by
Zoltán Krasznai
Preface
This section of the course aims to provide basic theoretical information on heredity, to give practical schemes for increasing fish production by selection or hybridization, and to present some of the latest results of genetic research.
To achieve the above goals, the following subject areas are considered:
Identification of genetic material (cellular division)
Transmission and distribution of genetic material (Mendelian principles, qualitative characters)
Quantitative inheritance
Selective breeding and intraspecific hybridization (Methods of selection)
Interspecific hybridization
Genetic manipulation
Identification of genetic material
The animal cell and cell division
The cells are the smallest individual units of the living organism. Internally, the cell structure consists of two distinct parts: the cytoplasm and the nucleus.
Cytoplasm: Within the cytoplasmatic matrix lie a number of structures concerned with active cell functions. The presence or size of these structures varies between different organisms and between different tissues.
Generally the following inclusions can be observed in the cytoplasm: mitochondria, entoplasmic reticulum, Golgi bodies, and ribosomes.
The nucleus: The nucleus is surrounded by a double-walled membrane. Under the light microscope, staining with Giemsa reveals a dark network called chromatin. During the process of cellular division, the chromatin appears organized into distinct bodies, the chromosomes. Some areas of the chromosomes stain very darkly (called heterochromatin) while others stain relatively lightly (euchromatin). In addition to chromosomes there are usually one or more rounded bodies attached to specific chromosome regions, called nucleolar organizers.
Cell division: There are two main types of cell division: mitosis and meiosis. Mitosis results in daughter cells having the same number of chromosomes as the mother cell had, while meiosis results in cells with a half set of chromosomes.
Mitosis can be divided into the following phases:
1. interphase: Between two divisions the cells are in interphase. The interphase period involves processes associated with growth, and preparation for mitosis. The interphase period in mitosis is usually many times longer than that of meiosis, but may be considerably shortened in rapidly developing embryonic tissues.
2. prophase: The first stage of mitosis. The preparation for cell division continues. The chemical constituents for the new chromosomes are being synthesized. The chromosomes coil, become more condensed and visible. Further prophase events are the splitting of the centrosome into two and the movement of each half to opposite sides of the nucleus, the disappearance of the nucleolus, and the beginning of the breakdown of the nuclear membrane.
3. metaphase: In this phase each condensed chromosome attaches at a point along its length called the centromere to a double-poled spindle-shaped structure, the spindle. Chromosomes whose centromeres are approximately midway between each end are described as metacentric (Fig. 1). Chromosomes with more terminally placed centromeres are called subtelocentric, whilst when the centromeres appear to be at their very tips the chromosomes are called telocentric. At this stage each chromosome is in a duplicated condition as a result of the previous synthesis, but a single centromere is functioning as the attachment point of both sister chromatids to the spindle fibres.
4. anaphase: This phase is the shortest of the mitotic phases, and is the point at which the centromeres of each pair of chromatids begin to function as double structures. Each centromere with its attached chromatid separates from its sister centromere and chromatid, and moves towards the opposite pole of the spindle.
Fig. 1 The most common types of chromosomes
5. telophase: During the telephase period each of the two groups of chromatids (or daughter chromosomes) are clustered at the poles of the spindle. The nuclear membrane is reestablished, the nucleoli reformed, and the two new cells go into interphase.
Meiosis: As a rule, meiotic divisions follow a standard pattern during which two successive divisions of the chromosome occur. One of these divisions is a reduction division, in which members of homologous pairs of chromosomes are separated into half daughter cells without duplication; i.e., their numbers are reduced to half. Although variation in this general pattern have been discovered, one of the basic attributes of meiosis is the initial pairing and subsequent separation of homologous chromosome pairs. The initial pairing occurs during the fairly lengthy first meiotic prophase, which has been subdivided into the following five stages: Leptotene, Zygotene, Pachytene, Diplotene, Diakinesis. This phase is then followed by the metaphase, anaphase and telophase of the first meiotic division. Meiosis is completed by the second meiotic division. This consists of the same four stages of division as mitosis: prophase, metaphase, anaphase and telophase; and results in four daughter cells which stay in interphase.
The structure of chromosomes, and variation
in chromosome size and number
Structure and composition of the chromosome
Each chromosome contains two identical parallel structures: the chromatids. Each chromatid in its turn consists of one or several thin filaments, the chromonemata or genonemata, containing characteristic condensed stainable regions called chromomeres. The chromonema represents a very long double filament helically wound in a dividing cell and despiralled to a large extent in the nuclei of non-dividing “resting” cells during the interphase. In resting cells the long thin chromonemata fill the nucleus completely, forming a random network.
Genes, the elementary units of heredity, are located throughout the whole length of the chromosomes. In fishes each chromosome is known to contain hundreds, perhaps thousands of genes. The basis of the chromonema is DNA. The structure of this substance plays a crucial role in biological processes. The DNA molecule takes the form of a double helix. In each of the two strands which together form the double helix, molecules of phosphoric acid and deoxyribose are repeated in a linear order, many thousand times. To each deoxyribose section, one of the following four nitrogen bases is covalently linked: adenine or guanine (purines), cytosine or thymine (pyrimidines). Bases located opposite each other in two adjacent DNA strands are associated by hydrogen bonds. The configuration of the bases is such that adenine may be bound only with thymine, and guanine only with cytosine.
The chromosome number of fishes is very variable. Primitive, less specialized fishes generally have a higher chromosome number than more developed species, and also more metacentric chromosomes. It is also known that Robertsonian evolution is a wide-spread phenomenon among fishes. Today hundreds of types of polyploid fish (both natural and induced) are known, though many of them are not viable. Generally if the fish has an odd number of chromosome sets (haploid, pentaploid, etc.) it is not viable. The triploid fishes are exceptions to this, being viable but sterile. In some cases polyploidy is very useful, e.g., in common carp (Cyprinus carpio), which is a naturally tetraploid species. Generally the tetraploids are viable and can reproduce, while hexaploids are viable but sterile. In fishes chromosomal polyploidy often occurs (e.g., in common carp and trout) and is possibly the result of translocations. Reduction in chromosome number through centric fusion is also known. Besides these major changes, minor changes like deletion and/or duplication of a specific locus or loci by unequal sister chromatid crossovers also occur. Unequal meiotic division will also result in minor chromosome changes. Major changes can be identified by karyotype analyses, and minor ones by studying the fine structure of the homologous chromosome pairs.
Changes in chromosome structure
Estimation of recombination frequency: It has been pointed out by Nagy and Csányi (1982) that gynogenesis provides an easy direct measure of recombination frequency and inbreeding level (see the discussion of gynogenesis and inbreeding in the section on “Genetic Manipulation” below). The probability for a gene of recombination frequency r to be heterozygous in the ith gynogenetic generation is r2, while the coefficient of inbreeding is Fi = l-r1.
The degree of uniformity or the degree of genotypic identity - the probability that two individuals selected from the ith gynogenetic generation are of the same genotype - either homozygous or heterozygous with respect to the gene in question is:
I(i) = 1-frac12; ri-1 - ri + 3/2 ri+1
Distances between market loci and centromeres can be calculated according to the mapping functions derived for tetrad analyses (Barrat et al., 1954).
The frequency of heterozygous progeny (r) is a direct measure of the recombination frequency (crossing over), because the diploid gynogenetic progeny has a set of chromosomes derived from sister chromatids of half of a meiotic tetrad, and heterozygous progeny will develop from a zygote containing sister chromatids with an odd number of exchanges between the centromere and the particular locus, if an F1 heterozygous female was used for gynogenesis.
The main principles of Mendelian inheritance
In 1866, Mendel formulated the main principles of the transmission of hereditary traits to the offspring. These principles are:
The uniformity of the first generation of hybrids (F1).
The segregation of traits.
Gamete purity.
The independent assortment of different traits in the offspring.
The Mendelian rules are based on the specificity of chromosomal behaviour during meiosis and fertilization. The most important part is played by the rule of gamete purity. Chromosomally it is based on the presence of just one of the two homologous chromosomes in the nucleus of each gamete. Each gene comes to the gamete together with the chromosome in which it is located. It is not affected by the genes located in another homologue. The different forms of one and the same gene are called alleles. If two different alleles of one gene (locus) are present in the hybrid in two homologous chromosomes, only one of them can come to each gamete after meiosis. After the fusion of gametes during fertilization, chromosomes with different alleles are combined according to the laws of statistics. In the absence of complicating circumstances four types of zygotes are formed in approximately equal numbers.
The inheritance of many traits in plants and animals follows the law of dominance. That means one of the alleles in the hybrid zygote is expressed more strongly than the other and suppresses the weaker allele in the course of development. In these cases in the second generation of hybrids can be observed the classical Mendelian ratio 3:1 [(25%AA+25%Aa+25%aA):(25%aa)]. When the dominance is complete the AA homozygotes are indistinguishable from Aa or aA heterozygotes. If AA homozygotes differ from Aa or aA heterozygotes the second generation of hybrids can be differentiated into three groups [AA (25%) Aa and aA (50%) and aa (25%)]. This proportion represents a more common universal consequence of the principle of gamete purity, because it directly reflects the main feature of chromosomal assortment in meiosis: the random segregation of the chromosomes of each pair in the course of meiosis and random fusion of gametes possessing different sets of allelic genes.
The Mendelian principles are fully applicable to fish. The hereditary variation in fishes is as diverse as in other animals. The following four groups of hereditary differences can be distinguished (Kirpichnikov, 1972).
1. Quantitative morpho-anatomical traits of alternative type inherited according to the laws of Mendel and yielding distinct and clear segregation in the offspring after crosses of individuals differing in these traits.
2. Quantitative differences with respect to various morphological and physiological traits having a polygenic inheritance. The expression of such traits not only depends on genetic factors but also on many variable environmental factors. Typical quantitative traits showing variation in most fish species include the number of vertebrae and fin rays, the number of scales and gill rakers, morphological features such as the length to weight ratio of the body, growing capacity, oxygen consumption, resistance to high and low temperature, age at maturity, food conversion, etc.
3. Biochemical differences expressed as a variation between blood groups, or the presence of several forms of one and the same protein synthesized under the control of different genes or different alleles of one gene. The presence of multiple forms of proteins, isozymes and isoforms, has been established for practically all genes coding for enzymes. Allelic differences in proteins are inherited in the same way as differences in the qualitative traits, in accordance with Mendelian laws.
4. Phenodeviants. This concept refers to malformations and other aberrations, which are generally poorly manifested and inherited in a complex fashion. Inbreeding and unfavourable environmental conditions contribute to the appearance of phenodeviants and stimulate the manifestation of such traits.
Quantitative inheritance
Normally there are no two completely identical individuals in
a fish population. The differences appear in biochemical, physiological
and morphological features, and often they are of a
quantitative nature. The origin of this variation is two-fold:
1. as a consequence of changes in the genotype of the fish, or
2. fluctuations caused by environmental factors.
The variation of many meristic or discontinuous quantitative traits in fishes follows the law of binomial distribution. The variation in number of elements may be described by the binomial coefficient of expansion (a+b)n.
The normal distribution described graphically by the symmetrical variation curve (Gaussian distribution) may be characterized by two easily calculated constants:
The mean (
) and the standard deviation (σ) or SD.
Fig. 2 The binomial distribution
Another way of expressing the variation is the variance (v or σ2), defined as the square of the standard deviation. In order to abolish dimensional values the standard deviation (SD) is expressed as a percentage of the mean (coefficient of variation or CV).
The variation in such quantitative traits as body weight, length, exterior indices, the size of different organs, the haemoglobin content in the blood, oxygen deficiency tolerance, rate of oxygen consumption and many others are continuous. It is therefore convenient to divide the range of variation into equal intervals corresponding to different classes. The calculation of biometric constants is then made on the basis of the mean value of a trait for one and the same class.
The variation in body weight frequently deviates from the normal distribution. The variation curve becomes irregular, multimodal or asymmetrical. The multimodality may result from the intermixing of different groups of fishes or from genetic segregation of a few genes affecting the growth rate. Negative asymmetry is also quite common, due to the strong dependence of the growth rate on the initial weight of the individual. Positive asymmetry is usually associated with malformations.
Because the whole population is not normally studied, but only a sample of it, the mean values and indices describing the variation are only approximations. The mean standard errors of these values are the following:
The magnitude of these errors is used to determine the range
of the real mean or the real variation of the indices. This
range, defined by confidence limits, may be established to a
certain, greater or lesser degree of probability. The addition
or the subtraction of two standard errors ( ± 2SE or σ ± 2SE σ)
corresponds to the determination of confidence limits with a
probability of p=0.95. The subtraction or addition of 3 standard
errors increases the probability to p=0.997.
The variance and its components
The variance (σ2) is a very important and useful tool in estimation of variation in a population. The variance generally refers to the total variance, or phenotypic variance (σ2 PH). The phenotypic variance has two main components: the genotypic (σ2G) and the environmental variance (σ2E).
σ2PH = σ2G + σ2E + 2rσaσE
where σPH2 and σ2 and σE2 correspond to the total phenotypic, genotypic and environmental variance respectively, while r is the correlation coefficient between the genotypic and environmental variation. The genotypic variance consists of two main components: the additive (σA2) and the non-additive (σN2) genetic variance.
Heritability
The heritability is defined as the hereditary fraction of the total variation. In a broad sense, the heritability (hG2) is equal to the ratio of the genotypic and phenotypic variance
For the fish breeder, it is more useful to define the fraction of the additive genetic variation, that is of the variation due to genes with a simple additive effect
Realized heritability (hR2)
If the difference between parental pairs with respect to some trait is S (selection differential) and the corresponding difference between their offspring is R (selection response), the R:S ratio will show what fraction of the difference between the parents is retained in the offspring: hR=R/S.
Determination of heritability from regression between parents and offspring
To determine heritability from regression, the calculation of the regression coefficient is required, i.e., the level of variation of the trait in the offspring per unit of change in the parents. Generally a linear regression can be used: y=a+bx, where y = the mean value of the trait for the parents, x = the mean value of the trait for the offspring, a = is a constant reflecting the overall differences in the expression of the trait between generations, and b = the regression coefficient. If the mean values of a trait characteristic of the parents are used, then heritability becomes equal to the regression coefficient (hr2 = b), but if the offspring is compared with only one of the parents the formula is hr2 = 2b. The multiplication by two is necessary because the offspring receives only half of its genes from the parent under consideration.
Selective breeding and intraspecific hybridization
There is a long tradition of improving production traits in farmed animals through selective breeding and/or hybridization. However, in aquaculture, even where culture traditions go back thousands of years, selective breeding has only rarely been practised. There are several reasons for this. For example, for many species artificial propagation and feeding of larvae have not been possible (e.g., for eel, yellowtail, pike). In many other species induced spawning is carried out in systems where several spawners are kept together, making the full control of matings impossible (e.g., cod, Indian carps). In addition, the introduction of a breeding system which includes more than phenotypic selection requires a system for tagging individuals, which is not available for larvae. However, fish have a very high reproduction capacity and potential for improvement through selective breeding. A good knowledge of genetic parameters is a pre-requisite for designing a good breeding scheme. The first step in formulating a breeding scheme is to define the breeding goals.
Breeding goals
Each trait included in a breeding goal should be defined precisely. Only those traits which can be accurately measured or judged directly or indirectly can be included. The traits must show genetic variance, and should be of economic importance. The breeding goals should be formulated by agreement with consumers and the processing and marketing industry. Most cultured fish species have long generation intervals, and therefore genetic improvements take time. Because of this only those traits which do not lose their value with time are suitable for inclusion in breeding goals (Refstie, 1987).
The following traits are of high economic importance and hence in most production systems they are included in the breeding goals:
Growth rate: One of the easiest traits to measure is increase in body weight or in body length over a certain period. In salmonids the genetic correlation between body weight and body length is very high (Refstie, 1980; Gjedrem, 1982). For this reason Gjedrem (1982) suggested that the trait with the highest heritability should be measured. There is a high genetic correlation between growth rate and food conversion, so that improvement in growth rate results in a rapid correlated improvement in food conversion (Gjedrem, 1982).
Food conversion: This is one of the most important traits, but very difficult to measure. Measurements on individual fish are impossible, and recordings from groups of fish are also laborious. However, due to the high percentage of total production cost attributable to food, conversion efficiency is becoming more and more important. It must also be taken into account that the food conversion rate may change if the food is changed, and the rank order of families for conversion performance can be different when a different when a different quality of food is given. On the other hand Edwards et al. (1977) and Refstie and Austreng (1981) showed no significant family:diet interaction in rainbow trout kept on feed of various carbohydrate levels. This indicated that selection for the utilization of cheaper, high carbohydrate food would not give worthwhile results. It was, however, concluded that net food conversion rate in young rainbow trout is heritable (h2=0.31+0.11) with a low coefficient of variation, CV = 6% (Gjedrem, 1982). Further refinement of methods to test this trait (by direct or indirect measurement) should be given high priority.
Mortality: This has a different economic importance as a trait according to whether it occurs early in the life cycle or close to harvesting. Because of the high reproductive capacity of fishes, the trait has no great economic value during the fish's early stages of development, while it is very important during the growing period. The trait is usually closely correlated with disease resistance, but because of the variety of diseases mortality is a complex trait. Measuring resistance to specific diseases may lead to improvements. Because there is a high correlation between stress conditions and susceptibility to diseases, a selection for better tolerance to stress may also result in improvement in survival rate.
Meat quality: Although the meat quality is an important characteristic, it is not recommended for inclusion in the breeding goals. This is because for many of the traits which determine meat quality (colour, taste, shape, meatiness) it is very difficult to make objective measurements. The dressing percentage generally shows low genetic variance.
Fecundity: While for other farmed animals fecundity is a very important trait, due to their very high reproductive rate the trait usually has low economic importance in fish.
Table 1
TRAITS AND THEIR RELATIVE ECONOMIC IMPORTANCE FOR
SOME SPECIES USED IN AQUACULTURE (Gjedrem, 1982)
| Traits | Salmonids | Warmwater fish | Bivalves | Prawns |
| Growth rate | +++ | +++ | +++ | +++ |
| Meat quality | ++ | + | + | ++ |
| Mortality early | + | + | + | + |
| Mortality late | ++ | ++ | ++ | ++ |
| Age at maturation | ++ | ? | ? | ? |
| Fecundity | 0 | 0 | 0 | 0 |
Genetic variation in quantitative traits
The additive genetic variance (σA2) has great importance when a selective breeding programme is planned. The magnitude of the additive genetic variance determines the success of (or response to) selection. It can be calculated by multiplying the heritability (h2) by the phenotypic variance (σPh2).
The difference between heritabilities estimated from dam and
sire components can be considered as an estimate of non-additive
genetic variance and possibly maternal effect. The magnitude of
variance is also a useful parameter. The coefficient of
variation (CV) expresses the ratio of phenotypic standard
deviation to the mean of the trait in question 
and gives information about the magnitude of variance.
Tables 2, 3 and 4 show that the heritability for body weight of young fish (juveniles) is rather low, while that for adults is moderate, promising a good response to selection. The meat quality, meatiness, meat colour, and dressing percentage all show low genetic variation, with the exception of the fat content of the body. The age at maturation shows moderate heritability in rainbow trout, and high heritability in Atlantic salmon. Mortality shows low heritability, but resistance to certain diseases shows high heritability.
Table 2
PHENOTYPIC AND GENETIC PARAMETERS FOR QUANTITATIVE TRAITS
(CV = coefficient of variation, x = mean,
σ = standard deviation, h2 = heritability)
| Species | Traits | x | σ | h2 | CV |
| Rainbow trout | 150 days weight (g) | 13.6 | 4.12 | 0.09±0.1 | 30 |
| 280 days weight (g) | 13.6 | 4.37 | 0.29±0.2 | 33 | |
| dead eyed eggs | - | - | 0.15±0.04 | - | |
| dead alevins | - | - | 0.13±0.03 | - | |
| growth rate (%/day) | 1.63 | 0.31 | 0.26±0.12 | 19 | |
| food consumed (%/day) | 1.37 | 0.27 | 0.41±0.13 | 20 | |
| net food conversion | 1.46 | 0.08 | 0.31±0.11 | 6 | |
| fat % | 9.71 | 1.00 | 0.47±0.34 | 10 | |
| O2 consumption | 3.73 | 0.68 | 0.51±0.12 | 18 | |
| Atlantic salmon | 180 days weight | - | - | 0.15±1.00 | - |
| tolerance to vibriosis | - | - | 0.12±0.05 | - | |
| age at maturity | - | - | 0.42±0.08 | - | |
| dressing % | 90 | 3.1 | 0.03±0.02 | 4 | |
| flesh colour score | 3.6 | 0.6 | 0.01±0.03 | 16 | |
| Pacific salmon | tolerance to IHN | - | - | 0.32 | - |
Table 3
AVERAGE VALUES FOR COEFFICIENT OF VARIATION (CV)
AND h2 BASED ON SIRE COMPONENT
| Economically important traits | Rainbow trout | Atlantic salmon | Common carp | Tilapia | Prawns | |||||
| CV | h2 | CV | h2 | CV | h2 | CV | h2 | CV | h2 | |
| Body weights | ||||||||||
juveniles | 33 | 0.12 | 78 | 0.08 | - | 0.15 | 26 | 0.04 | - | 0.12 |
adults | 22 | 0.17 | 27 | 0.36 | 22 | 0.36 | - | - | - | - |
| Body length | ||||||||||
juveniles | 14 | 0.24 | 23 | 0.14 | - | - | 8 | 0.06 | - | - |
adults | 9 | 0.17 | 8 | 0.41 | - | - | - | - | - | - |
| Mortality resistance | - | 0.14 | - | 0.11 | 28 | - | - | - | - | - |
| Meatiness | 20 | 0.14 | 19 | 0.16 | - | - | - | - | - | - |
| Meat colour | 23 | 0.06 | 16 | 0.01 | - | - | - | - | - | - |
| Fat % | 10 | 0.47 | - | - | - | 0.14 | - | - | - | - |
| Dressing % | 6 | 0.01 | 4 | 0.03 | - | - | - | - | - | - |
| Age at Maturation | - | 0.18 | - | 0.71 | - | - | - | - | - | - |
Table 4
INHERITANCE OF VIABILITY AND RESISTANCE IN FISH
(after Kirpichnikov, 1972)
| Species | Trait | Age group of fishes | Heritability, h2 | |
| Method of calculation | Mean values | |||
| Atlantic salmon | Viability | Eggs larvae | Dispersion analysis | 0.01±0.15 |
| Viability | Fry | Dispersion analysis | 0.11±0.34 | |
| Viability | Fry | Interfamiliar variance | 0.10±0.20 | |
| Resistance to vibrio disease | Fry | Dispersion analysis | 0.07±0.15 | |
| Brown trout | Viability | Eggs larvae | Dispersion analysis | 0.01±0.05 |
| Resistance to acid waters | Eyed eggs | Dispersion analysis | 0.09±0.27 | |
| Rainbow trout | Viability | Eggs larvae | Dispersion analysis | 0.06±0.14 |
| Splake (hybrids Salvelinus fontinalis x S. namaycush) | Viability | Fry | Dispersion analysis | 0.06±0.41 |
| Resistance to high temperature | Fry | Dispersion analysis | 0.38 | |
| Resistance to blue sack disease | Fry | Dispersion analysis | 0.41±0.60 | |
| Chinook salmon (Oncorhynchus tschawyscha) | Resistance to gas bubble disease | Fry | Dispersion | 0.04 |
| Sockeye salmon (Oncorhynchus nerka) | Resistance to infectious necrosis | Fry | Dispersion | 0.27±0.38 |
Methods of selection
Selection methods
After the components of variance have been examined and a
relatively good ratio of additive genetic variance has been
found, a selection programme can be worked out to achieve the
maximum possible genetic gain (
G).
The magnitude of genetic gain attainable when selection is applied is primarily dependent on four parameters (Falconer, 1960):
where G = genetic gain; i = selection differential; σ =
phenotypic deviation; σG = genetic standard deviation, and σPL =
generation interval.
1. G = Genetic gain (Fig. 3)
2. i = selection differential in standard deviation units. This shows by how many SD units the average of the brood stocks is higher than the population average. Due to the high fertility of trout and salmon, the selection differential can be very high. For example if only ≃ 0.1% of the population is chosen as brood stock, i = 3, but even in a relatively small population where the selected fish represent 1%, i is still equal to 2.66. In large populations an intensive selection can easily be carried out, but in small populations inbreeding may limit the intensity of selection possible in practice.
Fig. 3 Genetic gain
3. h2 = heritability. The heritability is the ratio of genetic variance to the total phenotypic variance. The level of heritability is very important, and determines the selection method to be used. If the heritability is high (higher than h2 = 0.5), an individual selection should be applied, whilst if lower than that family selection will be more effective (Fig. 4).
Fig. 4 Relative merit of fullsib family; (F) selection compared with individual; (I) selection (Falconer, 1964)
In salmonids the heritability varies greatly between different traits. Aulstad et al. (1972) estimated the inheritance growth rate in rainbow trout fingerlings as h2 = 0.09 ± 0.32, whilst for resistance to vibrio disease Gjedrem and Aulstad (1974) found h2 = 0.07 ± 0.1. For Atlantic salmon Gjerde and Gjedrem (1984) reported relatively high heritability for weight at 3.5 years of age (h2 = 0.54 ± 0.08) but the heritability of the dressing percentage was insignificant (h2 = 0.02 ± 0.02).
If family selection is applied the family size is decisive (Fig. 5).
Fig. 5 Response expected under family selection relative to that for individual selection, plotted against family size (Falconer, 1964)
From the above statements, the following general conclusions can be drawn:
If the heritability is high selection will be very effective, but no improvements can be achieved by selection if h2 = 0 or close to it.
If h2 is lower than about 0.5, family selection should be applied, but when higher individual selection is more efficient.
A combination of family and individual selection increases the efficiency.
The effectiveness of family selection can also be markedly increased by increasing the family size, especially if the heritability is lower than 0.4 (Gjedrem, 1982).
4. Variance
The variance is very important when genetic changes in a population are planned. It is convenient to consider variance as a coefficient of variation.
The variance in many traits is usually higher for fish than for other farmed animals
| (e.g., | Salmon body weight | CV | = | 102 | |
| Sea trout body weight | CV | = | 51 | ||
| Rainbow trout body weight | CV | = | 70 | ; Gjedrem, 1975). |
5. Generation interval
The generation interval is species-dependent. In rainbow trout it is generally 3 years, while in salmon a generation interval of 4 years is more common.
Selection indexes
In most cases selection must be done for several economic traits, not only one. Selection can be done in a more efficient way simultaneously for all the traits when based on a selection index.
Selection methods
Given the above considerations, the following general conclusions can be drawn:
Mass (or individual) selection is the easiest method to use. It can be applied if the heritability is high and the relationship between the animals is unknown. In salmonids this method can successfully be used only for growth rate.
For all the other traits (viability, carcass quality, disease resistance, etc.) other selection methods should be applied.
Progeny testing in salmonids is very time-consuming and greatly increases the generation interval, resulting in lower genetic gain.
Family selection based on full- and half-sib testing does not have these disadvantages (Gjedrem, 1976).
Gjedrem (1976) therefore suggested a breeding programme for salmonids which combines phenotypic and family selection.
Interspecific hybridization of salmonids
Many viable hybrids have been produced between different salmonid species, and interspecific hybridization has a long tradition. Recently Suzuki and Fukuda (1971) and Refstie and Gjedrem (1975) published reports on hybridization experiments with salmonids. Viable offspring were obtained following interspecific hybridization carried out by Refstie and Gjedrem (1975), and Purdom et al. (1976).
Table 5
MEAN WEIGHTS OF SALMONID HYBRIDS AT 11 MONTHS OF AGE (g)
(Refstie and Gjedrem, 1975)
| Male parent | Female parent | |||
| Char | Brown trout | Sea trout | Salmon | |
| Char | 55.2 | 58.2 | - | 96.5 |
| Brown trout | 73.3 | 41.8 | - | 7.7 |
| Sea trout | 58.3 | 24.9 | 31.8 | 6.1 |
| Salmon | 70.7 | 7.3 | 8.8 | 30.0 |
Refstie and Gjedrem (1975) reported considerable heterosis in the 11-month weight of some salmonid species crosses (Table 5). However, although heterosis was found in the 11-month weight of the char x salmon crosses, Kinghorn (1982) reported that by the time of slaughter pure-bred Atlantic salmon had attained higher weights than the hybrids.
When artificial interspecific hybridization is planned, it is very important to pay careful attention to the maintenance of the pure species. Negative heterosis has also been found in hatchability and survival of interspecific hybrids between salmonid species (see the review by Chevassus, 1979). In fact it is a general conclusion that most important traits cannot be improved by interspecific hybridization in salmonids.
Some of the traits studied have shown partial dominance. For example, Ord et al. (1976) reported dominance in VHS resistance in a cross between the susceptible rainbow trout and resistant coho salmon. Marketing can also be a major problem with interspecific hybrids.
There can however, be value in interspecific hybridization for utilization of non-heterotic effects (e.g., production of monosex fish or sterile fish (Krasznai, 1987).
Genetic manipulation
Genetic manipulations are usually carried out during the meiotic or mitotic division of the eggs. During normal fertilization, the sperm activates the second meiotic division of the egg, the second polar body is extruded, and a haploid female pronucleus is formed. The male and female pronuclei then fuse to form the first diploid nucleus of the embryo, which is then multiplied by a long series of mitoses.
There are basically two different types of genetic manipulations. In the first type only one of the parents' genomes will take part in the production of the progeny, while in the second type the genomes of both parents contribute to the development of the offspring.
Chourrout (1986) reviewed the methods of genetic manipulation. He listed twelve different techniques which can be successfully carried out in fishes, as follows:
I.1 Self-fertilization. This first requires the production of induced hermaphrodite fish.
I.2 Induced gynogenesis in practice involves two parents, but only the female genome forms the embryo. The genetic material of the sperm can be destroyed by various treatments. Diploid embryos are produced by inhibiting the first mitosis.
I.3 Induced androgenesis also involves two parents, but only the male genome forms the embryo. The genetic material of the egg is destroyed prior to fertilization. The diploid stage can again be achieved only by the suppression of the first mitosis.
II.1 Triploids result from the fusion of a diploidized female nucleus and a male nucleus of the same fish species. Interspecific hybridization can also result in the production of triploids by retention of the second polar body.
II.2 Tetraploids result from the first diploid nucleus of the embryo when the first mitosis is suppressed.
II.3 New triploids can be obtained by mating tetraploids and diploids.
II.4 New tetraploids are formed by mating tetraploids.
II.5 Pentaploids, hexaploids, and tetraploid gynogenetics result when gametes are provided by tetraploids and anti-meiotic or antimitotic treatments are applied.
II.6 In new gynogenetics and androgenetics, genomes come from the female diploid gametic set or from the male diploid one respectively.
II.7 Nuclear transplantation. Both gametic parental contributions are eliminated and replaced by an egg nucleus collected from a developing embryo.
II.8 Gene transfer. A foreign gene can be added to the genome of the species in active chromosome fragments.
II.9 Gene transfer can also be done by addition of a foreign gene to the genome of the species in plasmids.
Induced self-fertilization. There have been several attempts to induce self-fertilization in fish, but success has been achieved only once in rainbow trout (Chevassus et al. 1986). Jalabert et al. (1975) produced hermaphrodites by dietary administration of various oestrogens to young fingerlings. By this method YY males can be produced. The difficulty of producing hermaphrodites prevents the widespread use of the method.
Induced gynogenesis. Induced gynogenesis requires the destruction of the genetic material of the sperm. This is usually achieved by radiation or treatment with chemical mutagens.
The Hertwig effect. Hertwig (1911) found that subjecting the sperm to low doses of radiation prior to fertilization caused massive mortality of the embryos, but when higher doses were used some embryos survived until hatch. The explanation of the Hertwig effect is that low doses do not totally destroy the sperm genome, and the remaining portion makes a toxic contribution to the embryos. On the other hand high doses cause complete inactivation of the sperm and permit the development of haploid embryos.
Since that time many authors have used radiation (130 krad UV; X-ray; Y-ray), to inactivate the sperm (in salmonids: e.g., Purdom, 1969; Chourrout, 1980). In rainbow trout the chemical mutagen dimethylsulphate applied for 60–120 min caused Hertwig effect (Tsoi, 1969).
The destruction of the female genome and haploid gynogenesis. Induced androgenesis has been less studied than gynogenesis. The Hertwig effect has been described in salmonids by Parson and Thorgaard (1984). They used gamma irradiation of the egg (130 krad).
The suppression of the second meiotic division
Three types of treatments proved to be satisfactory: long cold shock, short heat shock, and short exposure to hydrostatic pressure.
Cold shock. In warmwater fish species this treatment works well. It has also been tried in coldwater species (Chourrout 1980) but the method has not been fully developed.
Short heat shock. Chourrout and Quillet (1982) determined the optimum temperature, duration and onset of short heat shock for trout as follows: duration 20 min at 26°C and 10 min at 29°C, onset: 0–30 min after fertilization.
However, shorter shocks at higher temperature can also be effective in rainbow trout. Such treatments were also successfully applied to Atlantic and Pacific salmon.
Pressure shock (7 000–10 000 PSI) suppresses the second meiotic division in many amphibian and fish species.
Several trials have been done with chemicals such as Cytochalasin-B or colchicine, but without really successful results.
The suppression of the first mitosis by heat
and pressure shocks
Heat and pressure shocks
In zebrafish the first mitosis was successfully suppressed by late pressure shock, resulting in about 20% survival. In experiments conducted in 1981 and 1982 multiple late heat shocks were tested on rainbow trout eggs, but resulted in a mixture of tetraploids, diplo-tetraploids, and diploids. Two more recent independent studies on salmonids showed that pressure shocks applied between 5 h 45 min and 5 h 50 min after fertilization were much more efficient at suppressing first mitosis. The optimal application time was determined by Chourrout (1984) as 5 h 20 min to 5 h 30 min at 10°C for rainbow trout. The technique was then successfully applied to chum and masou salmon.
Parson and Thorgaard (1985) succeeded in doubling the chromosome set of haploid androgenetic rainbow trout with pressure shocks (7 500–8 600 PSI lasting 6 or 2–3 min respectively, administered 5 h 20 min after fertilization.
Nuclear transplantation
Nuclear transplantation was first done on amphibians, and later success was achieved with Crucian carp in China. The yield of viable larvae is about 1%. The following procedure can be used:
Activate the host egg by pricking or by electric shock.
Destroy the genetic material in the egg by UV radiation.
Collect a diploid nucleus from an embryo (blastula stage) and inject it into the host egg.
Follow normal hatching techniques.
In contrast to amphibian studies, this type of work is not well documented for fish. The prospects for fish cloning by this method are not promising; and the reproduction of gynogenetic or androgenetic individuals seems much more appropriate.
Transmission of chromosome fragments after
incomplete sperm inactivation
Thorgaard et al. (1985) found some of the chromosome fragments remained active after irradiation with 50 krad.
In 1982 Chourrout and Quillet showed that fragments may be transmitted in irradiated coho salmon sperm into rainbow trout gynogenetics.
The goal of these transmissions was to transfer the resistance to IHN and VHS virus from coho salmon into rainbow trout. However, the efficiency of the method depends on several factors or assumptions:
The expressed fragment must be stable.
The genetic factors determining the disease resistance of coho salmon must be grouped together in the genome, and be situated not too far from the chromosomal centromere.
Transfer of cloned genes
The transfer of cloned genes has been described with the embryos of mice (Gordon and Ruddle, 1985). It was achieved by the microinjection of a thousand copies to the male pronucleus.
Later, the technique was applied to fish. Chourrout (1987) injected twenty million copies of a plasmid containing the human growth hormone into the rainbow trout egg, then after hatching detected the foreign genes in the hybrids by electrophoresis.
The method appears simple; it consists of two steps:
A small opening is drilled in the chorion manually and a pipette containing the plasmid is driven through it into the ooplasm. The hole closes itself again after injection.
The eggs are then incubated in the normal way. A hatching rate of about 75% can be anticipated.
However, it is very hard to evaluate this method, because the presence of the gene can only be detected by electrophoresis. This cannot provide information confirming the proper function of the transmitted genes.
From the theories and methods of genetic manipulation discussed above, the following conclusion can be drawn:
The efficient inhibition of the first mitosis, recently demonstrated in salmonids, gives rise to new genotypes (homozygous diploids and tetraploids) which may be very promising for the genetic improvement of the species. The fertility of the tetraploid adults in rainbow trout allows a considerable widening of the field of chromosome set manipulation.
Finally, much effort is likely to be put into gene transfer in the future, since fish are excellent subjects for such investigations.
References
Barrat, R.W., 1954 Neymeyer, D., Perkins, D.D. and L. Gajujobst. Map construction in Neurospora crassa. Adv.Gen., 6: 1–93
Chevassus, B.1979 Hybridization in salmonids: results and perspectives. Aquaculture, 17:315–20
Chevassus, B., Devaux, A., Chourrout, D. and B. Jalabert. 1986 Production of YY rainbow trout males by self-fertilization of induced hermaphrodites. Submitted for publication in Journal of Heredity
Chourrout, D. 1980 Thermal induction of diploid gynogenesis and triploidy in the eggs of the rainbow trout (Salmo gairdneri Richardson). Reprod.Nutr.Develop., 20: 727–33
Chourrout, 1984 D. Pressure induced retention of second polarbody and suppression of first cleavage in rainbow trout: production of all triploids, all tetraphloid and heterozygous and homozygous diploid gynogenetics. Aquaculture, 36:111–26
Chourrout, 1987 D. Genetic manipulation in fish: review of methods. In: Proc. World Symp. on Selection, Hybridization, and Genetic Engineering in Aquaculture. Bordeaux, 27–30 May 1986. Vol. II. Berlin, 1–26 p.
Chourrout, 1982 D. and E. Quillet. Induced gynogenesis in the rainbow trout: sex and survival of progenies production of all triploid populations. Theor.Appl.Genet., 63:201–5
Edwards, 1977 D.J., Austreng, E., Risa, S. and T. Gjedrem. Carbohydrate in rainbow trout diets. I. Growth of fish of different families fed diets containing different amounts of carbohydrate. Aquaculture, 11:31–8
Falconer, 1964 D.S. Introduction to quantitative genetics. Oliver and Boyd, Edinburgh and London, 365 p.
Gjedrem, 1975 T. Possibilities for genetic gain in salmonids. Aquaculture, 6:23–9
Gjedrem, 1976 T. Possibilities for genetic improvement in salmonids. J.Fish.Res.Board.Can., 33:1094–9
Gjedrem, T. 1982 Genetic variation in quantitative traits and selective breeding in fish and shellfish. Aquaculture, 33:51–72
Gjedrem, 1974 T. and D. Aulstad. Selection experiments with salmon. Differences in resistance to vibrio disease of salmon parr (Salmo salar). Aquaculture, 3:51–59
Gjerde, B. and T. Gjendrem. 1984 Estimates of phenotypic and genetic parameters for carcass traits in Atlantic salmon and rainbow trout. Aquaculture, 36:97–110
Gordon, J.W. and F.H. Ruddle. 1985 DNA-mediated genetic transformation of mouse embryos and bone marrow - A review. Gene, 33: 121–36
Hertwig. O. 1911 Die Radiumkrankheit tierrischer Kiemsellen. Arch.Mikrosk.Anat.Entwick., 77:1–97
Jalabert, 1975 B., Billard, R. and B. Chevassus. Preliminary experiments on sex control in trout: production of sterile fishes and simultaneous self-fertilizable hermaphrodites. Ann.Biol.Anim.Bioch.Biophys., 15: 19–28
Kinghorn, 1982 B.P. Genetic effect in crossbreeding. II. Multibred selection indices. Z.tierz.ZUechtungsbiol., 99:315–20
Kirpichnikov, 1972 V.S. In: Genetics selection and hybridization of fish (ed: Cherfas, B.I.) Acad.Sci.USSR, Israel Progr. for Sci. Translation, Jerusalem, 269 p.
Krasznai, 1987 Z.L. Interspecific hybridization of warmwater finfish. In: Proc. World Symp. on Selection, Hybridization and Genetic Engineering in Aquaculture. Bordeaux, 27–30 May 1986. Vol. II. Berlin, 36–45 p.
Nagy, 1982 A. and V. Csányi. Changes of genetic parameters in successive gynogenetic generations and some calculations for carp gynogenesis. Theor.Appl. Gen., 63: 105–10
Ord, 1976 W.M., Le Besse, M. and P. De Kinkelin. Viral hemorragic septicemia: comparative susceptibility of rainbow trout (Salmo gairdneri and hybrids Salmo gairdneri, Oncorhynchus kisutch) to experimental infection. J.Fish.Res.Board.Can., 33: 1205–8
Parson, 1984 J.E. and G. Thorgaard. Induced androgenesis in rainbow trout. J.Exp.Zool., 231:407–12
Parsons, 1985 J.E. and G. Thorgaard. Production of androgenetic rainbow trout. J. Hered., 76:177–81
Purdom, 1969 C.E. Radiation-induced gynogenesis and androgenesis in fish. Heredity, 24:431–44
Purdom, 1976 C.E., Thompson, D. and P.R. Dando. Genetic analyses of enzymes polymorphisms in plaice (Pleuronectes platessa). Heredity, 37: 193–206
Refstie, 1980 T. Genetic and environmental sources of variation in body weight and length of rainbow trout fingerlings. Aquaculture, 19:351–72
Refstie, T. 1987 Selective breeding and intraspecific hybridization of cold water finfish. In: Proc. World Symp. on selection, hybridization, and genetic engineering in aquaculture, Bordeaux, 27–30 May, 1986. Vol. 1. Berlin, 293–302 p.
Refstie, T. and T. Gjedrem. 1975 Hybrids between salmonidae species. Hatchability and growth rate in the freshwater period. Aquaculture, 6:333–42
Refstie, 1981 T. and E. Austreng. Carbohydrate in rainbow trout diets. III. Growth and chemical composition of fish from different families fed four levels of carbohydrate in the diet. Aquaculture, 25:35–49
Suzuki, 1971 R. and Y. Fukuda. Survival potential of F1 hybrids among salmonid fishes. Bull.Freshwater Fish.Res.Lab., 21:5–32
Thorgaard, 1985 G., Scheere, P.D. and J.E. Parson. Residual paternal inheritance in gynogenetic rainbow trout implication for gene transfer. Theor.Appl.Genet., 71:119–21
Tsoi, 1969 R.M. Action of nitrosolmethylurea and dimethylsulfate on sperm cells of the rainbow trout and the peled. Dokl.Akad.Nauk.SSSR, 189:411–414
