-162- Dr. Harald Skjervold Institute of Animal Genetics & Breeding The Agricultural University of Norway _s-nlh, Norway

Transcription

1 -162- BKEEDING E_PERI_S WITH SALMONIDS Dr. Harald Skjervold Institute of Animal Genetics & Breeding The Agricultural University of Norway _s-nlh, Norway

2 -163- BREEDING EXPERIMENTS WITH SALMONIDS In Norway during the last years we have been developing cooperative breeding schemes within populations of dairy cattle, pigs and sheep. In the process we have done extensive calculations based on estimated values of phenotypic and genotypic parameters. In evaluating alternative breeding schemes for pigs ten years ago, we faced a problem which called for a test population having large litters and many groups of half-sibs in each litter. This question directed our attention to fish. After some studies, we chose Salmonids for some preliminary breeding experiments. Since then the extent of this fish breeding experiment has increased considerably compared to the original plans. Instead of going into detail about results obtained from these experiments, _I will more generally deal with some breeding aspects in connection with sea farming of Salmonids. I shall also relate some of our findings to the possibilities of genetic gain in Salmonids. Fish Farmin_ - W_ Salmonids? Fish farming is an important industry worldwide, and many countries including major industrial nations such as Japan United States, France, Germar_ and the Soviet Union are farming fish for food on a large scale. The Food and Agriculture Organization (FAD)has estimated the output from fish farming to be 3.02 million tons in i196_and 7.33 million tons _in 1970, which is 5 and 10 percent, respectively, of the human consumption of fish. FAO forecast the production of fish farming in 1980 to be 1_.7 million tons, or as much as 16_oof the human consumption of fish. These predictions rely on an expected shortage of fish for human consumption (Table l). It is expected that most of this progress on a world basis will be in the development of freshwater fish farming. In the northern region where freshwater temperatures are rather low, our chief area of interest is the farming of fish in the sea or in brackish water. Thus we must exclude a large number of freshwater species for which there has been determined a complete production cycle (from eggs to sexually mature adults). Marine species usually require a high standard of management for their successful maintenance. Their natural dependence for survival is based on a balance between very high fecundity and very low survival rate. However,

3 -16_- complete production cycles also have been developed for some species of marine flatfish such as Plaice and Sole. Some cultivation work, such as artificial hatching and rearing of fingerlings, has been conducted using different species of Salmonids and for these species rather complete production cycles have been developed. The Salmonids suchas Salmon, Brown Trout, Sea Trout and Char are commercially important fish species in most countries in Western Europe. Animal Breedin_ and the Salmonids. Relative to farm animals there are certain breeding advantages in the Salmonids. Included among the most important of these advantages are: o Very high fertility. The average number of eggs per kilogram of female body weight ranges from 1500 in Brown Trout to 2500 in Char. The number of progeny ranges from about 1500 in Brown Trout to io-12,000 in the Atlantic Salmon. o Fertilization outside the body of the female./thus in mating it is possible to make several combinations. For example, it is possible to get many large half-sib families. o The very high fertility among females makes it possible to practice: - sometypes of family selection. Even though we are working with traits which show very low heritability, the large family groups will in practice result in rather accurate estimation of the breeding value. - progeny testing among females. This can be carried out by using a mixed sample of semen, or semen from sires of which the breeding value already is known. - very intensive selection among these females. o The combination of the high female fertility and an extrauterine fertilization improves the possibility of estimating nonadditive genetic components. o The extrauterine fertilization in most species of fish makes it possible to carry out some artificial manipulation of chromosome content as shown by Russian and English scientists. L i' o The very high female fertility means that these species are almost "tailor made" for some hybrid systems.

4 -165- o Compared with most other groups of animals, fish show a remarkable potential for hybridization. Many crosses between species occur in nature and many more are possible by using artificial fertilization. We also have some disadvantages in the use of fish for breeding, and the most important are: o A rather long generation interval. This interval may be 5 years in Atlantic Salmon and 3 in Rainbow Trout. However, by use of tempered water the generation interval can be reduced to 3 years in Salmon and perhaps down to 2 years in Rainbow Trout. These are still relatively long intervals and so the potential of these species as laboratory animals in breeding experiments is considerably reduced. By use of Mosquito fish (Gambusia), it is possible to achieve generation intervals as short as _-6 months. o Because the level of knowledge concernin@ the technology of fish farming is relatively low, we have a rathe_ high frequency of practical problems and "accidents" in experiments with fish. o It is well known that fish populations frequently develop hierarchies and this may be an important factor in size variation in some populations. For example, in flatfish such a hierarchy contributes greatly to the environmental variation for growth characteristics. However, roundfish seem less liable to this form of interaction. (Purdom 1972). Our Breedin_ Experiments. After recognizing how interesting the Salmonids could be for breeding experiments and appreciating the lack of previous work on them, we established an experimental station for fish breeding. Breeding experiments started in 1966 on a rather small scale. However, after a few years we had recognized that a very large amount of genetic variation existed for some of the most important traits. Thus there was the possibility for considerable genetic improvement of characteristics such as growth rate and food efficiency. Parallel to these discoveries, results from a few pilot marine fish farms proved very promising with Rainbow Trout and Atlantic Salmon. Later investigations have shown that on the West Coast of Norway we find rather good environmental conditions for marine fish farming because: - water temperature is almost ideal for farming the Atlantic Salmon. - there is good protection against storms (islands and rocks). - water depth is satisfactory. ) - there is constant water replacement (The Gulf Stream). - the seawater is not too polluted. - a great quantity of food exists (wastage from the fish industry). This increasing interest in sea farming also improved our chance of

5 -166- obtaining the financial support which was needed to establish a permanent experimental station. We obtained such support mostly from the agricultural organizations in 1971 and built the Fish Experimental Stgtion at Sunndals_a. The station is located near a big hydroelectric plant allowing access to large quantities 8f cooling water from the generators (this water has a temperature of lo C above the turbine outlet). In addition we have two other sources of fresh water, and a sea water pipeline. A hatchery was constructed with a capacity of & million eggs and 600 different groups. In fish barns I and II fiberglass plastic tanks were installed, 266 of two square meters and l&o of one square meter, each with an automatic feeder. These units are used to maintain progeny groups of fingerlings until tagging is performed at an age of about 9 months. After tagging the fingerlings are measured and transferred into the 36 concrete circular ponds, which have a diameter of lo meters. In 1972 the farmer organizations provided money to establish a research unit for sea farming. At this station we test families of Salmon and Trout as potential breeding stock. The tests are from fingerling stage to sexual maturity. Experimental Design. It was our object to achieve effective programs both for Salmon living in captivity and in the wild, and this influenced erimental design. To evaluate the potential of Norwegian Salmon mt was_nec_ssar_ to_.-_:_ compare populations from different rivers. Local fishing organizations were contacted for permission to collect Salmon eggs, and A1 different rivers covering almost the entire coast of Norway are represented in the experiments, Within each river two different mating systems were used. In one system the eggs from each of three females were _vided into three portions and each fertilized with sperm from three different males giving a 3 x 3 set with 9 sib-groups and termedthe "factorial mating system" For each river one to three sets were produced. This mating system was used in the 1971 collection, with the exception of Arctic Char and Sea Trout where the females yielded too few eggs to divide effectively. The second mating system was a simple hierarchical design, mating each male to three different females from the same river. In addition to these mating schemes, hybrids between Salmon from different rivers were produced. The foundation stocks of Rainbow Trout were collected in from different Trout farms in Denmark, Norway and Sweden.

6 In order to test for the existence of an interaction between the strain of Salmon and the environment, the Fish Experimental Station cooperated with fish farms along the coast (Figure l). The farms joining the project receive one part of the sib-group from each river strain or progeny group. In addition, a sib-group of all progeny groups is represented at the Fish Experimental Station at Avery(the seawater station), and at this station the Salmon are reared in floating nets in the sea. In cooperation with The Waterpower and Electricity Board, smolt from some river strains will be released into four rivers under investigation. The primary aim of this project is to record recapture frequencies, growth rate, age at recapture and to determine whether there is an interaction between river strain and locality. Possibilities for Genetic Gain in Salmonids. During the long period of domestication of farm animals a great change in the production traits has taken place. The genetic change has been particularly large during this century. Not much systematic breeding work has been carried out with Salmonids, and in Atlantic Salmon for example wild stocks are still used in fish farming. Some attempts-have been made with Rainbow Trout (Donaldson 1968 and... Limbach 1969) o The behavior of the wild fish in captivity suggests that they are living under stressed conditions. However, it is expected that systematic selection for performance will be partly a selection for behavior and so increase the process of domestication. are: The traits which are of greatest economic importance in fish farming --growth rate - feed efficiency - viability and resistance to disease - carcass quality - age at sexual maturity In connection with other species, genetic response to selection in fish is mainly dependent on the four parameters: - heritability - phenotypic standard deviation - selection differential, and - the generation interval ) Our investigations have provided information on the first two of these parameters.

7 -168- Heritability. In the literature very few estimates of heritability are given for fish. In Carp there are some estimates for growth rate (Moav et al. 1966, Nemashev 1968 and Kirpichnikov 1969) and these seem to be about O From our work heritability estimates for growth rate in Salmonids are between O r For resistance to vibro disease in Salmon, Gjedrem and Aulstad (197&) estimated the heritability to be O.lO. These data consisted for lo& thousand fingerlings and the vibrio attack varied from 0.9% to 29.7% between strains (Table 2). Variance. ' In order to be able to compare the size of the variance of different traits, it is convenient to look at the coefficient of variation. Some estimates for different traits in Salmenids are given in Table 3. The variance of growth rate in Salmonids is3 to 5 times that of growth rate in livestock, so that substantial genetic gain in growth rate of Salmonids should be possible. For carcass traits such as meat color, meat and carcass score, a coefficient of variation of about 30% has been obtained at Sunndals_ra. Genetic Gain. To demonstrate the potential for genetic gain in growth rate of Salmonids, an example is given, using our data. If we assume that only phenotypic mass selection is practiced, and that the animals of both sexes selected for breeding are the top 1 percent of the population, a selection differential of 2.66 units is expected. The standard deviation of the 1972 batch of Rainbow Trout was 0.95 kg, and the average body weight 2.5 kg. A generation interval of 3 years is possible in practice. Heritability for growth rate is considered to be 0.2. The expected genetic gain based on these estimates will be 0.17 kg/year or an increase of 8 percent. This is an exceptionally high genetic gain, which at first may look unrealistic. However, the selection intensity is less than can be practiced. If the heritability is as low as 0.I; the genetic gain will still be as much as& percent/year. This is about four times as high as what is common in populations of livestock when the most efficient selection schemes are used. Selection Methods..? Progeny testing of both sexes can be used in Salmonids, but because of the considerable increase in generation length produced, it is of less interest.

8 -169- Family selection based on full-and half-sib testing does not have the disadvantages associated with proger_ testing. There is no increase of the generation interval when using family selection. Further, by means of family selection a trait of economic importance can be recorded with a high degree of accuracy, although if considerable dominance variance exists, the ranking on additive breeding value of full-sib groups will be somewhat biased. The accuracy of the ranking can of course be increased by taking into account the ranking of half-sib groups. In a breeding scheme for Salmonids a combination of phenotypic and family selection should be used. Depression of inbreeding. Table & shows a rather high inbreeding depression on survival rate among fry and fingerlings. However, the depression due to inbreeding on growth rate is at an expected level. Artificial Manipulation of the Chromosome Number. Recently gyuogenesis has been demonstratedby Russian_ English and Norwegian scientists working with fish. High doses of racltation can be used to induce gynogenesis, the spermatozoa then being made genetically inactive. Such a treatment destroys the chromosomes, but does not kill the sperm cell. After the fertilization by such radiated sperm, the egg contains only a haploid set of chromosomes at the second phase of meiosis. Russian experiments with Carp, andenglish experiments with flatfish show that among the gynogenetic offspring there is a small fraction (1-2%) of normal-looking embryos, which are in fact diploid. Purdom (1972) has shown that by exposing eggs to a cold shock the frequency of diploid embryos in flatfish can be increased to as much as 8_. The elimination of one group of chromosomes as the second polar body is then prevented. The diploid gynogenetic offspring have only maternal chromosomes so this produces a result similar to self-fertilization. Experimentally such highly inbred lines would be of interest. Even with some crossing-over such diploid gynogenetic offspring will be reasonably homozygous. Dr. Gjedrem and his co-workers have produced 13 diploid gynogenetic Rainbow Trout, but only 4 have survived. Induced Polyoloid_. Such induced gynogenesis allows us to estimate the extent to which cold shock prevents the last part of the second phase of meiosis. If very effective, it should be possible to induce triploi_ by use of non-radiated sperm in combination with cold shock.

9 -170- Triploids have been prcduced in Plaice and in Plaice x Flounder hybrids, however, similar attempts to induce tetraploids have so far failed. Triploids cotlt]dbe of great importance in fish farming for at least two reasonso Their growth rate may be greater than that of diploids. o Triploid organisms are usually sterile. By use of triploids we therefore could avoid sexual maturation, which results in decreased growth rate and increased death rate. In fish farming saxual maturation therefore results in wastage of energy and makes it almost i_ossible to market matured fish at a standard weight. (Figure 2). We have not yet been able to induce triploids in Salmon or P_bow Trout. There has also been a problem in identifying triploids as this could only be done by counting chromosomes. However, it is now possible to make chromosome preparations from leukocyte culture (Grammeltvedt 197&). So the number of chromosomes can now be determined without slaughtering. Hybridizin_ Species of Salmonids. A considerable number of viable hybrids between different species of Salmonids have been reported. The main purpose of such studies has been to provide basic biological information on: - theability to produce such hybrids 7 _ < - the morphological features of them, and their potential for reproduction. At Sunndals_ra in 1972 we commenced a hybridization experiment with Salmonids. The main purpose of this experiment was to investigate the potential value of such hybrids for fish farming. We attempted to use Brown Trout, Sea TroUt, Salmon, Arctic Char and Rainbow Trout, but Rainbow Trout was subsequently eliminated because of difficulties in synchronizing its spawning with other species. In order to study differences between sire and dam species, diallel crossing was conducted in two phases, namely freshwater followed by saltwater crossing. Results for the freshwater phase only are available. All crosses of Salmon x Sea Trout and Salmon x Brown Trout have a growth rate below the average of their parental populations and this is shown from Table 5, and the Sea Trout x Brown Trout crosses had an average body weight below the control group. All combinations of crosses with Arctic Char had a rate of gain significantly above the average of their respective parental species and the most promising results are obtained by crossing Char x Salmon (Fig. 3). Char female x Salmon male, gives hatchability equal to that of J!

10 -171- Salmon, but the stu_ival rat9 is sigr_!ficantlybetter in the hybrids. It will be of great interestto study the extent of sterility in this combination of crossese Because of the great difference in number of chromosomes we should expect sterile hybrids and if this is so it is possible that we have an interesting animsl for fish farming. Hybridization would seem tobe of much interest in fish farming, and some other interesting species should be studied, particularly the Pink Salmon as its fry also survive in seawater. Hybridization in combination with manipulation of the chromosome numbers directs our thinking towards "genetic engineering." Literature Donaldson, L. R S_lectivgbree_ng 'C_f_sslmo_d fishes. - Selected papers fromthe Conference of Marine Agriculture, Oregon State University, Marine Science Center, Newport, Oregon May 23, 24, :1968. Oregon State University Press. Pg. 65-7&. Gjedrem, T. 197&. Oppdrett av laksefisk. Forelesninger ved Norges Landbruksh#gskole. Landbruksbokhandelen, ]432 _s-nlh, Norway. 1_3 pp. Gjedrem, T. and Aulstad, D. 197&. Selection experiments with salmon. I. Difference in resistance to vibrio disease of salmon parr. (Salmo salar), Aquaculture 3, Grammeltvedt, A. F. 197_. A method of obtaining chromosome preparations from rainbow trout (Salmo gairdneri) by leucocyte culture. Norw. J. Zool. 22, _. Eirpichnikov, V.S The present state of fish genetics. Genetics, selection and hybridization of fish. Academy of sciences of the USSR, Ministry of Fisheries of the USSR. Limbach, B. Von Progress in sport fishery research Fish Genetics Laboratory, Beulah, Wyoming. USA. Moav I R. and Wohlfarth, G.W Genetic improvement of yield in carp. FAO World Symposium on Warm-Water Pond Fish Culture. Rome, May 1966.

11 -172- Nemashcv, G.A Heritability of some selection characteristics in ropsha carp. Ref. in Genetics, selection and hybridization of fish. Academy of Sciences of the USSR._ M_try_ of Fisheries Of the USSE. Purdom, C. E Genetics and Fish Farming. Laboratory Leaflet (New Series) No. 25, Fisheries Laboratory, Lowestoft, Suffolk, England. FA0, yearbook of fishery statistics, Vol. 26, 1972.

12 -173 _Zm _m _ OPX m_ m z _ z _ m m - O 1 O_ m _0_ m F mmm m m_ z m _ m O _0 _ O z

13

14 -175-

15 -176- Table i World demand for fish. (Inmillion of Short Tons live weight) n n, For human consumption &O.9 A8.& 67.3 For fish meal /_ Total i00.6 n, Projected supply 92.0 Deficiency 8.6 Source: FAO

16 -177- ) Table 2 Number of salmon parr and percent dead from vibrio disease from different rivers. No. of fish Percent d_ad Locality (river) in hundreds of vibr!o disease Alta M_lselva Namsen llo 2.09 Fos_n 111!-58 Jordalsgrenda 4_1 2._2 Usma 19 &.59 Surna Driva _ Rauma Gaula Sunnfjord LSO 5.98 Laerdal _ Lone 20 _.09 Etne Sandvik 153 &.59 Lule_ (Sweden) _rt

17

18

19 -18n-

20 -181-! DR.._ SKJ_VOLD - "BR_,nlNG FXPERIMENTS._WITH SALMONIDS,, A. W. NORDSK_: (a) What is your opinion a% to the possibility of developing genetically superior performers when you put them out to sea? (b) Are they better performers because they are better harvesters of sea food? (c) Is this type of artificial selection in conflict with natural selection? (d) In testing your progeny groups in captivity, what was the principal food? SKJERVOLD: (a) So far we have no experience in developing superior strains for sea runs. Investigations indicate that there is genetic variation in traits such as disease resistance, recapture frequency and growth rate. (b) I do not know and I think it is very difficult to investigate. (c) It is not likely that selection for higher frequencies of return and higher productivity of the salmon will be in conflict with the natural selection. (d) Until the smolt stage, a dry pellet consisting of 50-60% fish meal is used. For larger fish kept in the sea, a wet diet of grained capelin, fish offal, vitamins plus a binder is common. GERRY FRIARS: To what extent is repeat spawning and years sea heritable inatlantic Salmon? spent at SKJ_IVOLD: I do not know of estimates of heritability for these traits, but we know that there exist considerable strain differences at least for number of years at sea before spawning. G. B. HAVENSTEIN: Due to differences that exist in temperature extremes between possible fish farming areas, do you foresee genetic x environmental interactions becoming important in your long range planning? In other words, will separate lines and crosses have to be developed for different climatic areas? SKJ_RVOLD: One of our experiments with Salmon is a studf of genetic x environmental interactions. The investigation is not completed. ROBOT ARVIDSON: (a) You mentioned knowledge of salmon breeding and fish management is very limited. Do you have contacts with other countries on salmon breedings - such as U.S., Canada, and Japan? (b) Do you exchange stocks between countries?

21 -182- SKJ_VOLD: Very few estimates of phenotypic and genetic parameters have been estimated for salmoni_s. Some work has been carried out With Rainbow Trout in the USK. Goncerning salmon, some selection projects have recently been started, one in St. AndreWs, Canada and one in Seattle, Washington. GEORGE GODFRIED: Did you expect the Arctic Char to perform better in crosses than_brown Trout and Salmon? SKJ_RVOLD: Arctic Char (Salvelinios alpinus) is considered to be more unrelated to Salmon (Salmo salar) and Trout (Salmo trutta) than these two species are to each other. F_arlier results indicate high performance of crosses between Brook Trout (Salvelinius fontinalis) and Salmon and Trout. WALT H_RVE_: W_at metnod is used to identify the individual fish?...skj_vold: We use cold branding for identifying the families, and we really have little use for individual identification. The main reason is that progeny testing is not practiced because of the long generation interval.