Study of the second generation of gilthead sea bream (Sparus aurata L.) within the genetic improvement program Progensa for commercial interest traits and implementation of new industrial enabling technologies under culture conditions

� UNIVERSIDAD DE LAS PALMAS f � DE GRAN CANARIA Anexo I ECOAQUA ._\l\\"tN(C�U Dr. Daniel Montero Vítores, COORDINADOR DEL PROGRAMA DE DOCTORADO ACUICULTURA: PRODUCCIÓN CONTROLADA DE ANIMALES ACUÁTICOS, DE LA UNIVERSIDAD DE LAS PALMAS DE GRAN CANARIA, CERTIFICA, Que la Comisión Académica del programa de doctorado, en su sesión, tomó el acuerdo de dar el consentimiento para su tramitación, a la tesis doctoral titulada "ESTUDIO DE LA SEGUNDA GENERACIÓN DE DORADA (Sparus aurata L.) DEL PROGRAMA DE MEJORA GENÉTICA PROGENSA ® PARA CARACTERES DE INTERÉS COMERCIAL E IMPLEMENTACIÓN DE NUEVAS TECNOLOGÍAS FACILITADORAS, BAJO CONDICIONES INDUSTRIALES DE CULTIVO" Presentada por el doctorando D ISLAM SAID ELALFY ELKHOLY y dirigida por D. JUAN MANUEL AFONSO LOPEZ y D. MANUEL MANCHADO CAMPAÑA. Y para que así conste, y a efectos de lo previsto en el Artº 6 del Reglamento para la elaboración, defensa, tribunal y evaluación de tesis doctorales de la Universidad de Las Palmas de Gran Canaria, firmo la presente en Las Palmas de Gran Canaria. ,., UNIVERSIDAD DE LAS PALMAS f 'lil( DE GRAN CANARIA Anexo I ECOAQUA W\VN'f t t J D. Rafael Robaina Romero, VICERRECTOR DE TITULOS Y DOCTORADO, DE LA UNIVERSIDAD DE LAS PALMAS DE GRAN CANARIA, CERTIFICA, Que la Comisión Académica del programa de doctorado ACUICULTURA: PRODUCCIÓN CONTROLADA DE ANIMALES ACUATICOS, tomó el acuerdo de dar el consentimiento para su tramitación, a la tesis doctoral titulada "ESTUDIO DE LA SEGUNDA GENERACIÓN DE DORADA (Sparus aurata L.) DEL PROGRAMA DE MEJORA GENÉTICA PROGENSA® PARA CARACTERES DE INTERÉS COMERCIAL E IMPLEMENTACIÓN DE NUEVAS TECNOLOGÍAS FACILITADORAS, BAJO CONDICIONES INDUSTRIALES DE CULTIVO" Presentada por el doctorando D. ISLAM SAID ELALFY ELKHOLY, y dirigida por D. JUAN MANUEL AFONSO LOPEZ y D. MANUEL MANCHADO CAMPAÑA. Y para que así conste, y a efectos de lo previsto en el Artº 6 del Reglamento para la elaboración, defensa, tribunal y evaluación de tesis doctorales de la Universidad de Las Palmas de Gran Canaria, firmo la presente en Las Palmas de Gran Canaria, a 20 de junio de dos mil dieciséis. ,..., UNIVERSIDAD DE LAS PALMAS f "" DE GRAN CANARIA Anexo 11 INSTITUTO UNIVERSITARIO DE ACUICULTURA SOSTENIBLE Y ECOSISTEMAS MARINOS (IU-ECOAQUA) Programa de Doctorado en Acuicultura: Producción Controlada de Animales Acuáticos Título de la Tesis "ESTUDIO DE LA SEGUNDA GENERACION DE DORADA (Sparus aurata L.) DEL PROGRAMA DE MEJORA GENÉTICA PROGENSA ® PARA CARACTERES DE INTERÉS COMERCIAL E IMPLEMENTACIÓN DE NUEVAS TECNOLOGÍAS F ACILIT ADORAS BAJO CONDICIONES INDUSTRIALES DE CULTIVO" Tesis Doctoral presentada por D. ISLAM SAID ELALFY ELKHOL Y Dirigida por el Dr. D. JUAN MANUEL AFONSO LO PEZ Dirigida por el Dr. D. MANUEL MANCHADO CAMPAÑA Las Palmas de Gran Canaria, a 10 de Junio de 2016 � UNIVERSIDAD DE LAS PALMAS / "' DE GRAN CANARIA Annex I ECOAQUA WWW ECOAQVA.f U Dr. Daniel Montero, COORDINATOR OF THE DOCTORAL PROGRAM, AQUACULTURE: CONTROLLED PRODUCION OF AQUATIC ANIMALS OF THE UNIVERSITY OF LAS PALMAS DE GRAN CANARIA, CERTIFIES, That the Academic Committee of the doctoral program, at its meeting, took the agreement to give consent for processing, the doctoral thesis entitled "STUDY OF THE SECOND GENERATION OF GILTHEAD SEA BREAM (Sparus aurata L.) WITHIN THE GENETIC IMPROVEMENT PROGRAM PROGENSA ® FOR COMMERCIAL IMPLEMENTATION OF NEW INTEREST INDUSTRIAL TRAITS AND ENABLING TECHNOLOGIES, UNDER CULTURE CONDITIONS" Presented by Mr. ISLAM SAID ELALFY ELKHOL Y and directed by Dr. JUAN MANUEL AFONSO LOPEZ and Dr. MANUEL MANCHADO CAMPAÑA. And for the record, and for the purposes of the provisions of the Artº 6 of Regulation for the preparation, defence, court and evaluation of doctoral theses at the University of Las Palmas de gran canaria, I sign this in Las Palmas de Gran Canaria, at June, 20th 2016 � UNIVERSIDAD DE LAS PALMAS r " DE GRAN CANARIA Annex I ECO, JUA '• ,'.�·.¡E· 4-., E - Mr. Rafael Robaina Romero VICECHANCELLOR OF DEGREE AND POSTGRADUATE OF THE UNIVERSITY OF LAS PALMAS DE GRAN CANARIA, CERTIFICAT, That the Academic Committee of the doctoral program AQUACULTURE: CONTROLLED PRODUCTION OF AQUATIC ANIMALS, took the agreement to give consent for processing, the doctoral thesis entitled "STUDY OF THE SECOND GENERATION OF GILTHEAD SEA BREAM (Sparus aurata L.) WITHIN THE GENETIC IMPROVEMENT PROGRAM PROGENSA ® FOR COMMERCIAL INTEREST TRAITS AND IMPLEMENTATION OF NEW INDUSTRIAL ENABLING TECHNOLOGIES, UNDER CULTURE CONDITIONS" Presented by Mr. ISLAM SAID ELALFY ELKHOLY, and directed by Dr. JUAN MANUEL AFONSO LOPEZ and Dr. MANUEL MANCHADO CAMPAÑA. And for the record, and for the purposes of the provisions of the Artº 6 of Regulation for the preparation, defence, court and evaluation of doctoral theses at the University of Las Palmas de Gran Canaria, I sign this at Las Palmas de Gran Canaria, 20111 June, two thousand and sixteen. � UNIVERSIDAD DE LAS PALMAS !"" DE GRAN CANARIA Annexll Onivir$ity of 1,e$ Pelme$ ®e G'R:({tt C'-({tt-({'R.111 UNIVERSITY INSTITUTE OF SUSTAINABLE AQUACUL TURE AND MARINE ECOSYSTEMS (IU-EcoAqua) Doctoral Program in Aquaculture: Controlled Production of Aquatic Animal Thesis title "STUDY OF THE SECOND GENERATION OF GILTHEAD SEA BREAM (Sparus aurata L.) WITHIN THE GENETIC IMPROVEMENT PROGRAM PROGENSA ® FOR COMMERCIAL INTEREST TRAITS AND IMPLEMENTATION OF NEW INDUSTRIAL ENABLING TECHNOLOGIES UNDER CULTURE CONDITIONS" Doctoral thesis presented by ISLAM SAID ELALFY ELKHOL Y Directed by Dr. JUAN MANUEL AFONSO LOPEZ Directed by Dr. MANUEL MANCHADO CAMPAÑA Las Palmas de Gran Canaria, 10th June 2016. To My sweety wife SELWAN, my handsome son ADAM and my beautiful daughter LAMAR Acknowledgment “All praise is due to Allah, the Lord of the Worlds, the Beneficent, the Merciful” I wish to express my gratitude to the culture affairs and missions sector (Ministry of Higher Education), Egypt, which gave me this opportunity to improve my knowledge in the aquaculture field through the grant that presented to me to study my doctorate at the University of Las Palmas de gran canaria. I would like to express my special appreciation and thanks to PhD advisors Professors, Juan Manuel and Manuel Manchado, you have been a tremendous mentor for me. I would like to thank you for encouraging my research and for allowing me to grow as a research scientist. Your advice on both research as well as on my career have been priceless. Juan is someone you will instantly love and never forget once you meet him. He’s the funniest advisor and one of the smartest people I know. I hope that I could be as lively, enthusiastic, and energetic as Juan and to someday be able to command an audience as well as he can. Manuel has been supportive and has given me the freedom to pursue various projects without objection, He has also provided insightful discussions about the research. Thanks Manuel for your help, your advices and your sympathy were always useful for me. All of you have been there to support me when I recruited patients and collected data for my Ph.D. thesis. I am also very grateful to María Jesús Zamorano Serrano for her scientific advice and knowledge and many insightful discussions and suggestions. She is my primary resource for getting my science questions answered and was instrumental in helping me cranked out this thesis. Content I Page Content .....................................................................................................................................I List of Figures ...................................................................................................................... III List of Tables ....................................................................................................................... IV Abstract .................................................................................................................................... 1 1 General Introduction ........................................................................................................ 5 1-1 Gilthead sea bream (Sparus aurata L.) .......................................................................... 7 1-1-1 Taxonomy ......................................................................................................... 7 1-1-2 Family descriptions .......................................................................................... 7 1-1-3 Biology and morphological characteristics of the gilthead sea bream ............. 8 1-1-4 Geographical distribution and habitat .............................................................. 9 1-1-5 Food and growth ............................................................................................... 9 1-1-6 Reproduction ..................................................................................................... 9 1-1-7 History of culture ............................................................................................ 10 1-1-8 Current status of the culture ............................................................................ 10 1-2 Genetic and its applications in aquaculture ................................................................. 13 1-2-1 Selective breeding ............................................................................................ 13 1-2-2 Genetic markers ................................................................................................. 15 1-2-3 Microsatellites markers ..................................................................................... 16 1-2-4 Multiplex polymerase chain reactions .............................................................. 17 1-2-5 Relationship matrix and family contribution ................................................... 19 1-2-6 Genotype x environment interaction ................................................................ 20 1-3 State of Selective breeding applications in aquaculture .......................................... 20 1-4 State of genetic improvement in gilthead sea bream ............................................... 23 1-5 Important economic traits in gilthead sea bream in genetic improvement programs ............................................................................................................................... 24 1-6 Justification and aims .............................................................................................. 26 2 Optimization of factorial familiar structure in Gilthead sea bream (Sparus aurata L.) by using consecutive mass-spawning, under industrial condition ......... 29 2-1 Abstract .................................................................................................................... 31 2-2 Introduction ............................................................................................................. 32 2-3 Materials and methods ............................................................................................. 35 Content II 2-3-1 Biological material and traits ........................................................................... 35 2-3-2 Genotyping and parental assignment ............................................................... 36 2-4 Results ...................................................................................................................... 38 2-4-1 Phenotyping ...................................................................................................... 38 2-4-2 Genotyping and parental assignment ............................................................... 38 2-4-3 Family contributions ........................................................................................ 39 2-5 Discussion ................................................................................................................ 44 2-5-1 Phenotyping ...................................................................................................... 44 2-5-2 Mass-spawning and parental assignment .......................................................... 45 2-5-3 Family contributions ........................................................................................ 46 3 Estimates of genetic parameters for new noninvasive technological traits, their relationships with essential biological traits and GxE interactions in oceanic cages and estuaries on gilthead sea bream (Sparus aurata L.) ................................................ 51 3-1 Abstract .................................................................................................................... 53 3-2 Introduction .............................................................................................................. 55 3-3 Materials and methods ............................................................................................. 61 3-3-1 Biological material ........................................................................................... 61 3-3-2 Analyzed traits ................................................................................................. 62 3-3-3 Genotyping and familial assignments .............................................................. 64 3-3-4 Data analysis .................................................................................................... 64 3-4 Results ...................................................................................................................... 66 3-4-1 Genotyping and parental assignment ............................................................... 66 3-4-2 Phenotyping ..................................................................................................... 66 3-4-3 Heritabilities and correlations .......................................................................... 67 3-4-4 Genotype - environment interaction ................................................................ 70 3-5 Discussion ................................................................................................................ 71 3-5-1 Genotyping and parental assignment ............................................................... 71 3-5-2 Phenotyping ..................................................................................................... 72 3-5-3 Heritabilities and correlations .......................................................................... 74 3-5-4 Genotype X environment interaction ............................................................... 79 4 Conclusions ....................................................................................................................... 81 5 References .......................................................................................................................... 85 Anexo. Resumen en español ............................................................................................ 109 List of Figures III List of Figures page  Figure 1-1. Gilthead sea bream specimen .................................................................. 8  Figure 1-2. Progression of gilthead sea bream aquaculture production in the Mediterranean area and in the rest of the World from 1984 to 2014 (APROMAR, 2015). .......................................................................................................................... 11  Figure 1-3. Main producer countries of gilthead sea bream (FAO 2015) ................. 11  Figure 1-4. Distribution of the gilthead sea bream productions in Spain for Autonomous regions in 2014 (%) (APROMAR, 2015) ............................................. 12  Figure 2-1. Family contributions under the 2DL model for PCTM2DL ..................... 42  Figure 2-2. Family contributions under the 4DL model for PCTM4DL ..................... 42  Figure 2-3. Family contributions under the 2DL model for CSUR2DL ...................... 43  Figure 2-4. Family contributions under the 4DL model for IFAPA4DL. ................... 43  Figure 3-1. Automatically detected points by IMAFISH_ML software. Lateral view, for the determination of noninvasive technological traits of gilthead sea bream: points x1,x2, x3, x4, x5, x6 of the anteroposterior axis will be used to determine longitudinal traits; y1, y2, y3, y4 and a, b, c, d, e are dorsoventral axes, that will be used to determine height traits. .................................................................................. 63 List of Tables IV List of Tables page  Table 1-1. Number of programs, number of families per program and average of number of traits selected in each program in different aquaculture species (modified from Neira, 2010, Rye et al., 2010 and Gjedrem et al., 2012). .................................... 22  Table 2-1. Phenotypic results for body weight, fork length, condition factor and specific growth rate (SGR) (mean ± standard deviation), and number of fish analyzed (N), for gilthead sea bream at harvest size in each facility (PCTM, Parque Científico Tecnológico Marino; IFAPA, Instituto de Investigación y Formación Agraria y Pesquera; PLV2001, Playa de Vargas 2001 S.L. Company; CANEXMAR, Canarias de Explotaciones Marinas S.L.; PIMSA, Pesquerías Isla Mayor S.A.; CULTMAR, Cultivos Marinos del Maresme S.A.; SADM, Servicios Atuneros del Mediterráneo S.L.). Three mass-spawning models were used: 1DS; one-day spawning, small scale. 1DL; one-day spawning, large scale.2DL; two-days spawning, large scale. 4DL; fourdays spawning ,large scale. ........................................................................................... 39  Table 2-2. The total number of breeders (tB), number of contributors breeders (cB) to the spawning in each batch and sex, and the number of Full-sib (FS) and Halfsib (HS) families. nd, not-determined ................................................................................. 40  Table 2-3. Statistical data for breeders and family production for 1DL, 2DL and 4DL models. Estimate of Effective size, Percentages of breeders contributing to the offspring. Number of families per breeder, average offspring per FS and HS family, offspring per HS paternal (p) and maternal (m)and the Coefficient of Variation (CV) for the number offspring per family under each spawning model. . ............................. 41  Table 3-1. The measured traits by IMAFISH_ML software. ...................................... 63  Table 3-2. Phenotypic result for growth, carcass, flesh composition and noninvasive technological traits (mean± standard error), for gilthead sea bream at harvest size after growth-out in oceanic-cage or in the estuary pond facilities ........................................ 67 List of Tables V  Table 3-3. Phenotypic correlations (below the diagonal), genetic correlations (above the diagonal, value ± standard error), and heritabilities (in bold at the diagonal, value ± standard error) for growth, carcass, flesh composition and technological traits in gilthead sea bream, at harvest size. . ............................................................................. 69  Table 3-4. Genotype-environment interaction for growth, carcass, flesh composition and technological traits in gilthead sea bream at harvest size, and reared in cage and estuary, quantified by genetic correlation (value ± standard error). ............................. 70 Abstract 1 Gilthead sea bream (Sparus aurata L.) is one of the most important species in the Mediterranean aquaculture. In the last quinquennium, implementation of genetic improvement programs has drawn the attention of gilthead sea bream producers to maintain a sustained growth in a competitive way. Although the organization of the industrial production based on genetic criteria increases the production costs, and at the beginning shows low genetic progress rate in short term, the genetic improvements provide a tool for a continuous, cumulative and permanent growth in the industry. The aim of the present study was to study of the second generation of gilthead sea bream, within the genetic improvement program PROGENSA®, for implementation and studying new non-invasive technological traits (Key Enabling Technologies; KET’s), and family contributions by managing of broodstocks through different models. Genetic correlations between KET’s and biological and commercial interest traits were considered, under industrial culture conditions, in order to identify new massive, numerous, precise and economical methodologies for genetic evaluations, at industrial scale. Genotype – Environment interactions were also studied for KET’s and biological traits, in two culturing systems; cages and estuaries. Three different spawning models were implemented, according their number of spawning days. One (small scale 1DS, and large scale 1DL), two (2DL) and four (4DL) consecutive days under industrial conditions, in order to optimize the factorial familiar structure by using consecutive mass-spawnings. All batches were cultured in industrial-like facilities with different numbers of breeders, sex ratio, and under similar culturing conditions. At sampling time, growth traits were estimated through body weight, fork length, condition factor and specific growth rate. Genotyping and parental assignment were inferred by using microsatellite markers in multiplexes PCR, inter-specific (RimA) and specific (SMsa1). In the 1DS model, three independent lots of breeders were constituted, from which only one batch was obtained from each one. A total of 18 breeders were used in the three broodsctoks. Only 17 breeders spawned (94%), to produce a total of 21 FS families and 16 HS families (maternal or paternal). At the 1DL model, with 66 breeders, only 28 fish (42 %), 17 females and 11 males, contributed to the spawning. A total 89 FS families were represented and 24 HS families were constituted (8 paternal and 16 maternal). In the 2DL model, 139 breeders (67%) contributed to the spawning and established 297 FS families and 105 HS families (52 paternal and 53 maternal). In the 4DL model, constituted for 123 breeders, 83 fish (67%), 40 females and 43 males, contributed to produce a total of 201 FS Abstract 2 families, where 59 were HS families (32 paternal and 27 maternal). The percentages of breeders contributing, in the second 2DL and 4DL, were similar, and clearly higher than the first model, 1DL. Regarding the total number of families per breeder, the 4DL model was the highest versus other tested models (1DL and 2DL). In order to study the existence of additive genetic variability, between biological traits and their estimated counterparts by engineering technologies or technological traits, and studying their genetic correlations with other traits such as growth, performance and carcass, and flesh quality, progenies obtained from the selected bloodstocks in two research centers (PCTM – ULPGC; Canary Island, and IFAPA; Andalusia), belonging to the second generation of the genetic improvement program PROGENSA® (named INNOTECSS), were tagged by Passive Integrated Transponder (PIT), and mixed at the two research centers and two Spanish companies (ADSA; Canary Island and PIMSA; Andalusia). At harvest size, fish were sampled for growth, carcass, flesh quality and technological traits, these last by IMAFISH software. Pieces of caudal fin were taken and conserved in ethanol for pedigree analysis. Genotyping and parental assignment were inferred by genetic characterization of all breeders and offspring, by using the SMsa-1 multiplex PCR (Super Multiplex Sparus aurata), containing 11 specific microsatellite markers. Genetic parameters (heritability and genetic and phenotypic correlations) for growth, performance, meat quality and morphology traits were estimated. Genotype–environment interactions (GxE), at harvest size, were estimated by genetic correlation between two facilities (ADSA and PIMSA). Heritability estimates ranged from 0.09 to 0.25 for growth traits, from 0.04 to 0.35 for carcass traits, from 0.08 to 0.27 for body composition traits, and the highest were reported by technological traits from 0.04 to 0.40. The genetic correlations for growth traits were mainly medium and high, for carcass traits were medium, and for body composition traits they were negative and high for moisture versus lipids, negative and low for proteins vs. moisture and lipids, and for the technological traits were high. Genetic correlations between carcass and growth traits, the estimations were medium and high, except for dressing (%) versus length which was low. Between body composition traits and growth traits were generally lowmedium except with moisture, which were medium and negative. For technological traits, all genetic correlations were medium-high with growth traits. For carcass traits, they were high and low for fillet weight and dressing%, respectively, with all technological traits. For body composition, were mainly medium between proteins or lipids versus technological traits, and Abstract 3 medium and negative for moisture. Genetic correlations between technological traits with fillet weight and dressing %, were in general medium and low, respectively. For technological and carcass traits, were mainly low, and between technological traits and body composition, were positives with proteins or lipids, and low and negative with moisture. Genetic-Environment interactions for all traits, at harvest size, were measured considering two culturing system (Cage in Canary island and estuary in Andalusia). They were mainly medium-high for all traits (between 0.46 and 1.00), except for FHA and FHB, which were low (0.05 and 0.16, respectively). All these results reveal that the use of egg batches from consecutive days, under a mass-spawning approach, allows maximize the factorial familiar structure by a high number of breeders contributing to the offspring, a high number of families (full and half sibs families), an optimal control of inbreeding rate, and genetic gain, and it is recommendable and benefit to the design of genetic programs in the gilthead sea bream industry. The data also suggest that the new non-invasive technological traits (Key Enabling Technologies; KET’s), present better additive genetic variation than biological traits, and higher precision, proposing new tools that can be used in breeding programs for gilthead sea bream, and that could lead a faster selection response to improve growth and morphology of the fish, through increased accuracy of breeding values and selection response rates. 1 General Introduction 10 by social and hormonal factors. The eggs are planktonic, spherical and transparent, 1 mm in diameter and with a single oil drop (Cejas et al., 1992: Moretti et al., 1999). 1.1.7 History of culture Traditionally, gilthead sea bream was cultured extensively in coastal lagoons and saltwater ponds, until intensive rearing systems were developed during the 1980s. The Italian 'vallicoltura' or the Egyptian 'hosha' are extensive fish rearing systems that act like natural fish traps, taking advantage of the natural tropic migration of juveniles from the sea into coastal lagoons. Gilthead sea bream is an excellent species to be produced under extensive aquaculture conditions in the Mediterranean basin due to its good market price, high survival rate and feeding habits (able to use different components of the food chain). Artificial breeding was in Italy in 1981-82 and large-scale production of gilthead sea bream juveniles was in 1988-1989 in Spain, Italy and Greece. The hatchery production and farming of this fish is one of the success stories in the aquaculture business. This species, very quickly, demonstrated a high adaptability to intensive rearing conditions, in ponds and cages, and its annual production increased regularly until reaching 173,024 mT in 2014 (APROMAR, 2015). 1.1.8 Current status of the culture Gilthead sea bream is the most important marine finfish species in Mediterranean aquaculture and its production is still on expansion. This species is produced in 20 countries, mainly in Greece, with 71,000 mT. (representing 41 % of total production), Turkey with 37,000 mT (21.4%) and Spain with 16,230 mT (9.4%). (Fig.1-2). Other countries that produced this species are Egypt, Tunisia, Italy, Cyprus, Croatia, Malta, Israel, France and Portugal, and there exist emerging productions in Albania, Algeria, United Arab Emirates and Bosnia. (APROMAR, 2015) (Fig.1-3). 1 General Introduction 11 Figure 1-2. Progression of gilthead sea bream aquaculture production in the Mediterranean area and in the rest of the World from 1984 to 2014 (APROMAR, 2015). Figure1-3. Main producer countries of Gilthead Sea bream (FAO, 2015) 1 General Introduction 12 Commercial production of this species in Europe triggered in the late 1980s and the first 1990s, due to an increased availability of juveniles to be reared in ponds and cages (Shields, 2001; Cañavate and Fernández-Diaz, 2001), thanks to the heavy investment in applied research (data from FEAP, 2008). Spain is one of the important producers of gilthead sea bream. In 2014, Valencia lead sea bream aquaculture production with 8,662 mT (53.4 % of the total), followed by Murcia (3,892 mT, 24%), Canary Islands (1,588 mT, 9.8 %), Andalusia (1,136 mT, 7%), and Catalonia (952 mT, 5.9%). Noteworthy, significant drop for almost half of the production occurred in the Canary Islands in the last decade (Fig.1-4). Figure1-4. Distribution of the gilthead sea bream productions in Spain for Autonomous regions in 2014 (%) (APROMAR, 2015) In other countries like Egypt, the aquaculture is currently the largest single source of fish supply accounting for almost 77 % of the total fish production of the country with over 98 % produced from privately owned farms. Most aquaculture activities are generally located in the Northern Nile Delta Region, with fish farms usually found clustered in the areas surrounding the four Delta Lakes. Recently, the majority of farmed fish are either freshwater species or those that can grow in brackish water. The production of fish and crustaceans in marine water is still in its early stages and its development is influenced by technical and economic problems. Gilthead sea bream is produced in limited amounts in marine fish farms. The total production of gilthead sea bream increased from 3,365 t in 2001 to 18,424 t in 2014 (GAFRD, 2016). 1 General Introduction 13 1.2 Genetic and its applications in aquaculture Genetic improvement programs are an essential tool to increase the competitiveness of the industry. It has been estimated that the use of genetically improved broodstocks for productive traits can contribute to a 50 % reduction of production costs (Gjedrem et al., 2012). The application of this research field in terrestrial animals and plants has made a substantial contribution to a sustain growth and viability of the agroindustry (Ponzoni et al., 2007; Neira, 2010; Gjedrem et al., 2012). However, compared to terrestrial farm animals, the application of the principles of quantitative genetics in fish has been limited and most of the fish species farmed in many countries, especially developing countries, are genetically similar or inferior to wild. Although in general, the response to selection is supposed to be higher in fish and shellfish than in terrestrial animals (Olesen et al., 2003), improvement programs has progressed slower than expected (Neira, 2010; Rye et al., 2010; Gjedrem et al., 2012.). Less than 10% of aquaculture production is based on genetically-improved stocks (Gjedrem et al., 2012). Although genetic strategies in aquaculture are not abundant, some genetic breeding programs have been developed for salmonids, carps, tilapia, catfish and sparids (Gjedrem, 1997; Gjedrem and Thodesen, 2005; Afonso et al., 2012). 1.2.1 Selective breeding Breeding programs improve fish quality and performance by selecting and mating only the best fish, in concordance with the idea that they transmit their superiority to their offspring. A selective breeding program is usually designed to improve productivity over several generations (a generation is the replacement of breeders by their offspring), and increase in this way the growth rates and yields gradually over many years (Ponzoni et al., 2007, 2008). There are two main reasons to use successful selective breeding in aquaculture; a) aquaculture species have very high fecundity allowing a strong selection intensity; b) many traits of interest in aquaculture have a high heritability (h2), which means that selecting for these traits can deliver substantial genetic improvements through offspring generations (Gjedrem and Thodesen, 2005; Gjedrem et al., 2012). Heritability is an essential concept for estimations of breeding values, genetic synthetic index and selection response (Falconer and Mackay, 1996). Thus, this parameter expresses how much phenotypic variation is explained by genetic additive variation, where environmental factors also play an important role in any breeding program depending on the trait and population. So, the 1 General Introduction 14 reduction, as much as possible, of environmental effects is important to discriminate genetic gain, superiority within the population, and reach reliable estimates of heritability and phenotypic and genotypic correlations under breeding programs (Gall and Huang, 1988). Different selection approaches are possible. The "individual selection" refers to the selection based on the individual’s phenotype with or without mating control (mass selection). This strategy is simple and inexpensive and commonly used in fish because it does not require strictly individual identification or the maintenance of pedigree records. Under this model, a cut-off value is established at a specific point so that all fish below this value are discarded, whilst those above that are retained. Individual selection is more effective in traits with high heritability, which imply high selection accuracy. For improving two or more traits, an effective approach has to be used, by using multi-trait synthetic index (Tave, 1995). However, fish breeding programs based on individual selection have a risk of inbreeding because a few parents produce large progenies. Some experiments were unable to generate performance improvements using mass selection because of the lack of genetic variation within the starting population (TeichertCoddington and Smitherman, 1988) or a large reduction in heritability in subsequent generations (Dunham and Brummett, 1999). Hulata et al. (1986) observed no improvement over two generations of mass selection for growth rate in Nile Tilapia (Oreochromis niloticus), due to inbreeding and genetic drift. An important factor with great effect on individual’s phenotype and the efficiency of selection process is the environmental variance, which has to be minimized to ensure that genetically superior individuals are detected based solely on the measurement of their phenotype (Kirpichnikov, 1981). All these data indicate that simple and unstructured mass selection will result in problems unless the number of parents is large (Gjerde et al., 1996; Villanueva et al., 1996), and even so, chance could have a negative effect. Other selection approaches, widely used in fish aquaculture, are based on family selection. These methodologies use between-family and within-family selection or a combination of both types. Between-family, selection depends on arrangement the families in a rank and the best families are retained to be propagated to the next generation. Either the whole family or a random subsample is kept for next generation. In the within-family selection, individuals are selected within family groups, resulting in each family contributing the best individuals from that group. The criterion of this type of selection is the 1 General Introduction 15 deviation of each individual from the mean of the family to which it belongs. Finally, a combination of both information sources, within and between, becomes a more robust method since the best fish from the best families are selected and, therefore, the greater selection intensity should result in a higher response in each generation (Gjerde and Rye, 1998). Family selection allows for the selection of traits with low heritabilities as well as for traits only measurable after death such as disease resistance and carcass characteristics (Bentsen, 1990). Selecting from families also counteracts much of the variation due to environment and thus makes it easier to identify genetic differences in the population (Herbinger and Newkirk, 1990). The use of family information offers greater selection accuracy, but requires the identification of pedigree structure to establish family average (Aleandri and Knibb, 1999). Hence, family-selection until development of molecular tools was based on the ability to separate the families, by rearing them separately or, if tagging was possible, to tag them as soon as possible to be communally reared (Wohlfarth and Moav, 1985). 1.2.2 Genetic markers The development of DNA-based molecular markers had a revolutionary impact on animal genetics and, particularly, in aquaculture. With molecular markers, it was theoretically possible to observe and exploit genetic variation through the entire genome (Liu and Cordes, 2004). Their application has allowed for a rapid progress in aquaculture exploitation of genetic variability and inbreeding, parentage assignments, species and strain identification, the construction of high-resolution genetic linkage maps and identification of marker genetic loci associated with quantitative trait loci (QTL), and use them in selection programs (Marker Assisted Selection; MAS). Several types of molecular markers have been highly popular in aquaculture genetics (Liu and Cordes, 2004; Hashimoto et al., 2009, 2010), such as allozyme (Wilson et al., 1995) and mtDNA (Funkenstein et al., 1990), Restriction Fragment Length polymorphism (RFLP) (Botstein et al., 1980), Randomly Amplified Polymorphic DNA (RAPD) (Welsh and McClelland, 1990), Amplified Fragment Length Polymorphism (AFLP) (Vos et al., 1995), microsatellite (Tautz and Renz, 1984), Single Nucleotide Polymorphism (SNP), and Expressed Sequence Tag (EST) markers. Particularly microsatellites have become the most popular markers for genetic breeding in aquaculture due to the low costs, reproducibility and speed for analysis (Magoulas, 1999). 1 General Introduction 16 1.2.3Microsatellite markers Microsatellites or Simple Sequence Repeat (SSR), are molecular markers composed of tandem repeated units of short sequences (1-6 nucleotides) that are abundantly distributed across the genomes (Tautz, 1989; Litt and Luty, 1989). Microsatellites are able to detect high levels of allele polymorphism (Chistiakov et al., 2006). The alleles differ in length due to the variation in the number of repeats. Estimated mutation rates of microsatellites range between 10-2 and 10-6 per locus per generation (Ellegren, 2000). This high evolutionary rate is caused by the polymerase 'slippage' during DNA replication and unequal crossing over during recombination that lead to outcome differences in the number of repeats (Tautz, 1989). The most abundant microsatellites are dinucleotides (30–67%). Particularly, the (AC)n motif is the most common repeat motif, 2.3-fold more abundant than (AT)n in the vertebrate genome (Toth et al., 2000). Microsatellites behavior as codominant markers with a simple mendelian inheritance. They are relatively small and can be easily amplified by PCR using primers, which are designed from their flanking highly conserved regions. The products of amplification are separated by size through electrophoresis to detect the alleles and the polymorphism of microsatellite. In the field of fisheries and aquaculture, microsatellites are useful for studies of genetic variability and inbreeding, parentage assignment, differentiation of genetic stocks, genetic breeding programs, constructing dense linkage maps, QTL determination and marker-assisted breeding programs. Thus, a large number of microsatellite markers has been described in aquaculture fish species, including Atlantic salmon (Sletten et al., 1997; Skaala et al., 2004), catfish (Liu et al., 1999; Liu et al., 2001; Serapion et al., 2004), tilapia (Lee and Kocher, 1998; Palti et al., 2001; Carleton et al., 2002; Cnaani et al., 2002; Streelman and Kocher, 2002; Rowena et al., 2004), common carp (Crooijmans et al., 1997; Tanck et al., 2001; Liang and Sun, 2003; Kohlmann et al., 2003; Lal et al., 2004; Sun and Liang, 2004; Yue et al., 2004; Li et al., 2007), chinook salmon (Williamson et al., 2001, Naish and Park, 2002), rainbow trout (Rexroad et al., 2001, 2002a, 2002b; Nathan et al., 2007), and Senegalese sole (Funes et al.,2004). In sparid species, numerous microsatellite markers, with a high polymorphism, have been used in gilthead sea bream (Magoulas, 1999; Batargias et al., 1999; Launey et al., 2003; De Innocentiis et al., 2004, 2005; Brown et al., 2005, Oliva et al., 2005; Senger et al., 2006; Castro et al., 2007; Navarro et al., 2008; Porta et al., 2010; Lee-Montero et al., 2013; 1 General Introduction 17 Negrín-Báez et al., 2015), red sea bream (Takagi et al., 1997), snapper (Adcock et al., 2000), black spot sea bream (Stockley et al., 2000; Piñera et al., 2006), red porgy (Navarro et al., 2008), common pandora (Ramšak et al., 2003), and red banded sea bream (Ponce et al., 2006; Navarro et al., 2008). 1.2.4 Multiplex polymerase chain reactions In the modern selective breeding programs, microsatellites allow for the identification of the parents of high-performing progeny in communal rearing environments and, therefore, assist in the selection process (García de León et al., 1998). The large number of microsatellites needed for many of these applications can often generate high costs due to materials and staff time. However, cost reduction can be achieved through multiplexing microsatellites processing. The multiplex PCR approach is based on the combination several primer sets amplifying different loci in the same PCR reaction, to be later processed simultaneously in a single lane of an electrophoretic gel or sequencer (Olsen et al., 1996; Chamberlain et al., 1988; Neff et al., 2000). Multiplex PCR requires that primers lead to amplification of unique regions of DNA, both in individual pairs and in combinations of many primers, under a single set of reaction conditions. Data analysis should allow the analysis of each individual amplification from the mixture of all products (Markoulatos et al., 2002). For a successful multiplex PCR assay, the relative concentration of the primers, concentration of the PCR buffer, balance between the magnesium chloride (MgCl2) and deoxynucleotide (dNTPs) concentrations, cycling temperatures, and amounts of DNA template and Taq DNA polymerase are important. An optimal combination of annealing temperature and buffer concentration is essential in multiplex PCR to obtain highly specific amplification products. MgCl2 concentration needs only to be proportional to the amount of dNTPs, while adjusting primer concentration for each target sequence is also essential (Markoulatos et al., 2002). One of the most important concepts in multiplex PCR is the optimal primer-totemplate ratio. The optimization of multiplex PCR should aim to minimize such nonspecific interactions. Parameters such as homology of primers with their target sequences, length, GC content, and concentration has to be considered for the primer design and Tm threshold (Brownie et al., 1997). The final primer concentrations may vary considerably among loci 1 General Introduction 18 and need to be empirically established. When there is uneven amplification, primer rations need to be changed increasing the concentration of primers with a “weak” signal and decreasing the concentration of the “strong” loci (Navarro et al., 2008). Multiplex PCR reactions are an optimized cost-effective rate and robust tool for genotyping (Neff et al., 2000; Wesmajervi et al., 2006; Navarro et al., 2008). Renshaw et al. (2006) reported a lower material cost of multiplex reactions than single ones. These authors showed that sample genotyping on tetraplex or octaplex reactions were 3.5 and 6.7 times lower than individual genotyping, respectively. They also reported similar money saving values in staff time. Neff et al. (2000) required one-fourth the PCR consumables for genotyping four microsatellite markers with a tetraplex reaction than with single ones. Navarro et al. (2008) compared the staff, consumable and running costs of genotyping with ten microsatellite markers in a multiplex PCR versus separated PCR reactions demonstrating that multiplex reactions were no more than one-sixth that of single reactions even when these were performed in a unique run. On the other hand, multiplex reactions also minimize genotyping errors since they reduce steps during the sample analysis process and introduce automation. Bonin et al. (2004) analyzed twice eighteen microsatellite markers with single PCR on 34 samples of brown bear, and reported a 0.8% of genotyping errors due to human factors. Moreover, multiplex PCR reduces the risk of contamination and avoids the mix of genotypes from different samples (Navarro et al., 2008). However, in sparids, few multiplex reactions with a high number of microsatellites have been proposed. For common pandora, Ramšak et al. (2003) developed a multiplex PCR with three microsatellite markers by means of the touchdown technique. For gilthead sea bream, Launey et al. (2003) proposed two multiplex PCR reactions of three and five markers. Brown et al. (2005) developed the amplification of a multiplex PCR containing four microsatellite markers, whereas Porta et al. (2010) proposed two multiplex reactions of four and six markers as genotyping tools by combining microsatellite markers of identical annealing temperatures. Navarro et al. (2008), also in gilthead sea bream, developed two multiplex PCR reactions of ten and seven interespecific microsatellites markers. This multiplex assays also showed successful cross-amplification in red porgy (six markers in each multiplex) and in red banded sea bream (eight and five markers, respectively). Lee Montero et al. (2013) developed the first standardized panel of two new multiplex PCRs containing 11 markers each one, named SMsa1 and SMsa2 (Super Multiplex Sparus 1 General Introduction 19 aurata). Later, Negrín-Báez et al. (2015) established a set of 13 multiplex PCRs containing 106 specific microsatellite markers in total as a tool for QTL detection in gilthead sea bream (Sparus aurata L.). 1.2.5 Relationship matrix and family contribution Many fish species, including gilthead sea bream, allow for the establishment of viable individual crosses under a factorial design (Knibb et al., 1998; Montero et al., 2001), however its implementation within the context of the industry has significant problems due to facilities requirements and manpower efforts that trigger production costs. Under industrial conditions, this species is propagated through mass-spawning from broodstocks containing approximately 40 - 60 specimens. From a genetic point of view, this strategy shows the advantage that common environmental sources of variation are reduced, thus increasing the precision for genetic parameters estimation (Herbinger et al., 1999). However, under this strategy, genealogy of offspring remains unknown, a factor absolutely necessary for the estimation of genetic parameters. Hence, this strategy requires physical tagging to identify all individuals in conjunction with DNA analysis to reconstruct pedigree information (Navarro et al., 2008). Also, most mass-spawning models use spawns from only one or just a few days to build the fish population to be evaluated. Hence, there exists a tendency to reduce the effective number of breeders that contribute to next generation regarding effective size (Brown et al., 2005; Fessehaye et al., 2006; Piñera, 2009; Porta et al., 2010). The matrix of relatedness between individuals is a key to estimate genetic parameters exactly those traits of economic interest for industry requirement (Navarro et al., 2009a, b). The representative of family contribution is a key element for genetic breeding programs, and particularly important in gilthead sea bream for a successful implementation of genetic improvement in the production system at hatchery level. The high fertility rate of gilthead sea bream can bias the offspring representation and batches of fishes from just a couple of breeders under mass-spawning can occur. It is a major disadvantage that can prevent the maximization of the response to selection. Hence, to monitor and maximize the contribution of the breeder score, under industrial mass-spawning conditions, is of a great practical importance for companies. 1 General Introduction 26 1-6 Justification and aims Gilthead sea bream industry needs to improve its competitiveness to sustain a longterm activity. Although important advances have been achieved in nutrition, reproduction or disease control in the last years, breeding programs are still unexplored. This is due in part to the high cost of companies to organize their production with genetic criteria. Although the genetic progress rate could appear small in comparison with other management factors, this offers a complementary, continuous, cumulative and permanent way to improve the production values (López-Fanjul and Toro, 2007). Under industrial conditions, gilthead sea bream is propagated through a mass-spawning approach, requiring physical tagging devices to identify all individuals, in conjunction with DNA analysis, to reconstruct the genealogy (Navarro et al., 2008). However, more information about the strategies to get a maxim representation of families, and how to manage the fish population for genetic evaluation, is necessary to get optimal estimates of genetic variation. To maximize the genetic progress in any selective breeding program, there are various aspects that must be taken into account, (a) Minimum size of families, (b) Equal number of sexes in broodstock, and (c) Equal family contribution (Falconer y Mackay, 1996). The establishment of directed crosses with homogeneous family sizes requires extensive facilities, only practicable in large multinationals companies with large scale production. However, for small producers, it is not easy to face such costs, so they cannot compete unless they make highly specialized products. Thus, although the items listed above on the family contribution are true, a combination of low consanguinity and increased the number of family can offer an alternative for those companies that they cannot establish large-scale programs, it does allow their to implement genetic improvement programs. In a genetic selection program, the animals are ranked according their genetic breeding values, in a given trait, in a way that whose mattings in mass-spawning produce minimal inbreeding. But what do about those breeders that are not mated, when they are "ideal" for producing new genotypes of greater genetic talent?. If it happens, the opportunity to get better genetic combinations per unit of time will decrease. Before starting a breeding program, the improvement objectives (biological traits of commercial interest for companies), and selection criteria (traits through which is carried out) need to be defined (Gjedrem, 2000). Another essential aspect is how the traits are measured, because the standardization of measurement methods allows for comparisons 1 General Introduction 27 between and within companies and programs, and at the same time increases the accuracy of genetic parameters estimations (Afonso et al., 2012). When breeding programs are developed, one of the most important aspects is the perfect definition of selection criteria, in an objective and precise manner (without personal bias), and at a low cost (Gjedrem, 1997). Since large data matrices obtained simultaneously in different locations, which makes differences between measuring methods, introduce sources of environmental variation and it is difficult to clean in some cases. All of this could lead to decrease the estimates of heritability and increased environmental variance. So one of the most important ways to maximize the estimates of heritability of a trait, and therefore genetic progress between generations, is the development of new technological tools and derivate traits, which, from the genetic point of view, are highly correlated with the important biological traits, interesting for the industry. Given all this, the objectives of the current study are: 1To optimize the factorial familiar structure by using consecutive massspawnings, under industrial conditions. 2To study the existence of additive genetic variability of biological traits in the second generation of PROGENSA breeding program (named INNOTECSS). 3To study the existence of additive genetic variability of new non-invasive technological traits (Key Enabling Technologies; KET’s). 4To study the genetic correlations between biological and non-invasive technological traits, in order to identify new massive, numerous, precise and economical methodologies for genetic evaluations, at industrial scale. 5To quantify the GxE interactions for biological and non-invasive technological traits, through the evaluation of descendants from the same families in different locations-culturing systems. 2. Optimization of factorial familiar structure in Gilthead sea bream ( Sparus aurata L.) by using consecutive massspawnings, under industrial conditions 2-1. Abstract 31 One of the most important aspects of the fish breeding programs success is the production of new recombinants per time unit between selected animals according their breeding values, which depend of familial contribution. Maximization of familial contribution, and minimization of the average coancestry between mates per generation, under Mass-spawning model, would allow increasing the genetic gain. In this study, three different spawnings in gilthead sea bream (Sparus aurata L.) were implemented, according their number of spawning days; one (small scale 1DS, and large scale 1DL), two (2DL) and four (4DL) consecutive days under industrial conditions, in order to maximize the contribution under Mass-spawning production system, in terms of the number of families. All batches were cultured in industrial-like facilities with different numbers of breeders, sex ratio (male:female), and under similar culturing conditions. At sampling time, growth traits were estimated through body weight, fork length, condition factor and specific growth rate. Genotyping and parental assignment were inferred by using microsatellite markers in multiplexes PCR. In the 1DS model, composed for 18 breeders, a total of 17 spawned in the three batches (94%) to produce a total of 21 FS families and 16 HS families (maternal or paternal). At the 1DL model, with 66 breeders, only 28 fish (42 %), 17 females and 11 males, contributed to the spawning. A total 89 FS families were represented and 24 HS families (8 paternal and 16 maternal). In the 2DL model, 139 breeders (67%) contributed to the spawning and established 297 FS families and 105 HS families (52 paternal and 53 maternal). In the 4DL model, constituted for 123 breeders, 83 fish (67%), 40 females and 43 males, contributed to produce a total of 201 FS families, where 59 were HS families (32 paternal and 27 maternal). The percentages of contributed breeders at the second and third tested model (2DL and 4DL) were similar, and clearly higher than the first model. Regarding the total number of families per breeder, the 4DL model was the highest versus others tested models (1DL and 2DL). These data suggest that the use of consecutive Mass-spawning allows to combine the industrial culture model with obtaining a high number of families and number of sibs per family for genetic selection. Keywords Gilthead sea bream, mass-spawning, factorial familiar structure, familial contribution. 2.2 introduction 32 Gilthead sea bream (Sparus aurata L.) is one of the most important species in Mediterranean aquaculture that reached a total world production of 173.024 metric tons in 2014 (APROMAR, 2015). This volume of production has been reached thanks to significant advances in different fields such as nutrition, reproduction, disease control and genetic improvement, even though not all these research fields have progressed at the same speed, and hence contributing unequally to the gilthead sea bream production business model. In the last quinquennium, implementation of genetic improvement programs has drawn the attention of gilthead sea bream producers to maintain a sustained growth in a competitive way. Although the organization of the industrial production based on genetic criteria increases the production costs, and at the beginning shows low genetic progress rate in short term, the genetic improvements provide a tool for a continuous, cumulative and permanent growth in the industry (López-Fanjul and Toro, 2007). Most of these programs have been mainly established on two different approaches in order to generate enough families for a robust evaluation: i) mass-spawning, with random mating of breeders in communal tanks, and ii) directed-crosses, based on artificial fertilization with a manual controlled mating scheme. The main advantages of mass-spawning are the feasibility to be easily integrated in an industrial production routine, without demanding high investments in facilities and staff, the reduction of biases in genetic parameter estimations associated with common environment sources (Herbinger et al., 1999), and the maximization of family crosses in a factorial familiar structure manner. On the contrary, directed-crosses maximize the familial control, but also the production costs since individual families have to be generated by stripping and later kept separately until tagging (3 - 15 g in weight), increasing the common environment sources. Thus, mass-spawning has become the most popular approach for family production pursued in public and private breeding programs in gilthead sea bream (Chavanne et al., 2016), coupled to molecular tools for efficient parentage assignment and pedigree reconstruction for inbreeding control (Navarro et al., 2008; LeeMontero et al., 2013). An important issue for the success of fish breeding programs is the generation rate of new recombinants between selected animals according to their breeding values. In the gilthead sea bream industry, most of the hatcheries distribute their breeders in tanks, for massspawning, containing approximately 50 - 60 animals (sex ratio 1:1) that release eggs daily under controlled photoperiod conditions. Although these broodstocks are available to produce 2. Optimization of factorial familiar structure in Gilthead sea bream 33 from thousands to millions of eggs per day (normally referenced per kilogram of female along spawning season), most producers prefer those spawns with the best eggs quality and quantity evaluated by fertility, hatching and larval survival rates (Fernández-Palacios, 2005). Under this production model, is difficult to establish a precise estimation of families’ representation and number of offspring per family, which really is a disadvantage of the mass-spawning approach, for evaluation of some traits like survival or disease resistance. Moreover, the high fertility rate of gilthead sea bream can also easily bias family production, becoming a major drawback of this approach for a proper evaluation of genetic merit. Brown et al. (2005), found a high variance in family sizes, with a large number of non-contributing fish to the offspring that reduced the effective population size (Falconer and Mackay, 1996). Astorga (2005), reported 88.8 % of variation coefficient for fertilized eggs per female and day, indicating a significant variation in daily eggs production between and within breeders. In other species, Frost et al. (2006) studied the genetic diversity loss in two independent commercial hatcheries of barramundi (Lates calcarifer) over three mass-spawning events finding that broodstock’s contribution and effective population size in multiple spawns differ between daily egg batches. So, the success of breeding programs based on mass-spawning is strongly dependent on the way that breeders, under evaluation, are managed to maximize family production and deal with inbreeding (number of breeders and families, relatedness of breeders, etc...), in order to preserve genetic variation (Pante et al., 2001; Aho et al., 2006). The selection of optimal breeders requires high genetic variation levels in the broodstocks to maintain stable responses over time (Falconer and Mackay, 1996). The maximization of genetic gain should be compatible with low inbreeding rates restricting crosses between related parents (Meuwissen, 1997; Grundy and Hill, 1993). Nirea et al. (2012), described that, under BLUP schemes, maximum genetic gain per generation and minimum rates of inbreeding in percentage are reached under a factorial familial structure when Minimum coancestry mating (Sonesson and Meuwissen, 2000) or Minimized covariance of ancestral contributions mating (Henryon et al., 2009) are implemented. Moreover, Sonesson et al. (2012), reported that designed inbreeding rate in a breeding program can be reached when inbreeding control with estimation of breeding values are implemented under pedigree data. Thus, the maximization of familial contribution and minimization of the average coancestry between mates per generation would allow increase of selected seed into the market, actually estimated between 31-44% (Chavanne et al., 2016), for gilthead sea bream. Hence, the main objective of this study was to analyze the familial 2.2 introduction 34 contribution of some broodstocks under different egg collection models in a mass-spawning approach. The information obtained is useful to assist in the generation of family crosses in the nuclei of breeders and harmonize the commercial interests of the industry to evaluate genetic merits during their production cycles. 2.3. Materials and methods 35 2.3.1 Biological material and traits Three different spawning contribution models were considered in this study according to the number of consecutive days in which eggs were collected and pooled: one (1D), two (2DL) and four (4DL) days. All egg batches, were incubated under the same conditions and cultivated similarly as described in Roo et al. (2009). Samplings for growth traits were carried out according to AquaExcel-TOL (AquaExcel Project 2013) for body weight (ATOL:0000351), fork length (ATOL:0001658), condition factor (ATOL:0001653), and specific growth rate (ATOL:0001662). One-day spawning model (1D) In this model, two broodstocks with different number the specimens were evaluated: Small and Large. A Small broodstock, referred as 1DS, included six breeders (4 males: 2 females) into the same tank at the Marine Science and Technological Park of the University of Las Palmas de Gran Canaria (PCTM, Gran Canaria, Spain). Three different and independent broodstocks of 1DS were constituted (18 breeders in total). Each one produced a batch of eggs collected from only one day, and named as PCTM1DS-1, PCTM1DS-2, PCTM1DS-3. The Large broodstock (referred as 1DL), consisted in 66 breeders (sex ratio  2♂:1♀) into the same tank at Tinamenor S.A. Company (San Vicente de la Barquera, Cantabria, Spain), from where only one egg batch was evaluated. Batches of 1DS and 1DL were cultured at the PCTM facilities, under industrial production conditions. The three 1DS batches were sampled (45, 47 and 53 specimens for PCTM1DS-1, PCTM1DS-2, PCTM1DS-3, respectively) at 194 days post-hatching (dph), to estimate the breeders’ contribution. For 1DL, 200,000 juveniles were sent to a cage located at Playa de Vargas 2001 S.L. Company (PLV20011DL, Gran Canaria, Spain), and sampled at 130 dph. Also, a subsample of 479 specimens was retained to the ongrowing facilities of PCTM (named as PCTM1DL) for on-growing, as described in Navarro et al. (2009). At 509 dph, fish from PCTM1DL and PLV20011DL (479 and 395 specimens, respectively) were slaughtered, growth traits were quantified, and a piece of fin per fish was stored in ethanol until analysis. Two-days spawning model (2DL) This model corresponds to the experimental design explained in Lee Montero et al. (2015). Shortly, in this model, egg batches from two consecutive days were collected and 2.4 results 42 Figure 2-1.Family contributions under the 2DL model for PCTM2DL Figure 2-2.Family contributions under the 4DL model for PCTM4DL 2. Optimization of factorial familiar structure in Gilthead sea bream 43 Figure 2-3.Family contributions under the 2DL model for CSUR2DL Figure 2-4.Family contributions under the 4DL model for IFAPA4DL 2.5 discussion 44 2.5.1 Phenotyping In this study, new data for some growth traits in gilthead sea bream were obtained in different production systems including cages, ponds and tank facilities. All these animals were obtained under mass-spawning from small and large broodstocks after pooling egg batches constituted with different consecutive days (1DS, 1DL, 2DL and 4DL). Growth data provided in the 4DL model indicate that average weight at harvest was higher in tank-based (PCTM4DL, IFAPA4DL) than in ocean-cage (ADSA4DL) or estuarine pond (PIMSA4DL) facilities. These differences between tank (PCTM4DL) and cage (ADSA4DL) facilities in the Canary Island region can be explained at least partially by the type of feeding systems (self-feeder vs manual, respectively), temperature and density, which were similar to those described by LeeMontero et al. (2015). Other important factor is the feeding frequency, which is clearly higher intense in culturing tank-based (PCTM4DL) than in ocean-cage (ADSA4DL), where, at the same time, bad weather plays an essential role during the feeding processes, decreasing the feeding periods. Similarly, the differences between tank (IFAPA4DL) and pond (PIMSA4DL) facilities in South of Spain can be explained by rearing conditions (510 m3 and 2,000 m3tanks, respectively), water quality (filtered pumped and estuarine water, respectively), and physicchemical characteristics (small and high fluctuations in salinity or temperature conditions for IFAPA and PIMSA, respectively). Moreover, in the estuary is quite more difficult to adjust the amounts of feed to the existing biomass. Despite these growth differences, these results are in concordance with previously reported at the same Spanish regions (Ginés et al., 2004; Navarro et al., 2009; García-Celdrán et al., 2015a; Lee-Montero et al., 2015). Concerning SGR, relevant significantly differences between specimens grown under 4DL over 2DL models were observed. Animals from the 4DL and 2DL models belong to the same public Spanish breeding program, PROGENSA® (Afonso et al., 2012), in which selection pressure on growth traits and absence of body deformities (García-Celdrán et al., 2015a; Lee-Montero et al., 2015) were done. The 2DL and 4DL descendant samples belong to two consecutive generations of selection, (F1 and F2, respectively). Similar genetic progress for SGR have been obtained in other commercial breeding programs in gilthead sea bream such as in CULMAREX (Fernandes et al., 2010), NIREUS (Thorland et al., 2015) and ANDROMEDA (Kostas Tzokas, personal communication). 2. Optimization of factorial familiar structure in Gilthead sea bream 45 2.5.2 Mass-spawning and parental assignment The on-growing companies demand progressively a constant supply of high-quality fingerlings during whole year, in order to be produced under different production systems. To attend this demand, most of the hatcheries synchronize reproduction and design larval production using a mass-spawning approach, the most extended reproduction system in gilthead sea bream production sector (Chavanne et al., 2016). Most of broodstocks are composed by 50-60 breeders per tank, with a sex ratio1:1 or 1:2 (females:males), and synchronized by artificial photoperiod. Under this breeding structure, the number and weight of breeders play an essential role for overcoming some drawbacks at early larval stages such fertilization and larval mortality rates or deformities prevalence (Fernández-Palacios, 2005). Thus, mass-spawning lets to the hatcheries an economical balance between cost and profit, and it represents a design that minimizes environment effects on phenotypic and genetic parameters estimations (Herbinger et al., 1999). However, the high number of eggs per kilogram of female during gilthead sea bream spawning season (Cejas et al., 1992), the unknown of breeders’ sex, and the factorial familial structure, do difficult to know the relationships between descendants and breeders. Accuracy of parental assignments and high number of families are especially important when breeders are under a genetic selection program, and genetic evaluation, in terms of cost minimization, maximum genetic gain and minimum rate of inbreeding per generation. In this study the SMsa-1 multiplex of microsatellite was used (Lee-Montero et al., 2013), confirming its high efficiency to reconstruct parental assignments with a success of 87.5 - 100% in an unambiguous way at least a single-parent pair (100% of success for at least on single parent), under different scenarios of mass-spawning (small and large spawns from one, two or four consecutive days). It was available on fish coming from four consecutives spawns (4DL), spite of it represented a 8% of more families than those coming from two consecutives spawns (2DL; Lee-Montero et al., 2015), and 23% of more families than only one spawn (1DL; Navarro et al., 2009a,b). Success of 85% to 100% in parental assignments have also reported by using multiplex reaction of microsatellites on commercial species (Fishback et al., 2002; Saillant et al., 2006, 2007; Dupont-Nivet et al., 2008). For spawns from four consecutive days with a 95.8% of unambiguously assigned offspring (Batargias et al., 2015), or by using single reactions in gilthead sea bream (Castro et al., 2008; Fernandes et al., 2010), but at higher costs. 2.5 discussion 46 2.5.3 Family contributions There are two critical aspects for genetic breeding programs: the number and size of families. Under a mass-spawning approach, the breeders contributing to descendants and the number of sibs are fully at random, prone to bias the family representation. In this study, the best results in terms of breeders contribution and family representation were reported by 1DS model (97%), which used the lowest number of breeders (six per broodstocks PCTM1DS-1, PCTM1DS-2 and PCTM1DS-3). In all broodstocks, all females were mated with all males under mass-spawning (a perfect factorial familial structure), except in PCTM1DS-3 broodstock, where only one male (M4) did not mate with both females. At small scale is much easier the interaction among all breeders than at larger scale, due to management factors that can modify fish interactions and behavior into the tanks including aspects of dominance (Knight, 1985). These results are agree with Sonesson and Nielsen (2012), who estimated that rate of inbreeding decreased and genetic gain increased, as expected, with increasing number of contributing sires and dams. In mass-spawning models at large-scale, the total number of families produced (corrected by the total number of breeders) is positively correlated with the number of consecutive days of pooled batched eggs. Thus, the 2DL model was clearly better than 1DL, while 4DL presented the best behavior in terms of total number of families. Concerning the number of offspring per type of family (FS and HS), values increased from 1DL to 2DL. For 4DL model, it appeared to decrease, but if these values are corrected, according the number of descendants sampled in 2DL, they would be higher (Table 2-3). Comparison between PCTM2DL and PCTM4DL, showed clearly a higher number of breeders contributing to the descendant, when the number of consecutive days increased (Fig. 2-1 and 2-2). This tendency was not corroborated by CSUR2DL and IFAPA4DL facilities. However, should be considered that CSUR2DL had 71% more of breeders. Due to both sceneries (PCTM2DL - PCTM4DL and CSUR2DL - IFAPA4DL), the final average of percentage of breeders contributing to the offspring were similar between 2DL and 4DL models, probably because during 2 or 4 days the breeders under sexual activity are the same. In any case, it seems that a high number of breeders and of consecutives mass-spawnigs, at the same time, could increase the number of breeders contributing of descendants with positive impacts on genetic gain (Sonesson and Nielsen, 2012). However, from a practical point of view, most of hatcheries tend to constitute broodstocks without a very high number of breeders to avoid risks associated to accidental deaths, and eggs production program through all year. Hence, our data demonstrate that factorial familial structure can be improved through the number of 2. Optimization of factorial familiar structure in Gilthead sea bream 47 spawning days while maintain the number of descendants per family, which affects positively to the offspring relationships, genetic gain, and inbreeding rate, avoiding genetic variability loss. Genetic variability is an important attribute of the species under domestication process or commercial exploitation. Thus, those with higher levels of variation are most likely to present high additive genetic variance for productive traits (Alarcón et al., 2004). To preserve genetic variation as such as possible, there are different ways like maintaining several broodstocks and increasing the breeder’s mating. Both aspects are closely related with the inbreeding rate, which has to be established previously in each generation. However, this prediction can be range widely or shortly respect to the observed one depending as breeders mated, and constituted and managed broodstocks. In the gilthead sea bream industry, it is easily solved, according its production system. Large gilthead sea bream hatcheries maintains a wide number of breeders under spawn, organized by different broodstocks, in order to response market demand along the year (around nine – ten broodstocks of 55 breeders each one, as average, i.e., around 500 breeders). Thus, spite broodstocks are subjected to different photoperiod regimes to maintain a continue fingerling production along the year, hatcheries with breeding programs in running can mix adequately the selected breeders genetics by mating, taking advantage of gilthead sea bream males mature at two years old and females at three years old (Ginés et al., 2003). The results of this study, carried out under framework of INNOTECSS project (name of the second generation of PROGENSA®), are in concordance with this last scheme. It, supports the idea that a genetic breeding program is also possible between small gilthead sea bream hatcheries when they coordinate and share information (phenotypic and genetic), about their hundreds of breeders structured in broodstocks (Afonso et al., 2012). When hundreds of breeders are selected, from thousands of selected fish candidates previously evaluated, inbreeding rate and genetic gain can be maximized (Sonesson et al., 2012). Inbreeding rate can also be traced by effective population size (Ne), mainly when genealogical matrix is not available, like in mass-selection (Falconer and Mackay, 1996). To use large effective population sizes is important, in order to design a breeding program (Aho et al., 2006). The optimal effective population size to maintain genetic diversity may be in the order of several hundred. Meuwissen and Woolliams (1994) suggested that a Ne between 31 and 250, for preventing a decline in fitness traits. Tave (1993) recommends a Ne of 45-250 for food fish farming. Over the medium term, a constant Ne of 100 is the most applicable to 2.5 discussion 48 fish farming operations that are likely to be constrained in the size of facilities available for maintaining broodstock. A Ne of this magnitude should prevent substantial inbreeding and loss of alleles over a period of approximately ten generations (Tave, 1993). The results of this study were in concordance with the expected effective population size per generation, higher than one hundred for lots 2DL and 4DL (two consecutives generations of PROGENSA breeding program). Both estimations are agreed with Gjerde et al. (1983) and Jørstad & Nævdal (1996), who concluded that Ne of 100 per generation is sufficient to minimize inbreeding in salmonid breeding programs. Methods to increase the effective population sizes are needed, in order to manage the genetic variation within the farm populations studied (Brown et al., 2005). Many techniques, such as the manual stripping of fish and pair mating is an ideal scenario, but also difficult to implement in gilthead sea bream, because of the need of a critical number of fish to stimulate natural spawning behavior (Gorshkov et al., 1997). Spite of it, this species allows the establishment of viable individual crossings (Knibb et al., 1998; Montero et al., 2001). However, its implementation within the context of the industry has presented problems because it increases production costs in terms of human resources and infrastructure (extensive facilities). However, mass-spawning is widely extended in the gilthead sea bream, because it optimized costs. In this study, mass-spawning based, the number of breeders contributing is increased from consecutives batches, and thus a greater Ne. Moreover, an alternative strategy to improve it, at industrial scale, could be use consecutives 4DL massspawnings with at least one month of difference, where supposedly different groups of breeders would be under sexual activity, without any effect on spawning quality parameters (Ferrnández-Palacios, 2005). Some opportunities exist, for example maximizing the number of fish spawning within a group by synchronizing the ovulation in females through the application of hormone treatments (Zohar and Mylonas, 2001). Batargias (1998) calculated the Ne of an experimental stock of gilthead sea bream over two successive seasons, based on the variance of parental contribution. The census size of the population was 32, and estimated Ne was 12.0 and 9.1m, from mass-spawning in two consecutive seasons. The low Ne observed in the offspring population was attributed to the high variation in the contribution of individual fish, an unbalanced sex ratio (especially in the second season) and no participation of several fish. Brown et al. (2005) found that the effective population size, based on sampling from a single mass-spawning, was moderately low, given the number of fish actually contributing and the number of fish in the broodstock as a whole. García-Celdrán et 2. Optimization of factorial familiar structure in Gilthead sea bream 49 al. (2016), reported similar results on three independent broodstocks of gilthead sea bream, belonging to the original broodstocks to establish the base population of PROGENSA Spanish breeding program in later generations, where differences in breeders contributing and unbalanced sex ratio affected to expected inbreeding rate, estimated from estimated population sizes. However, when genealogy data matrix under mass-spawning is known, and it is combined with optimized parent selection from BLUP evaluation with minimum inbreeding mating, is possible to manage inbreeding over generations within small breeding programs with 30-40 males and 30-40 females (Hely et al., 2013). Moreover, inbreeding control is possible in selection breeding programs, i.e. desired and observed rates of inbreeding can be very similar when the same basis are used for pedigree control and breeding values estimations, i.e. pedigree-based inbreeding control with traditional pedigree-based BLUP estimated breeding values (Sonesson et al., 2012). 3. Estimates of genetic parameters for new noninvasive technological traits, their relationships with essential biological traits and GxE interactions in oceanic cages and estuaries on gilthead sea bream ( Sparus aurata L.) 3. Estimates of genetic parameters for new noninvasive technological traits 58 conditions, Knibb et al. (1997), based on the response to a divergent mass selection after a generation, estimated high (0.51) and intermediate (0.29) realized heritability values for weight at harvest size, in low and high selection lines, respectively. In industrial lots, Navarro et al. (2009a), reported also intermediate heritability values (0.28-0.34) at different relevant ages in days post-hatching (130 days, tagging size; 165 days, transferring to cage size; 330 days, first incipient or unfuntional sexual maturation; 509 days, harvest size), with genetic correlations ranging between low-high (0.11-0.93). Fernandes et al. (2010), estimated higher heritability values for weight at two sizes (0.43 and 0.45, at 345 and 496 days age, respectively), and weight gain between both ones (0.41), under only one culture system and region. In the framework of PROGENSA breeding program, medium heritability values have been estimated for weight at harvest size, 0.25 (García-Celdrán et al., 2015a) and 0.29 (LeeMontero et al., 2015), and specific growth rate along on-growing period, 0.20 (Lee-Montero et al., 2015), across environmental sceneries (four Spanish regions and cages and estuaries). However, when estimations were carried out in the same region-culturing systems, cages in Canary Islands where fixed factors are well controlled, estimates for weight are also high, 0.43 (Lee-Montero, 2012). For stress resistance, Montero et al. (2001) reported a low heritability for plasma cortisol after confinement, 0.06, and no selection response after selection, as expected. The second trait in importance for gilthead sea bream industry is the morphology (Chavanne et al., 2016; Janssen et al., 2016), because it is sold as whole fish. Astorga et al. (2004) estimated for first time for this species a high heritability (>0.70) for occurring of any deformity (38 classes). It was not observed by Castro et al. (2008), when lordosis and operculum deformations were separately determined, in spite of using the same methodology, which can be explained because these authors sieved deformed fish. Recently, García-Celdrán et al. (2015c) and Lee-Montero et al. (2015) described low-medium heritabilities for deformities in head, operculum and spinal column. Thorland et al. (2015) reported a low heritability of 0.16 for jaw deformity in an industrial population. For resistance disease, Antonello et al. (2009) estimated a low heritability of 0.12 for resistance to Photobacterium damselae subsp. piscicida, which requires complex and well focused approaches. Similarly, for body composition (carcass and flesh quality), at industrial scale, low-medium heritability values have been reported by Navarro et al. (2009b) and GarcíaCeldrán et al. (2015a,b). So, improve traits with low heritability or difficult for measuring, other indirect selection methods, like genome selection, are available (Pérez-Enciso and Toro, 2007). 3.2 Introduction 59 Measurement methods When breeding programs are carried out, one of the most important aspects is an appropriate definition of selection criteria to quantify and measure variables in a precise manner (without personal bias), and at a low cost (Gjedrem, 1997). Since large data matrices are usually built through simultaneously samplings in different locations, where uncontrolled sources of environmental variation occur, which are difficult to clean in some cases. This leads to a significant reduction of heritability estimates and an increasing of environmental variances. This reveals the importance of searching a way of standardization of measurement methods in breeding genetic, in order to let data comparisons between and within facilities through some generations. A relevant effort has been done in the AQUAEXCEL European project, to standardize measurement methods for 62 traits related with animal welfare, growth and meat, nutrition and reproduction and hosted in the AQUAEXCEL-ATOL. This database has been built following the criteria of 255 researchers, by establishment of definition, measurement method, material (biological, reagent and instrumental), units and value ranges, parameters to measure, synonyms and references (AquaExcel project, 2013). One of the most important ways to normalize measurement methodologies for a target trait is the automatic analysis based on non-invasive technological traits, which has been demonstrated to increase the accuracy of heritability estimates for a trait and, hence, the genetic progress between generations. Known as technological traits, they are an engineeringmediated data acquisition in easy, friendly and scalable way. From genetic point of view, it is essential that these new technological traits are highly correlated with biological traits of commercial interest. Nowadays, there are some examples of traits that can be estimated using some technological ways. The percentage of lipids in the muscle can be estimated by using the device Distell Fish Fat Meter - FFM (Distell.com, West Lothian, Escocia), a non-invasive method that uses a microstrip sensor sensitive to the water content of the sample able to determine the percentage of fat in muscle. Afonso et al. (2012) studied muscle fat content in gilthead sea bream at harvest, in a population belonging to the PROGENSA breeding program, by direct (chemical) and indirect (FFM) measurement methods, on the same pieces of meat. These authors found a heritability 25% higher in FFM than in the former with a genetic correlation of 94%. Muscle chemical composition, can be also estimated by Infrared Spectroscopy (NIR), by using the FOODSCAN LAB equipment (FOSS, Denmark). For fat content, exists a phenotypic correlation of 0.96 between chemical and FOODSCAN for the same group of samples (Rafael Ginés, personal communication). 3. Estimates of genetic parameters for new noninvasive technological traits 60 Morphological traits are very important economic traits that affect product sales particularly when they are marketed as a whole piece like gilthead sea bream (Blonk et al., 2010; Costa et al., 2011). Chavanne et al. (2016) and Janssen et al. (2016) reported that morphology is the second most frequently selected trait gilthead sea bream, sea bass and common carp in Europe. In this sense, Navarro et al. (2016) developed fast and automated software for image analysis (named IMAFISH_ML) to measure 27 fish morphometric traits (technological traits) on three commercially relevant fish species: gilthead sea bream (Sparus aurata L.), meagre (Argyrosomus regius,) and red porgy (Pagrus pagrus). There are some potential publications that estimated the genetic parameters for some morphological characteristics in fish, such as rainbow trout (Kause et al, 2003), Nile tilapia (Rutten et al., 2005), common carp (Kocour et al., 2007), sea bass (Costa et al., 2010) and common sole (Solea solea) (Blonk et al., 2010). Lee-Montero (2012) did a deep study about heritability of different deformities in gilthead sea bream, under selection process, and their genetic relationships with noninvasive technological traits, reporting genetic determination for some of them, defined by Navarro et al. (2016): Fish Maximum Height (FMH), Total Dorsal Area (TDA), Maximum Width (MW), Fillet Area (FilA), Fillet Volume (FilV), Tail-Excluded Length (TaEL), and Head Height (HeH). These traits presented medium-high (0.35 – 0.51) heritability estimates, and low-high genetic correlations with presence-absence of deformities, and high estimations with growth and carcass traits. However, their genetic relationships with body composition and genotype-environment interactions among them are known, when would be recommendable to do indirect selection through traits with higher heritabilities. In this regard, the morphometric measurements can be used to evaluate the carcass quality, in genetic improvement programs, and to get correlated response on fish body component yield (Rutten et al., 2004). The main objective of the current study is to determine the additive genetic variability for biological traits (growth, carcass and body composition) and new noninvasive technological traits, and compare those with their counterpart values obtained using, when adequate, by engineering technologies. The estimation of genetic correlations among some traits such as growth, carcass and flesh quality traits, by using a non-invasive manner, would let faster, more accurate and cheaper economic evaluations in breeding programs, reducing, on the other hand, animal suffering and minimizing the number of animals needed to obtain robust information for evaluations. 3.3 Materials and methods 61 3.3.1 Biological material In this study, genetic parameters of the second generation of PROGENSA breeding program (Afonso et al., 2012), named and executed as INNOTECSS (Spanish national project, Ref.: RTA2013-00023-C02), were estimated. F0 (starting generation): It was built by commercial and research broodstocks, with different numbers of breeders and sex ratio: Marine Science and Technological Park of the University of Las Palmas de Gran Canaria (PCTM, Gran Canaria, Spain) in the Canary Islands (n=59, 1.81♂:1♀), CULMASUR S.A. (CSUR) in Andalusia (n=99, 1♂:1♀), and PISCIMAR S.L. (PMAR) in Valencia (n=48, 1♂:1.23♀). To produce the first generation, all stocks were synchronized for eggs releasing and collection for two consecutive days. The eggs were simultaneously incubated at the facilities of aquaculture research centers, from three Spanish regions: PCTM in the Canary Islands, Instituto de Investigación y Formación Agraria y Pesquera (IFAPA) in Andalusia, and Institut de Recerca i Tecnologia Agroalimentàries, (IRTA) in Catalonia. F1 (first generation): Larvae were reared in the conditions described by Roo et al. (2009). At 84 days post-hatching (dph), a random sample of 1,400 fish was exchanged between research centers, including a new center from another region: Instituto Murciano de Investigacion y Desarrollo Agrario y Alimentario (IMIDA) in Murcia. At day 179 dph, fingerlings were individually tagged with a passive integrated transponder (PIT; Trovan Daimler-Benz), in abdominal cavity, following the tagging protocol described by Navarro et al. (2006). At the same time, a piece of caudal fin was taken and conserved in ethanol until DNA was extracted for genetic analysis, and reconstruction of genealogy. A sample of fish remained in the on-growing facilities at each research center as potential pre-elite broodstocks (the second generation), and the remaining fish were transported to the growth-out facilities: CANEXMAR S.L. (CANEXMAR; Canary Islands), PIM S.A. (PIMSA; Andalusia), CULTIMAR S.A. (Catalonia), and Servicios Atuneros del Mediterraneo S.L. (SADM; Murcia). At marketing size, 3,743 fish were slaughtered and analyzed for growth, yield, and meat and fish quality traits. The reconstruction of the relationships among the descendants and breeders was carried out following the indications described by Lee-Montero et al. (2013). 3. Estimates of genetic parameters for new noninvasive technological traits 62 F2 (second generation): From the surviving F1 pre-elite broodstocks in PCTM and IFAPA, two elite broodstocks of breeders were built, one per each research center, to produce the second generation of fish composed of families with the higher breeding index that comprise length and the absence of body malformation after BLUP evaluation. Two control broodstocks were randomly selected, one per each research center. The number of breeders and sex ratio were similar in both elite (PCTM, n= 65 , 1♂:1.43♀, IFAPA , n=58, 1♂:1.42♀), and control (PCTM, n = 67 , 1♂: 1♀; IFAPA , n= 60, 1♂:1.4♀) broodstocks. As indicated above, these broodstocks were synchronized for egg releasing and embryos collected from four consecutive days, which were incubated at the same time, and cultured similarly as described in the F1 generation. Similarly, a random sample of fish was remained in the ongrowing facility at each research center, as potential pre-elite broodstocks to constitute the third generation, and the remaining were transported to the collaborating fish farms in each region (ADSA; Canary Islands and PIM S.A. in Andalusia). 3.3.2 Analyzed traits At harvest size (700 dph), a sample of fish from PCTM (56 fish), ADSA (1,034 fish) and PIMSA (589 fish) were slaughtered and growth, carcass as well as meat and fish quality traits were measured. Alive broodstocks (464 from PCTM, and 336 from IFAPA) were also sampled for growth traits. At this age, some males appeared were fluent. Traits were characterized according the standardized methodologies defined in AQUAEXCEL-ATOL (AQUAEXCEL project, 2013). The condition factor (CF) was calculated according to the ATOL:0001653 protocol (100 x body weight[g] x fork length−3[cm]). The deformities were visually assessed by direct observation (ATOL: 0000087). Specific growth rate by ATOL: 0001662 (SGR=100 x ln(final body weight) - ln(initial body weight) x time interval-1) also was calculated. For percentage of lipids in the muscle was estimated by using the device Distell Fish Fatmeter - FFM (Distell.com, West Lothian, Scotland) making four readings, two on each side of the fish. Then, viscera were removed manually and gutted body weight was recorded (ATOL: 0001057), and expressed as a percentage of the body weight (dressing percentage, ATOL: 0000548). Fish were manually skinned and filleted without including the nape and the belly flap. Both fillets were weighed together. Fillets were vacuum packaged and frozen at -20 until analysis of the composition of meat, for which the fillets were homogenized and analyzed by the indirect method of near-infrared spectroscopy (near infrared spectroscopy, NIR) using equipment FOODSCAN LAB (FOSS, Denmark) to obtain proteins, moisture and lipids content, as a percentage of fresh muscle. Each fish was 3.3 Materials and methods 63 photographed with a digital camera, side view, in a dark room with controlled light as according to Navarro et al. (2016). Later, all pictures were analyzed, in a standardized way, by applying the automatic image analysis IMAFISH_ML software, developed in MatLab v.7.5 (The Math -Works Inc, Massachusetts, USA) (Navarro et al., 2016). All traits are shown in Table 3-1, and a fish picture indicating length and height measures are depicted Figure 3-1. Table 3-1. The measured traits by IMAFISH_ML software. Trait Abbreviation Description Total Lateral Length (cm) TLL Distance from x1 to x6, within the horizontal axis (Figure 1). Caudal Peduncle Height (cm) CPH Axis y3 (Figure 1). Fillet Maximum Length FilML Distance from x2 to x3, within the horizontal axis (Figure 1). Tail-Excluded Length TaEL Distance from x1 to x3, within the horizontal axis (Figure 1). Standard Length (cm) SL Distance from x1 to x4, within the horizontal axis (Figure 1). equidistant Fish Heights FHA FHB FHC FHD FHE Total lateral length (TLL) is divided into six equal parts, then, heights of each one of these five points are measured. FHA, FHB, FHC, FHD and FHE are the axes a, b, c, d and e, respectively (Figure 1). Figure. 3-1. Automatically detected points by IMAFISH_ML software. Lateral view, for the determination of noninvasive technological traits of gilthead sea bream: points x1,x2, x3, x4, x5, x6 of the anteroposterior axis will be used to determine longitudinal traits; y1, y2, y3, y4 and a, b, c, d, e are dorsoventral axes, that will be used to determine height traits. 3. Estimates of genetic parameters for new noninvasive technological traits 64 3.3.3 Genotyping and familial assignments DNA was extracted from a fragment of the caudal fin preserved in 1 ml of absolute ethanol by using a DNeasy kit Nucleospin® 96 tissue (MACHEREY-NAGEL), using a TECAN robot (Tecan Schweiz AG, Switzerland), and Freedom Evowar® Standard v2.5 software, following the manufacturer’s instructions. Extracted DNA was stored at -20°C in the provided buffer until use. DNA quantity and quality were determined using a NanoDrop 8000 spectrophotometer v.3.7 (Thermo Fisher Scientific), and quality checked by agarose gel at 1% to 10 v/cm. All breeders and offspring were characterized genetically using the SMsa-1 multiplex PCR (Super Multiplex Sparus aurata), developed by Lee-Montero et al. (2013) containing 11 specific microsatellite markers. Genotypes were determined by GENEMAPPER v.3.7 software, and using the bin set kit (SMsa1 kit). Family relationships between breeders and offspring were determined by the exclusion method using VITASSIGN (v8.2.1) software (Vandeputte et al., 2006), considering the breeders’ gender as unknown. 3.3.4 Data analysis Prior to do the genetic parameters estimation, all quantitative data were analyzed for normality and variance homogeneity. The effects of fixed factors (tracking, culturing system, origin), were studied by General Linear Model using SPSS v.18 software (SPSS, Chicago, IL, USA). Variance components of all traits considered in this study, to obtain genetic parameters (heritabilities and correlations), were estimated by the Restricted Maximum Likelihood Method following the model, y = X  + Zu + e y is the recorded data recorded on the studied traits, β the fixed effects (tracking, culturing system, origin), u the random animal effect, and e the error. Genotype–environment interactions (GxE) at harvest size (700 days), were estimated by genetic correlation between two facilities (ADSA and PIMSA), by considering the particular trait of interest as a different trait at each facility (Falconer and Mackay, 1996). All genetic estimates were carried out using VCE (v 6.0) software (Neumaier and Groeneveld, 1998; Groeneveld et al., 2010). The magnitude of estimated heritability was established following the classification suggested by Cardellino and Rovira (1987) as: low (0.05–0.15), medium (0.20–0.40), high 3.3 Materials and methods 65 (0.45–0.60) and very high (>0.65). The magnitude of correlation was established following the classification indicated by Navarro et al. (2009a) as: low (0–0.40), medium (0.45–0.55) and high (0.60–1), regardless if positive or negative. 3.4 Results 66 3.4.1 Genotyping and parental assignment The use of multiplex SMsa1 PCR using the exclusion method, with a maximum of two tolerated errors, provided successful parental assignment of 87 % for descendants cultured in cages at Canary Islands, and 88% in the estuary at Andalusia. The remaining of descendants (13% and 12% from cage and estuary, respectively), were assigned to two possible breeder couples, but they were not excluded from the genetic analyses, being included in the relationship matrix as ‘just one known parental breeder’. Regarding the breeders’ contribution, from total elite breeders (123), 83 breeders (67%), 40 females and 43 males, contributed to produce a total of 201 full-sib families, and 59 half-sib families (32 of paternal and 27 maternal). From 127 control breeders, 83 breeders (65%), 42 females and 41 males, contributed to produce a total of 123 full sib families, and 50 half-sib families (27 paternal and 23 maternal). 3.4.2 Phenotyping The phenotypic results for growth (body weight, fork length and condition factor), carcass (dressing percentage and fillet weight), body composition (proteins, fat and moisture in muscle), and noninvasive technological traits in oceanic-cage and estuary are shown in Table 3-2. Fish cultured in estuary showed higher values for most of growth, carcass, flesh composition and noninvasive technological traits than those cultured in cage, except for dressing percentage, SGR and Moisture values variables. 3. Estimates of genetic parameters for new noninvasive technological traits 67 Table 3-2 Phenotypic result for growth, carcass, flesh composition and noninvasive technological traits (mean± standard error), for gilthead sea bream at harvest size after growth-out in oceanic-cage or in the estuary pond facilities Trait Cage Estuary Weight 313.14 ± 3.90 439.91 ± 3.10 Length 24.34 ± 0.10 27.01 ± 0.07 CF 2.04 ± 0.00 2.21 ± 0.01 SGR 0.75 ± 0.00 0.66 ± 0.00 Dressing (%) 92.20 ± 0.23 90.91 ± 0.13 FilletW 98.09 ± 1.24 150.41 ± 1.57 Proteins (%) 19.38 ± 0.03 20.84 ± 0.05 Moisture (%) 73.18 ± 0.08 68.77 ± 0.09 Lipids (%) 6.55 ± 0.06 8.71 ± 0.11 FFM (%) 9.36 ± 0.11 NM TLL 25.98 ± 0.09 29.38 ± 0.08 CPH 2.27 ± 0.01 2.97 ± 0.03 FilML 13.21 ± 0.05 15.90 ± 0.06 TaEL 19.49 ± 0.07 22.17 ± 0.07 SL 21.98 ± 0.08 25.08 ± 0.08 FHA 7.15 ± 0.03 8.20 ± 0.05 FHB 8.95 ± 0.03 10.72 ± 0.05 FHC 7.87 ± 0.03 9.36 ± 0.03 FHD 4.77 ± 0.03 5.94 ± 0.05 FHE 3.34 ± 0.03 3.92 ± 0.05 (SGR) Specific growth rate, (CF) Condition factor, (FFM) Fish Fat Meter, NM: not measured. 3.4.3 Heritabilities and correlations Estimates of heritabilities for each trait and genetic and phenotypic correlations for traits are shown in Table 3-3. Heritabilities ranged from 0.09 to 0.25 for growth traits, from 0.04 to 0.35 for carcass traits, from 0.08 to 0.27 for body composition traits, and from 0.04 to 0.40 for noninvasive technological traits. Within group of traits The genetic correlations between growth traits were mainly medium and high. Between carcass traits were medium, and between body composition traits: for moisture 3. Estimates of genetic parameters for new noninvasive technological traits 74 standard length by IMAFISH (SL) showed a coefficient of variation lower, 11%, than the manual method. 3.5.3 Heritabilities and correlations Heritability The growth traits (weight and length) are highly dependent on factors related with feeding and rearing conditions but they usually have a medium-high genetic determination in most livestock animals (Cardellino y Rovira, 1987), where additive genetic variation play a major rule. In gilthead sea bream, Knibb et al. (1997) estimated realized heritabilities for weight at harvest size, ranged between medium (up-selection) and high (down-selection), after a single generation of divergent selection against a control population. Navarro et al. (2009a), in an industrial lot fatted in two facilities, described intermediate values of heritability for weight and length, from fingerling to adult. Lee-Montero (2012), also for an industrial lot grown in only one locality (Canary Island) belonging to the first generation of PROGENSA breeding program, and to the same sizes range, reported low and high heritability values (fingerling and adult, respectively). Similarly, Fernandes et al. (2010) described high heritability value (0.45), at adult, for industrial lot cultured in only one facility. When lots of fish were constituted with the same families and grown at different localities along Spanish coast, heritability estimations for adults were lower, medium values (LeeMontero, 2012). García-Celdrán et al. (2015c), also in an industrial lot, found low and medium heritability values for weight and length, in fingerling and adult (at harvest size). Data of the present study, which belong to the second generation of PROGENSA and where selection pressure was carried out through the weight at harvest size in two different localities, showed similar and consistent medium heritabilities. It is in concordance with inbreeding control of both generations, because effective population sizes were similar (157 and 155, for first and second generations, respectively).For compactness or condition factor (CF), heritability of this study was low (0.09), which is agree with values reported by Navarro et al. (2009a) (0.05-0.13) from fingerling to adult. Once again, in lots cultured in only one locality showed higher heritabilities according Lee-Montero (2012) and García-Celdrán et al. (2015a), who reported moderate values (0.20 and 0.18, respectively). All of these results were lower than the results obtained by other authors in other species, such as in red drum (Sciaenops ocellatus L.) by Saillant et al. (2007) (0.28), in common carp by Vandeputte et al. (2004) (0.37), and in sea bass by Dupont-Nivet et al. (2008) (0.19). The general tendency of gilthead 3-5 Discussion 75 sea bream to exhibit lower heritabilities for CF in comparison with other species could be related with the positive allometry towards the height (Ginés, 1997). Concerning carcass traits, heritability estimates described in different species oscillated in a wide range. For dressing percentages, values ranged from 0.02 (Powell et al. 2008) to 0.45 (Kause et al. 2002). In gilthead sea bream, Navarro et al. (2009a) obtained medium values (0.31) whereas Lee Montero (2012) and García-Celdrán et al. (2015a) estimated low values (0.06 and 0.07, respectively), like in this study (0.04). This wide difference between estimations of Navarro et al. (2009a) and other authors, including the present study, can be explained because lots of Lee Montero (2012), García-Celdrán et al. (2015a) and this work came from the same genetic background of PROGENSA. For fillet weight, the results also show great variation coming from, possibly, because there is a great variability in terms of measure methods. In gilthead sea bream, Navarro et al. (2009a), García-Celdrán et al. (2015a) and Lee Montero (2012) reported low and high values (0.15, 0.17 and 0.41, respectively). In this study, medium values for fillet weight were estimated (0.35) reporting a robust genetic basis and high potential of filleting market (Luna et al., 2006). In gilthead sea bream, Janssen et al., (2016) reported that product quality traits are the third most frequently selected traits in breeding programs. That is because the industry and consumers are paying high attention to meat quality since aquaculture products are always compared with those from wild fisheries. The medium heritabilities values of lipids content and moisture obtained by this study agree with the estimates reported by Lee-Montero (2012) (0. 17 and 0.20, respectively) and García-Celdrán et al. (2015a) (0.31 and 0.24, respectively). Navarro et al. (2009b) reported lower values (0.05 and 0.09, respectively). The highest values in this study, similar to by Lee-Montero (2012) and García-Celdrán et al. (2015a), are probably due to technical methods and genetic background. Thus, Navarro et al. (2009b) determined these parameters by chemical and manual methods, while in the latter studies; they were determined by the NIRs methodology. Moreover, fish populations used by LeeMontero (2012), García-Celdrán et al. (2015a), and this study belong to different samples belonging to PROGENSA breeding program. Regarding the percentage of protein, the heritability of this study was low (0.08) that coinciding with those estimated by Lee-Montero (2012) (0.06), and García-Celdrán et al. (2015b) (0.03). It reflects the difficulty presented by this trait to improve by direct methods. 3. Estimates of genetic parameters for new noninvasive technological traits 76 In gilthead sea bream, morphological traits are related with growth, carcass, meat and fish quality (Lee-Montero, 2012). They are relevant because this species is mainly commercialized as whole fish, and the visual perception of foods in general and aquatic foods in particular, is extremely important for the consumer (Gumus et al., 2011). Image processing techniques have been used increasingly for food quality evaluation in recent years (Du and Sun, 2004). Since, image analysis reduces the sources of variation between measurements, if methodology is correctly standardized, and increases the accuracy of genetic parameter estimates (Blonk et al., 2010). In the present study, ten out of the 27 variables determined by image analysis program developed by Navarro et al. (2016), and previously do not studied, were selected in order to know their additive genetic variation and relationships with essential biological traits, actually selected in populations of selection programs (Chavanne et al., 2016; Janssen et al., 2016). Heritabilities of ten noninvasive technological traits (TLL, CPH, FilML, TaEL, SL, FHA, FHB, FHC, FHD, and FHE) were studied in two fish lots at industrial scale, in different culturing systems (cage and estuary). All noninvasive technological traits, connected with different fish heights in gilthead sea bream (FHA, FHB, FHC, FHD and FHE), reported heritabilites ranged between moderate (0.17 for FHA) and high (0.40 for FHC), except for FHE trait that it was low (0.09). These results suggest that is possible to change the gilthead sea bream shape, by selection, according to the commercial interests of companies, to fulfill market demands. Regarding length-related traits using noninvasive measurements by IMAFISH (TLL, FilML, TaEl and SL), values were similar to those obtained by manual methodology (24.3 and 27.0 cm for length in cage and estuary, respectively, estimated manually, versus 22.0 and 25.1 cm for standard length in cage and estuary, respectively, measured by IMAFISH), but heritability was improved, as average 24% by IMAFISH software. Genetic correlation Genetic correlation between growth traits (weight, length, CF, SGR) obtained in this study were high and positive, as expected and in concordance with as previously described (Navarro et al. 2009a; Fernandes et al., 2010; Lee-Montero et al. 2015; García-Celdrán et al. 2015a). Only genetic correlations between length and CF have a tendency to be lower, like also described by Navarro et al. (2009a), Lee-Montero et al. (2015) and García-Celdrán et al. (2015a), probably affected because CF is a derivate trait where length suffers a cubic transformation. Growth traits were genetically correlated with all studied noninvasive technological traits (from 0.60 for FHA-SGR to 1.00 for weight-TLL, weight-TaEL, weight- 3-5 Discussion 77 SL, length-TLL, length-FilML, length-TaEL and length-SL). High heritabilities and genetic correlations of noninvasive technological traits, between them and versus growth traits, suggest that indirect selection for growth traits through noninvasive technological traits is available, at the same time that higher number of fish candidates can be evaluated with minimum cost. Furthermore, it increases the genetic gain and reduces the inbreeding rate (Sonesson et al., 2012) Regarding the genetic correlations between carcass and growth traits; fillet weight and weight/length values were high and positive, as previously observed in gilthead sea bream and other species (Gjerde and Gjedrem, 1984; Kause et al., 2002; Neira et al., 2004; Doupé and Lymbery, 2005; Rutten et al., 2005; Kause et al., 2007; Kocour et al., 2007; Powell et al., 2008; Navarro et al. 2009a; Lee-Montero, 2012; García-Celdrán et al., 2015a). Correlations between dressing percentages and weight/length were medium (0.26-0.67), in agreement with previous studies (Doupé and Lymbery, 2005; Rutten et al., 2005; Kocour et al., 2007). So, direct selection for weight or length, which are less invasive that fillet weight (with higher heritability than weight and length), should increase, in any sense, the feed efficiency. Intriguingly, other works in gilthead sea bream found low or negative genetic correlation estimates between carcass and growth traits (Navarro et al. 2009a; Lee-Montero, 2012; García-Celdrán et al., 2015a). This discrepancy between studies can be due fish size at harvest, since fish produce proportionally more bone than meat with age in some populations, due to high negative genetic correlation between gutted weight and fillet weight-% (Navarro et al., 2009a), or more meat proportionally, as denoted by high and positive genetic correlation between gutted weight and fillet weight-% (Lee-Montero, 2012). Corroboration of absolute and signs for these values should be confirmed by analysis of fish populations at higher sizes, like one kilo, discriminating between both (meat and bone) along development. In this study, carcass traits reported high and positive or low and positive values of genetic correlations between dressing-% or fillet weight versus noninvasive technological traits, respectively. Bearing in mind that FHC showed the highest heritability value (14% higher than fillet weight and 10 fold higher than dressing-%), and both carcass traits were positively correlated, noninvasive direct selection for FHC trait would generate a similar indirect selection response on fillet weight that direct selection for this last trait (0,4% higher indirect over direct selection). However, indirect selection of dressing-% through fillet weight would produce a selection response 41% higher than direct selection for dressing-%. So, direct selection for FHC, in noninvasive way, would produce similar selection response for fillet 3. Estimates of genetic parameters for new noninvasive technological traits 78 weight but a linked response for dressing-% trait, through its genetic correlation with fillet weight, because directly it would be scarce due to its very low heritability (0.04). Respect to the genetic correlations of quality traits, values between lipids and moisture content were high and negative, coinciding with those described in both gilthead sea bream and other species (Iwamoto et al., 1990; Kause et al., 2002; Quinton et al., 2007; Navarro et al., 2009b; Lee-Montero, 2012). Lee Montero (2012) found a high and negative correlation (-0.96) between protein and lipids content, however, in this study were negative but low and with a high standard error (-0.03±0.23). So, this result should be consolidated with other studies and populations. However, positive genetic correlations between quality with growth traits of this study are in agreement with the results obtained by Lee-Montero (2012). Moreover, in this study, noninvasive technological traits reported high and negative genetic correlations with moisture, medium and positive with lipids content, and low and positive with protein content. These genetic correlations and heritability estimated values would produce higher indirect selection response for quality traits thorough direct selection for technological traits with high heritability, like FHC, but at least it would generate an expected, linked and noninvasive selection response for quality traits. The use of image analysis is an objective for measuring fish morphological traits. In terrestrial animals, where genetic improvement is well established, the studies reveal that the morphological traits are genetically correlated with production traits (Simm, 1998; Conafe, 2012). On the other hand, the selection for growth traits can alter the shape of the fish through indirect selection responses (Blonk et al., 2010), and also can be accompanied by an increase in the prevalence of body malformations in animal (Gjedrem, 2005), and also in fish depending of their genetic correlations. Gjerde et al. (1986; 2005) reported vertebral deformations in Atlantic salmon (0.00-0.36), with genetic correlation between deformity and fish weight, from -0.22 to -0.42, indicating that high genetic growth potential not causes deformity. Kolstad et al. (2006) described spinal deformities (kyphosis, lordosis, and scoliosis) in Atlantic cod, 0.27, and their genetic correlations estimated between weight and occurrence of spinal deformity, 0.50, indicating that fast growth should be recognized as a risk factor for deformity incidence. Kettunen and Fjalestad (2007), in Atlantic cod, reported bend on spinal column and pelvic fin deformities, with heritability ranges of 0.33-0.42 and 0.23-0.28, respectively. So these authors recommended to include these traits in genetic selection index. Ando et al. (2008) described a number of abdominal vertebrae, heritability of 0.65, and caudal vertebrae, heritability of 0.84, in Masu salmon. The genetic correlation 3-5 Discussion 79 between both traits was calculated as -0.92. Kocour et al. (2006) reported an anal fin deformity with low heritability of 0.07, in common carp. Genetic correlation with others skeletal deformities (mouth and caudal fin) were high, 0.78-0.88. Bardon et al. (2009), in European sea bass, estimated heritabilities of 0.33 and 0.12 for lordosis and scoliosis, respectively. Positive genetic correlation between weight and these malformations were detected, from 0.21 to 0.40, where deformities have a negative phenotypic impact on growth. Kause et al. (2005) reported low heritability, 0.02, for skeletal deformity (head, neck, back or tail), in rainbow trout. Positive genetic correlation between weight and deformation was detected, 0.18, indicating that when selection is applied for rapid growth an increase in the frequency of developmental disorders is expected. Similar results have been reported in gilthead sea bream by Astorga et al. (2004), describing morphological malformations (38 types, including skeletal deformities) iwith a heritability of 0.74-0.85, for presence-absence of any deformity, and considering that major genes can be playing an important role in abnormalities. Thorland et al. (2006; 2015) reported spinal deformity (include several) and jaw deformity in this species, with 0.12 and 0.16 heritability estimations, respectively. Suggesting that malformations traits have to be included in selection processes. Recently, wide studies have been done in this species under the framework of PROGENSA project by Lee-Montero et al. (2015) and García-Celdrán et al. (2015c). These authors described relevant additive genetic variation for lordosis, operculum and head, genetically and positively correlated with growth traits, suggesting the importance to include their prevalence in selection indexes due to the depreciation suffered for deformed fish, in hatcheries and ongrowing companies, during marketing processes. So, consider and develop new noninvasive technological traits for deformity identification, in an automatic way seems really important for the industry, in terms of economic losses prevention. 3.5.4 Genotype x environment interaction (GxE) Gilthead sea bream is produced by using different production systems with large environmental differences among them. So, it is very important in breeding programs, to estimate GxE interactions to avoid their possible detrimental effects on the genetic selection process (Falconer and Mackay, 1996). Interactions may involve changes in rank order for genotypes between environments and changes in the absolute and relative magnitude of the genetic, environmental and phenotypic variances between environments (Bowman, 1972). In fish breeding, the GxE interaction is probably not significant when the genetic correlation of a 3. Estimates of genetic parameters for new noninvasive technological traits 80 trait measured in different environments is higher than 0.7 (Sae-Lim et al., 2013). In gilthead sea bream, Navarro et al. (2009a), reported genetic correlations between two facilities (tank and oceanic cage, in the same region) of 0.8 for growth traits, and 0.93 for carcass traits. Navarro et al. (2009b), for the same facilities, described genetic correlations of 0.81±0.22 for visceral fat, 0.15±0.94 for muscle fat, and 0.22±0.98 for moisture. These results indicated low or absent genetic–environment interaction for growth and carcass traits, including visceral fat. While genetic–environment interaction cannot be discarded for quality traits, which presented the highest standard errors. Lee-Montero et al. (2015) reported, also in gilthead sea bream, a high genetic correlations for presence-absence of any deformity among the five facilities (0.65-0.99), including estuaries and oceanic cages, and indicating that G x E interactions were non-relevant. These authors also reported genetic correlation values ranging between 0.63 to 0.94 for length, and from 0.05 to 0.99 for SGR. These results agree with results of this study, where medium-high genetic correlations between oceanic-cage and estuary for all traits GxE (0.46 - 1.00) were estimated, except for FHA and FHB, which were low (0.05 and 0.16 respectively), and GxE interactions cannot be discarded. 4. CONCLUSIONS 4. Conclusions 83 1. Estimates of growth traits (weight, length, condition factor and SGR or specific growth rate), for mass contribution models at small and large scale studied in this work (1DS, 4DL), showed mean values and phenotypic correlations according with fish age at the slaughtering. 2. Growth traits (weight, length, condition factor and SGR), within the mass contribution model 4DL of four consecutive spawns, fish cultured in tanks (PCTM, IFAPA) showed higher values than fish cultured in ocean cage (ADSA) or estuary (PIMSA). 3. Mass contribution model 4DL, belonging to the second generation of PROGENSA breeding program, for "slaughtering weight" criterion, showed mean values of SGR 18% higher than the model 2DL, belonging to the first generation of PROGENSA. 4. The multiplex PCR of 11 microsatellite markers called SMsa1 showed high efficiency inferring the relationship matrix, between breeders and descendants of mass contribution model 4DL of four consecutive spawns, despite presenting a 8 % more families than model of two consecutive spawns (2DL), and 23% more families than model of one spawning (1DL). 5. Mass contribution model 4DL of four consecutive spawns, belonging to the second generation of PROGENSA breeding program, allowed to keep effective population size values higher than 100 (155), and therefore controlling rate of inbreeding per generation below 1% (0.32%). 6. Mass contribution model 4DL of four consecutive spawns, allows a high number of breeders contributing to the descendants, and a high number of families (full and half sibs, maternal and paternal), maximizing the factorial family structure. 7. Estimates of heritability for growth, carcass, body composition and non-invasive new technological traits of the second-generation of PROGENSA breeding program, ranged between low and medium values. 8. Genetic correlations within group of traits were medium and high for growth traits. For carcass traits they were medium. For body composition traits they were: negative and high for humidity versus against lipid content, whereas the protein content from moisture and / or lipids were negative and low. For new non-invasive technology traits, genetic correlations between them were high. 9. The genetic correlations between groups of traits were medium and high between growth and carcass traits, with the exception of length with the dressing %, which was low. Between growth and body composition traits genetic correlations were generally low - medium, except for growth traits with moisture, which were medium and negative. Genetic correlations between non-invasive technological and growth traits were high for fillet weight, and low for dressing-%. Between non-invasive technological and body composition traits were mainly medium with protein or lipid content, and medium and negative with moisture. 5. reference 90 Cañavate, P.J., Fernandez-Díaz, C., 2001. Pilot evaluation of freeza-dried microalgae in the mass rearing of gilthead sea bream Sparus aurata larvae. Aquaculture 193, 257-269. Cardellino, R.,Rovira, J., 1987. Mejoramiento genético animal. Hemisferio Sur, Buenos Aires, 253 pp. Carleton, K.L., Streelman, J.T., Lee, B.Y., Garnhart, N., Kidd, M., Kocher, T.D., 2002. Rapid isolation of CA microsatellites from the tilapia genome. Anim. Genet. 33, 140-144. Castelló, F., 1993. Biology of teleost marine fish farming more interesting. In: Marine Aquaculture: biological foundations and technology of production. Castelló.F Eds . University of Barcelona, Spain pp 61-67. Castro, J., Pino, A., Hermida, M., Bouza, C., Chavarrías , D., Merino, P., Sánchez, L., Martínez P., 2007. A microsatellite marker tool for parentage assessment in gilthead sea bream (Sparus aurata). En: M., Vandeputte, Chatain B., Hulata G. (Eds.), Genetics in Aquaculture: Proceedings of the Ninth International Symposium, 26-30 June 2006, Montpellier, France. Aquaculture, 272, Supplement 1, 210-216. Castro, J., Pino-Querido, A., Hermida, M., Chavarrías, D., Romero, R., García-Cortés, L.A., Toro, M.A., Martínez, P., 2008. Heritability of skeleton abnormalities (lordosis, lack of operculum) in gilthead sea bream (Sparus aurata) supported by microsatellite family data. Aquaculture 279, 18–22. Cejas, J., Villamandos, J., Samper, M., 1992. Estudio sobre la reproducción de la dorada (Sparus aurata) en Canarias. Influencia del peso/edad de las hembras sobre la calidad de la puesta. Inform. Tecnic.Inst. Esp. Ocean. 126, 1-38. Chamberlain, J.S., Gibbs, R.A., Ranier, J.E., Nguyen, P.N., Caskey, C.T., 1988. Deletion screening of the Duchenne muscular dystrophy locus via multiplex DNA amplification. Nucleic Acids Res. 16, 11141-11156. Chavanne, H., Janssen, K., Hofherr, J., Contini, F., Haffray, P., Consortium, A., Komen, H., Nielsen, E.E., Bargelloni, L., 2016. A comprehensive survey on selective breeding programs and seed market in the European aquaculture fish industry.Aquaculture International. pp 1-21. 5. Reference 91 Chistiakov, D.A., Hellemans, B., Volckaert F.A.M., 2006. Microsatellites and their genomic distribution, evolution, function and applications: A review with special reference to fish genetics. Aquaculture 255, 1-29. Cnaani, A., Ron, M., Lee, B.Y., Hulata, G., Kocher, T.D., Seroussi, E., 2002. Mapping the transferrin gene in tilapia. Anim. Genet. 33, 78-80. CONAFE, 2012. http://www.conafe.com/. Costa, C., Antonucci, F., Pallottino, F., Aguzzi, J., Sun, D.W., Menesatti, P., 2011. Shape analysis of agricultural products: a review of recent research advances and potential application to computer vision. Food Bioprocess Technol. 4, 673–692. Costa, C., Vandeputte, M., Antonucci, F., Boglione, C., Costa, Menesatti, P., Cenadelli, S., Parati, K., Chavanne, H., Chatain, B., 2010. Genetic and environmental influences on shape variation in the European sea bass (Dicentrarchus labrax). Biological Journal of the Linnean Society, 101, 427–436. Crooijmans, R.P.M.A., Poel, JJ, Groenen, MAM, Bierbooms, VAF, Komen, J., 1997. Microsatellite markers in common carp (Cyprinus carpio L.). Anim Genet. 28, 129-134. Crow, JF., Denniston, C., 1988. Inbreeding and variance effective population numbers. Evolution 42, 482–495. De innocentits, S., Lesti, A., Livi, S., Rossi, A.r., Crosetti, D., Sola, L., 2004. Microsatellite markers reveal population structure in gilthead sea bream Sparus aurata from the Atlantic Ocean and Mediterranean Sea. Fisheries Science 70, 852-859. De innocentits, S., Miggiano, E., Ungaro, A., Livi, S., Sola, L., Crosetti, D., 2005. Geographical origin of individual breeders from gilthead sea bream (Sparus aurata) hatchery broodstocks inferred by microsatellite profiles. Aquaculture 247, 227-232. Doupé, R.G., Lymbery, A.J., 2005. Genetic covariation in production traits of sub-adult black bream Acanthopagrus butcheri after grow-out. Aquaculture Research, 36, 1128-1132. Du, C.J., Sun, D.W., 2004.Recent developments in the applications of image processing techniques for food quality evaluation. Trends Food Sci. Technol. 15, 230–249. Dunham, RA., Brummett, RE., 1999. Response of two generations of selection to increased body weight in channel catfish, Ictalurus punctatus, compared to hybridization with blue catfish, I. furcatus, males. Journal of Applied Aquaculture 9, 37-45. 5. reference 92 Dupont-Nivet, M., Vandeputte, M., Vergnet, A., Merdy, O., Haffray, P., Chavanne, H., Chatain, B., 2008. Heritabilities and GxE interactions for growth in the european sea bass (Dicentrarchus labrax L.) using a marker-based pedigree. Aquaculture 275, 81–87. Ellegren, H., 2000. Microsatellite mutations in the germline: implications for evolutionary inference. Trends Genet. 16, 551-558. Falconer D.S., Mackay T.F.C., 1996. Introducción a la genética cuantitativa. Acribia, Zaragoza, pp. 469. FAO. 2015. http://www.fao.org/fishery/culturedspecies/Sparus_aurata/en. FEAP, 2008.The Federation of European Aquaculture Producers, production and price report of member associations of the F.E.A.P, 20012008. Fernandes, T. F., Miguel, A. T., Jesús, F., Ania, P. Q., Miguel, H., Marine, H., Guillermo, B. D., María, D. L. B., Paulino, M., 2010. Estimación de parámetros genéticos para caracteres de crecimiento en Dorada (Sparus aurata L.). XV Reunión nacional de mejora genética animal, 16-18 junio, Vigo. Fernández-Palacios, H., 2005. Efecto de determinados nutrientes en la composición de dietas para reproductores de dorada (sparus aurata) sobre la calidad de sus puestas. Ph.D. Thesis, University of Las Palmas de Gran Canaria, Spain. Fessehaye, Y., El-Bialy, Z., Rezk, MA., Crooijmans, R., Bovenhuis, H., Komen, H., 2006. Mating systems and male reproductive success in Nile tilapia (Oreochromis niloticus) in breeding hapas: a microsatellite analysis. Aquaculture 256, 148–158. Fishback, A.G., Danzmann, R.G., Ferguson, M.M., Gibson, J.P., 2002. Estimates of genetic parameters and genotype by environment interactions for growth traits of rainbow trout (Oncorhyncus mykiss) as inferred using molecular pedigrees. Aquaculture 206, 137– 150. Flos, R., Reig, L., Oca, J., Ginovart, M., 2002. Influence of marketing and different landbased system on gilthead sea bream (Sparus aurata) quality. Aquac. Int. 10, 189-206. Frost, L. A., Brad S. Evans, Dean R. Jerry., 2006. Loss of genetic diversity due to hatchery culture practices in barramundi (Lates calcarifer). Aquaculture 261: 1056–1064. 5. Reference 93 Funes, V., Zuati, E., Catanese, G., Infante, C., Manchado, M., 2004. Isolation and characterization of ten microsatellite loci for Senegal sole (Solea senegalensis Kaup). Molecular Ecology Notes, 4, 339– 341. Funkenstein, B., Cavari, B., Stadie, T., Davidovitch, E., 1990. Restriction site polymorphism of mitochondrial DNA of the gilthead sea bream (Sparus aurata) broodstock in Eilat, Israel. Aquaculture 89, 217-223. GAFRD, 2016.General Authority for Fish Resources Development. Egypt. Gall, GAE., Huang, N., 1988. Heritability and selection schemes for rainbow trout: body weight. Aquaculture73, 43-56. García de León, F. J., M. Cannonne, E. Quillet, F. Bonhomme y B. Chatain., 1998. The application of microsatellite markers to breeding programmes in the sea bass, Dicentrarchuslabrax. Aquaculture 159: 303-316. García-Celdrán, M., Ramis, G., Manchado, M., Estévez, A., Afonso, J.M., María-Dolores, E., Peñalver, J., Armero, E., 2015a. Estimates of heritabilities and genetic correlations of growth and external skeletal deformities at different ages in a reared gilthead sea bream (Sparus aurata L.) population sourced from three broodstocks along the Spanish coasts. Aquaculture 445: 33–41. García-Celdrán, M., Ramis, G., Manchado, M., Estévez, Afonso, J.M., Armero, E., 2015c. Estimates of heritabilities and genetic correlations of carcass quality traits in a reared gilthead sea bream (Sparus aurata L.) population sourced from three broodstocks along the Spanish coasts. Aquaculture 446; 175–180. García-Celdrán, M., Ramis, G., Manchado, M., Estévez, Navarro, A., Armero, E., 2015b. Estimates of heritabilities and genetic correlations of raw flesh quality traits in a reared gilthead sea bream (Sparus aurata L.) population sourced from broodstocks along the Spanish coasts. Aquaculture 446; 181–186. García-Celdrán, M., Ramis, G., María-Dolores, E., Peñalver, J., Borrell, Y. J., Manchado, M., Estévez, A., Afonso, J.M., Armero, E., 2016. Genetic assessment of three gilthead sea bream (Sparus aurata L.) populations along the Spanish coast and of three broodstocks managements.Aquaculture International (in press). DOI 10.1007/s10499016-9998-8. 5. reference 94 Ginés, R., 1997. Posibilidades de manipulación del fotoperíodo durante el engorde intensivo de dorada (Sparus aurata). PhD Thesis, University of las Palmas de gran canaria, Spain. Ginés, R., Afonso, J.M., Argüello, A., Zamorano, M.J., López J.L., 2004. The effects of longday photoperiod on growth, body composition and skin colour in immature gilthead sea bream (Sparus aurata L.). Aquaculture Research, 35, 1207-1212. Ginés, R., Afonso, J.M., Argüello, A., Zamorano, M.J., López, J.L., 2003. Growth in adult gilthead sea bream (Sparus aurata L.) as a result of interference in sexual maturation by different photoperiod regimes. Aquaculture Research, 34, 73-83. Gjedrem, T., 1997. Flesh quality improvement in fish through breeding. Aquaculture International, 5, 197-206. Gjedrem, T., 2000. Genetic improvement of cold-water fish species.Aquacult. Res. 31, 25–34. Gjedrem, T., 2005. Selection and Breeding Programs in Aquaculture. Ed. Gjedrem, T., Springer, Dordrecht, The Netherlands. 364 pp. Gjedrem, T., J. Thodesen., 2005. Selection. In: Gjedrem (ed.), Selection and breeding programs in aquaculture. Springer, Heidelberg, Germany. Gjedrem, T., Robinson, N., Morten, R., 2012. The importance of selective breeding in aquaculture to meet future demands for animal protein: A review. Aquaculture, 350353, 117-129. Gjerde, B., Gjedrem, T., 1984. Estimates of phenotypic and genetic parameters for carcass traits in Atlantic salmon and rainbow trout. Aquaculture 36, 97-110. Gjerde, B., Gjoen, HM., Villanueva, B., 1996. Optimum designs for fish breeding programmes with constrained inbreeding. Mass selection for a normally distributed trait. Livest Prod Sci 47, 59–72. Gjerde, B., Gunnes, K., Gjedrem, T., 1983. Effect of inbreeding on survival and growth in rainbow trout. Aquaculture 34, 327-332. Gjerde, B., Pante, M.J., Baeverfjord, G., 2005. Genetic variation for a vertebral deformity in Atlantic salmon (Salmo salar). Aquaculture, 244, 77-87. Gjerde, B., Rye, M., 1998. Design of breeding programmes in aquaculture species: possibilities and constraints. In, Bartley, D and Basurco, B (Eds) Genetics and Breeding of Mediterranean Aquaculture Species.Proceedings of the Seminar of the CIHEAM 5. Reference 95 Network on Technology of Aquaculture in the Mediterranean (TECAM), April 1997, CIHEAM, Zaragoza, Spain.pp181-192. Cahier Options Mediterrennes, vol 34. Gjerde, B., Schaeffer, L.R., 1989. Body traits in rainbow trout: II. Estimates of heritabilities and of phenotypic and genetic correlations. Aquaculture, 80, 25-44. Gordin, H., Zohar, Y., 1978. Induced spawning of Sparus aurata (L.) by means of hormonal treatments. Ann. Biol. Anim. Bioch. Biophys. 18, 985-990. Gorshkov, S., Gordin, H., Gorshkova, G., Knibb, W., 1997. Reproductive constraints for family selection of the gilthead sea bream (Sparus aurata). Isreali Journal of Aquaculture - Barmidgeh 49, 124-134. Greenwood, P. H., Rosen, D. E., Weitzman, S. H. Myers, G.S., 1966. Phyletic studies of the teleostean fishes, with a provisional classification of living forms. Bulletin American Museum Natural History 131, 4:340-455. Grigorakis, K., 2007. Compositional and organoleptic quality of farmed and wild gilthead sea bream (Sparus aurata) and sea bass (Dicentrarchus labrax) and factors affecting it: A review. Aquaculture, 272, 55–75. Groeneveld, E., Kovaˇc, M., Mielenz, N., 2010.VCE User’s Guide and Reference Manual Version 6.0 (ftp://ftp.tzv.fal.de/pub/vce6/doc/vce6-manual-3.1-A4.pdf). Grundy, B., W. G. Hill., 1993. A method for reducing inbreeding with best linear unbiased prediction. Anim. Prod. 56:427. Gümüş, B., Balaban, M.Ö., Ünlüsayın, M., 2011. Machine Vision Applications to Aquatic Foods: A Review. Turkish Journal of Fisheries and Aquatic Sciences 11: 171-181. Hashimoto, D.T., Mendonça, F.F., Senhorini, J.A., Bortolozzi, J., Oliveira, C., Foresti, F., Porto-Foresti, F., 2010. Identification of hybrids between Neotropical fish Leporinus macrocephalus and Leporinus elongatus by PCR-RFLP and multiplex-PCR: Tools for genetic monitoring in aquaculture. Aquaculture 298, 346-349. Hashimoto, D.T., Parise-Maltempi, P.P., Laudicina, A., Bortolozzi, J., Senhorini, J.A., Foresti, F., Porto-Foresti, F., 2009. Repetitive DNA probe linked to sex chromosomes in hybrids between Neotropical fish Leporinus macrocephalus and Leporinus elongatus (Characiformes, Anostomidae). Cytogenet. Genome Res. 124, 151–157. 5. reference 96 Hely, F. S., Amer, P. R., Walker, S. P., Symonds, J. E., 2013. Optimised parent selection and minimum inbreeding mating in small aquaculture breeding schemes: a simulation study. Animal, 7:1, pp 1–10. Henryon, M., Sørensen, AC., Berg, P., 2009. Mating animals by minimising the covariance between ancestral contributions generates less inbreeding without compromising genetic gain in breeding schemes with truncation selection. Animal, 3:1339-1346. Herbinger, C., O’Reilly, P., Doyle, R., Wright, J., Flynn, F., 1999. Early growth performance of Atlantic salmon full-sib families reared in single family tanks versus in mixed family tanks. Aquaculture, 173, 105-116. Herbinger, CM., Newkirk, GF., 1990. Sources of family variability for maturation incidence in cultivated Atlantic salmon. Aquaculture 85, 153-162. Hulata G., Wohlfart G.W., Halevy., 1986. Mass selection for growth rate in the Nile tilapia (Oreochromis niloticus). Aquaculture 57:177-184. Iwamoto, R.N., Myers, J.M., Hershberger, W.K., 1990. Heritability and genetic correlations for flesh coloration in pen-reared coho salmon. Aquaculture 86, 181-190. Janssen. K., Chavanne H., Berentsen P., Komen H., 2016. Impact of selective breeding on European aquaculture. Aquaculture (in press). Jorstad, KE., Naevdal, G., 1996. Breeding and Genetics.In, Penner, W and Barton, BA (Eds) Principles of Salmonid Culture.Developments in Aquaculture and Fisheries Science, vol 29, pp.665-725. Elsevier, Amsterdam. Jover, M., 2007. Alternativas de futuro para la producción de la dorada (Sparus aurata). XI Congreso Nacional de Acuicultura, Vigo, 1337-1342. Kause, A., D. Tobin, D. F. Houlihan, S. A. M. Martin, E. A. Ma¨ntysaari, O. Ritola, and K. Ruohonen., 2006. Feed efficiency of rainbow trout can be improved through selection: Different genetic potential on alternative diets. J. Anim. Sci. 84:807–817. Kause, A., Paananen, T., Ritola, O., Koskinen, H., 2007. Direct and indirect selection of visceral lipid weight, fillet weight, and fillet percentage in a rainbow trout breeding program. Journal of Animal Science, 85, 3218-3227. Kause, A., Ritola, O., Paananen, T., Eskelinen, U., Mäntysaari, E., 2003. Big and beautiful?Quantitative genetic parameters for appearance of large rainbow trout. J. Fish Biol. 62, 610–622. 5. Reference 97 Kause, A., Ritola, O., Paananen, T., Mäntysaari, E., Eskelinen. U., 2002.Coupling body weight and its composition: A quantitative genetic analysis in rainbow trout. Aquaculture 211, 65–79. Kause, A., Ritola, O., Paananen, T., Wahlroos, H., Mäntysaari, E.A., 2005. Genetic trends in growth, sexual maturity and skeletal deformations, and rate of inbreeding in a breeding programme for rainbow trout. Aquaculture, 247, 177–187. Kettunen, A., Fjalestad, K.T., 2007. Genetic parameters for important traits in the breeding program for Atlantic cod (Gadus morhua L.). Aquaculture 272 (Supplement 1), S276. Kinghorn, B., 1983. Genetic variation in food conversion efficiency and growth in rainbow trout. Aquaculture 32: 141–155. Kirpichnikov, V.S., 1981.Genetic Basis of Fish Selection. Springer Verlag, Berlin. Knibb, W. R., 2000. Genetic improvement of marine fishwhich method for industry? Aquaculture, 31: 11-23. Knibb, W., Gorshkova, G., Gorshkov, S., 1998. Selection and crossbreeding in Mediterranean cultured marine fish (Review). Cahiers Options Méditerranéennes, 34, 47-60. Knibb, W.R., Gorshkova, G., Gorshkov, S., 1997. Selection for growth in the gilthead sea bream (Sparus aurata). The Israeli Journal of Aquaculture-Bamidghe, 49, 57-66. Knight, B., 1985. Feeding behaviour and fish culture.En "Nutrition and feeding in fish" (C.B. Cowey, A.M. Mackie y J.G. Bell, eds.), Academic Press, Londres, 1985, pp. 223-241. Kocour, M., Linhart, O., Vandeputte, M., 2006. Mouth and fin deformities in common carp: is there a genetic basis?. Aquaculture Research 37, 419–22. Kocour, M., Mauger, S., Rodina, M., Gela, D., Linhart, O., Vandeputte, M., 2007. Heritability estimates for processing and quality traits in common carp (Cyprinus carpio L.) using a molecular pedigree. Aquaculture, 270, 43–50. Kohlmann, K., Gross, R., Murakaeva, A., Kersten, P., 2003. Genetic variability and structure of common carp (Cyprinus carpio) populations throughout the distribution range inferred from allozyme, microsatellite and mitochondrial DNA markers. Aquat Living Resour 16, 421-431. 5. reference 98 Kolstad, K., Thorland, I., Refstie, T., Gjerde, B., 2006. Genetic variation and genotype by location interaction in body weight, spinal deformity and sexual maturity in Atlantic cod (Gadus morhua) reared at different locations off Norway. Aquaculture259, 66–73. Lal, K.K., Chauhan, T, Mandal, A., Singh, R.K., Khulbe, L., Ponniah, A.G., Mohindra V., 2004. Identification of microsatellite DNA markers for population structure analysis in Indian major carp, Cirrhinus mrigala (Hamilton Buchanan, 1882). J Appl Ichthyol 20, 87-91. Launey, S., Krieg, F., Haffray, P., Bruant, J.S., Vanniers, A., Guyomard, R., 2003. Twelve new microsatellite markers for gilthead sea bream (Sparus aurata L.): characterization, polymorphism and linkage. Mol. Ecol. 3, 457-459. Lee, W.J., Kocher, T.D., 1998. Microsatellite mpping of prolactin locus in tilapia genome.Amin. Genet. 29, 68-69. Lee-Montero, 2012. Desarrollo de un programa piloto a nivel nacional de mejora genética en dorada (Sparus aurata L.): estimación de parámetros genéticos de caracteres de crecimiento y calidad e interacción genotipo ambiente. PhD Thesis, University of las Palmas de gran canaria, Spain. Lee-Montero, A. Navarro, D. Negrın-Baez, M. J. Zamorano, C. Berbel, J. A. Sanchez, M. Garcıa-Celdran, M. Manchado, A. Estevez, E. Armero, J. M. Afonso., 2015. Genetic parameters and genotype–environment interactions for skeleton deformities and growth traits at different ages on gilthead sea bream (Sparus aurata L.) in four Spanish regions. Animal Genetics 46, 164–174. Lee-Montero, A. Navarro, Y. Borrell, M. Garcıa-Celdra´n, N. Martın, D. Negrın-Baez, G. Blanco, E. Armero, C. Berbel, M. J. Zamorano, J. J. Sanchez, A. Este´vez, G. Ramis, M. Manchado, J. M. Afonso., 2013. Development of the first standardised panel of two new microsatellite multiplex PCRs for gilthead sea bream (Sparus aurata L.). animal genetics, 44, 533–546. Li, D., Kang, D., Yin, Q., Sun, X., Liang, L., 2007. Microsatellite DNA Marker Analysis of Genetic Diversity in Wild Common Carp (Cyprinus carpio L.) Populations. Genetics and Genomics 34, 984-993. Liang, L.Q., Sun, X.W., 2003. Mapping cold tolerance strain on genetic linkage map of common carp. J Dalian Fish Univ, 18, 278-281. (In Chinese with an English abstract). 5. Reference 99 Litt, M., Luty, J.A., 1989. A hypervariable microsatellite revealed by in vitro amplification of dinucleotide repeat within the cardiac muscle actin gene. Am. J. Hum. Genet. 44, 397401. Liu, Z., Li, P., Kocabas, A., Karsi, A., Ju, Z., 2001. Microsatellite containing genes from the channel catfish brain: evidence of trinucleotide repeat expansion in the coding region of nucleotide excision repair gene RAD23B. Biochem.Biophys. Res. Commun. 289, 317324. Liu, Z., Tan, G., Li, P., Dunham, R.A., 1999. Transcribed dinucleotide microsatellites and their associated genes from channel catfish Ictalurus punctatus.Biochem.Biophys. Res. Commun. 259, 190-194. Liu, Z.J., Cordes, J.F., 2004. DNA marker technologies and their applications in aquaculture genetics. Aquaculture 238, 1-37. López-Fanjul C., Toro, M.A., 2007. Fundamentos de la mejora genética en acuicultura. En: Martínez, P., Figueras, A. (Eds.), Genética y genómica en acuicultura. Madrid, pp. 155182. Lumare, F., Villani, P., 1973. Matuita sessuale indotta e feconfazione artificial in Sparus aurata (L.). Inv. Pesq. 37, 57-71. Luna, L., Fernández, J., Incera, M.T., Duque, D, Fernández, J., García, R., Fernández, C., 2006. La demanda de filete de dorada en el mercado Español. Ministerio de Agricultura, Pesca y Alimentación. ISBN: 84-491-0707-5. Lythagoe, J.N., Lythageo, G., 1992. Fishes of the sea: the North Atlantic and Mediterranean. The MIT Press, Cumberland, Rhode Island, U.S.A. pp 113-114. Magoulas, A., 1999. Application of molecular markers to aquaculture and broodstock management with special emphasis on microsatellite DNA. Cahiers Options Mediterrannes 34, 153-168. Markoulatos, P., Siafakas, N., Moncany, M., 2002. Multiplex Polymerase Chain Reaction: A Practical Approach. Clinical Laboratory Analysis 16, 47-51. McKay, LR., Ihssen, PE., Friars, GW., 1986. Genetic parameters of growth in rainbow trout, Salmo gairdneri, as a function of age and maturity. Aquaculture 58, 241-254. 5. reference 106 Sonesson, AK., Meuwissen, THE., 2000. Mating schemes for optimum contribution selection with constrained rates of inbreeding.Genet Sel Evol, 32:231-248. Soula, M., 2012. Estimación de parámetros genéticos en corvina, Argyrosomus regius, para caracteres de crecimiento, rendimiento y calidad de la carne y del pez. PhD Thesis, University of las Palmas de Gran Canaria, Spain. Souza, A.F.L., 2008. Rendimento, composição química e perfil de minerais das principais espécies de peixes comercializadas no estado do Amazonas. 132f. Dissertação (Mestrado em Ciências Agrárias) Universidade Federal do Amazonas, Manaus. Standal, M., Gjerde, B., 1987. Genetic variation in survival of Atlantic salmon during the sea rearing-period. Aquaculture 66, 197–207. Stockley, B.M., Rogers, A.D., Iyengar, A., Menezes, G., Santos, R., Long, A., 2000. Ten microsatellite loci isolated and developed for the blackspot seabream, Pagellus bogaraveo (Brunnich 1768). Mol. Ecol. 9, 999-1000. Streelman, J.T., Kocher, T.D., 2002. Microsatellite variation associated with prolactin expression and growth of salt-challenged tilapia. Physiol. Genomics 9, 1-4. Suau, P., Lopez, J., 1976. Contribution to the studio of the gilthead sea brem, Sparus aurata L. Inv. Pesq. 40, 160-199. Sun, X.W., Liang, L.Q., 2004. A genetic linkage map of common carp (Cyprinus carpio L.) and mapping of a locus associated with cold tolerance. Aquaculture 238, 165-172. Takagi, M., Taniguchi, N., Cook, D., Doyle, R.W., 1997. Isolation and characterization of microsatellite loci from red sea bream Pagrus major and detection in closely related species. Fish. Sci. 63, 199-204. Tanck, M.W., Palstra, A.P., van der Weerd, M., Leffering, C.P., van der Poel, J.J., Bovenhuis, H., Komen, J., 2001. Segregation of microsatellite alleles and residual heterozygosity at single loci in homozygous androgenetic common carp (Cyprinus carpio L.). Genome 44, 743-751. Tautz, D., 1989. Hypervariability of simple sequences as a general source for polymorphic DNA markers. Nucleic Acids Res. 17, 6463-6471. 5. Reference 107 Tautz, D., Renz, M., 1984. Simple sequences are ubiquitous repetitive components of eukaryote genomes. Nucleic Acids Res. 17, 4127-4138. Tave, D., 1986. Genetics for Fish Hatchery Managers. AVI Publishing Company, Inc., Westport, Connecticut, p. 299. Tave, D., 1993. Genetics for fish hatchery managers (2nd Ed). Kluwer Academic Publishers, Boston Tave, D., 1995. Selection Schemes in Medium Sized Fish Farms. FAO Technical Series 352, Rome. Teichert-Coddington, DR., Smitherman, RO., 1988. Lack of response by tilapia to mass selection for rapid early growth. Transactions of the American Fisheries Society 117, 297-300. Thorland I., Papaioannou, N., Kottaras, L., Refstie, T., Papasolomontos, S., Rye, M., 2006. Family based selection for production traits in gilthead sea bream (Sparus aurata) and European sea bass (Dicentrarchus labrax) in Greece. Genetics in Aquaculture IX, p104 (presentación oral). Montpellier, France. Thorland, I., Kottaras, L., Refstie T., Dimitroglou, A., Papaharisis L., Rye M., 2015. Response to selection for harvest weight in a family based selection program of gilthead sea bream (Sparus aurata). XII international symposium on genetics in aquaculture, Santiago de Compostela, Spain. Toth, G., Gaspari, Z., Jurka, J., 2000. Microsatellites in different eukaryotic genomes: survey and analysis. Genome Res. 10, 967-981. Vandeputte, M., Kocour, M., Mauger, S., Dupont-Nivet, M., De Guerry, D., Rodina, M., Gela, D., Vallod, D., Chevassus, B., Linhart, O., 2004. Heritability estimates for growth-related traits using microsatellite parentage assignment in juvenile common carp (Cyprinus carpio L.). Aquaculture 235, 223-236. Vandeputte, M., Mauger, S., Dupont-Nivet, M., 2006. An evaluation of allowing for mismatches as a way to manage genotyping errors in parentage assignment by exclusion. Molecular Ecology Notes, 6, 265–267. 5. reference 108 Villanueva, B., Wooliams, JA., Gjerde, B., 1996. Optimum designs for breeding programmes under mass selection with an application in fish breeding. Anim Sci 63, 563–576. Vos, P., Hogers, R., Bleeker, M., Reijans, M., van de Lee, T., Hornes, M., Frijters, A., Pot, J., Pelemant, J., Kuiper, M., Zabeay, M., 1995. AFLP: a new technique for DNA fingerprinting. Nucleic Acids Res. 23, 4407-4414. Welsh, J., McClelland, M., 1990. Fingerprinting genomes using PCR with arbitrary primers. Nucleic Acids Res. 18, 7213-7218. Wesmajervi, M.S., Westgaard, J.I., Delghandi, M., 2006. Evaluation of a novel pentaplex microsatellite marker system for paternity studies in Atlantic cod (Gadus morhua L.). Aquacult. Res. 37, 1195-1201. Williamson, K.S., Cordes, J.F., May, B., 2001. Characterization of microsatellite loci in Chinook salmon (Oncorhynchus tshawytscha) and cross-species amplification in other salmonids. Mol. Ecol. Notes 2, 17-19. Wilson, F.I., Bourke, E.A., Cross, T.F., 1995. Genetic variation at traditional and novel allozyme loci, applied to interactions between wild and reared salmo salar L. Atlantic salmon. Heredity 75, 578-588. Wohlfarth, GW., Moav, R., 1985. Communal testing, a method of testing the growth of different genetic groups of common carp in earthen ponds. Aquaculture 48, 143-157. Yue, G.H., Ho, M.Y., Orban, L., Komen, J., 2004. Microsatellites within genes and ESTs of common carp and their applicability in silver crucian carp. Aquaculture 234, 85-98. Zohar, Y., Mylonas, CC., 2001. Endocrine manipulations of spawning in cultured fish: from hormones to genes. Aquaculture 197, 99-136. ANEXO RESUMEN EN ESPAÑOL Resumen español 111 Resumen La dorada (Sparus aurata L.), es una de las especies más importantes de la acuicultura Mediterránea. En el último quinquenio, la implementación de programas de mejora genética ha atraído la atención de productores al objeto de contribuir al crecimiento sostenible y competitivo de las empresas. La implementación de programas de mejora genética en dorada introduce costes a través de la racionalización de la producción, a la vez que muestra poco progreso a corto plazo, pero sin embargo es sabido que se trata de una herramienta que tiene beneficios continuos, acumulativos y permanentes para la industria. El principal objetivo del presente estudio ha sido estudiar la segunda generación del programa de mejora genética PROGENS A®, para implementar y estudiar caracteres tecnológicos no invasivos (KET’s), y diferentes modelos de gestión de reproductores que permitan mejorar la contribución familiar. Se estudian las correlaciones genéticas entre las KET’s y caracteres biológicos de interés comercial, bajo condiciones industriales de cultivo, con el fin de identificar metodologías nuevas para la evaluación genética masiva, preciso y económica de animales. Las interacciones Genotipo-Ambiente son consideradas para los caracteres KET’s y caracteres biológicos, en dos sistemas de cultivo: jaulas y esteros. Se han estudiado tres diferentes modelos de puesta masal, de acuerdo con el número de días de puesta. Un día (a pequeña escala 1DS, y a gran escala 1DL), dos días (2DL) y cuatro días (4DL) consecutivos bajo condiciones industriales, con el fin de optimizar la estructura de cría de manera factorial. Todas las puestas fueron cultivas de manera similar utilizando diferente número de reproductores y proporción de sexos. En el muestreo, se estimaron valores de crecimiento mediante el peso, la longitud, el factor de condición y la tasa específica de crecimiento. Los genotipos y la matriz de parentesco fueron estimados mediante el uso de marcadores microsatélites en reacciones múltiples de PCR, interespecíficas (RimA) y específicas (SMsa1). En el modelo 1DS, se constituyeron tres lotes de reproductores independientes, de cada uno de los cuales se obtuvo un puesta. En total, se utilizaron 18 reproductores entre los tres lotes (6 reproductores por cada uno), de los cuales 17 contribuyeron a alguna de las puestas masales (94%), para producir un total de 21 familias de hermanos completos y 16 familias de medios hermanos (maternas y paternas). En el modelo 1DL, compuesto por 66 reproductores, sólo 28 peces (42%), 17 hembras y 11 machos, contribuyeron a la puesta. Se Resumen español 112 produjo un total de 89 familias de hermanos completos de las que 24 fueron de medios hermanos (8 paternas y 16 maternas). En el modelo 2DL, contribuyeron 139 reproductores (67%), y se establecieron 297 familias de hermanos completos y 105 de medios hermanos (52 paternas y 53 maternas). En el modelo 4DL, constituido por 123 reproductores, 83 peces (67%), 40 hembras y 43 machos, contribuyeron a producir un total de 201 familias de hermanos completos, de las que 59 fueron de medios hermanos (32 paternas y 27 maternas). Se estudió la existencia de variabilidad genética aditiva en caracteres biológicos y sus equivalentes desarrollados mediante tecnologías de ingeniería o caracteres tecnológicos, así como sus correlaciones genéticas con otros caracteres de crecimiento, rendimiento, carcasa y calidad de la carne, a través de poblaciones de descendientes provenientes de reproductores establecidos en dos centros de investigación (PCTM en las Islas Canarias, IFAPA en Andalucía). Estos pertenecen a la segunda generación del programa de mejora genética PROGENSA (denominada INNOTECSS), y fueron marcados individualmente mediante PIT, y remitidos a ambos centros de investigación y empresas del sector (ADSA y PIMSA). A la talla de sacrificio, los peces fueron muestreados para crecimiento, carcasa, calidad de la carne y los caracteres tecnológicos, estos últimos mediante el programa IMAFISH. Muestras de las aletas caudales de los peces fueron cogidas y conservadas en etanol para los análisis de pedigrí. El genotipado y la asignación de parentesco fueron inferidos mediante la caracterización genética de todos los reproductores y los descendientes, usando la múltiple de PCR SMsa-1 (Super Múltiple Sparus aurata), conteniendo 11 marcadores microsatélites específicos. Se estimaros los parámetros genéticos (heredabilidad y correlación genética y fenotípica) para caracteres de crecimiento, rendimiento, calidad de la carne y la morfología. A la talla de sacrificio, se estudió también la interacción genotipo-ambiente (GxE), entre los dos sistemas de engorde (ADSA, jaula; PIMSA, estero). Las heredabilidades estimadas oscilaron entre 0,09 y 025 para el crecimiento, 0,04 y 0,35 para la carcasa, 0,08 y 0,27 para la composición corporal y 0,04 y 0,40 para los caracteres tecnológicos, que fueron las estimas más altas. Las estimas de las correlaciones genéticas para los caracteres de crecimiento fueron principalmente medias y altas, siendo medias para la carcasa. Para la composición corporal, fueron negativas y altas para la humedad frente al contenido en lípidos, negativa y baja para el contenido en proteínas y la humedad o lípidos, y altas para los caracteres tecnológicos. Las correlaciones genéticas entre Resumen español 113 la carcasa y los caracteres de crecimiento fueron medias y altas, excepto para el rendimiento del filete frente a la longitud que fue baja. Dichos valores fueron bajo – medios entre los caracteres de composición y los de crecimiento, excepto con la humedad que fueron medias y negativas. Para los caracteres tecnológicos, todas las correlaciones genéticas con el crecimiento fueron media-altas. Para los caracteres tecnológicos con los de la carcasa los valores fueron altos, mientras que para con el peso del filete y el rendimiento filete fueron bajos. Los caracteres tecnológicos mostraron correlaciones genéticas medias con los contenidos en proteína y lípidos, y medias y negativas con la humedad. Las interacciones genotipo-ambiente de todos los caracteres, a la talla de sacrificio, fueron medidas considerando los dos sistemas de cultivo de dorada, jaula y esteros en las Islas Canarias Andalucía, respectivamente. Las estimas fueron principalmente medias-altas para todos los caracteres (entre 0,46 y 1), excepto para FHA y FHB, los cuales fueron bajos (0,05 y 0,16, respectivamente). Todos los resultados revelaron que el uso de puestas masales de días consecutivos permite maximizar la estructura familiar factorial, a través del incremento del número de reproductores que contribuyen a la descendencia y de familias de hermanos completos y medios hermanos, el control de la consanguinidad y la ganancia genética, por lo que es recomendable su uso en el diseño de programas de mejora genética en dorada. Los datos también ponen de relieve que los caracteres tecnológicos no invasivos (KET’s), mostraron mejor variación genética aditiva que los caracteres biológicos, a la vez que una mayor precisión, lo que los erige como herramientas que pueden ser utilizadas en programas de mejora genética de dorada, permitiendo así un mayor respuesta a la selección para el crecimiento y la morfología de los peces, a través del incremento de la exactitud de la estima para los valores mejorantes y las tasas de respuesta a la selección. Resumen español 114 1 Introducción general 1.1 Dorada (Sparus aurata L.) 1.1.1 Taxonomía La dorada Sparus aurata (Linnaeus 1758) tiene la siguiente clasificación taxonómica, según Greenwood et al. (1966): Phylum : Chordata Subphylum : Gnathostomata Clase : Actinopterygii Subclase : Teleostei Superorden : Neognathi Orden : Perciformes Familia : Sparidae Género : Sparus Especie : Sparus aurata 1.1.2 descripciones familiares Los peces de la familia Sparidae ha oblonga cuerpo y por lo general lateralmente comprimido. Ellos son de agua salada y la mayoría de ellos son demersales, y se encuentran en aguas poco profundas. Sus huevos son pelágicos, esféricos (con un diámetro de alrededor de 1 mm) y tienen una gota de aceite. La mayoría de los espáridos son hermafroditas: son primeros machos y se convirtieron a hembras al alcanzar la madurez sexual (hermafroditismo protándrico) o son primeras hembras y se convirtieron en machos (hermafroditismo protógino). Debido a su excelente carne, muchos representantes de esta familia tienen un alto valor comercial. Espáridos incluyen muchos géneros y un gran Resumen español 115 número de especies (más de 100) que viven en todos los mares tropicales y templadas, incluidas el agua excepcionalmente frías y salobres. Sólo en el Mediterráneo, hay once géneros: Dentex, Sparus, Diplodus, Pagellus, Pagrus, Lithognatus, Spondyliosoma, Oblada, Crenidens, Boops y Sarpa (FAO, 2009). 1.1.3 características morfológicas y Biología La dorada es un pez teleósteo de agua salada. Presenta un cuerpo con una forma ovalada, muy alto y comprimido lateralmente. El perfil de la cabeza es convexo con pequeños ojos. Las mejillas están cubiertas de escamas, y el hueso pre-opérculo es sin escamas. La boca tiene la mandíbula más corta que el maxilar superior. Ambas mandíbulas muestran caninos (4-6) y dientes con forma molar, en series de 2-4 en el maxilar superior y en series de 3-4, de las cuales 1-2 son notablemente más grande, en el maxilar inferior. Las hendiduras branquiales son cortas, 11-13 en el primer arco branquial y 7-8 en la parte baja. La línea lateral tiene 75-85 escamas. La aleta dorsal presenta 11 radios duros y 13 radios blandos, la aleta anal tiene 3 radios duros y 11-12 radios blandos. Las aletas pectorales son largas y puntiagudas, mientras que las ventrales son cortas. La aleta caudal tiene lóbulos puntiagudos. Todas las vértebras y las costillas presentan parapófisis y ausencia de costillas sésiles (FAO, 2009). El color de la dorada es de color gris plateado con una gran mancha oscura en el comienzo de la línea lateral que cubre también la parte superior del hueso del opérculo. Una banda dorada y negra se encuentra entre los ojos, la dorada siempre se estrecha en la parte central. La aleta dorsal es de color azul-gris con una línea mediana de negro. La aleta caudal es de color blanco gris verdoso con puntas negras (Figura 1-1). 1.1.4 Distribución geográfica y hábitat La dorada tiene una importante distribución geográfica. Se puede encontrar en el Mar Mediterráneo y el Océano Atlántico desde Gran Bretaña a Senegal, del mismo modo que su presencia es rara en el Mar Negro (Lythagoe, 1992; Moretti et al., 1999). Debido a sus hábitos eurihalinos y euritérmicos, esta especie se encuentra en ambos ambientes de agua salobre, tales como lagunas costeras y estuarios, en particular durante las primeras etapas de su ciclo de vida marina. Nacido en el mar abierto durante octubre- Resumen español 122 acuicultura, y desde 1999, y los años subsiguientes, han dominado este tipo de estudios (Magoulas, 1999). 1.2.3 Los marcadores microsatélites Los microsatélites o repeticiones de secuencias simples (SSR), son marcadores moleculares formados por unidades repetidas en tándem de secuencias cortas de 1-6 nucleótidos, y se distribuyen en abundancia través de los genomas (Tautz, 1989; Litt y Luty, 1989). Los microsatélites muestran altos niveles de polimorfismo alélico (Chistiakov et al., 2006). Los alelos difieren en la longitud de los fragmentos de ADN que contiene el microsatélite, debido a la variación en el número de repeticiones de la región consenso. La tasa de mutación de los microsatélites se estima entre 10-2 y 10-6 por locus por generación (Ellegren, 2000). Parece causada por del deslizamiento de la polimerasa durante la replicación del ADN, dando lugar a diferencias de resultado en el número de repeticiones de la región de consenso (Tautz, 1989). La mayoría de los microsatélites son dinucleótidos (3067%). (AC)n es el motivo dinucleótido más común en el genoma de los vertebrados. Es 2,3 veces más frecuente que (AT)n, que es el segundo tipo más general de dinucleótidos (Toth et al., 2000). Los microsatellites son marcadores codominantes con la herencia mendeliana simple. Ellos son relativamente pequeños y pueden amplificarse fácilmente mediante PCR utilizando cebadores, que están diseñados a partir de sus regiones flanqueantes. Los productos de amplificación se separan por tamaño a través de electroforesis para detectar los alelos y el polimorfismo de los microsatélites. En el ámbito de la pesca y la acuicultura, los microsatélites son útiles para el estudio de la variabilidad genética y la consanguinidad, las asignaciones de paternidad, identificación genética de las poblaciones, la selección y los programas de mejora genética, la construcción de mapas de ligamiento densos, determinación de QTL y los programas de mejora asistida por marcadores. Por lo tanto, un gran número de marcadores de microsatélites se ha descrito en las especies de peces de acuicultura. Al igual que en el salmón del Atlántico (Sletten et al., 1997; Skaala et al., 2004), el pez gato (Liu et al., 1999, 2001; Serapion et al., 2004), tilapia ( Lee y Kocher, 1998; Palti et al., 2001; Carleton et al., 2002; Streelman y Kocher, 2002; Cnaani et al., 2002; Rowena et al., 2004), la carpa común (Crooijmans et al., 1997; Tanck et al., 2001; Liang et al., 2003; Kohlmann et al., 2003; Yue et al., 2004; Lal et al., 2004; Sun et al., 2004; Li et al., 2007), salmón chinook (Williamson Resumen español 123 et al., 2001; Naishy Park., 2002), la trucha arco iris (Rexroad et al., 2001, 2002a, 2002b; Nathan et al., 2007), y el lenguado senegalés (Funes et al.,2004). En las especies de espáridos, numerosos marcadores de microsatélites, con un alto polimorfismo, se han utilizado en la dorada (Magoulas, 1999; Batargias et al., 1999; Launey et al., 2003; De Innocentiis et al, 2004, 2005; Brown et al, 2005; Oliva et al., 2005; Senger et al., 2006; Castro et al., 2007; Navarro et al., 2008; Porta et al., 2010; Lee-Montero et al., 2013;Negrín-Báez et al., 2015), Dorada japonesa (Takagi et al., 1997), Dorada australiana (Adcock et al., 2000), la mancha negra dorada (Stockley et al., 2000; Piñera et al., 2006), breca (Ramšak et al., 2003) y sama roquera ( Ponce et al., 2006; Navarro et al., 2008). 1.2.4 Reacciones Múltiple en cadena de la polimerasa En los programas de selección modernos, los marcadores microsatélite permiten la identificación de descendientes criados en un mismo ambiente y además ayudan en los procesos de selección (García de León et al., 1998). El uso de un número de microsatélites alto habitualmente genera altos costos debido a los materiales y el personal requerido. Sin embargo, la reducción de costes se puede lograr a través de la implementación de reacciones múltiples. Es la combinación de los productos de amplificación de la reacción en cadena de la polimerasa (PCR simples) de varios marcadores en una misma reacción, que luego son cargados en un mismo pocillo de gel de electroforesis o secuenciador (Olsen et al., 1996; Chamberlain et al., 1988; Neff et al., 2000). Esta es una variante de la PCR en la que dos o más secuencias diana se pueden amplificar mediante la inclusión de más de un juego de cebadores en la misma reacción. La PCR múltiplex requiere que los cebadores amplifiquen las regiones únicas de ADN, tanto de manera individual como en combinación con otros juegos de cebadores, bajo las mismas condiciones de reacción. Además, los métodos deben estar disponibles para el análisis de cada producto de amplificación individual de la mezcla de todos los productos (Markoulatos et al., 2002). Para el éxito de un ensayo de PCR multiplex es importante la concentración relativa de los cebadores, la concentración del tampón de PCR, el equilibrio entre el cloruro de magnesio (MgCl2) y las concentraciones de desoxinucleótidos (dNTPs), las temperaturas de hibridación, y las cantidades de ADN molde y Taq-polimerasa. Una combinación óptima de la temperatura de hibridación y la concentración del tampón es esencial en la PCR multiplex Resumen español 124 para obtener productos de amplificación altamente específicos. La concentración de MgCl2 sólo tiene que ser proporcional a la cantidad de dNTPs, mientras que el ajuste de concentración de cebador para cada secuencia diana también es esencial (Markoulatos et al., 2002). La presencia de más de un par de cebadores en la PCR multiplex aumenta la posibilidad de la obtención de productos de amplificación no deseables (Markoulatos et al., 2002). Uno de los conceptos más importantes de la PCR multiplex es la óptima relación entre los cebadores. En la optimización de la PCR múltiple debe reducirse al mínimo las interacciones inespecíficas. Los parámetros como la homología de cebadores con sus secuencias objetivo, duración, contenido de GC y la concentración, tiene que ser considerado para el diseño de cebadores (Brownie et al., 1997). La concentración final de los cebadores puede variar considerablemente entre los loci y establecido empíricamente. Cuando hay amplificación desigual, se requiere cambiar las proporciones de varios cebadores en la reacción, el aumento de la concentración de cebadores de los loci "débil" y la disminución de la concentración de los loci "fuerte" (Navarro et al., 2008). El desarrollo de las reacciones multiplex presenta la principal ventaja de optimizar la relación coste-eficacia (Wesmajervi et al., 2006; Navarro et al., 2008), sin perder la consistencia de genotipado y la resolución de las bandas (Neff et al., 2000). Renshaw et al. (2006) reportaron un coste inferior del material en reacciones múltiples frente a reacciones individuales. Estos autores demostraron que la determinación del genotipo de una muestra en reacciones tetraplex y Octaplex fueron 3,5 y 6,7 veces menor que mediante reacciones individuales, respectivamente. También informaron de los valores de ahorro similares en el tiempo del personal. Neff et al. (2000) requiere una cuarta parte de los consumibles de PCR para el genotipado de cuatro marcadores de microsatélites con una reacción tetraplex que con las reacciones indviduales. Navarro et al. (2008) compararon el personal, consumibles y los gastos de funcionamiento para el genotipado con diez marcadores de microsatélites en una sola PCR múltiple frente a las PCR individuales y por separado, y los resultados revelaron que en la reacción múltiple los costes eran la sexta parte que en las reacciones individuales, incluso cuando estos se realizaron en una carrera única. Por otra parte, las reacciones multiple también minimizan errores de genotipado ya que reducen pasos durante el proceso de análisis de la muestra e introducen automatización. Bonin et al. (2004) analizaron el doble de dieciocho marcadores microsatélites con solo una reacción de PCR en Resumen español 125 34 muestras de oso pardo, y encontró sólo un 0,8% de los errores de genotipado debidos a factores humanos. Por otra parte, la PCR multiple reduce el riesgo de contaminación y evita la mezcla de los genotipos de diferentes muestras (Navarro et al., 2008). Sin embargo, en espáridos, se han propuesto pocas reacciones múltiples con una gran cantidad de microsatélites. En la breca, Ramšak et al. (2003) desarrollaron una PCR múltiple con tres marcadores microsatélites por medio de la técnica de touchdown (amplificación gradual a diferentes temperaturas). Para la dorada, Launey et al. (2003) propusieron dos reacciones múltiples de tres y cinco marcadores. Brown et al. (2005a) desarrolló y comprobó la amplificación de una PCR multiple de cuatro marcadores microsatélites, y Porta et al. (2010) propusieron dos reacciones múltiplex de cuatro y seis marcadores como una herramienta de genotipado. Todos estas multiples de PCR fueron diseñadas mediante la combinación de marcadores microsatélites con idénticas temperaturas de hibridación. Navarro et al. (2008), también en la dorada, desarrollaron dos reacciones múltiples de diez y siete marcadores microsatélites interespecíficos rediseñados, y mostró una amplificación cruzada exitosa en bocinegro, sama roquera y dorada con seis, ocho y cinco marcadores, respectivamente. Lee-Montero et al. (2013) desarrollaron el primer panel normalizado de dos nuevas PCR múltiples with11 marcadores en cada una, denominadas SMsa1 y SMsa2 (Super Multiplex Sparus aurata). Negrín-Báez et al. (2014) establecieron un conjunto de 13 PCR múltiples de marcadores microsatélites específicos (106) como una herramienta para la detección de QTL en la dorada (Sparus aurata L.). 1.2.5 Matriz de relación y contribución familiar Muchas especies de peces, incluyendo la dorada, permiten la realización de cruces individuales viables (Knibb et al., 1998; Montero et al., 2001). Sin embargo, su aplicación en el contexto de la industria presenta problemas, ya que aumenta los costos de producción en términos de recursos humanos e infraestructura. En condiciones industriales, para garantizar la calidad y cantidad de la producción, esta especie se propaga a través de puestas masales, de aproximadamente 40 - 60 reproductores. Desde un punto de vista genético, esta estrategia muestra la ventaja de que las fuentes de parecido provenientes del ambiente común se reducen, aumentando así la precisión de la estima de parámetros genéticos (Herbinger et al., 1999). Por otro lado, las puestas masales evitan el conocimiento de la genealogía de los peces en condiciones de cultivo, lo que es absolutamente necesario en la Resumen español 126 estimación de parámetros genéticos y el desarrollo de programas de selección. Debido a eso, la introducción de programas de mejora genética con sistemas de puestas masales a escala industrial requiere del etiquetado físico, para identificar todos los individuos en relación con el análisis de ADN y reconstruir la genealogía (Navarro et al., 2008), ya que la matriz de relaciones entre los individuos es un requisito fundamental para estimar parámetros genéticos de los caracteres de interés económico para la industria. El uso de marcadores genéticos, es una herramienta de reconstrucción de pedigrí o matriz de parentesco dentro de un grupo de los descendientes, por lo general provenientes de desoves masales de sólo uno o varios días. Sin embargo, el desove masal tiende a reducir el número de reproductores que contribuye a la generación siguiente, y por tanto también el censo efectivo (Brown et al., 2005; Fessehaye et al., 2006; Pinera, 2009; Porta et al., 2010). La matriz de parentesco entre los individuos es clave para estimar parámetros genéticos de los caracteres de interés económico para la industria (Navarro et al., 2009a, b). Por otro lado, la contribución es un acontecimiento trascendental en la genética, pero particularmente importante en la dorada, cuando se quiere implementar la mejora genética en su sistema de producción a nivel de criadero. Sin embargo, la alta tasa de fecundidad de dorada permite producir lotes de peces a partir de sólo un par de reproductores bajo puesta masal, lo cual es una desventaja importante desde el punto de vista de la mejora, ya que ello podría llevar a que los reproductores de alto valor genético no contribuyan en tales circunstancias, lo que evitaría maximizar la respuesta a la selección. Así, es de una gran importancia práctica para las empresas el maximizar la contribución de los reproductores, bajo desove masal a escala industrial. 1.2.6 Interacción genotipo x ambiente El estudio de la interacción genotipo-ambiente (G x E) es un problema importante en la etapa final de la cría, lo que representa un aspecto de gran importancia en los programas de mejora genética. Las interacciones pueden implicar cambios en el orden de importancia para los genotipos entre los ambientes y los cambios en la magnitud absoluta y relativa de las varianzas genéticas, ambientales y fenotípicas entre los ambientes (Bowman, 1972). Por lo tanto, la importancia de este factor puede aparecer sobre todo en las especies de peces que su sistema de producción tiene diferencias ambientales significativas, como sucede en la dorada. Resumen español 127 La dorada es producida por diferentes sistemas de producción, con grandes diferencias ambientales entre ellos. En los criaderos, los alevines se producen utilizando sistemas intensivos o semi-intensivos, y en la etapa de engorde los peces pueden cultivarse en diversas formas: en estanques y lagunas costeras, con métodos extensivos y semiintensivos, o en instalaciones terrestres y en jaulas marinas, con sistemas de cría intensiva. Estos sistemas e instalaciones de producción tienen efecto sobre los valores fenotípicos, especialmente en el crecimiento y caracteres de calidad del pescado (Navarro et al., 2009a, b). Por lo tanto, las diferencias ambientales pueden aumentar o disminuir los valores fenotípicos, interactuar y producir la interacción GxE (Falconer y Mackay, 1996). Por lo tanto, para desarrollar un programa de mejora genética eficiente, es esencial la determinación de las interacciones genotipo-ambiente con el fin de conocer la capacidad de cualquier expansión industrial con la minimización de su impacto sobre la ganancia genética. GxE interacciones pueden ser estimadas por la reclasificación de los individuos de las mismas familias en dos ambientes diferentes, estableciendo las correlaciones genéticas del mismo carácter en los dos ambientes, bajo la consideración de que fuesen dos caracteres distintos (Falconer y Mackay, 1996). Navarro et al. (2009a, b) estimaron la interacción GxA en la dorada para caracteres de crecimiento a través de las correlaciones genéticas de los mismos en dos ambientes (jaulas y tanques), y encontraron reclasificaciones débiles entre las jaulas marinas y los tanques (con correlaciones genéticas entre 0,70 y 0,99), que fueron consideradas como poco importantes. Lee-Montero et al. (2015) estimaron valores altos de correlaciones genéticas en la dorada para el crecimiento y deformidades esqueléticas, a la edad de sacrificio en diferentes condiciones ambientales, lo que revela que las GxE son escasas. Lo que está en consonancia con los resultados de Sae-Lim et al. (2013), quienes demostraron que las correlaciones genéticas con los valores más altos de 0,7 probablemente no constituyen interacciones GxE importantes. 1.3 Estado de las aplicaciones de la selección genética en acuicultura Durante los últimos veinte años, la acuicultura juega un papel importante en la reducción de la brecha alimentaria para cubrir las necesidades humanas en términos de proteínas de origen animal, sobre todo por los beneficios positivos que tiene para la salud humana, debido principalmente al efecto protector de los lípidos de peces frente a enfermedades cardiovasculares. En 2013, por primera vez, la producción acuícola en todo el mundo superó el sector de la pesca con 97,2 millones de toneladas, frente a los 93,8 Resumen español 128 millones de toneladas de pesca. Así, más del 50% de los productos acuáticos procede de la acuicultura (APROMAR, 2015). En general, los programas de cría modernos para las especies de acuicultura se han establecido más tarde que en las plantas y los animales terrestres. Por lo tanto, muchos productores de peces, moluscos y crustáceos todavía utilizan poblaciones silvestres o poblaciones de producción a sólo unas pocas generaciones de distancia de su origen silvestre (Tave, 1986). Aunque, en general, la respuesta a la selección suele ser mayor en los peces, moluscos y crustáceos que en los animales domésticos terrestres, y su potencial de ganancia genética está bien documentado, el desarrollo de programas de selección genética ha progresado lentamente (Neira, 2010; Rye et al., 2010; Gjedrem et al., 2012). Actualmente, existe una creciente necesidad de las empresas de acuicultura por mejorar su competitividad mediante la optimización de su producción y/o aumentar la calidad de sus productos. En este sentido, la genética es una herramienta de gran potencial, de la que se ha estimado que su uso, en el contexto de la mejora genética para lotes de reproductores en caracteres de interés productivo, puede contribuir a una reducción del 50% de los costes de producción (Gjedrem et al., 2012). Aunque la tasa de progreso genético puede parecer pequeña, con respecto a las atribuibles a los avances en los factores de manejo, debe tenerse en cuenta que éste es continuo, acumulativo y permanente (López-Fanjul y Toro, 2007). Como se muestra en la tabla 1-1, la carpa común, la trucha arco iris, el salmón del Atlántico, tilapia, bagre de canal y peces ornamentales, son las especies que recibieron el esfuerzo genético más importante. Sin embargo, hay otras especies de alto valor, tales como, la lubina, la dorada y el rodaballo, que recientemente han incorporado programas de cría selectiva aplicada (Chavanné et al., 2016). 1.4 Estado de la mejora genética de la dorada Al igual que en otras especies de la acuicultura, la mejora genética en dorada está muy poco extendida, y el crecimiento de esta industria ha invertido en la optimización de sus producciones a través de otros factores, principalmente de gestión, tales como la nutrición, la reproducción o la prevención de enfermedades. Las razones se deben en parte a la falta de personal especializado, el alto costo que supone para las empresas organizar su producción con criterios genéticos, las características biológicas de la dorada, y a la falta de una metodología que combine los intereses de la producción con la explotación de la Resumen español 129 variación genética. Aunque hay algunos programas a nivel de empresa de cría de esta especie (Rye et al., 2010; Chavanné et al., 2016; Janssen et al., 2016), sólo se han publicado algunos parámetros genéticos, y menos aún se han estimado bajo condiciones industriales. Knibb et al. (1997) estimaron una heredabilidad de 0,34 ± 0,02 para el peso al sacrificio utilizando los coeficientes de regresión de las líneas de selección. Thorland et al. (2006), estimaron las heredabilidades para el peso al sacrificio (250-400 g) (0.61±0.06), el color de la piel evaluados por una escala discreta (0.20±0.02), y la presencia o ausencia de deformidad de la cabeza (0.05±0.02) y deformidades de la columna (0.12±0.02). Antonello et al. (2009) estimaron los parámetros genéticos para los caracteres de crecimiento (0,38±0,07 para la longitud) y la resistencia a la pasteurelosis en dorada (0.12±0.04 to 0.45 ±0.04 para la resistencia a enfermedades). Navarro et al. (2009a, b), fueron los primeros autores en estudiar y estimar las heredabilidades para un gran número de caracteres (crecimiento, rendimiento y de la composición corporal), bajo condiciones industriales, así como sus características fenotípicas y relaciones genéticas. Estos autores describieron valores que oscilaron entre 0,28 – 0,34 para el peso, 0,27 – 0,35 para la longitud, 0,05-0,13 para el factor de condición, 0,02-0,5 para caracteres de composición corporal y 0,12-0,41 para caracteres de rendimiento. Lee-Montero et al. (2015) estimaron, por primera vez en esta especie, los parámetros genéticos de las deformidades esqueléticas (desde 0,07 hasta 0,26) y las características de crecimiento a diferentes edades comercialmente importantes (desde 0,20 to 0.,41 para la longitud, y 0,17 a 0,43 para el factor de condición), teniendo en cuenta las interacciones GxE, que fueron escasas. Estas estimaciones de heredabilidades son muy alentadoras y son de la misma magnitud que las descritas en otras especies cultivadas intensivamente como el salmón del Atlántico (Standal y Gjerde, 1987), y la trucha arco iris (Mckay et al., 1986). Esto indica que hay un gran potencial de mejora en algunas características importantes, como el peso, a través de la cría selectiva de dorada. En esta especie, los programas de cría que están en funcionamiento llevan un número de generaciones de selección comprendidas entre 1 y 5. Las estimaciones de ganancia genética para el peso al sacrificio son del 29% por Brown (2003), del 19% por Navarro et al. (2009) y Lee-Montero et al. (2015), del 13% por Thorland et al. (2015) después de 2,6 generaciones de selección, y de entre 5-10% en la tasa de crecimiento por Knibb (2000). Actualmente, alrededor de 31- Resumen español 130 44% la producción de dorada es llevada a cabo con semillas seleccionadas a nivel comercial (Chavanne et al., 2016). 1.5 Caracteres de importancia económica en dorada en los programas de mejora genética Dentro de los posibles caracteres a mejorar en dorada a nivel industrial, son especialmente interesantes todos aquellos cuya optimización conduce a una producción económica más rentable. Es muy importante que antes de comenzar cualquier programa de cría se deban definir las metas u objetivos de la mejora (Gjedrem, 2000). Los costos de la alimentación en los proyectos de producción de animal, incluidos los peces, representa más del 65% del coste total, por lo que el carácter de conversión alimenticia (FCR) es uno de los rasgos económicos más importantes en este tipo de proyectos (Kause et al., 2006). Sin embargo, debido a la dificultad de medir este carácter en peces en condiciones industriales, su mejora generalmente se realiza indirectamente a través de las características de crecimiento (peso, longitud y SGR) (Gjedrem y Thodesen, 2005; Navarro et al., 2009a, b), ya que tiene una correlación genética alta y positiva con estos caracteres (McPhee et al., 1979; Afonso, 1996). Los caracteres de rendimiento y canal son también caracteres de mucha importancia económica, debido a que las canales y la carne de pescado se utilizan para el consumo humano (Souza, 2008). La fabricación de productos elaborados, como el fileteado, representa un sector muy interesante para la industria y el mercado de consumo. Si bien, este tipo de comercialización está sin explotar en la dorada, estudios de mercado muestran que de implementarse habría una demanda interesante, rentable y de alta (Luna, 2006). Por lo tanto, es importante utilizar características de la canal como criterios de selección relacionados principalmente con el rendimiento filete (Silva et al., 2009). En este sentido, las mediciones morfométricas se pueden utilizar para evaluar la calidad de la canal como criterio de selección en los programas de mejora genética y para obtener respuesta correlacionada en el rendimiento (Rutten et al., 2004). Hoy en día, los consumidores están interesados en los aspectos relacionados con la calidad de la carne, como sabor, jugosidad, textura y apariencia que se asocia significativamente con la composición corporal (Navarro et al., 2009b). Los caracteres de Resumen español 131 calidad de la carne y de su composición, tales como el contenido de lípidos, son también caracteres muy importantes, que tienen potencial como caracteres indicadores de mejora selectiva de eficiencia de la alimentación (Pym, 1990; Archer et al., 1999). El contenido de grasa en el músculo es un componente muy importante, con una correlación positiva con el sabor y la jugosidad (Grigorakis, 2007). La apariencia es uno de los caracteres económicos más importantes que afecta a la comercialización del producto (Blonk et al., 2010; Costa et al., 2011). La importancia de este carácter es cada vez mayor en productos animales comercializados enteros, como la dorada. Además, las características morfológicas de los peces determinan dramáticamente otros caracteres tales como la compacidad y características de la canal. Por lo tanto, el peso de filete está estrechamente relacionado con la longitud y la altura de filete. Para aumentar la eficacia de cualquier programa de selección, los caracteres económicos importantes se deben medir de manera objetiva y precisa, y a un bajo coste (Gjedrem, 1997). En este sentido, Navarro et al. (2016) desarrollaron un programa para el análisis rápido y automatizado de imágenes (IMAFISH_ML), que mide 27 caracteres morfométricos (caracteres no invasivos) en tres especies de peces de interés comercial: dorada (Sparus aurata L.), corvina (Argyrosomus regius) y bocinegro (Pagrus pagrus). Estos caracteres están relacionados con el crecimiento, la canal y la composición de la carne, que son los caracteres más importantes en la mayoría de los programas de mejora genética en peces. Hay algunas publicaciones potenciales que estimaron los parámetros genéticos para algunas de las características morfológicas de los peces. En trucha arco iris, Gjerde y Schaeffer, (1989) y Kause et al. (2003), demostraron que caracteres relacionados con la morfología corporal, el color de la piel y su distribución pueden ser modificados a través de la selección genética. Así, estimaron unas heredabilidades de entre 0,46-0,61 y 0,29-0,5 para las escalas subyacente y fenotípica, respectivamente. En Tilapia del Nilo, Rutten et al. (2005) encontraron que hubo una fuerte relación entre las medidas corporales y el peso del filete. Kocour et al.(2007), en carpa común, estimaron parámetros genéticos para caracteres de crecimiento y de procesamiento, encontrando que la longitud estándar, el peso corporal, el porcentaje de grasa y la longitud relativa de la cabeza mostraban heredabilidades altas (>0,5), mientras que la altura relativa del cuerpo, el ancho relativo del cuerpo, el rendimiento canal y filete tenían heredabilidades medias (0,2-0,5). Costa et al. (2010), en lubina, estimaron efectos genéticos y ambientales para la forma, encontrando altas heredabilidades (0,4-0,55), habiendo con una buena correlación entre las diferencias en la forma con las distancias