Genetic parameters for survival during the grow-out period in the GIFT strain of Nile tilapia (Oreochromis niloticus) and correlated response to selection for harvest weight

Aquaculture Research, 2015, 1–9 doi:10.1111/are.12859 Genetic parameters for survival during the grow-out period in the GIFT strain of Nile tilapia (Oreochromis niloticus) and correlated response to selection for harvest weight Azhar Hamzah1,2, Wagdy Mekkawy3,4, Hooi Ling Khaw3, Nguyen Hong Nguyen5, Hoong Yip Yee3, Khairul Rizal Abu Bakar3, Siti Azizah Mohd Nor1 & Raul W Ponzoni6 1 School of Biological Sciences, University Science Malaysia, Minden, Penang, Malaysia 2 National Prawn Fry Production & Research Center, Kota Kuala Muda, Kedah, Malaysia 3 World Fish, Penang, Malaysia Faculty of Agriculture, Animal Production Department, Ain Shams University, Cairo, Egypt 4 5 School of Science, Education and Engineering, University of the Sunshine Coast, Maroochydore, QLD, Australia 6 Facultad de Agronomia, Departamento de Produccion Animal, Universidad de la Rep
ublica, Montevideo, Uruguay Correspondence: W Mekkawy, World Fish, Jalan Batu Maung, 11960 Penang, Malaysia. E-mail: w.mekkawy@cgiar.org Abstract The aims of this study were the estimation of genetic parameters for survival rate from tagging until harvest and the evaluation of the correlated response in survival rate to selection for harvest weight in the genetically improved farmed tilapia (GIFT) strain. The heritability for survival rate was low (0.038), and so was its genetic correlation with harvest weight (0.065), suggesting that selecting for the latter trait would have had no effect on survival. The calculation of the probability of survival by spawning season and line, fitting a model that included the random effects of individual animal and common environment, confirmed this prediction. There were very small and variable between line differences in the probability of survival, which generally favoured the selection line. We conclude that the focus of the GIFT programme on improving harvest weight was not detrimental to the survival of the fish during the grow-out phase. Keywords: survival, GIFT, heritability, genetic correlation, selection response Introduction Increasing the production of high quality seed from genetically improved strains is a key factor in © 2015 John Wiley & Sons Ltd the development and long term sustainability of aquaculture industries. The overall reproductive rate of such strains impacts upon the economic efficiency of both selective breeding programmes and hatchery operations. In Nile tilapia, Ponzoni, Nguyen and Khaw (2007) showed that the reproductive rate in tilapia had the greatest impact on the economic benefit derived from the genetic improvement programmes. The genetically improved farmed tilapia (GIFT) strain of Nile tilapia (Oreochromis niloticus) has been under selection for improved growth rate for 16 generations (six generations in the Philippines and 10 generations in Malaysia; Ponzoni, Nguyen, Khaw & Hamzah 2011). In the grow-out operations both the growth rate and survival of the fish have a major impact on profitability. There is limited but encouraging information (Santos, Ribeiro, Vargas, Mora, Filho, Fornari & Oliveira 2011; Ninh, Thoa, Knibb & Nguyen 2014) on the effects of selection for greater growth rate on survival during grow-out in tilapia. Anyhow, in long term breeding programmes, fitness-related traits such as survival rate may decline (Rauw, Kanis, Noordhuizen-Stassen & Grommers 1998) and their monitoring should be an integral part of such programmes. The aim of this study was to examine genetic variation in the survival rate during grow-out (stocking of the fish in ponds to harvest) and to 1 Genetic parameters for survival of GIFT at harvest A Hamzah et al. estimate the correlated response in this trait to the selection for high growth rate in the GIFT strain of Nile tilapia (O. niloticus). Materials and methods The fish and data structure The study was based on the data and pedigree information of the GIFT breeding programme in Malaysia, where it was established in 2001. To date, its main aim has been to improve growth rate. Details on the development of the strain and achieved genetic gains can be found in Ponzoni et al. (2011). Table 1 shows the number of sires, dams and progeny by spawning season and selection line (Control and Selection). Family production, rearing and selection procedures Two lines were created from the 2002 progeny, one selected for average breeding values (Control line) for live weight, and another one selected for high breeding values (Selection line) for that trait. Further details on the establishment of the GIFT strain in Malaysia, on selection procedures and mate allocation strategy are given in Hamzah Table 1 Number of sires, dams and progeny by spawning season by line Spawning season (generation) 2002 (1) 2003 (2) 2004 (3) 2005 (4) 2006 (5) 2007 (6) 2008 (7) 2009 (8) 2010 (9) 2011 (10) Line Sires Dams Progeny – C S C S C S C S C S C S C S C S C S 52 19 35 17 54 13 42 10 49 15 41 14 52 9 51 8 52 10 55 54 19 65 22 84 20 76 15 88 15 71 14 76 11 69 8 70 10 66 4261 1885 6171 2453 9938 804 3092 589 3473 1084 5073 988 5233 792 5106 474 4121 658 4424 C, control line; S, selection line. 2 Aquaculture Research, 2015, 1–9 (2006), Ponzoni, Hamzah, Saadiah and Kamaruzzaman (2005), Ponzoni, Khaw, Nguyen and Hamzah (2010), Ponzoni et al. (2011) and Hamzah, Ponzoni, Nguyen, Khaw, Yee and Azizah (2014). Mating of selected breeders in each spawning season was performed in 1 m3 nylon hapas of 2-mm mesh size, installed in an earthen pond. One male was mated to two females in the Selection line (nested mating design), whereas in the Control line one male was mated to one female (single pair mating). The females were placed in the breeding hapas before the males. Only ‘ready to spawn’ (Longalong, Eknath & Bentsen 1999) females were paired with a male in the hapa. After a week of mating, fertilized eggs were collected from the mouth of the female and immediately transferred to hatching jars where they remained 3–5 days until hatching. The date of spawning was recorded for each individual pair mated. In the Selection line, males were then paired with a second female in another hapa. The hatched fry of each family were transferred from the incubators to nursery hapas (1 m3 with 2 mm mesh size), stocked at a density of 200 fry per m3. At least three nursery hapa replicates of each family were maintained in the same pond. When the fingerlings reached an average weight of 5 g, 100 individuals per family were randomly sampled and tagged. The base population was identified using a passive integrated transponder (PIT) tag. In the 2002 and 2003 spawning seasons, Floyâ tags were used, whereas in the 2004 spawning season Floyâ tags (100 individuals per family) and T-bar anchor tags (20 individuals per family) were used. Due to the low retention rate of the Floyâ and T-bar tag, PIT tags were used from the 2005 spawning season onwards. The low retention rate of Floyâ and Tbar tags resulted in the confounding of tag losses with mortality. For that reason in this study we only use data from the spawning seasons where PIT tags were used (i.e. from 2005 onwards). In all spawning seasons, the tag number, sex and live weight were recorded at harvest. The tagged individuals of the 2005 and subsequent spawning seasons were stocked in ponds for performance testing. The fish were harvested at 200–450 g live weight (after 120 days grow-out). Data analysis Survival from stocking until harvest was treated as a binary trait where the fish that were present at © 2015 John Wiley & Sons Ltd, Aquaculture Research, 1–9 Aquaculture Research, 2015, 1–9 Genetic parameters for survival of GIFT at harvest A Hamzah et al. harvest were coded as ‘1’, whereas they were coded as ‘0’ if not present and presumed dead. Records from a total of 35 910 individuals, corresponding to spawning seasons 2005–2011, were used for the analysis of survival. As earlier mentioned in Section 2.2, the data from the first three spawning seasons were discarded because of the low retention rate of the Floyâ tags which resulted in a confounding between fish mortality and tag losses. Harvest weight and survival between stocking and harvest were routinely recorded in both the Control and Selection lines, which enabled the estimation of the genetic correlation between both traits. Binary logistic regression was one of the procedures used in the analysis of survival from stocking to harvest. SPSS software (Statistical Package for the Social Sciences, 2011) was used for this purpose. The statistical model included line (Control and Selection), spawning season (seven levels, 2005–2011), their two-way interaction and age at stocking nested within the line and spawning season as a linear covariate. The correlated response (to selection for harvest weight) in survival was estimated from the differences between least squares means of the Selection and Control lines. For the estimation of genetic parameters for harvest weight and survival, a bivariate Bayesian linear-threshold model was fitted. The threshold probit model assumes that survival is determined by an underlying continuous variable (liability). A threshold point links the liability and the categorical expression of the trait. The assumptions of the probit model of survival were that the threshold point was equal to 0 and that the residual variance was fixed at 1. To satisfy the assumptions of normality and homogeneity of variance, a square root transformation of harvest weight was carried out. The model for survival included the same ‘fixed effects’ as those fitted in the binary logistic regression described above (namely, line, spawning season, their interaction and age at stocking nested within the line and spawning season as a linear covariate). For harvest weight the model fitted included: line (Control, Selection), spawning season (seven levels, 2005–2011), sex (Female, Male), their two-way interactions and harvest age nested within the line and spawning season as a linear covariate. The random effects fitted were the same for both traits, namely, the additive genetic effect of the fish and the common environmental effect in full sib groups. In matrix notation the model is as follows: © 2015 John Wiley & Sons Ltd, Aquaculture Research, 1–9 y ¼ Xb þ Za þ Wc þ e where y is a vector that includes the underlying continuous variable of survival and the square root of the observed phenotypes of harvest weight; X, Z and W are incidence matrices related to the ‘fixed’, additive genetic and common environmental full sib effects, respectively, b is the vector of ‘fixed effects’, a is the additive genetic effect of individual animal for the traits studied, c is the vector of common environmental full sib effects (i.e. the effect of separate rearing of full sib families in the nursing hapas) and e is the vector of the residual environmental effect. The variance– covariance structure can be written as: 0 1 0 a AbG V@ c A ¼ @ 0 e 0 1 0 0 IbC 0 A 0 IbR where A is the additive relationship matrix, G is the genetic variance–covariance matrix, C and R are the variance–covariance matrices of the common environmental effect and residual environmental effect, respectively, I is the identity matrix and ⊗ denotes Kronecker product. Flat priors were assigned to all the ‘fixed effects’ with b / constant. The prior distributions of the additive genetic, common environmental and residual effects were assumed to follow multivariate normal distributions with ajA; G  Nð0; AbGÞ cjI; C  Nð0; IbCÞ ejI; R  Nð0; IbRÞ respectively. Conjugate priors were assumed for the variance–covariance matrices of G, C and R using inverse Wishart distribution (IW) with Gjva ; Va  IWðva Va ; va Þ Cjvc ; Vc  IWðvc Vc ; vc Þ Rjve ; Ve  IWðve Ve ; ve Þ respectively. The Va, Vc and Ve are the scale parameters of IW for additive genetic, common environmental 3 Genetic parameters for survival of GIFT at harvest A Hamzah et al. and residual variance-covariance, respectively, whereas ma, mc and me are degrees of freedom, also known in a Bayesian context as degrees of belief. We assumed ma = mc = me = 6 for vague priors. Gibbs sampling was used to obtain the marginal posterior distributions of the genetic and environmental parameters. We used a single chain of 150 000 iterations, 20 000 iterations were considered as burn-in and the lag between iterations was 20. Bayesian analysis was carried out using the software THRIGIBBSF90 (Misztal, Tsuruta, Strabel, Auvray, Druet & Lee 2002). The convergence of the analysis was checked using the Raftery and Lewis (1992) algorithm. The expected breeding values of survival from the threshold model were used to estimate the probability of survival of each fish through the cumulative function of the normal distribution. Then, the correlated response in survival was estimated from the differences between posterior means of the probabilities of survival for the Control and Selection lines. The correlated changes in survival as a consequence of selection for increased harvest weight were examined in two ways, namely, by comparing the least squares means for survival in the Control and Selection lines, and by calculating the probability of survival from the estimated breeding values using the cumulative function of the normal distribution in both lines within each spawning season. Results Descriptive statistics Descriptive statistics for harvest weight, survival from stocking to harvest, stocking age and harvest age are shown in Table 2. Genetic parameters for survival and harvest weight The posterior means of the genetic parameters and their posterior standard deviation for the survival and harvest weight are presented in Table 3. The heritability of survival was low (0.038). The genetic correlation between the survival and harvest weight was positive but very low (0.065). The common environmental effect made a greater contribution to the total variation in survival than the additive genetic effect. There was a positive and 4 Aquaculture Research, 2015, 1–9 Table 2 Descriptive statistics for HW, survival rate between SH, SA and HA in the GIFT strain Traits Line N Mean SD CV (%) HW (g) C S C S C S C S 3813 20 091 5389 30 521 5389 30 521 3813 20 091 167.60 236.70 70.80 65.80 97.10 98.98 229.20 236.90 71.98 78.59 45.50 47.40 18.70 19.47 32.58 25.89 42.95 33.20 64.30 72.11 19.22 19.67 14.21 10.93 SH (%) SA (days) HA (days) HW, harvest weight; SH, stocking and harvest; SA, stocking age; HA, harvest age; GIFT, genetically improved farmed tilapia; N, number of observations; C, control line; S, selection line; SD, standard deviation; CV, coefficient of variation. Table 3 Posterior means and standard deviations of the genetic parameters for harvest weight and survival fitting a bivariate linear-threshold model Trait component Harvest weight r2a 1.065 0.014 0.226 0.065 2.025 0.089 Genetic covariance h2 Genetic correlation r2c Common environmental covariance c2 Common environmental correlation       0.136 0.006 0.026 0.026 0.149 0.028 0.430  0.019 0.148  0.045 Survival 0.047  0.002 0.038  0.002 0.178  0.011 0.145  0.008 r2a = additive genetic variance, r2c = common environmental variance, h2 = heritability, c2 = common environmental effect as a proportion of the phenotypic variance. very low common environmental correlation between the survival and harvest weight. Phenotypic selection response The significance levels of the fixed effects fitted in the logistic regression model and least squares means of the survival rate are presented in Tables 4 and 5 respectively. The least squares mean for survival rate in the Selection line was lower (67%) than in the Control line (71%). Genetic selection response The correlated genetic selection responses in the survival rate across generations are presented as the posterior means of the probability of survival by generation by line in Table 6. Across all spawning © 2015 John Wiley & Sons Ltd, Aquaculture Research, 1–9 Aquaculture Research, 2015, 1–9 Genetic parameters for survival of GIFT at harvest A Hamzah et al. Table 4 Fixed effects fitted in the logistic regression analysis and the level of significance Effect d.f. Wald v2 P > v2 Spawning season Line Spawning season 9 line Stocking age within spawning season and line 6 1 6 14 63.18 13.24 57.31 275.40 0.001 0.001 0.001 0.001 Table 5 Least squares means (LSM) and standard errors (SE) of the survival rate between stocking and harvest for the spawning season and line effects Effect Spawning season (generation) 2005 (4) 2006 (5) 2007 (6) 2008 (7) 2009 (8) 2010 (9) 2011 (10) Line C S LSM SE 0.72c 0.87a 0.82b 0.55e 0.52e 0.60d 0.64d 0.01 0.01 0.01 0.01 0.01 0.02 0.02 0.71a 0.67b 0.01 0.00 The least squares means with a common superscript do not differ significantly (P > 0.05). C, control line; S, selection line; LSM, least squares means; SE, standard errors. Table 6 Posterior means and their posterior standard deviations (within parentheses) of the probability of survival by spawning season for the control and selection lines Spawning season (generation) Control line Selection line 2005 2006 2007 2008 2009 2010 2011 0.666 0.673 0.675 0.678 0.684 0.688 0.690 0.672 0.677 0.680 0.680 0.687 0.692 0.697 (4) (5) (6) (7) (8) (9) (10) (0.014) (0.013) (0.010) (0.010) (0.011) (0.010) (0.012) (0.019) (0.018) (0.018) (0.019) (0.023) (0.027) (0.023) seasons there were no significant differences between the Selection line and the Control line. Discussion Overall survival In a breeding programme, survival is an important fitness trait because it affects the number and the © 2015 John Wiley & Sons Ltd, Aquaculture Research, 1–9 total weight of fish at harvest. The average survival rate of 67% recorded in the GIFT strain across spawning seasons was low but comparable to the survival rate of Nile tilapia reported in other studies (Bolivar & Newkirk 2002; Charo-Karisa, Komen, Rezk, Ponzoni, van Arendonk & Bovenhuis 2006; Trọng, Mulder, van Arendonk & Komen 2013). Although steps were taken to reduce variation in ambient temperature and water parameters in the grow-out ponds, fluctuations between spawning seasons could not be avoided. Most likely, this contributed to the large coefficient of variation (CV) of survival. In an aquaculture population where individuals are grown in a pond, the highly competitive ones will have advantages such as their capability in gaining access to food resources and spaces. This could result in a large CV of body weight in the population which in turn could result in a reduction in the productivity. There is evidence showing that selection for high growth rate could have increased aggressiveness in the population (Lahti, Laurila, Enberg & Piironen 2001; Weber & Fausch 2003). Therefore, in this study, the large CV and low mean survival may have been due to competition effects among individuals (Jobling 1995; Adams, Huntingford, Turnbull, Arnott & Bell 2000). A large scale experiment to investigate the genetic basis for social interaction was conducted using the GIFT population (H.L. Khaw, R.W. Ponzoni, H.Y. Ye, M.A. Aziz & P. Bijma, unpubl. data). The authors found heritable competitive effects for harvest weight in this population. Genetic variation The low heritability estimate we obtained for survival in the GIFT strain indicates that improving the trait through selective breeding would be difficult. The estimate was lower than the results reported by Charo-Karisa et al. (2006) who observed heritabilities of 0.35–0.77 for survival till harvest in Nile tilapia. Note, however, that CharoKarisa’s estimates are most likely overestimates for some reason (one would expect low heritability values for a fitness trait). Nevertheless, additive genetic variation in survival was also observed in other farmed aquaculture species such as rainbow trout (Vehvil€ ainen, Kause, Quinton, Koskinen & Paananen 2008; Vehvil€ ainen, Kause, Koskinen & Paananen 2010), common carp (Nielsen, Ødeg ard, 5 Genetic parameters for survival of GIFT at harvest A Hamzah et al. Olesen, Gjerde, Ardo, Jeney & Jeney 2010), oyster (Ernande, Clobert, Mccombie & Boudry 2003) and shrimp (Kenway, Macbeth, Salmon, McPhee, Benzie, Wilson & Knibb 2006). The heritability estimates for survival in these studies varied with age of the fish and culture environment (e.g. 0.0– 0.16). The dam and common environmental full sib effect on survival trait was significant, as typically observed in fish (Rana 1988; Marteinsdottir & Steinarsson 1998). The effect in Nile tilapia is due to the incubation of the fertilized eggs in the females’ mouth until hatching, and to the separate rearing of different full sib families in their respective hapas until the fish can be individually tagged. Thus, not accounting for these effects may result in an upward bias in heritability. Correlations between harvest weight and survival The genetic correlation between harvest weight and survival was positive but very low, suggesting that selection for high growth rate would not affect the survival rate during the grow-out period. Positive genetic correlations were reported in other studies in Nile tilapia (Charo-Karisa et al. 2006; Maluwa & Gjerde 2007; Luan, Olesen, Ødegard, Kolstad & Dan 2008; Rezk, Ponzoni, Khaw, Kamel, Dawood & John 2009; Santos et al. 2011; Ninh et al. 2014). The same Aquaculture Research, 2015, 1–9 trend was observed in salmonids (Rye, Lillevik & Gjerde 1990; Jonasson 1993), common carp (Nielsen et al. 2010; Vehvil€ ainen, Kause, Kuukka-Anttila, Koskinen & Paananen 2012), oysters (D
egremont, Ernande, B
edier & Boudry 2007) and shrimp (Gitterle, Rye, Salte, Cock, Johansen, Lozano, Su
arez & Gjerde 2005; Krishna, Gopikrishna, Gopal, Jahageerdar, Ravichandran, Kannappan, Pillai, Paulpandi, Kiran, Saraswati, Venugopal, Kumar, Gitterle, Lozano, Rye & Hayes 2011). By contrast, negative genetic correlations between survival and body weight were reported in the rainbow trout (Rye et al. 1990), Pacific oyster (Evans & Langdon 2006) and black tiger shrimp (Kenway et al. 2006). The reasons for the lack of agreement among studies are difficult to establish precisely, but are most likely due to a combination of species-specific factors and the particular set of circumstances in which the studies were conducted. Correlated changes in survival to selection for increased harvest weight Genetic gain in the harvest weight was continuous throughout the study period (Ponzoni et al. 2011). The least squares means for survival in the Control and Selection lines are presented in Table 5, whereas the probabilities of survival from 1 Control 0.9 Selection 0.8 0.7 Survival 0.6 0.5 0.4 0.3 0.2 0.1 Spawning season (generation) 6 2011 (10) 2010 (9) 2009 (8) 2008 (7) 2007 (6) 2006 (5) 2005 (4) 0 Figure 1 Least squares means ( SE), by spawning season (generation) and line, for survival between stocking and harvest fitting a logistic regression model. © 2015 John Wiley & Sons Ltd, Aquaculture Research, 1–9 Aquaculture Research, 2015, 1–9 Genetic parameters for survival of GIFT at harvest A Hamzah et al. the estimated breeding values using the cumulative function of the normal distribution in both lines within each spawning season are shown in Table 6. With the first approach (least squares means) only fixed effects were fitted. Because the interaction between the spawning season and line was significant, the line effect alone is not informative. An examination of the least squares means in the spawning season by line subclasses shows that both the rank and the magnitude of the differences between lines vary between the spawning seasons (Fig. 1), accounting for the significant spawning season by line interaction in Table 4. The model fitted with the second approach was more complex, it included two random effects (in addition to the error term), namely, the individual animal effect and the common environmental effect in full sib groups. Hence, one may argue that with this approach one may be better able to gauge the effect the selection for harvest weight may have had on survival. The probability of survival was only marginally superior in the Selection than in the Control line, indicating that, consistent with the genetic parameter estimates, there had not been an undesirable correlated response in survival (Table 6). Concluding remarks The heritability of survival during the period spanning from stocking in pond to harvest was low. The genetic correlation between survival in that period and harvest weight was positive but very low. Consistent with these genetic parameter values we found no indication of an undesirable correlated response in survival when we fitted a model that included the random effects of the individual and the common environment. We conclude that the focus of the GIFT programme on improving harvest weight was not detrimental to the survival of the fish during the grow-out phase. The genetic gain in harvest weight during the period studied was continuous generation after generation, and of the order of 100% (Ponzoni et al. 2011). Note, however, that our study was circumscribed to survival between stocking in pond and harvest, so that we do not know if survival in other phases of development (e.g. hatching to stocking, post-harvest) may have been affected by selection for harvest weight. © 2015 John Wiley & Sons Ltd, Aquaculture Research, 1–9 Acknowledgments The GIFT breeding programme in Malaysia is a collaboration between the Department of Fisheries, Malaysia and WorldFish and is funded by the European Union and the Department of Fisheries, Malaysia. This work was partially funded by the CGIAR Research Program on Livestock and Fish. References Adams C.E., Huntingford F.A., Turnbull J.F., Arnott S. & Bell A. (2000) Size heterogeneity can reduce aggression and promote growth in Atlantic salmon parr. Aquaculture International 8, 543–549. Bolivar R.B. & Newkirk G.F. (2002) Response to within family selection for live weight in Nile tilapia (Oreochromis niloticus) using a single-trait animal model. Aquaculture 204, 371–381. Charo-Karisa H., Komen H., Rezk M.A., Ponzoni R.W., van Arendonk J.A.M. & Bovenhuis H. (2006) Heritability estimates and response to selection for growth of Nile tilapia (Oreochromis niloticus) in low-input earthen ponds. Aquaculture 261, 479–486. D
egremont L., Ernande B., B
edier E. & Boudry P. (2007) Summer mortality of hatchery produced Pacific oyster spat (Crassostrea gigas). I. Estimation of genetic parameters for survival and growth. Aquaculture 262, 41– 53. Ernande B., Clobert J., Mccombie H. & Boudry P. (2003) Genetic polymorphism and trade-offs in the early lifehistory strategy of the Pacific oyster, Crassostrea gigas (Thunberg, 1795): a quantitative genetic study. J. Evol. Bio. 16, 399–414. Evans S. & Langdon C. (2006) Direct and indirect responses to selection on individual body weight in the Pacific oyster (Crassostrea gigas). Aquaculture 261, 546–555. Gitterle T., Rye M., Salte R., Cock J., Johansen H., Lozano C., Su
arez J.A. & Gjerde B. (2005) Genetic (co)variation in harvest body weight and survival in Penaeus (Litopenaeus) vannamei under standard commercial conditions. Aquaculture 243, 83–92. Hamzah A. (2006) Genetic improvement of tilapia (Oreochromis niloticus) through selective breeding and crossbreeding. MSc thesis, Universiti Sains Malaysia, Penang, Malaysia 77pp. Hamzah A., Ponzoni R.W., Nguyen N.H., Khaw H.L., Yee H.Y. & Azizah S.M.N. (2014) Genetic evaluation of the Genetically Improved Farmed Tilapia (GIFT) strain over ten generations of selection in Malaysia. Pertanikan J. Trop. Agric. Sci. 37, 411–429. Jobling M. (1995) Simple indices for the assessment of the influences of the social environment on growth performance, exemplified by studies on Arctic charr. Aquaculture International 3, 60–65. 7 Genetic parameters for survival of GIFT at harvest A Hamzah et al. Jonasson J. (1993) Selection experiments in salmon ranching: I. Genetic and environmental sources of variation in survival and growth in freshwater. Aquaculture 109, 225–236. Kenway M., Macbeth M., Salmon M., McPhee C., Benzie J., Wilson K. & Knibb W. (2006) Heritability and genetic correlations of growth and survival in black tiger prawn Penaeus monodon reared in tanks. Aquaculture 259, 138–145. Krishna G., Gopikrishna G., Gopal C., Jahageerdar S., Ravichandran P., Kannappan S., Pillai S.M., Paulpandi S., Kiran R.P., Saraswati R., Venugopal G., Kumar D., Gitterle T., Lozano C., Rye M. & Hayes B. (2011) Genetic parameters for growth and survival in Penaeus monodon cultured in India. Aquaculture 318, 74–78. Lahti K., Laurila A., Enberg K. & Piironen J. (2001) Variation in aggressive behaviour and growth rate between populations and migratory forms in the brown trout, Salmo trutta. Animal Behaviour 62, 935–944. Longalong F.M., Eknath A.E. & Bentsen H.B. (1999) Response to bidirectional selection for frequency of early maturing females in Nile tilapia (Oreochromis niloticus). Aquaculture 178, 13–25. Luan T.D., Olesen I., Ødegard J., Kolstad K. & Dan N.C. (2008) Genotype by environment interaction for harvest body weight and survival of Nile tilapia (Oreochromis niloticus) in brackish and fresh water ponds. In: Proceedings of 8th International Symposium on Tilapia in Aquaculture, Cairo, Egypt, pp. 231–240. Maluwa A.O. & Gjerde B. (2007) Response to selection for harvest body weight of Oreochromis shiranus. Aquaculture 273, 33–41. Marteinsdottir G. & Steinarsson A. (1998) Maternal influence on the size and viability of Iceland cod Gadus morhua eggs and larvae. Journal of Fish Biology 52, 1241– 1258. Misztal I., Tsuruta S., Strabel T., Auvray B., Druet T. & Lee D.H. (2002) BLUF90 and related programs (BGF90). Proceedings of 7th World Congress on Genetics Applied to Livestock Production, Communication No. 28– 07, Montpellier, France. Nielsen H.M., Ødeg ard J., Olesen I., Gjerde B., Ardo T., Jeney G. & Jeney Z. (2010) Genetic analysis of common carp (Cyprinus carpio) strains I: genetic parameters and heterosis for growth traits and survival. Aquaculture 304, 14–21. Ninh N.H., Thoa N.P., Knibb W. & Nguyen N.H. (2014) Selection for enhanced growth performance of Nile tilapia (Oreochromis niloticus) in brackish water (15–20 ppt) in Vietnam. Aquaculture 428, 1–6. doi:10.1016/ j.aquaculture.2014.02.024. Ponzoni R.W., Hamzah A., Saadiah T. & Kamaruzzaman N. (2005) Genetic parameters and response to selection for live weight in the GIFT strain of Nile Tilapia (Oreochromis niloticus). Aquaculture 247, 203–210. 8 Aquaculture Research, 2015, 1–9 Ponzoni R.W., Nguyen N.H. & Khaw H.L. (2007) Investment appraisal of genetic improvement programs in Nile tilapia (Oreochromis niloticus). Aquaculture 269, 187–199. Ponzoni R.W., Khaw H.L., Nguyen N.H. & Hamzah A. (2010) Inbreeding and effective population size in the Malaysian nucleus of the GIFT strain of Nile tilapia (Oreochromis niloticus). Aquaculture 302, 42– 48. Ponzoni R.W., Nguyen N.H., Khaw H.L. & Hamzah A. (2011) Genetic improvement of Nile tilapia (Oreochromis niloticus) with special reference to the work conducted by the WorldFish Center with the GIFT strain. Reviews in Aquaculture 3, 27–41. Raftery A.E. & Lewis S.M. (1992) How many iterations in the Gibbs sampler? In: Bayesian Statistics IV (ed. by J.M. Bernardo, J.O. Berger, A.P. Dawid & A.F.M. Smith), pp. 763–774. Oxford University Press, Oxford, UK. Rana K. (1988) Reproductive biology and the hatchery rearing of tilapia eggs and fry. In: Recent Advances in Aquaculture, Vol. 3 (ed. by J.F. Muir & R.J. Roberts), pp. 343–406. Croom Helm/Timber Press, London, UK/ Portland, OR, USA. Rauw W.M., Kanis E., Noordhuizen-Stassen E.N. & Grommers F.J. (1998) Undesirable side effects of selection for high production efficiency in farm animals: a review. Livestock Production Science 56, 15–33. Rezk M.A., Ponzoni R.W., Khaw H.L., Kamel E.A., Dawood T.I. & John G. (2009) Selective breeding for increased live weight in a synthetic breed of Egyptian Nile tilapia, Oreochromis niloticus: response to selection and genetic parameters. Aquaculture 293, 187–194. Rye M., Lillevik K.M. & Gjerde B. (1990) Survival in early life of Atlantic salmon and rainbow trout: estimates of heritabilities and genetic correlations. Aquaculture 89, 209–216. Santos A.I., Ribeiro R.P., Vargas L., Mora F., Filho L.A., Fornari D.C. & Oliveira S.N. (2011) Bayesian genetic parameters for body weight and survival of Nile tilapia farmed in Brazil. Pesquisa Agropecuaria Brasileira 46, 33–43. Statistical Package for the Social Sciences (SPSS) Inc. (2011) IBM SPSS Statistics for Windows, Version 19.0. IBM Corp, Armonk, NY, USA. Trọng T.Q., Mulder H.A., van Arendonk J.A.M. & Komen H. (2013) Heritability and genotype by environment interaction estimates for harvest weight, growth rate, and shape of Nile tilapia (Oreochromis niloticus) grown in river cage and VAC in Vietnam. Aquaculture 384– 387, 119–127. Vehvil€ ainen H., Kause A., Quinton C.D., Koskinen H. & Paananen T. (2008) Survival of the currently fittest-genetics of rainbow trout survival across time and space. Genetics 180, 507–516. © 2015 John Wiley & Sons Ltd, Aquaculture Research, 1–9 Aquaculture Research, 2015, 1–9 Genetic parameters for survival of GIFT at harvest A Hamzah et al. Vehvil€ainen H., Kause A., Koskinen H. & Paananen T. (2010) Genetic architecture of rainbow trout survival from egg to adult. Genet. Res. Camb. 92, 1–11. doi:10.1017/S0016672310000017. Vehvil€ainen H., Kause A., Kuukka-Anttila H., Koskinen H. & Paananen T. (2012) Untangling the positive genetic correlation between rainbow trout growth and © 2015 John Wiley & Sons Ltd, Aquaculture Research, 1–9 survival. Evolutionary Applications 5, 732–745. doi:10.1111/j.1752-4571.2012.00251.x. Weber E.D. & Fausch K.D. (2003) Interactions between hatchery and wild salmonids in streams: differences in biology and evidence for competition. Canadian Journal of Fisheries and Aquatic Science 60, 1018–1036. 9