水产动物育种学资料 (11).pdf
ReviewGenetic improvement for the development of efficient global aquaculture:A personalopinion reviewTrygve GjedremNofima and Akvaforsk Genetics Center(AFGC),P.O.Box 210,1431 s,Norwaya b s t r a c ta r t i c l ei n f oArticle history:Received 12 January 2012Accepted 7 March 2012Available online 14 March 2012Keywords:Breeding programsSelectionAchievementsPotentialIt has been exciting to follow the rapid development of aquaculture production in Norway,and internation-ally,since 1971.As an animal breeder I am particularly impressed with the genetic gain obtained for growthrate,and also for disease resistance in several aquatic species,which is five to six times higher than what hasbeen achieved in terrestrial farm animals.This is illustrated in five selected projects I have been involved in.The sad story is,however,that only less than 10%of the worlds aquaculture production is based on geneti-cally improved stocks.The big challenge for the future is to develop more selective breeding programs forexisting and new emerging aquaculture species in order to increase the production of this nutritious foodsource and to improve the efficiency of the use of feed,water,land and labor resources.2012 Elsevier B.V.All rights reserved.Contents1.Introduction.132.Knowledge from livestock genetics.133.Transfer to aquaculture research.133.1.Offer to move into a new research world.133.2.General view on selective breeding in aquaculture.143.3.Research facilities.143.4.Problems to be solved.143.5.Organizing the work.144.Breeding program for Atlantic salmon.144.1.Experiments.144.2.Developing the breeding program.154.3.Achievements.155.Breeding program for Nile tilapia.165.1.Status and general opinion in the field.165.2.Experiments with tilapia.165.3.Achievements.176.Breeding program for rohu carp.176.1.Status and general opinion in the field.176.2.Experiments with rohu.176.3.Achievements.187.Breeding program for Litopenaeus vannamei.187.1.Status and general opinion in the field.187.2.Experiments with shrimp.187.3.Achievements.188.Breeding program for sea ranching of Atlantic salmon.188.1.Background.188.2.Experiments in sea ranching.188.3.Achievements.199.Documentation and education.19Aquaculture 344-349(2012)1222E-mail address:trygve.gjedremnofima.no.0044-8486/$see front matter 2012 Elsevier B.V.All rights reserved.doi:10.1016/j.aquaculture.2012.03.003Contents lists available at SciVerse ScienceDirectAquaculturejournal homepage: for future aquaculture production.1910.1.The good news.1910.2.The bad news.2011.My advice.20Acknowledgments.21References.211.IntroductionIn 1972,Elsevier Science Publishers B.V.established a journaldevoted to aquaculture.The total global production of culturedorganisms at that time was less than five million tons while in 2008it reached 52.5 million tons(FAO,2010).Elsevier Science PublishersB.V.therefore showed great foresight when they established thejournal Aquaculture which has been the key scientific journal in thisfield over the last 40 years.It is difficult to determine when selective breeding and cross-breeding was first systematically applied to aquatic species as littlepublished information is available.It is,however,most likely that in-dividual(mass)selection was practiced at the time of domesticationto improve growth rate together with body shape and externalcolor.In this connection the high fecundity of fish can create prob-lems because it increases the likelihood of selecting close relatives,unless the number of breeding candidates per family is equalized(Gjerde et al.,1996)or some type of walk-back selection is applied(Sonesson,2005).It is well known that in the past farmers usuallyused only a few parents each year,and that after a few generationsthe animals became inbred and showed signs of depression of fitnessand performance.With these experiences,some farmers lost confi-dence in selective breeding and continued to recruit breeders fromwild stocks which were easily accessible and inexpensive.The inheritance of qualitative traits in common carp(Cyprinuscarpio)and aquarium fish species received attention early.Zhang(1994)describes the development of colored carp strains whichbegan three hundred years ago when the Emperor sent red carps toJiangxi province.Long term inbreeding and isolation resulted inlarge variation among strains.In ancient China,goldfish,which is anornamental variety,was developed from crucian carp.Koi carp withits diverse color varieties were developed in Japan,also from commoncarp.Both fish types are now commonly found around the world andkept for decorative purposes.Later,genes for scale cover in commoncarp were described and named scaled,mirror,linear and leather(Kirpichnikov,1937).One of the first documented experiments investigating selectionin fish was initiated in the USA in 1919(Embody and Hyford,1925),in which brook trout(Salvelinus fontinalis)were selected for in-creased survival to furunculosis.Over three generations survivalrate increased from 2%to 69%.Large response to selection for in-creased survival to furunculosis in common carp was reported fromGermany(Schaperclaus,1962).Ilyassov(1987)summarized the re-sults from selection of common carp against dropsy disease in theUkrainian ropsha strain which began in 1953 by concluding that“Mass phenotypic selection within different breeds has given varyingresults”.By the fourth and fifth generations of selection,the improve-ment in survival was 30 to 40%over non-selected control carp.Thefirst work on common carp selection in USSR dates back to the1920s and culminated in the 1950s with the development of twohighly productive strains,Ukrainian scaly and frame carps(Kuzema,1971).Moav and Wohlfarth(1963,1973,1976)applied individual se-lection for growth rate over five generations in common carp withoutobtaining any response when selecting for fast growth rate,but a pos-itive response for slow growth rate.They concluded that overdomi-nance played a role in the inheritance of growth rate in commoncarp and that there was no genetic variation in the trait.However,Kinghorn(1983)commented that the report of no response to selec-tion for high growth rate is not conclusive in this case.Limbach(1969)reported response to selection for growth rate inrainbow trout(Oncorhynchus mykiss)and inbreeding depression forgrowth in progeny after mating close relatives.In 1932,Lauren R.Donaldson started to select rainbow trout for increased growth rate,increased number of eggs and early sexual maturation and achieveda remarkable response over many generations(Donaldson andOlson,1955).Donaldson started selection on return rate in sea ranch-ing with chinook salmon(Oncorhynchus tshawytscha)in 1949 andreported response to selection(Donaldson and Menasveta,1961).The heritability for fingerling weight in common carp was esti-mated to be 0.21 by Nenashev(1966),while in rainbow trout it was0.16 at an age of 150 days and 0.32 at 280 days(Aulstad et al.,1972).In oyster(Crossostrea gigas),Lannan(1972)estimated aheritability of 0.33 for body weight.2.Knowledge from livestock geneticsI was born in 1928 in Bjerkreim,Rogaland and my parents werefarmers.I studied Animal Sciences at the Agricultural University ofNorway(from 2005 the Norwegian University for Life Sciences),where I obtained my undergraduate degree in 1956,Licentiate degreein 1962 and later a Master of Science degree at University of Wiscon-sin,USA in 1963.At the Department of Animal Genetics and Breeding,Agricultural University of Norway,I led extensive breeding projects insheep and studied phenotypic and genetic variation in productiontraits.I was responsible for developing a new breeding program forsheep based on progeny testing of rams in Ram circles which stillis used by the Norwegian sheep industry(Gjedrem,1969a).The the-sis for my doctorate degree,which I defended in 1970,was entitled:Studies related to progeny testing of rams and selection indexes(Gjedrem,1969b).In the early 1970s efficient breeding programs had been devel-oped world wide for all main terrestrial farmed animal species(Hagedoorn,1950)and it was generally accepted that selection wasan efficient method to improve the productivity and production effi-ciency of farm animals,as it was for plants.The discussions weremore directed towards which breeding and selection methodswere most efficient and how they could be improved and bestimplemented.3.Transfer to aquaculture research3.1.Offer to move into a new research worldIn 1970,the late Prof.Dr.Harald Skjervold,then Head of Depart-ment of Animal Genetics and Breeding,asked me to take responsibil-ity for research in aquaculture and the building of necessary facilities.With my background in agriculture and research in terrestrial farmanimals,I felt that this task and position was difficult and challenging.At that time I had no experience and knowledge in aquaculturefarming or about the culture of aquatic species.In spite of my fathersadvice,I decided to jump into the water and accepted the offer andentered the field of aquaculture in February 1971.13T.Gjedrem/Aquaculture 344-349(2012)12223.2.General view on selective breeding in aquacultureAs I tried to build my competence in aquaculture and contactswith national and international researchers and experts in the field,I was shocked to learn about the generally negative views on selectivebreeding.The main arguments were as follows:Dont you know thatselective breeding does not work for aquatic species?Dont you knowthat the well known professor in Israel,Rom Moav,did not get any re-sponse after many generations of selection for growth rate in com-mon carp?Dont you know that he concluded that there is nogenetic variation for growth rate in common carp?This was followedwith reference to Deputy Director Colin E.Purdom in Lowestoft,En-gland who repeatedly declared that selective breeding does notwork in fish,because fish are different from terrestrial farm animals.These views were discouraging and came at a time when the pub-lications referred to above(Section 1)which had found positive re-sponses to selection in fish species were not well known,at leastnot to me.However,for us at the Department of Animal Geneticsand Breeding the results obtained from our own initial experimentswith rainbow trout in freshwater during 19671970(Aulstad et al.,1972)demonstrating genetic variation in body weight at 280 daysof age,were convincing and stimulated further studies in quantitativetraits in fish.Still,knowledge of the magnitude of the genetic varia-tion for economically important traits in aquatic species was very lim-ited and no estimates of genetic correlations between productiontraits were available.3.3.Research facilitiesIn 197071,researchers at Department of Animal Genetics andBreeding were convinced that to ensure future success for the emerg-ing Atlantic salmon(Salmo salar)and rainbow trout industries,theproductivity of these species should be improved through domestica-tion and selective breeding.How this could be done,and how tostructure an effective breeding program were still open questions atthe time when there was no systematic and efficient breeding pro-gram for aquatic species.In addition to reliable estimates of pheno-typic and genetic parameters for traits of economic importance foreach candidate species in question,we needed basic knowledgeabout reproduction characteristics.Some limited information wasavailable for rainbow trout,but none for Atlantic salmon.To obtainthis information we needed facilities to produce individual full-sibfamilies,to keep the fertilized eggs of each family separated in ahatchery,and to separately rear the fry of each family until theycould be physically tagged.Many questions needed answers,for ex-ample:which would be feasible tagging methods and what type ofgrow-out facilities and feeds would be needed both in the freshwaterand the in the sea-rearing phase?During 197173,a freshwater aquaculture research station wasbuilt at Sunndalsra,located approx.200 km south of Trondheim,Norway.The facility included a hatchery containing 300 trays,abarn for Atlantic salmon with 216 2 m2tanks and one barn for rain-bow trout with 192 mostly 1 m2tanks,and 36 78 m2concreteponds outdoors to rear the fingerlings before release into floatingnet cages in the sea.Sunndalsra was selected as the site for the re-search station because it was close to a large hydroelectric power fa-cility.The temperature of the large volumes of water used to cool theturbines was high enough during the winter for production of 11/2 year old smolts which could be transferred into sea water cagesin MayJune.In addition,the research station was located close to ariver with freshwater and a fjord with sea water.In 19731974,a ma-rine research station with floating net cages was established atAvery(approx.100 km west of Sunndalsra).Looking back,it is im-pressive that Dr.Skjervold succeeded to raise enough money to buildthese large research stations in 1971 when Norwegian aquacultureproduction totaled only 100 t of Atlantic salmon and 540 t of rainbowtrout.When the operations at the new facilities were initiated,onlythe manager Arne Kittelsen(who was trained by Prof.Lauren Donald-son at College of Fisheries,University of Washington,USA)hadhands-on experience with fish culture.The rest of the team,re-searchers and field workers,had to learn on the road.There was littleknowledge and experience and much skepticism,as can be seen fromthe following anecdote:Dr.Skjervold was once asked what the pur-pose of the 36 outdoor ponds was,and he replied that he was plan-ning to produce 100,000 smolts.The reply was that he mightsucceed in that,but that he would never manage to sell them.Well,he did;in 2011 Norway produced 300 million smolts.3.4.Problems to be solvedOne of the first questions asked was which species would be bestfor farming in Norway.At that time we focused on salmonids andcompared Atlantic salmon,brown trout(Salmo trutta)specimensfrom large lakes,rainbow trout,Arctic char(Salvelinus alpinus),pinksalmon(Oncorhynchus gorbuscha)(Gjedrem and Gunnes,1978)andcrosses between these species except for pink salmon(Refstie,1983).It was concluded that Atlantic salmon and rainbow trout hadthe highest potential for farming in Norway.There were many challenges to overcome before we could pro-duce the planned 200 Atlantic salmon and 200 rainbow trout familieseach generation.Atlantic salmon will be used as an example below.In1971 a dry pellet became available for start feeding of fry and auto-matic feeders were attached to each of the tanks and worked satisfac-tory.However,at that time no dry pellets were suitable for grow outof salmon in the sea.A moist pelleted feed,which disintegrated easilyin sea water and caused pollution problems,was produced during the1970s.Identification of individual fish for determining genetic rela-tionship was also a problem.Several methods were evaluated(Refstie and Aulstad,1975),and freeze branding using liquid nitrogencombined with fin clipping w
