagging. Any systematic environmental effects common to members of a family during the period of separate rearing will be confounded with the respective families’ additive genetic effect and mask the true genetic differences between the families and hence reduce the accuracy of the estimated breeding values. A dense additive genetic relationship matrix among the families tested facilitates separation of the additive genetic and common environmental effects. Nevertheless, under any circumstances, the time when the families are raised in separate environments should be kept as short as possible and the environment for all family groups should be made as uniform as poss­ible in those periods.

Separate rearing of individual families in the early stages may also introduce a major disparity between the initial rearing conditions in the breeding programme and those in commercial production. A direct result of this is a lack of selection pressure on the populations for good performance, whether it is for growth or survival, under commercial conditions in the critical early period after stocking ponds.

Recent developments in molecular techniques (see section 3.3.2) may offer the possibility to determine parentage of a high number of individual shrimp at reasonable cost, and thus eliminate the need to rear individual families separately until they can be physically tagged. If this becomes possible a main disadvantages of current schemes of sib selection (and combined selection) would be removed.

Progeny testing

Progeny testing is based on the selection of breeders according to the performance of their progeny. In shrimps progeny testing is particularly difficult as it is not easy to maintain animals alive for sufficient time to produce enough progeny and test them, analyse the results and then use the animals as selected breeders. In other species such as cattle this problem is obviated by cryopreservation of sperm and embryo, but this is not currently possible with shrimp. Progeny testing is the only selection method for which the ac­curacy of selection can be 1.0 but suffers the disadvantage of extending the generation interval, as records from the breeding candidates’ offspring must be available before final selection can be performed. This latter drawback and the fact that sib and combined selection generally offers high selection accuracies for highly fecund species has lead to minimal use of progeny testing schemes in aquaculture.

In the case of diseases in which progeny can be screened in a challenge test at a very early age, progeny testing may be possible in shrimp. The progeny of young broodstock can be tested for disease resistance and the results of this progeny test can then be used to select the parents, especially males, for future crosses.

Combined selection

Combined selection, as the name suggests, combines all available phenotypic data recorded on the breeding candidates and their relatives (sibs and progeny). Combined selection utilises both the within family and between family records and maximises the rate of genetic gain. It is considered to be the optimal selection method when it can be applied. However, when sib records are used to estimate the breeding values, the breeding values for siblings will tend to be more similar than under individual selection. Therefore, with truncated selection (selection of only those breeders with a value greater than a fixed value) there is a high probability of selecting the breeders from a limited number of families. Consequently, with combined selection the need to restrict the number of selected individuals from each family is more important than with mass selection. For a given population size and a given restriction on the rate of inbreeding, the optimum design when applying truncation selection will therefore have to consist of more and smaller full- and half-sib groups than with mass selection.

Whilst combined selection is considered the optimal method when it can be effectively applied, as currently used it has the same disadvantage as sib selection with reference to marking of animals and the consequent delay before rearing different families in a common environment coupled with the limited number of animals that can be tagged. As noted above these restrictions may shortly be relieved by the use of molecular techniques that can determine the pedigree of progeny.

The two most widely used breeding schemes in shrimps are mass selection and combined selection. The advantages and disadvantages of these two schemes are summarised in table 1.

Back crossing

Backcrossing is used to introgress a particular trait, normally controlled by very few genes, that is in a poor genetic background into a population, which has generally desirable traits, but lacks the specific trait in question. Backcrossing can be envisaged as potentially useful to introduce a resistance gene that occurs at very low frequency and is only found in a poor genetic background into a population without resistance to the disease but with otherwise excellent overall performance.

Protecting genetic material inbreeding

Breeders invest in the development of improved populations, but they will not profit from their investment if others propagate and sell the improved populations without rewarding the breeders. Currently in animals there is much less legal protection of breeders’ rights than occurs, for example, in plants under the International Union for the Protection of New Varieties of Plants (UPOV). Currently, the traditional intellectual property regimes cannot properly protect innovations and allow recuperation of research and development expenses in the animal genetics industry (for aquaculture see Rosendal et al., 2006). The best alternatives seem to lie in contract and licensing agreements (Ogden and Weigel, 2007) or technology, such as triploids or single sex populations that make reproduction of improved populations difficult. Shrimp breeders may enter into material transfer agreements (MTAs) in which they stipulate the restrictions on the multiplication and distribution of populations derived from the original material supplied to a grower. However, in practice these agreements are likely to be difficult to enforce and to police.

Breeders do however have various mechanisms by which they may at least make it more difficult or less worthwhile for others to propagate and disseminate improved populations derived from the breeder populations.


Doyle et al., (2006) suggest that breeders can effectively protect their genetic material by distribution of broodstock that has a very narrow genetic base, but that in the first generation for distribution is not highly inbred. However, secondary propagation of broodstock will be highly inbred and the resulting populations are likely to have poor survival, particularly under harsh conditions. Whilst this scheme is theoretically attractive it still has not been tested under commercial conditions to demonstrate its effectiveness. At the same time the mere threat of inbreeding becoming a major problem may be sufficient to induce hatcheries to continually purchase improved stock, rather than attempting to propagate and disseminate lines which may perform poorly, particularly under adverse conditions, due to inbreeding.

Negative relation between desired traits and prolificacy and desired traits

Some desired traits, such as disease resistance, may be negatively correlated with growth and, of greater interest for protecting genetic material, with reproductive performance. Breeders could develop two lines, one with superior growth, but low prolificacy and the other with low growth but extremely high prolificacy. The breeders could then distribute a mixture of the two lines with a small proportion of highly prolific animals with low performance. These lines would perform well, but unauthorised multiplication would rapidly lead to populations dominated by low performance but highly prolific animals.

Control of fertility

Distribution of populations which cannot be easily reproduced, is an attractive option for protecting genetic material. There are several possible means of making it difficult to reproduce a population. Amongst these are the production of sterile animals, such as triploids, or populations with only one sex. Triploids are normally infertile. Although significant efforts have been made to develop protocols and methods for stable production of triploids in penaeid shrimp (Dumas and Campos Ramos, 1999; Li et al 2003; Morelli and Aquacop, 2003; Norris et al., 2005; Sellars and Preston, 2005), regular production of triploids is not yet reported. Induction of triploidy for production of sterile and monosex populations is common in several fish species (reviewed by Pandian and Koteeswaran, 1998), but has not been achieved in shrimps.

Commercial superiority

The genetic gain that can be achieved in shrimps in growth rate is of the order of 5-10% per generation in a well-designed breeding programme. If this genetic gain can be maintained in the broodstock distributed to the hatcheries then those hatcheries that use the latest available genetic material will be able to supply growers with populations that grow 5-10% faster than animals reproduced, without selection for growth, from previous generations. We suggest that this level of improved performance may be sufficient for hatcheries to charge higher prices for the latest generations of broodstock and also to charge more for the populations they disseminate. If this is the case then it does not make sense for hatcheries to illegally propagate earlier generations with lower performance.

Adaptability and stability of performance

The aquatic environments in which shrimps are produced are extremely varied. The growing conditions of each individual production event1 vary depending on the there physical location, the particular moment in time and the management scheme imposed by the producer. In general the shrimp producer desires populations that are well adapted to and produce well across the variation that exists in his particular location using his preferred management practices. Furthermore, the producer wishes for populations that provide relatively stable but continuously increasing productivity over time: he abhors the possibility of having a bumper harvest followed by an almost complete loss of productivity or the occasional exceptional production event in a few ponds and a series of extremely low productivities in others. Growers generally prefer a stable guaranteed yield to a higher average yield with wild fluctuations about the mean.

The variations in productivity within a particular farm may be due to spatial, temporal and management variation. It is generally difficult for growers to eliminate spatial variation within their farm, although it may be possible. Observations on ponds in individual farms indicate that there are differences in the productivity even in neighbouring ponds with similar management grown at the same time and these variations can be large. Using commercial data from 10 years of production of P. vannamei, Gitterle et al., (2009) ran a mixed model analysis where they found that for yield (kg/Ha/cycle) 20% of the total variation among ponds corresponds to farming conditions (spatial and management differences) not related to differences in the season, and a further 17% of the variation was due to temporal (year and season) effects. The temporal variation is principally of two types, fluctuations in the conditions or trends. Fluctuations are typically related to variation in such factors as the weather with phenomena like El Niño, when an abnormal weather pattern occurs and then subsequent years return to normal until the next El Niño occurs. In the particular case of El Niño on parts of the Pacific Coast of South America the water temperature increases and this suppresses development of White Spot infestation leading to better survival and increased productivity in Niño years. Trends in temporal variation tend to be due to cumulative effects and can be positive or negative. Although examples in shrimp culture are difficult to find, possibly due to a lack of long term records and analysis of shrimp productivity, in agriculture these long term effects are related to such factors as soil degradation, build up of diseases in intensive production systems and improvement of the genetic stock. Similar trends are likely to occur in shrimp production. A lack of stability in production due to variation in management practices is the one source of in farm variation over which the grower has a large degree of control. It is through management that growers can, and do, attempt to minimise the variation in productivity and to provide apt conditions for high and stable productivity of shrimps of good quality.

The individual grower is primarily interested in how genetic stock performs in his farm, and only in special cases, such as disease susceptibility in neighbouring farms which could lead to epidemics in the whole region, is he concerned with the performance in other sites. On the other hand breeders, particularly those in large modern selective breeding programmes, need to amortise the costs of their breeding programmes over a large area and period of time, and hence need to produce stock capable of performing well over the range of conditions and management practices found in the target area. The alternative of developing lines or strains for individual producers or limited areas is not economically viable although it may be able to customise stocks for particular farms in the dissemination phase of a breeding programme (see the “dissemination schemes” section of this chapter). This philosophy of providing well adapted stock is succinctly described by Gjedrem (2005) “We want robust animals, which will tolerate fluctuations in environmental conditions.” However, it may be too much to ask for a single population of any one species of shrimps to be well adapted to all the varied environments and management conditions it may face. Many pig and poultry lines have been developed and are now raised in sites with immensely diverse environments, however, the management is such that the environment in which these lines are raised is relatively uniform. The growing conditions are effectively manipulated so as to suit the requirements of the pig and poultry lines in question. In the case of cattle the situation is radically different. Their environmental conditions in which cattle are raised, particularly when there main feed source is grazing pastures, are not easily modified: the environments they perceive are much more varied than in the case of modern poultry and pig operations. Consequently, a whole series of cattle breeds have been developed with specific breeds or populations apt for a given range of environmental conditions. Similarly, in those crops which are widely grown in a range of environmental conditions, plant breeders have developed varieties that are suitable for specific conditions. In order to amortise their costs over a reasonably wide area the plant breeders have developed the concept of mega-environments. These mega-environments encompass a range of well defined environmental and management conditions within which the breeders estimate that they can provide varieties that will have high and relatively stable performance levels. Varieties are then developed that are targeted for each of these mega-environments. Shrimp breeders need to define their targeted mega-environments.

Breeders estimate the adaptability of their varieties, genotypes or populations by testing them under conditions representative of commercial practice in the varied conditions under which they are likely to be grown. An analysis of variance of the two way classification of variance of genotypes and environments will provide an estimate of the variance attributable to the genotype by environment interaction (GxE). In general when the ranking of genotypes in various environments remains similar the breeder does not face a major problem. He simply selects the best genotypes. However, when genotypes are ranked differently across sites the breeder has to decide whether it is necessary to redefine the mega-environments for which populations will be developed, or if specialised lines will have to be developed for specific conditions within the mega-environment. Gitterle et al (2009) analysed production data for P. vannamei in 25 farms on the Colombian Atlantic coast. Management and environmental conditions vary substantially: stocking dates ranged from 5 animals/m2 to superintensive 140 animal/m2. Thirty percent of the total variation in yield (harvested fresh weight of shrimps per unit pond area) was attributable to GxE interaction, indicating that in commercial production GxE is potentially important.

Shrimp breeders, until now, have paid little attention to GxE and adaptation to specific conditions. If the variation in environmental conditions within locations or sites is greater or similar to the variation between sites breeders will have to breed for broad adaptability as it is most unlikely that it will be feasible to manage separate stock specifically adapted to the varying conditions within a particular site. However, there is strong evidence that in some cases the differences between sites or locations is much greater than the within site variation and this may lead to strong GxE interactions. For example, the Atlantic coast of Northern South America does not experience WSSV epidemics, whereas in the Pacific coast WSSV is now endemic and causes severe losses. Negative genetic correlations have been demonstrated between increased tolerance to WSSV infection and growth rate (Gitterle et al., 2005b) and high ranked families for tolerance tend to have low growth rates. Hence, a strong GxE interaction growth is to be expected between the Pacific and Atlantic coasts, and in fact the high growth rate populations developed and widely used on the Atlantic coast are not used on the Pacific coast. Thus, with diseases that are dependent on the environmental conditions or management practices that exclude them in some cases and not in others, it is highly likely that GxE interactions exist for desired traits and that separate mega-environments should be defined for these cases and separate populations developed for each of the mega-environments.

It may be possible to detect GxE interactions related to variation in grow out conditions that are influenced by management. For example, there is some evidence that there is genetic variation in the tolerance to low oxygen levels. If this is the case then it is highly likely that there is a GxE interaction for growth at different oxygen levels. However, this does not necessarily mean that it would be a wise decision to establish a low oxygen mega-environment and breed low oxygen tolerant populations. Even though there may be genetic variation in tolerance of low oxygen it may not be useful genetic variance in the sense that the levels of tolerance are not sufficient to allow high levels of productivity under low oxygen conditions. If this is the case then management through lower population densities or more aeration may be a much more effective approach than developing a separate population with an added trait.

Whilst there may be GxE interactions in certain traits, there are likely to be inherent traits that are desirable and useful in all environments. For example, a priori, it would seem likely that the ability to grow fast and convert feed efficiently under good conditions would be a generally desirable characteristic of shrimp populations. Taking the WSSV example from above, with the need for resistance to WSSV in the Pacific Coast of Northern South America, it would seem obvious that it would be advantageous to select for the same characteristics of fast growth and high feed conversion efficiency that are desired in the Atlantic Coast. Now, the conventional wisdom for animal breeders, who normally work largely with quantitatively controlled traits, is that the most efficient means of obtaining genetic gain is to simultaneously select for all desired traits in the same population. This may well be true if only one population is being developed for one mega-environment. However, in shrimps it is likely that more than one population will be required due the wide variety of conditions and the GxE interaction. If that is the case then we suggest that it may be more efficient for a breeding programme to develop a nucleus population that has universally desired traits such as high growth rate and efficient feed conversion. Then from this high performance nucleus population separate populations or sub-populations can be developed by selecting from within the nucleus or by introduction of genotypes with specific traits that are required for those populations to be well adapted to specific mega-environments. This strategy is likely to be particularly effective with characteristics which are controlled by a small number of genes, and is the same approach as that used by cattle breeders to introduce the poll trait into already improved populations, rather than starting from scratch again.

One possible means of developing populations specifically adapted to a particular environment is to practice intense selection for that environment in the dissemination phase (see section 8). This, on its own, will not improve the adaptation of the nucleus for the specific conditions. However, if some superior animals that have been selected in the dissemination programme are returned to the nucleus, then it may also be possible to improve the adaptation of the nucleus to those specific conditions.

Dissemination schemes

Obtaining the full benefits from investments in systematic genetic improvement work requires efficient systems for transferring the genetic gains obtained in the breeding nucleus to the target shrimp sector(s) (Ponzoni et al., 2008). The dissemination strategy has a major impact on the overall cost-benefit ratio of the breeding programme, since the specific operational costs of the breeding nucleus involving production of breeding candidates, performance testing and selection are fixed and only marginally influenced by the total output of improved seed. Dissemination programmes may cross national borders adding further complications due to regulations and quarantine restrictions.

Due to the limited number of broodstock usually kept for the breeding nucleus operation and the difficulties associated with combining genetic improvement operations with large scale seed production, dissemination of genetically improved seed for most species involve one or more collaborating hatcheries serving as multiplication units (figure 1). These hatcheries receive genetically improved material (usually as nauplii or PLs), rear them till maturation and then in turn offer improved seed to the end-users. It should be noted that selection, whether conscious or not, will occur in this rearing phase as less than ten percent of the nauplii delivered to the hatcheries will survive and be selected as breeders. This offers the possibility to continue the selection process in the dissemination phase and if necessary to select for specific adaptation for a particular producer or group of producers with special requirement. At the very least it is necessary to ensure that multipliers do not lose the painstakingly obtained genetic gain in the process of dissemination. Multipliers should be carefully chosen based on their capacity to adequately handle genetic material, their strategic location in target mega-environments and market regions and their ability to produce large volumes of high quality seed when it is needed by the growers.

The selection conducted in the breeding nucleus aims to maximise the long term selection response for a broadly defined breeding goal, normally involving several traits. Maintenance of genetic variance in the nucleus population is essential, and this imposes significant restrictions on the selection intensity that can be applied in the breeding nucleus. This restriction may significantly reduce the short term genetic gains in the nucleus, but it maintains genetic variability in the nuclear population which facilitates rapid response to changes in the environment or the appearance of threats such as epidemics of previously unknown or unreported diseases. These restrictions do not apply, or at least are less severe, in dissemination populations. Hence, there is an opportunity to obtain rapid short term genetic gain in the dissemination process. Furthermore, if specific adaptation is required to a particular environment due to a GxE interaction it may be possible in the dissemination process to select populations specifically adapted to a particular environmental condition.

Well designed dissemination schemes open the way for effective differentiation of the genetic product from the improvement programme. The selection of dissemination animals can therefore be made according to short term priorities, i.e with emphasise on traits of immediate importance. Selection of dissemination animals may be made more narrowly with high selection pressure to maximise genetic gain with less attention to reduction of inbreeding. While selection of the breeding nucleus often targets multiple traits assigned with individual economic weights that reflect the long term priorities the end-users’ short term priorities may deviate substantially. Effective dissemination schemes should take this into account and allow large individual producers or groups of producers to define their own short-term breeding goal and provide grow-out seed accordingly. This was done in the breeding programme for P. vannamei in Colombia (Erazo et al., 2005). Selection in the breeding nucleus simultaneously targeted harvest weight, general pond survival and resistance to WSSV, whereas for dissemination WSSV resistance was not included as a desired trait for producers in the Atlantic coast where WSSV is not a problem.

In recent years, the rapid spreading of viral and bacterial diseases and to some extent also biodiversity concerns has strongly limited international dissemination of live shrimp material including genetically improved material. In the Americas many countries have effectively have sealed their borders for this trade. However in Asia the other major shrimp farming region, legal regulations or their enforcement have been more lax as witnessed by the dramatic increase in the production of a species like P. vannamei, originally from the Pacific coast of the Americas. Nevertheless interest in specific pathogen free (SPF) material for major species has grown in Asia in recent years (Lightner, 2005, Pantoja et al., 2005, Hening et al., 2005). The development of reliable methods for cryopreservation of shrimp material could significantly facilitate safer and more cost effective international dissemination of genetically improved seed between breeding nuclei and regional multipliers situated in other countries by allowing for more rigorous testing of health status and significantly reduce the volume of biological material to be transported. The limited number of reports on successful cryopreservation in marine shrimp (Anchordoguy et al., 1988; Dumont et al.,1992; Dong et al., 2004; Gwo 2000; Lezcano et al., 2004) however, suggests that it still may take long before reliable methods are available for the major penaeid species.

We suggest that an optimal dissemination process for shrimps should have the following features. The initial broodstock used to produce the nauplii or post larvae for the mutlipliers should be selected for traits that the end producer desires. Their progeny should then be raised in conditions as close to commercial practice as possible and at commercial harvest weight the largest survivors should be selected with as high a selection pressure as is commensurate with supplying the producers with enough seed to meet their needs. This process effectively provides selection for pond survival and growth. The selected materials should then be raised to maturity and used as broodstock either to produce nauplii for commercial grow out or to pass through a second round of selection before dissemination as commercial stock. In the Colombian programme we have observed that intense selection pressure in this dissemination phase can provide up to 2g per animal of extra growth after approximately 15 weeks when animals are produced in the same environment and with the same management as that used in selection process, but this increase is much less when transferred to distinct management systems and environmental conditions.

A critical aspect of effective dissemination schemes is the capacity to monitor the performance at the commercial level (see the “methods to measure genetic gain” section of this chapter). If the breeding programme compiles the data collected by commercial operations on the performance shrimp in the ponds, coupled with information on management, genetic stock used, weather conditions and the pond characteristics it is possible to demonstrate superior performance under commercial conditions and also to rapidly detect any deficiencies in the performance of improved genetic stock disseminated to commercial operations.


The number of shrimp breeding programmes has grown rapidly since techniques were developed to close the reproductive cycle in captivity. Most breeding programmes focus on high growth rates and good survival under the intense commercial grow out conditions that are becoming the norm. Disease resistance is an important selection trait in various programmes, but only as a means to increase pond survival. Due to their high prolificacy and the large number of animals that can be handled it is possible to apply intense selection pressure to shrimp population and obtain rapid genetic gain. Although inbreeding is potentially a serious problem, it appears that in highly prolific species such as shrimp deleterious genes are rapidly eliminated from the population. Nevertheless, loss of genetic variation and inbreeding are likely to limit selected populations’ ability to face new threats such as changing environmental or management conditions or the appearance of previously unimportant diseases. Consequently, although mass selection programmes often provide excellent short term results many programmes now use combined selection so as to manage in breeding and maintain genetic variability. Currently major limitations on combined selection are related to marking or tagging animals: the animals have to be reared separately before tagging and only a limited number of animals can be tagged, thus reducing the potential for very intense selection and the chance of picking up rare mutants or recombinants with exceptional performance. DNA fingerprinting offers exciting opportunities to remove these restrictions and greatly improve the long term genetic gains that can be obtained from combined selection. At present few breeding programmes take into account genotype x environment interactions. However, due to the heterogeneous environment in which shrimps are raised, the interactions appear to be substantial and it is suggested that breeders should define mega-environments which they target with their nucleus breeding programmes. Although the benefit cost ratio of breeding programmes is generally very large, this can only be achieved if effective dissemination programmes provide growers with animals suited to their particular conditions. The dissemination programmes can be designed in such a manner that during the multiplication process selection pressure is applied to provide growers with material particularly suited to their conditions.


Amaya E, Davis A and Rouse D (2007) Replacement of fishmeal in practical diets for the Pacific white shrimp (L. vannamei) reared under pond conditions Aquaculture 262 393–401

Anchordoguy T, Crowe JH, Griffin FJ and Clark WH (1988) Cryopreservation of sperm from the marine shrimp Sicyonia ingentis Cryobiology 25 238–243

Aranguren F and Lightner D (2009) Reverse gavage: A new standardized technique for infecting shrimp suitable for challenge tests. Book of Abstracts Aquaculture America pp 407 Seattle USA

Argue BJ, Arce SM, Lotz JM and Moss SM (2002) Selective breeding of Pacific white shrimp (Litopenaeus vannamei) for growth and resistance to Taura Syndrome Virus Aquaculture 204 447– 460

Bakos J, Varadi L, Gorda S, and Jeney Z (2006) Lessons from the Breeding Program on Common Carp in Hungary In Development of aquatic animal genetic improvement and dissemination programs: current status and action plans pp 27-33 Eds RW Ponzoni, BO Acosta and AG Ponniah ,WorldFish Center Conference Proceedings

Bao Z, Zhang Q, Wang H and Dai J (1994) Cytochalasin B induced triploidy in Penaeus chinensis Acta Oceanoligica Sinica 13 261-267

Barnett NL, Purvis IW, van Hest B, and Franklin IR (1999) The accuracy of current dam pedigree recording strategies employed by stud Merino breeders In Proceedings of the Association for the Advancement of Animal Breeding and Genetics pp 373–376. Mandurah, Western Australia

Bentsen HB and Olesen I (2002) Designing aquaculture mass selection programs to avoid high inbreeding rates Aquaculture 204 349-359

Benzie JAH, Kenway M, Ballment E, Frusher S and Trott L (1995) Interspecific hybridization of the tiger prawns Penaeus monodon and Penaeus esculentus Aquaculture 13 103-111

Benzie JAH, Kenway M and Ballment E (2001) Growth of Penaeus monodon x Penaeus esculentus tiger prawn hybrids relative to the parental species Aquaculture 193 227-237

Benzie J (2005) Marine Shrimp Genomics. (accessed june 10 2009)

Bierne N, Beuzart I, Vonau V, Bonhomme F, Bedie E and Aquacop GC (2000) Microsatellite-associated heterosis in hatchery-propagated stocks of the shrimp Penaeus styliostris Aquaculture 184 203-219

Blasiola G (2005) Koi: A complete Pet Owner`s Manual Barrons Educational Services, New York

Bray WA, Lawrence AL, Lester LJ, Smith LL (1990) Hybridization of Penaeus setiferus Linnaeus, 1767 and Penaeus schmitti Birkenroad, 1936 Decapoda Journal of Crustacean Biology 10 278–283

Browdy CL and Warr G (2005) Double-Stranded RNA Induces Sequence-Specific Antiviral Silencing in Addition to Nonspecific Immunity in a Marine Shrimp: Convergence of RNA Interference and Innate Immunity in the Invertebrate Antiviral Response? Journal of Virology 79 13561-13571

Brown CR, Wolliams JA and McAndrew BJ (2005) Factors influencing effective population size in commercial populations of gilthead seabream, Sparus aurata. Aquaculture 247 214 – 219

Caballero A and Toro MA (2000) Interrelations between effective population size and other pedigree tools for the management of conserved populations Genetics Research 75 33-343

Castillo Juárez H, Montaldo HH, Arechavaleta ME, Alonso RA, Campos-Montes GR (2005) Programa de Mejora Genética del camarón blanco en México In Proceedings Memorias 2ndo foro internacional de Acuicultura pp , Hermosillo, Sonora, México.

Carr WH, Fjalestad KT, Godin D, Swingle J, Sweeney JN, Gjedrem T (1997) Genetic variation in weight and survival in a population of specific pathogen-free shrimp, Penaeus vannamei In Diseases in Asian Aquaculture, vol. III. Fish Health Section pp 265– 271 Eds TW Flegel and IH MacRae. Asian Fisheries Society, Manila, Philippines

Charoensapsri W, Amparyup P, Hirono I, Aoki T and Tassanakajon A (2009) Gene silencing of a prophenoloxidase activating enzyme in the shrimp, Penaeus monodon, increases susceptibility to Vibrio harveyi infection Developmental and Comparative Immunology 33 811-20

Clavero-Salas A, Sotelo-Mundo RR, Gollas-Galvan T, Hernandez- Lopez J, Peregrino-Uriarte AB, Muhlia Almazan A, Yepiz-Plascencia G (2007) Transcriptome analysis of gills from the white shrimp Litopenaeus vannamei infected with White spot syndrome virus Fish& Shellfish Immunology 23 459–472

Cock J, Gitterle T, Salazar M, Rye M (2009) Breeding for disease resistance in shrimp Aquaculture 286 1-11

Coltman DW, Pilkington JG, Smith JA, and Pemberton JM (1999) Parasite-mediated selection against inbred Soay sheep in a free-living, island population Evolution 53 1259–1267

Coltman DW and Slate J (2003) Microsatellite measures of inbreeding: a meta-analysis. Evolution 57 971–983

Coulson TN, Albon S, Slate J and Pemberton JM (1999) Microsatellite loci reveal sex-dependent responses to inbreeding and outbreeding in red deer calves Evolution 53 1951–1960

Dai JX, Bao ZM and Zhang QQ (1993) Studies on triploid induction in Penaeus orientalis, I: temperature shocks Hereditas 15 15–18

Darwin C (1883) The Variation of Animals and Plants under Domestication. 2nd edition, Appleton. (accessed mayo 15 2009)

DeDonato M, Manrique R, Ramírez R, Mayer L and Howell C (2005) Mass selection and inbreeding effects on a cultivated strain of Penaeus (Litopenaeus) vannamei in Venezuela. Aquaculture 247 159-167

de Lorgeril J, Gueguen Y, Goarant C, Goyard E, Mugnier C, Fievet J, Piquemal D and Bachere E (2008) A relationship between antimicrobial peptide gene expression and capacity of a selected shrimp line to survive a Vibrio infection Molecular Immunology 45 3438–3445

Dixon TJ, Coman GJ, Arnold SJ, Sellars MJ, Lyons RE, Dierens L, Preston NP, and Yutao LI (2008) Shifts in genetic diversity during domestication of Black Tiger shrimp, Penaeus monodon, monitored using two multiplexed microsatellite systems Aquaculture 238 1-6

Dong Q, Lin J and Huang C (2004) Effects of cryoprotectant toxicity on the embryos and larvae of pacific white shrimp Litopenaeus vannamei Aquaculture 274 655-670

Dong S, Kong J, Zhang T, Meng X and Wang R (2006) Parentage determination of Chinese shrimp (Fenneropenaeus chinensis) based on microsatellite DNA markers Aquaculture 258 283-288

Doyle RW and Herbinger CM (1994) The use of DNA fingerprinting for high-intensity within-family selection in fish breeding In Proceedings of the 5th World Congress on Genetics Applied to Livestock Production pp 7-12 Department of Animal and Poultry Science, University of Guelph, Guelph

Doyle RW, Moss DR and Moss SM (2006) Shrimp Copyright: Inbreeding Strategies Effective Against Illegal Copying of Genetically Improved Shrimp. Global Aquaculture Advocate April/May 76-79

Dumas S and Campos-Ramos R (1999) Triploidy induction in the Pacific white shrimp Litopenaeus vannamei (Boone) Aquaculture Research 30 621-624

Dumont P, Levy P, Simon C, Diter A and Francois S (1992) Freezing of sperm ball of the marine shrimp Penaeus vannamei In Workshop on Gamete and Embryo Storage and Cryopreservation in Aquatic Organisms, Marry Le Roi, France

Dunham RA (2004) Aquaculture and fisheries biotechnology. CABI. 372pp

Dunham RA, Majumdar K, Hallerman E, Bartley D, Mair G, Hulata G, Liu Z, Pongthana N, Bakos J, Penman D, Gupta M, Rothlisberg P and Hoerstgen-Schwark G (2001) Review of the status of aquaculture genetics. In Aquaculture in the Third Millennium. Technical Proceedings of the Conference on Aquaculture in the Third Millennium pp. 137-166 Eds RP Subasinghe, P Bueno, MJ Phillips, C Hough, SE, McGladdery and JR Arthur Bangkok, Thailand NACA, Bangkok and FAO, Rome. (Accessed May 18 2009)

Eknath AE, Dey MM, Rye M, Gjerde B, Abella TA, Sevilleja RC, Tayamen MM, Reyes TA and Bentsen HB (1998) Selective breeding of Nile tilapia in Asia In Proceedings 6th World Congress on Genetic Applied to Livestock Production pp 11-16 Armidale, Australia,

Embody GC and Hayford CO (1925) The advantage of rearing brook trout fingerlings from selected breeders. Transactions of the American Fisheries Society 55 135-138

Erazo C, Suarez A, Lozano C, Cock J, Suarez JA, Faillace J, Salazar M and Gitterle T (2005) Esquema de diseminación de la semilla mejorada de Penaeus vannamei, utilizando una ronda de selección masal y sistemas de cria superintensiva [Dissemination schemes of the improved seed of L vannamei using a round of mass selection and super intensive raising systems] V Seminario Internacional de Acuicultura. Universidad Nacional de Colombia. Bogota, 21-25 de noviembre de 2005.

Estoup A, Gharbi K, San Cristobal M, Chevalet C, Haffray P and Guyomard R (1998) Parentage assignment using microsatellites in turbot (Scophthalmus maximus) and rainbow trout (Oncoryhnchus mykiss) hatchery populations Canadian Journal of Fisheries and Aquatic Sciences 55 715–725.

Falconer DS and Mackay TFC (1996) Introduction to Quantitative Genetics Longman Group Ltd. Essex, England

FAO (1995) Code of Conduct for Responsible Fisheries, Rome, FAO ISBN 92-5-103834-1

Fast AW and Wyban J (2001) Polyploidy in shrimp; advanced technology for improved performance, genetic and environmental security Global Aquaculture Advocate 4 16-17

Freitas PD, Calgaro MR and Galetti Jr PM (2007) Genetic diversity within and between broodstocks of the white shrimp Litopenaeus vannamei (Boone, 1931) (Decapoda, Penaeidae) and its implication for the gene pool conservation Brazilian Journal of Biology 67S 939-943

Gall GAE and Bakar Y (2002) Application of mixed-model techniques to fish breed improvement: analysis of breeding-value selection to increase 98-day body weight in tilapia Aquaculture 212 93-113

Garcia JC, Reyes A, Salazar M and Granja CB (2009) Differential gene expression in the Pacific white shrimp Litopenaeus vannamei with different levels of resistance to White Spot Syndrome Virus infection Aquaculture 289 253-258

Gitterle T (1999) Evaluación de la Resistencia de diferentes poblaciones del camaron marino Litopenaeus vannamei (Boone 1931) al Virus del Sindrome del Taura (TSV) bajo condiciones controladas [Evaluation of the resistance of different populations of the marine shrimp L vannamei (Boone 1931) to the Taura syndrome virus (TSV) under controlled conditions]. Thesis Universidad Jorge Tadeo Lozano, Bogota, Colombia 74 pp

Gitterle T, Rye, M, Salte R, Cock J, Johansen H, Lozano C, Suárez JA. and Gjerde B (2005a) Genetic (co)variation in harvest body weight and survival in Penaeus (Litopenaeus) vannamei under standard commercial conditions Aquaculture 243 83-92

Gitterle T, Salte R, Gjerde B, Cock J, Johansen H, Salazar M, Lozano C and Rye M (2005b) Genetic (co)variation in resistance to White Spot Syndrome Virus (WSSV) and harvest weight in Penaeus (Litopenaeus) vannamei Aquaculture 246 139–149

Gitterle T, Johansen H, Erazo EC, Lozano C, Cock J, Salazar M and Rye M (2007) Response to multi-trait selection for harvest weight, overall survival, and resistance to white spot syndrome virus (WSSV) in Penaeus (Litopenaeus) vannamei Aquaculture 272 S262

Gitterle T, Martinez W, Johanssen H, Marimon F, Salazar M, Faillace J, Suarez CA, Rye M and Cock, JH (2009) Commercial field performance as a measure of genetic improvement in the Pacific White Shrimp Penaeus (Litopenaeus) vannamei. X International Symposium of Genetic in Aquaculture Bangkok June 2009

Gjedrem T (1996) Selection and breeding programs in aquaculture Springer, Netherlands

Gjedrem T (1997) Selective breeding to improve aquaculture production World Aquaculture 28 33-45

Gjedrem T (2005) Genotype environment interaction In Selection and Breeding Programs in Aquaculture pp 233-242 Ed T Gjedrem. Springer, Netherlands

Gjerde B, Gjøen HM and Villanueva B (1996) Optimum designs for fish breeding programmes with constrained inbreeding. Mass selection for a normally distributed trait. Livestock Production Science 47 59-72

Gjerde B (2005) Design of fish breeding programs. In Selection and breeding programs in Aquaculture pp. 173-195 Ed T Gjedrem. Springer, Netherlands

Gjerde B (1986) Growth and reproduction in fish and shellfish Aquaculture 57 37-55

Gjerde, B. (2006) State of the art in selective breeding of aquacultured species. BARD Workshop Aquaculture Genetics – Status and Prospects February 20-23. 2006 – Eilat, Israel. The Israeli Journal of Aquaculture, Bamidgeh 58(4), 230.

Gjøen HM and Bentsen HB (1997) Past, present and future of genetic improvement in salmon aquaculture. ICES Journal of Marine Science 54 1009-10014

Glenn KL, Grapes L, Suwanasopee T, Harris DL, Li Y, Wilson K and Rothschild MF (2005) SNP analysis of AMY2 and CTSL genes in Litopenaeus vannamei and Penaeus monodon shrimp Animal Genetics 36 235–236

Goyard E, Patrois J, Peignon J, Vanaa MV, Dufour R, Viallon J, Bedier E (2002) Selection for better growth of Penaeus stylirostris in Tahiti and New Caledonia. Aquaculture 204 461–468

Goyard E, Goarant C, Ansquer D, Brun P, de Decker S, Dufour R, Galinié C, Peignon J-M, Pham D, Vourey E, Harache Y and Patrois J (2008) Cross breeding of different domesticated lines as a simple way for genetic improvement in small aquaculture industries: Heterosis and inbreeding effects on growth and survival rates of the Pacific blue shrimp Penaeus (Litopenaeus) stylirostris Aquaculture 278 43-50

Gross PS, Bartlett TC, Browdy CL, Chapman RW and Warr GW (2001) Immune gene discovery by expressed sequence tag analysis of hemocytes and hepatopancreas in the Pacific White Shrimp, Litopenaeus vannamei, and the Atlantic White Shrimp L setiferus Developmental and Comparative Immunology 25 565–577

Gwo JC (2000) Cryopreservation of aquatic invertebrate semen Aquaculture Research 31 259–271

Harris DL and Newman S (1994) Breeding for Profit: Synergism Between Genetic Improvement and Livestock Production (A Review) Journal of Animal Science 72 2178-2200

Havenstein B, Ferket PR and Qureshi MA (2003) Growth, Livability, and Feed Conversion of 1957 Versus 2001 Broilers When Fed Representative 1957 and 2001 Broiler Diets Poultry Science 82 1500–1508

Hazel N (1943) The genetic basis for constructing selection indexes Genetics 8 476-490

Henderson CR (1975) Best linear unbiased estimation and prediction under a selection model Biometrics 31 423

Hennig OL, Arce SM, Moss SM, Pantoja CR and Lightner DV (2005) Development of a specific pathogen free population of the Chinese fleshy prawn, Fenneropenaeus chinensis: Part II. Secondary quarantine Aquaculture 250 579-58

Henson-Apollonio V (2006) Maintaining the “Public Good” Nature of Improved Fish Strains: Dissemination of knowledge and materials In Development of aquatic animal genetic improvement and dissemination programs: current status and action plans pp 45-50 Eds RWPonzoni, BO Acosta and AG Ponniah WorldFish Center Conference Proceedings

Hetzel DJS, Crocos PJ, Davis GP, Moore SS and Preston NC (2000) Response to selection and heritability for growth in the Kuruma prawn, Penaeus japonicus Aquaculture 181 215-223

Hill WG and Bunger L (2004) Inferences on the genetics of quantitative traits from long-term selection in laboratory and domestic animals Plant Breeding Reviews 24 169-210

Hinrichs D, Wetten M and Meuwissen THE (2006) An algorithm to compute optimal genetic contributions in selection programs with large numbers of candidates Journal of Animals Science 84 3212-3218

Hoffman AA and Merilä J (1999) Heritable variation and evolution under favourable and unfavourable conditions Trends in Ecology and Evolution 14 96-101

Hollebecq MC and Haffrey P (1999) Genetic Improvements In Carp: biology and culture p 107 Ed. B Roland Springer, Netherlands

Holtsmark M, Sonesson AK, Gjerde B and Klemetsdal G (2006) Number of contributing subpopulations and mating design in the base population when establishing a selective breeding program for fish Aquaculture 258 241-249

Holtsmark M, Klemetsdal G, Sonesson AK, Woolliams JA (2008) Establishing a base population for a breeding program in aquaculture, from multiple subpopulations, differentiated by genetic drift: II. Sensitivity to assumptions on the additive genetic relationships of base animals Aquaculture 274 241-246

Huang CM and Liao IC (1990) Response to mass selection for growth rate in Oreochromis niloticus Aquaculture 85 199-205

Hulata G (2001) Genetic manipulations in aquaculture: a review of stock improvement by classical and modern technologies Genetica 111 155–173

Ibarra AM, Perez-Rostro C, Tamirez JL and Ortega-Estrada E (2007) Genetics of the resistance to hypoxia in postlarvae and juveniles of the Pacific white shrimp Penaeus (Litopenaeus) vannamei (Boone 1931) Aquaculture Research 38 838-846

Jeney Z, Bakos J, Demeter-Pedery T, Varadi L and Gorda S (2008) Carp Breeding Genetics in Hungary. Genetics and biodiversity: Consortium on freshwater finfish genetics and breeding Research Institute for Aquaculture No. 2 (RIA 2), Ho Chi Minh City, Vietnam (accessed June 16 2009)

Jerry DR, Evans BS, Kenway M and Wilson K (2006) Development of a microsatellite DNA parentage marker suite for black tiger shrimp Penaeus monodon Aquaculture 255 542-547

Jerry DR, Preston NP, Crocos PJ, Keys S, Meadows JRS and Li Y (2004) Parentage determination of Kuruma shrimp Penaeus (Marsupenaeus) japonicus using microsatellite markers (Bate) Aquaculture 235 237-247

Jiang D, Torremorell M, Gong H and van de Steen H (2005a) Specific pathogen free and/or resistant broodstock: a conflict, compromise or opportunity? In Proceedings “Memorias 2ndo foro internacional de Acuicultura”, December 1-3, Hermosillo, Sonora, México

Jiang D, Rocha J and van de Steen H (2005b) Is there any value in selection for carcass and/or meat quality in shrimp? In Proceedings “Memorias 2ndo foro internacional de Acuicultura”, December 1-3, Hermosillo, Sonora, México

Johnson NA, Vallejo RL, Silverstein JT, Welch TJ, Wiens GD, Hallerman EM and Palti Y (2008) Suggestive Association of Major Histocompatibility IB Genetic Markers with Resistance to Bacterial Cold Water Disease in Rainbow Trout (Oncorhynchus mykiss) Marine Biotechnology 10 429–437

Jørgensen SV, Afanasyev S and Aleksei Krasnov (2008) Gene expression analyses in Atlantic salmon challenged with infectious salmon anemia virus reveal differences between individuals with early, intermediate and late mortality BMC Genomics 9 179

Kenway M, Macbeth M, Salmon M, McPhee C, Benzie J, Wilson K, and Knibb W (2006) Heritability and genetic correlations of growth and survival in black tiger prawn Penaeus monodon reared in tanks Aquaculture 259 138-145

Keys SJ, Crocos PJ, Burridge CY, Coman GJ, Davis GP and Preston NP (2004) Comparative growth and survival of inbred and outbred Penaeus (marsupenaeus) japonicus, reared under controlled environment conditions: indications of inbreeding depression. Aquaculture 241 151-168

Klimpel KR (2001) Shrimp Genomics: Making the Most of Limited Resources ARS-OI Biotechnology-Aquaculture Workshop may 15 2009)

Lawrence AL. Bray WA, Wilkenfield JS, Lester LJ (1984) Successful interspecific crosses of two species of marine shrimp, Penaeus stylirostris - Penaeus setiferus pp39 Abstracts Annual Conference World Aquaculture Society March 1984 Vanocuver

Le Moullac G, Goyard E, Saulnier D, Haffner P, Thouard E, Nedelec G, Goguenheim J, Rouxcel C, and Aquacop GC (2003) Recent improvements in broodstock management and larviculture in Marine species in Polynesia and New Caledonia: genetic and health approaches Aquaculture 227 89-106

Lezcano M, Granja C and Salazar M (2004) The use of flow cytometry in the evaluation of cell viability of cryopreserved sperm of the marine shrimp (Litopenaeus vannamei).
Cryobiology 48 349-56

Li F, Xiang J, Zhang X, Zhou L, Zhang C and Wu C (2003) Gonad Development Characteristics and Sex Ratio in Triploid Chinese Shrimp (Fenneropenaeus chinensis) Marine Biothechnology 5 528-535

Li FH, Zhou LH, Xiang JH, Liu XD and Zhu JZ (1999) Triploidy induction with heat shocks to Penaeus chinensis and their effects on gonad development Chinese Journal Oceanological Limnology 17 57–61

Li Z, Li J, Wang Q, He Y and Liu P (2006) AFLP-based genetic linkage map of marine shrimp Penaeus (Fenneropenaeus) chinensis Aquaculture 261 463-472

Lightner DV (2005) Biosecurity in Shrimp Farming: Pathogen exclusion through use of SPF stock and routine surveillance Journal of the World Aquaculture Society 36 229-248.

Lotz JM (1997) Disease control and pathogen status assurance in an SPF-based shrimp aquaculture industry, with particular reference to the United States. In Diseases in Asian Aquaculture III, Fish Health Section pp 243-2254 Eds TW Flegel and IH MacRae Asian Fisheries Society, Manila

Lynch M and Walsh B (1998) Genetics and Analysis of Quantitative Traits Sinauer Associates, Sunderland, Massachusetts USA.

Lyons RE, Dierens LM, Tan SH, Preston NP and Li Y (2007) Characterization of AFLP markers associated with growth in the Kuruma Prawn, Marsupenaeus japonicus, and identi?cation of a candidate gene Marine Biotechnology 9 712-721

McVey JP (1993) CRC Handbook of Mariculture: Crustacean aquaculture 2nd ed. CRC Press, USA

Meuwissen THE (1997) Maximizing the response of selection with a predefined rate of inbreeding Journal of Animals Science 75 934-940

Meuwissen THE, Hayes BJ and Goddard ME (2001) Prediction of Total Genetic Value Using Genome-Wide Dense Marker Maps Genetics 157 1819-1829

Misamore M and Browdy CL (1997) Evaluating hybridization potential between Penaeus setiferus and Penaeus vannamei through natural mating, artificial insemination and in vitro fertilization Aquaculture 150 1-10

Moav R and Wohlfarth GW (1974) Magnification through competition of genetic differences in yield capacity in carp Heredity 33 181-202

Moav R, Brody T, Hulata G (1978) Genetic Improvement of Wild Fish Populations Harvested wild organisms can be preserved by hybridization with “tailor-made” selected breeds Science 22 1090-1094

Moav R and Wohlfarth GW (1976) Two-way selection for growth rate in the common carp (Cyprinus carpio) Genetics 82 83–101

Morelli M and Aquacop (2003) Effects of heat-shock on cell division and microtubule organization in zygotes of the shrimp Penaeus indicus (Crustacea, Decapoda) observed with confocal microscopy. Aquaculture 216 39–53

Moss DR, Arce S and Moss S (2004) O.I. Research Reveals Inbreeding Depression in Pacific White Shrimp Global Aquaculture Advocate 7 50-51

Moss DR, Arce SM, Otoshia CA, Doyle RW and Moss SM (2007) Effects of inbreeding on survival and growth of Pacific white shrimp Penaeus (Litopenaeus) vannamei Aquaculture 272 S30-S37

Nielsen, JL (1998) Population genetics and the conservation and management of Atlantic salmon (Salmo salar) Canadian Journal of Fisheries and Aquatic Science 55145–152

Norris BJ, Coman FE, Sellars MJ and Preston NP (2005) Triploid induction in Penaeus japonicus (Bate) with 6-dimethylaminopurine Aquaculture Research 36 202-206

Ogata HY, Oku H and Murai T (2002) Growth, feed efficiency and feed intake of off-spring from selected and wild Japanese flounder (Paralichthys olivaceus) Aquaculture 211 183-193

Ogden ER and Weigel K (2007) Can you shrinkwrap a cow? Protections available for the intellectual property of the animal breeding industry Animal Genetics 38 647–654

O’Leary NA, Trent III HF, Robalino J, Peck MET, McKillen DJ and Gross PS (2006) Analysis of multiple tissue-specific cDNA libraries from the Pacific white leg shrimp. Litopenaeus vannamei International Comparative Biology 46 931–939

Pandian TJ and Koteeswaran R (1998) Ploidy induction and sex control in fish Hydrobiologia, Golden Volume 384 167-243

Pante MJR, Gjerde B, McMillan I and Misztal I (2002) Estimation of additive and dominance genetic variances for body weight at harvest in rainbow trout, Oncorhyncus mykiss Aquaculture 204 383-392

Pantoja CR, Song X, Xia L, Gong H, Wilkenfeld B, Noble B and Lightner DV (2005) Development of a specific pathogen-free (SPF) population of the Chinese fleshy prawn Fenneropenaeus chinensis: Part 1: Disease Pre-screening and Primary Quarantine Aquaculture 250 573-578

Pardey PG and Beintema NM (2001) Slow Magic: Agricultural R&D a Century After Mendel International Food Policy Research Institute Washington D.C.

Pogson HG and Zouros E (1994) Allozyme and RFLP heterozygotes as correlates of growth rate in the scallop Placopecten magellanicus: a test of the associative overdominance hypothesis Genetics 137 221–231

Ponzoni RW (2006) Genetic improvement and effective dissemination: keys to prosperous and sustainable aquaculture industries In Development of aquatic animal genetic improvement and dissemination programs: current status and action plans pp 1-7 Eds RW Ponzoni, BO Acosta and AG Ponniah WorldFish Center Conference Proceedings

Ponzoni RW, Nguyena NH and Khaw HL (2007) Investment appraisal of genetic improvement programs in Nile tilapia (Oreochromis niloticus) Aquaculture 269 187-199

Ponzoni RW, Khaw HL and Ninh NH (2008) Accounting for genotype by environment interaction in economic appraisal of genetic improvement programs in common carp Cyprinus carpio Aquaculture 285 47-55

Preston NP, Croscos PJ and Keys S (2002) Improving the growth rates of farm stocks of Panaeus japonicus through selective breeding Aquaculture 204 239

Purdom CE (1993) Genetics and Fish Breeding Springer, Netherlands

Reddy PVGK, Gjerde B, Tripathi SD, Jana RJ, Mahapatra KD, Gupta SD, Saha JN, Sahoo M, Lenka S, Govindassamy P, Rye M and Gjedrem T (2002) Growth and survival of six stocks of rohu (Labeo rohita, Hamilton) in mono and polyculture production systems Aquaculture 203 239-250

Reed DH and Frankham R (2001) How closely correlated are molecular and quantitative measures of genetic variation? A meta-analysis Evolution 55 1095-1103

Robalino J, Browdy CL, Prior S, Metz A, Parnell P, Gross P and Warr G (2004) Induction of Antiviral Immunity by Double-Stranded RNA in a Marine Invertebrate Journal of Virology 78 10442-10448

Robalino J, Bartlett T, Shepard E, Prior S, Jaramillo G, Scura E, Chapman RW, Gross PS, Browdy CL, Warr GW (2005) Double-Stranded RNA induces sequence-specific antiviral silencing in addition to nonspecific immunity in a marine shrimp: Convergence of RNA interference and innate immunity in the invertebrate antiviral response? Journal of Virology 79 13561–13571

Robalino J, Almeida JS, McKillen D, Colglazier J, Trent HF, ChenYA, Peck ME, Browdy CL, Chapman RW, Warr GW and Gross PS (2007) Insights into the immune transcriptome of the shrimp Litopenaeus vannamei: tissue-specific expression profiles and transcriptomic responses to immune challenge Physiological Genomics 29 44–56

Robinson N and Hayes B (2008) Modelling the use of gene expression profiles with selective breeding for improved disease resistance in Atlantic salmon (Salmo salar) Aquaculture 285 38-46

Rocha J, Guerrelhas AC, and Teixeira AK (2008) Shrimp Genetic Improvement in Brazil, now with SPF Lines. (accessed June 15 2009)

Rodhouse PG, McDonald JH, Newell RIE and Koehn RK (1986) Gamete production, somatic growth and multiple-locus enzyme heterozygosity inMytilus edulis Marine Biology 90 209-214

Rosendal GK, Olesen I, Bentsen HB, Tvedet MW and Bryde M (2006) Access to and legal protection of aquaculture genetic resources – Norwegian perspectives The Journal of the World Intellectual Property 9 392-412

Rye M and Mao IL (1998) Nonadditive genetic effects and inbreeding depression for body weight in Atlantic salmon (Salmo salar L.) Livestock Production Science 57 15-22

Rye M and Gjedrem M (2005) Measuring Genetic Change In Selection and breeding programs in aquaculture pp 243-250 Ed T Gjedrem. Springer, Netherlands

Sellars MJ and Preston NP (2005) The effects of ionizing radiation on the reproductive capacity of adult Penaeus (Maruspenaeus) japonicus (Bate) Aquaculture Research 36 1144–1147.

Shockey JE, O’Leary NA, de la Vega E, Browdy CL, Baatz JE and Gross PS (2009) The role of crustins in Litopenaeus vannamei in response to infection with shrimp pathogens: An in vivo approach Developmental and Comparative Immunology 33 668-73

Siegel PB, Dodgson JB and Andersson L (2006) Progress from Chicken Genetics to the Chicken Genome Poultry Science 85 2050–2060

Silverstein JT, Hostuttler M and Blemings KP (2005) Strain differences in feed efficiency measured as residual feed intake in individually reared rainbow trout, Oncorhynchus mykiss (Walbaum) Aquaculture Research 36 704-711

Slate J, Kruuk LEB, Marshall TC, Pemberton JM, Clutton-Brock TH (2000) Inbreeding depression influences lifetime breeding success in a wild population of red deer (Cervus elaphus) Proceedings of the Royal Society, London series B 267 1657-1662.

Slate J and Pemberton JM (2002) Comparing molecular measures for detecting inbreeding depression Journal of Evolutionary Biology 15 20–31

Sonesson AK (2005) A combination of walk-back and optimum contribution selection in fish: a simulation study Genetic Selection and Evolution 37 587–599

Sorensen DA and Kennedy BW (1984a) Estimation of Response to Selection Using Least-Squares and Mixed Model Methodology Journal of Animals Sciences 58 1097-1106

Sorensen DA and Kennedy BW (1984b) Estimation of Genetic Variances from Unselected and Selected Populations. Journal of Animal Science 59 1213-1223

Teichert-Coddington DR and Smitherman RO (1988) Lack of Response by Tilapia nilotica to Mass Selection for Rapid Early Growth Transactions of the American Fisheries Society 117 297–300

Thelen GC and Allendorf FW (2001) Heterozygosity-fitness correlations in rainbow trout: effects of allozyme loci or associative overdominance? Evolution 55 1180–1187

Thodesen J, Grisdale-Helland B, Helland SJ and Gjerde B (1999) Feed intake, growth and feed utilization of off-spring from wild and selected Atlantic salmon (Salmo salar) Aquaculture 180 237-246

Thodesen J, Gjerde B, Grisdale-Helland, B, Storebakken T (2001) Genetic variation in feed intake, growth and feed utilization in Atlantic salmon (Salmo salar) Aquaculture 194 273-281

Thodesen J and Gjedrem T (2006) Breeding Programs On Atlantic Salmon In Norway– Lessons learned In Development of aquatic animal genetic improvement and dissemination programs: current status and action plans pp 22-26 Eds RW Ponzoni, BO Acosta and AG Ponniah ,WorldFish Center Conference Proceedings

Thorgaard GH (1986) Ploidy manipulation and performance Aquaculture 57 57-64

Tian Y, Kong J and Cuihua Y (2006) Comparative growth and viability of hybrids between two populations of Chinese shrimp (Fennropenaeus chinensis) Chinese Science Bulletin 51 2369-2374

Tian Y, Kong J, Li W, Luan S, Yangand C, Wang Q (2008) Genetic improvement on Chinese shrimp (Fenneropenaeus chinensis): growth and viability performance in F1 hybrids of different populations Chinese Journal of Oceanology and Limnology 26 369-374

Vandeputte M (2003) Selective breeding of quantitative traits in the common carp (Cyprinus carpio): a review Aquatic Living Resources 16 399–407

Vidal OM, Granja CB, Aranguren F, Brock JA and Salazar M (2001) A profound effect of hyperthermia on survival of Litopenaeus vannamei juveniles infected with white spot syndrome virus Journal of the World Aquaculture Society 32 364–372

Visscher PM, Woolliams JA, Smith D and Williams JL (2002) Estimation of pedigree errors in the UK dairy population using microsatellite markers and the impact on selection. Journal of Dairy Science 85 2368–2375

Wang HX, Wu CG, Zhang LS and Xiang JH (2006) The application of microsatellite markers for parentage determination in selective breeding of Pacific white shrimp (Litopenaeus vannamei) Yi Chuan 28 179-183

Weber K (2004) Population size and long-term selection Plant Breeding Reviews 24 249-268

Wilson K, Li Y, Whan V, Lehnert S, Byrne K, Moore S, Pongsomboon S, Tassanakajon A, Rosenberg G and Ballment E (2002) Genetic mapping of the black tiger shrimp Penaeus monodon with amplified fragment length polymorphism Aquaculture 204 297-309

Wray NR and Thompson R (1990) Prediction of rates of inbreeding in selected populations Genetics Research 55 41– 54.

Wray NR and Goddard ME (1994) Increasing long-term response to selection Genetic Selection and Evolution 26 431-451

Xiang JH, Clark WH, Griffin F and Hzerzler P (1991) A study on feasibility of chromosome set manipulation in the marine shrimp, Sicyonia ingentis In Proceedings of the 1990 International Crustacean Conference pp277 Eds PJF Davie and RH Quinn. Brisbane, Australia

Xiang JH, Zhou LH, Liu RY, Zhu JZ, Li FH and Liu XD (1993) Induction of the tetraploids of the Chinese shrimp Penaeus chinensis. In Biotechnology in Agriculture pp 841-846 Eds C You and ZL Chen China Science and Technology Press, Beijing

Xiang JH, Zhou LH and Li FH (1998) Reproductive and genetic manipulation in the Chinese shrimp, Penaeus chinensis (Osbeck, 1765) In Proceedings of the Fourth International Crustacean Congress pp 987–995

Yu M, Cheng Y and Rothschild MF (2006) SNP analysis of Molting related genes in Penaeus monodon and Litopenaeus vannamei shrimp (Brief report) Archiv Fur Tierzucht-Archives of Animal Breeding 49 411-412

Zeng D, Chen X, Li Y, Peng M, Ma N, Jiang W, Yang C, and Li M (2008) Analysis of Hsp70 in Litopenaeus vannamei and detection of SNPs Journal of Crustacean Biology 28 727-730

Zhang L, Yang C, Zhang Y, Li L, Zhang X, Zhang Q and Xiang J (2007) A genetic linkage map of Pacific white shrimp (Litopenaeus vannamei): sex-linked microsatellite markers and high recombination rates Genetica 131 37-49

1 In shrimp grow out we use a production event as the moment from which nauplii are produced through to the moment when the shrimps are harvested in a particular pond, raceway or other production unit.

Table 1. Summary of advantages and disadvantages of family and mass selection programmes.
Advantages Disadvantages
Racks pedigrees and facilitates control of the rate of accumulation of inbreeding in the target populations, limiting possible detrimental inbreeding depression effects. Broodstock are not normally evaluated under commercial conditions, and animals supposedly tested under commercial grow out conditions are not subjected to commercial management through all life stages.
Currently the most widely accepted method with widely proved efficiency. Considered industry standard for aquaculture programmes. High costs of establishing infrastructure and maintaining programme.
Makes use of sib information, providing more accurate selection of individual breeders. Expected results from tests not always reflected in commercial response.
Effective also for traits of low heritability (e.g. survival). Enables selection for traits that can only be recorded through destructive tests (e.g. carcass quality) or tests that cannot be undertaken in the breeding nucleus (e.g. disease challenge tests). May extend generation interval.
Effective for traits under polygenic control
Provides information on possible genotype x environment interaction and on genetic make up of traits and estimation of genetic gain.
Populations developed are likely to be more tolerant of extreme changes in the environment.
Suitable for the development of SPFs and eradication of diseases in the nucleus.
Individual or Mass
Advantages Disadvantages
Breeders are selected under commercial conditions including the hatchery and larval stages and natural copula can be used. Danger of selection for single gene resistance to pathogens, which may break down.
Facilitates high selection intensity and all of the genetic variance can be used.

Inevitably and potentially rapid loss of genetic variance if no specific measures taken to maintain genetic variation.

No need to identify parentage of individuals.

Genetic gain is difficult to estimate.

Phenotypic records obtained reflect performance under commercial production conditions.

Provides no information on genetic correlations and correlated response on the relationship between traits, particularly lack of knowledge of negative genetic correlations.

Dissemination of material is simple and there is a high level of confidence in its performance.

Management of diseases in the nucleus is difficult

Large size of population makes it possible to obtain new variants due to mutation and rare recombinants.

Can only be used for traits that can be evaluated on the live potential breeders. Efficient only for traits with high heritability, primarily growth.

Potentially shorter generation interval and hence faster response to selection.

Simple and effective in improving performance under conditions similar to those used in the selection process.


Figure 1. Dissemination of improved seed to end-users

Our Sponsors


Sustainable Aquaculture Modular Courses AVAILABLE NOW - University of St Andrews

Seasonal Picks

Know Your Freshwater Fishes - 5m

Aquacullture Webinars from TheFishSite - Sign Up Now