Genetic improvement for important farmed aquaculture

Genetic improvement for important farmed aquaculturespecies with a reference to carp,tilapia and prawns in Asia:achievements,lessons and challengesNguyen Hong NguyenUniversity of the Sunshine Coast,Maroochydore,Qld 4558,AustraliaAbstractThis study provides an overview of successful genetic improvement programmes for important farmed aquaculture species in Asia,with a focus on lessons and experi-ences gained as well as challenges remaining.In both fish and prawns (Macrobrach-ium rosenbergii ),conventional selective breeding approaches have resulted in significant improvement in productivity,with genetic gains ranging from 8to 12%per generation.Selection for high growth has also brought about beneficial changes in fillet weight of fish and edible meat in prawns without detrimental effects on flesh quality attributes and fitness-related traits.Genetically improved animals show remarkable vigour and high adaptation to a range of culture envi-ronments/conditions in Asian countries.Despite these successes,however,the con-duct and practical implementation of such breeding programmes still present several challenges.These include the expansion of breeding objectives,manage-ment of inbreeding in closed-selection populations,controlling the effects of geno-type by environment interactions,simultaneous production of large number of full-and half-sib families for species with asynchronous spawning behaviour,maintain-ing pedigree records,dissemination of the improved strains for widespread produc-tion,as well as a reluctance by many to carry out systematically designed genetic improvement for aquatic animal species.There are also challenges with regard to the application of genomic information in genetic enhancement programmes and the development of genetically improved strains in response to climate and envi-ronmental changes.In this study,each of these challenges is discussed and solu-tions are proposed to increase efficiency of future genetic improvement programmes for economically important aquaculture species.Keywords Breeding,genetics,heritability and genetic improvement,selection responseCorrespondence:Nguyen HongNguyen,University of the SunshineCoast,Locked Bag 4,Maroochydore DC,Qld 4558,Australia Tel.:+61754565138Fax:+61754565150E-mail:NNguyen@.auReceived 7Nov 2013Accepted 14Apr 2015Introduction2Examples of successful genetic improvement in carp,tilapia and prawns 3Carp 4Tilapia5Giant freshwater prawn 6Challenges6Biological constraints7©2015John Wiley &Sons Ltd DOI:10.1111/faf.121221F I S H and F I S H E R I ESFamily production in tilapia and prawns7 Mating design in carp,tilapia and prawns7 Family and individual identification in prawns7 Molecular-based pedigree in carp,tilapia and prawn8 Social interactions and indirect genetic effects9 Correlated responses to selection for high growth9 Multitrait selection10 Genotype by environment interaction11 Management of inbreeding13 Estimation of genetic gain14 Climate change14 Gene(maker)-assisted selection and genomic selection16 Dissemination of improved strains17 Capacity to run genetic improvement programmes17 Other issues18 Conclusions18 Acknowledgements19 References19IntroductionAsian aquaculture accounts for about80%of total world production,but it will need to grow sub-stantially to meet the demands of a rapidly expanding human population(FAO2012).One way to increase the production of cultured species, per unit of land and water use,is through genetic improvement.By way of comparison,most of the improvement in land agriculture(livestock and crops)over the last half century has been as a result of the development and use of genetically superior breeds or varieties(Hill2008).For exam-ple,growth rate in chickens increased by about 300%with more than90%of that improvement being directly attributable to genetic selective breeding(Havenstein et al.2003).On the other hand,there are only a few examples where aqua-culture production has benefited from genetically improved strains such as Atlantic salmon(Gjed-rem2012).There is also the issue of loss of productivity from poor genetic management and inbreeding. Aquaculture seedstocks are often collected periodi-cally from the wild and,if bred in captivity over generations,are seldom adequately managed from a genetic viewpoint(Eknath and Doyle1990; Ponzoni et al.2010).Without proper manage-ment,the usual outcome for highly fecund aqua-culture species is loss of genetic diversity as a consequence of inbreeding,leading to a decline in productivity,and there are many instances in which‘domesticated stock’have proved to be less productive than their counterparts from the wild, due tofitness reductions(Araki et al.2008). Across the sector,the lack of improved strains capable of producing high-quality seed is consis-tently identified as the most widespread and persis-tent technical obstacle to the sustainable development of aquaculture among both small and medium enterprises(FAO2012).In response to this situation,a number of breeding pro-grammes have been initiated to develop genetically improved strains and enhance quality seed produc-tion for important farmed aquatic species in Asia (Nguyen and Ponzoni2006).Despite intermittent reports of success(Hung et al.2013b;Ninh et al. 2013;Hamzah et al.2014b),there are many remaining problems to be resolved prior to the implementation of improved selective breeding pro-grammes.Key challenges include biological limita-tions,genetic selection and mate allocation strategies,the effects of the environment on geno-type stability,particularly in response to climate change,as well as the adoption of the improved strains by end-users(N.H.Nguyen,unpublished data).This study focuses on three species(carp, tilapia,giant freshwater prawn Macrobrachium ro-senbergii,Palaemonidae)chosen due to their high socioeconomic importance and large production volumes in Asia(FAO2012).We will emphasize the many existing challenges to the industry and2©2015John Wiley&Sons Ltd,F I S H and F I S H E R I E S Genetic selection in carp,tilapia and prawns N H Nguyenattempts to provide solutions to overcome these using breeding programmes.Many of these issues apply to other aquaculture genetic selection pro-grammes.Suggestions for future studies are pre-sented.Examples of successful genetic improvement in carp,tilapia and prawnsFormal genetic improvement programmes have been established for carp,tilapia and Giant fresh-water prawn in a number of Asian countries (Ponzoni et al.2011;Hung et al.2013a;Ninh et al.2013;Hamzah et al.2014b).The general experimental design for these species is presented in Fig.1,but it is usually adjusted to fit biology and production cycle of individual species during practical implementation.Briefly,in each genera-tion,between 50and 100full-and half-sib fami-lies are produced,following hierarchical nested (e.g.one male mated to two females)or partial factorial mating design (e.g.male 1with females 1and 2,male 2with females 2and 3,and so on to male n with female 1and n ).Natural spawning (taking place in hapas (ponds)without outside intervention or induced breeding (with hormonal treatment)is applied to synchronize reproduction within as short a time interval as possible.Individ-ual families are reared separately until the fry reach a suitable weight to be physically tagged (5–10g for fish and 2–5g for prawns).Communal grow-out for performance testing is then con-ducted in the prevailing culture environments:normally earthen ponds.At harvest,body mea-surements,in addition to other important traits (sexual maturity,survival),are recorded.The data are then processed,using best linear unbiased pre-diction (BLUP)methodology to determine genetic merits (or estimated breeding values,EBV)for each individual in the pedigree.The highest EBV animals are selected to become parents of the next generation.A combined between-and within-fam-ily selection was applied,and matings made among genetically unrelated brood stock,based on an individual’s EBV and genetic relationship to other animals in the pedigree.The sameTagging 50 –70 pcs/ family5000 fish4 – 6 monthsGeneƟc evaluaƟonand selecƟonSelected breedersNursery 4–8 weeksHatchery CondiƟoning of breedersMate allocaƟonGrow-out system60 –100 families♂♀with DNA taggingEarly communal rearingFigure 1General design of genetic improvement programmes for farmed aquaculture species.The production cycle oftilapia (about 1year per generation)is given as an illustrative example.Photograph courtesy:Azhar Hamzah and WorldFish.©2015John Wiley &Sons Ltd,F I S H and F I S H E R I E S3Genetic selection in carp,tilapia and prawns N H Nguyenprocedures regarding family production,genetic evaluation and selection process are repeated in subsequent generations during the course of the project.In all of the above programmes,genetic improvement has focussed on the enhancement of production performance(i.e.body weight).In some cases,as the programme proceeds,economi-cally important traits(carcass trait,flesh quality,fitness-related traits–such as survival and repro-duction)are also investigated and possibilities for broadening the breeding objectives of target spe-cies are considered.The experimental design described above is generally also used for other aquatic animal species(Rye et al.2010).CarpCarp is the most important freshwaterfish,con-tributing about41.6%to total world aquaculture production(FAO2012).Several species of carp have been widely cultured in South and South-East Asian countries.Selective breeding pro-grammes have been carried out in important spe-cies such as silver barb(Puntius gonionotus, Cyprinidae)in Bangladesh and Thailand(Hussain et al.2002),rohu(Labeo rohita,Cyprinidae)in India(Gjerde et al.2002;Mahapatra et al.2007) and common carps(Cyprinus carpio,Cyprinidae)in China,Indonesia and Vietnam(Ninh et al.2011). The majority of these countries have established systematic genetic improvement programmes,and fully pedigreed populations are maintained byfit-ting an individual animal model to estimate genetic merit and select replacements.This enables rigorous conduct of the genetic improvement pro-gramme as well as the necessary control of inbreeding to secure long-term response to selec-tion.Across carp species examined in this study, genetic gain ranged from8to20%per generation (Table1).There is also considerable genetic varia-tion in these populations,with the heritability for body weight ranging from0.20to0.40(Table1), providing valuable stock for future selection. Although selection was primarily for harvest weight,correlated increases in body length,height and width were also achieved in carp species (Ninh et al.2013).The establishment of a fully pedigreed population is fundamental in broadening the breeding objectives for some carp species,for instance by including additional new traits such as survival or disease resistance in rohu(Mahapat-ra et al.2008)or applying new molecular genetic technologies for parental assignment in common carp in Vietnam(Ninh et al.2011).Parentage assignment using microsatellite markers enables the early communal rearing of all families from birth as a means of reducing common environ-mental effects,achieving faster growth rates(thus shortening the time to selection)and,as a conse-quence,accelerating the development of beneficial genetic traits in the population as a whole(Ninh et al.2013).The selected carp strains developed from these programmes have demonstrated significantly greater growth performance than local stocks of the same species,ranging from20to40%across agro-ecological regions and farming systems in Bangladesh,Thailand and Vietnam(Nguyen and Ponzoni2008).The marked improvement of the ‘Jayanti’rohu carp relative to local strains(up to 75%)was also recorded in a range of culture envi-ronments in India(Mahapatra et al.2007).Due to the superior performance of the improved strains, there has been a growing demand for the improved carp seed in major carp producing coun-tries where these improved varieties have signifi-cantly increased the productivity and profitability of carp producers(Dey et al.2010).The economic benefit and benefit-to-cost ratio from the genetic improvement programme in carp has been highly beneficial(Ponzoni et al.2008)to national econo-mies,especially in countries where a pyramidTable1Heritability(Æaverage standard errors)for body weight and genetic gain(%)achieved in carp species in Asia.Latin name CommonnameGenerationof selection HeritabilityGenetic gainper generation(%)ReferencesCyprinus carpio,Cyprinidae Common carp30.25–0.32(Æ0.06)10.0Ninh et al.(2013) Labeo rohita,Cyprinidae Rohu50.23Æ0.0617.0Mahapatra et al.(2007) Puntius gonionotus,Cyprinidae Silver barb30.207.2Hussain et al.(2002)4©2015John Wiley&Sons Ltd,F I S H and F I S H E R I E S Genetic selection in carp,tilapia and prawns N H Nguyenbreeding structure (two or three tiers including nucleus,multiplication and production)(Nguyen and Ponzoni 2006)is well established to dissemi-nate the improved genes from the nucleus,either directly or indirectly,to commercial producers such as in India and Vietnam (K.Mahapatra and N.H.Ninh,personal communication).TilapiaTilapia is the second most important freshwater fish after carp,and a number of improved strains have been developed by national aquaculture institutions in Asia.In this study,I present the example of the genetically improved farmed tilapia (GIFT)strain that has involved major international and regional efforts by The WorldFish Center as well as Norwegian,Philippine and Malaysian insti-tutions.The GIFT base or foundation population was established in early 1990s (Eknath et al.1993,2007;Bentsen et al.1998).Prior to 1996,these fish went through six generations of selec-tion for high growth in the Philippines (Bentsen et al.2012).In 2001,the GIFT strain was trans-ferred to a research station operated by the Department of Fisheries,Malaysia.Since then,the selection programme has been continuing (Ponz-oni et al.2005),and by 2013,the GIFT fish had undergone 16generations of selection (Nguyen et al.2010b;Ponzoni et al.2011;Hamzah et al.2014b).There has been substantial improvement in growth performance of the GIFT strain over the course of selection,with an average gain of 10%per generation (one generation per year in tilapia)combining to produce body weights more than double when compared with the control over 10generations of selection (Fig.2)(Hamzah et al.2014b).Moreover,restricted maximum likelihood (REML)estimates of heritability for body weight (h 2=0.28)in the latest generation (2012)indi-cated that there is still genetic variation in the population to encourage additional selective breed-ing to achieve further increases in body weight.To date,the GIFT strain has been distributed and cultured in 14different Asia –Pacific (China,Bangladesh,Indonesia,Philippines,Papua New Guinea,Sri Lanka,Thailand,Vietnam)and Latin American (Brazil,Costa Rica,Ecuador and Mexico)nations.The fish have performed well in many dif-ferent farming systems.For example,in Sri Lanka,the GIFT strain exhibited a 112%increase in body weight at harvest compared to local varieties across three different production environments:ponds at the government research station,ponds of farmers and seasonal reservoirs.In addition,survival rates of GIFT were 8.2–23.8%higher than local stocks (Nguyen et al.2011a).Under commercial production,culture of monosex (sex reversed or YY male)GIFT strain is widely prac-tised in the region (Kamaruzzaman et al.2009).Not surprisingly,the GIFT strain has had a very large impact on production volumes.As a conse-quence,the livelihoods of farmers in many Asian countries has been improved and there has been a considerable positive socioeconomic impact accord-ing to the Asian Development Bank (ADB 2005)0123456020406080100120200320042005200620072008200920102011G e n e t i c s t a n d a r d d e v i a t i o n U n i tP e r c e n t a g e o f t h e c o n t r o l , %Genetic SD Unit%Figure 2Genetic trend for body weight in the genetically improved farmed tilapia (GIFT)strain over 10generations of selection in Malaysia (Hamzah et al.2014b).Genetic standard deviation unit =estimated breeding values in actual measurement unit/r A where r A is the square root of the additive genetic variance.©2015John Wiley &Sons Ltd,F I S H and F I S H E R I E S5Genetic selection in carp,tilapia and prawns N H Nguyenand the International Food Policy Institute(Yosef 2009).This is a highly successful example of genetic improvement in farmed aquaculture.GIFT has become a model species,and the associated technologies(GIFT Manual)have been applied to improve other indigenous tilapia such as Oreochr-omis sharinus,Cichlidae in Malawi(Maluwa et al. 2006),Nile tilapia Oreochromis niloticus,Cichlidae in Egypt(Rezk et al.2009)and red tilapia Oreochr-omis spp.,Cichlidae in Malaysia and Thailand (Hamzah et al.2008;Pongthana et al.2010; Nguyen et al.2011b).Freshwater prawnGenetic improvement of the giant freshwater prawn (GFP),M.rosenbergii,was only initiated in2007, after the successes of breeding genetically improved strains of farmed tilapia and carp became better known.GFP is one of the most important crusta-ceans in inland aquaculture in the(sub)-tropics and fits in well to the typical Asian smallholders’system of prawn polyculture along with carp or tilapia (Zimmermann et al.2010).Selective breeding has begun in China(Luan et al.2012),India(Pillai et al.2011)and Vietnam(Thanh et al.2009;Hung et al.2013b).The main aim of such programmes is to develop high-yielding strains with good adapta-tion and high survival rates under culture condi-tions in Asia.Implementation of the programmes started with the collection and evaluation of geo-graphically discrete populations in each country (India,Malaysia or Vietnam).We performed a full diallel cross involving three different strains to cre-ate a synthetic base population for systematic stock improvement.Statistical analysis of the diallel crosses showed that the additive genetic and reci-procal effects were the major significant sources of variation for all growth traits,whereas the heterosis effect was not significantly different from zero (Thanh et al.2010;Pillai et al.2011).A synthetic base population was formed from the best perform-ing individuals regardless of their genetic make-up. Selection based on additive genetic effects has been practised in these populations because cross-breed-ing among existing strains is likely to result in only marginal genetic superiority due to the small mag-nitude of the heterotic effects(Thanh et al.2010). Mixed model estimation of genetic parameters showed that there is heritable genetic variation in the selection population,with heritability ranging from8to41%across a number of body traits (Hung et al.2013a).Selection also resulted in significant direct response in harvest weight(aver-aging7%per generation i.e.per year)and simulta-neously brought about beneficial gains to carcass traits such as abdominal length,abdominal weight, telson-off weight and skeleton-off weight(3–4%per generation)(Hung et al.2013b)(Box1).ChallengesDespite the successes,there are several challenges remaining relating to the practicalimplementation 6©2015John Wiley&Sons Ltd,F I S H and F I S H E R I E S Genetic selection in carp,tilapia and prawns N H Nguyenof genetic improvement programmes for carp, tilapia and prawns.In the following sections,each of the challenges is discussed with particular empha-sis on how they can be accommodated to improve the efficiency of genetic breeding programmes.Biological constraintsFamily production in tilapia and prawn Simultaneous production of a large number of families still remains a challenge for species with asynchronous spawning behaviours and in species where induced breeding is not possible,for exam-ple tilapia and freshwater prawn.In these species, females do not often reach sexual maturity or spawn at the same time.In addition,for M.rosen-bergii,there is a high rate of egg clutch abortion, about30–40%(Thanh et al.2009,2010),and in breeding programmes,it often takes1–2months to produce between50and100families:the sam-ple size required to constrain the inbreeding rate to<1%per generation.Prolonged spawning inter-vals,accompanied by the rearing of separate fami-lies,can lead to large differences,not only in stocking weight,but also to differences in ambient environmental conditions among full-sib groups (Nguyen et al.2007;Ponzoni et al.2011;Hamzah et al.2014b).To obtain a large number of families per generation within a short time,it is advisable to use excessive numbers of mating hapas(net cage)or tanks,although the minimum quantity will vary by species.In species with asynchronous spawning behaviours,such as tilapia,in-vitro fer-tilization(IVF)can help to shorten the spawning interval,allowing the design of alternative advanced mating schemes.Mating design in carp,tilapia and prawns Application of advanced mating schemes,such as factorial design,is still difficult in tilapia and GFP. Classical nested mating design(one male to two females)is usually practised in tilapia and prawns. However,at the same effective population size (N e),factorial mating can achieve a greater accu-racy in parameter estimation and selection response compared with conventional single pair or nested mating design(Dupont-Nivet et al. 2006).When compared with nested designs,facto-rial matings increase the number of half-sib fami-lies while decreasing the number of full-sib families,thus reducing the risk of selecting many individuals from the same parents(Sorensen et al.2005).The single pair mating design is best in terms of maximizing effective population size(N e), which is necessary to secure long-term genetic response to selection(Dupont-Nivet et al.2006). However,in shallow full-sib pedigreed populations, it is difficult to separate the additive genetic com-ponent of variance from dominance and common environmental effects in single pair matings. There are four main types of factorial mating designs:complete,incomplete,by set and rectan-gular factorial(Berg and Henryon1998);each has advantages and disadvantages.For instance, complete factorial mating can produce a popula-tion with both paternal and maternal information, but results in a very large number of families (s9d families where s is the number of sires and d is the number of dams),which may be difficult to accommodate in practical programmes.The fac-torial mating by set compromises the number of families produced(from a reasonably large number of parents)and the availability of facilities allo-cated to breeding programmes.In all schemes,fac-torial matings require the use of in-vitro fertilization procedures and artificial incubation systems.However,the facilities are relatively sim-ple and the procedures have been successfully implemented in the common carp breeding pro-gramme in Vietnam(Ninh et al.2011).Family or individual identification in prawnsIn addition to the production of families in a con-trolled manner(pair mating or external fertiliza-tion),the progeny of the different families must then be identified in some way so that they can be communally stocked and tested for genetic evalua-tion and selection purposes.Passive integrated transponder(PIT)tags have been used widely and successfully for individual identification in carp and tilapia,but their use is more difficult with crustaceans.For example,in the giant freshwater prawn,M.rosenbergii,there are two major issues, especially with regard to individual identification. In prawns,physical tagging is much more difficult than infish species,due to small body size,and periodical moulting.Hung et al.(2012)have eval-uated different tagging methods:VIE(visible implant elastomer),VIA(visible implant alpha-numeric tag),eye tags and PIT tags.VIE is from Northwest Marine Technology(NMT)and is pro-vided in liquid form with10different colours.Five fluorescent colours(orange,green,blue,yellow and pink)are suitable for tagging freshwater©2015John Wiley&Sons Ltd,F I S H and F I S H E R I E S7Genetic selection in carp,tilapia and prawns N H Nguyenprawn.We have successfully tagged1–2g prawns at four different positions:on the left and right of thefirst andfifth abdominal segments.After a grow-out period of120day,the retention rate was98%and readability was100%(Hung et al. 2012).Given thesefive colours,four tagging posi-tions and multiple tags,we can tag up to256 families.If all10colours are used,600families can be tagged.The second tagging system tested was visible implant alpha(VIA),also from NMT. VIA is manufactured from soft polyester with unique numbers(the standard format being ~1.592.5mm).Tagging is straightforward but time-consuming(100prawns per hour for a skilled worker).The tag can be implanted through the hinge membrane of the second abdominal somite of1–2g prawns.The retention rate after grow-out(78%)was lower than VIE tags although readability was98%(Pillai et al.2009).Eye tags are generally unsuitable due to the short eye stalk of freshwater prawns;however,they can be used as a secondary tagging system to mark individuals at harvest.In combination with VIE,a full pedi-gree population has been maintained for this spe-cies in Vietnam.PIT tags are excellent for individual identification infish,but there are four major problems when tagging freshwater prawn: (i)tag migration from the abdomen to the head, (ii)low retention rates,(iii)injury and pigmenta-tion and(iv)difficulty in tagging small prawns.In summary,VIE and VIA can be used effectively to maintain pedigrees for M.rosenbergii because they meet the four essential criteria:(i)individual iden-tification at an early age(1–2g),(ii)high reten-tion(98%for VIE and78%for VIA)and readability rates(100and98%,respectively),(iii) easy and inexpensive to apply and(iv)harmless to the prawn and the consumer.Both VIE and VIA have been found to have negligible effect on growth and survival(Pillai et al.2009;Hung et al. 2012).Molecular-based pedigree in carpMaintaining pedigree information is very impor-tant in aquatic animals.Due to the small size of newly bornfish and prawns,it is not possible to tag them after hatching and families need to be maintained separately until they reach a suitable size for physical tagging.Separate rearing of indi-vidual families before tagging also induces envi-ronmental(i.e.hapa or tank)effects common to full-sib families.This problem can be overcome using DNA markers for parentage assignment. With genetic tagging,all families can be pooled and sent to communal rearing as soon as practica-ble after hatching.Parentage testing and pedigree verification with genetic tagging has four main advantages:(i)increasing the number of families tested without the need for extra tanks and ponds, (ii)reducing the effects common to full-sibs,(iii) shortening generation intervals(fish attain sexual maturity earlier than expected under early non-communal rearing as early communal rearing reduces the need for hapas that do not,in general, provide a favourable growing environment for juvenilefish,especially carp species)and(iv)mini-mizing the interaction between the selection and production environments as the animals can be grown under full commercial conditions.Conse-quently,DNA tagging has the potential to assist in increasing genetic gain(Ninh et al.2013).Both experimental and theoretical results show that using8–14microsatellite markers,progenies can be assigned to parents with a high degree of accu-racy(90–99%)for many aquatic species(Villanu-eva et al.2002).DNAfingerprinting is increasingly employed for genetic tagging in com-mercial breeding programmes of farmed aquacul-ture species(Estoup et al.1998;Fishback et al. 2002;Vandeputte et al.2011;Whatmore et al. 2013).Ninh et al.(2013)compared communal early rearing(CER)and separate early rearing(SER) schemes for common carp.Thefish under the CER scheme grew faster and achieved greater genetic gain than under SER(Ninh et al.2011,2013). The superiority of CER relative to SER demon-strated potential benefits of molecular parentage assignment as a useful tool in practical selective breeding programmes.Although the current costs of genotyping are high,they are expected to decline significantly with the development of DNA chips where thousands of single nucleotide poly-morphisms(SNPs)can be genotyped simulta-neously.SNPs can be used successfully to replace microsatellites for the purposes of parentage assignment and pedigree verification due to their lower error rate in comparison to microsatellite markers(Trọng et al.2013b).However,two to eight times more SNPs than microsatellites are required to obtain the same power of successful assignment for traceability(Weller et al.2006;Ha-user et al.2011).Despite the advantages of genetic tagging,the technique does not completely8©2015John Wiley&Sons Ltd,F I S H and F I S H E R I E S Genetic selection in carp,tilapia and prawns N H Nguyen。