Studies On Genetic Parameters Of Growth And Flesh Traits In Pacific Oyster (Crassostrea Gigas)

The Pacific oyster(Crassostrea gigas) is a major global aquaculture species, ranking first in the production among all other aquaculture fish and shellfish species. Oyster was supplying good protein source for human beings, the total production of oyster in our country was reaching 421.9 million tones, ranking first in the world. Genetic breeding on important economic traits are and will be carried out. Genetic parameters are basic information for genetic breeding program decision making, selective response prediction and individual genetic evaluation, in the process of animal genetic improvement.In this research, genetic parameters of growth trait and flesh trait were evaluated, as well as genotype by environment interaction (G by E) was tested, with the methods of crossbreeding and family selection method. During the process of analyzing flesh trait of pacific oyster, we used Fourier transform near infrared (FT-NIR) spectrum technology to establish two analyzing models with characters of fast and high throughout to test numerous nutritional ingredients of flesh trait on pacific oyster.From this research we hope to find out the major factor of influencing growth and flesh trait is genetic factor or environment factor. To verify whether or not using indirect method to genetically improve the flesh trait from growth trait aspect which could undamage oysters. The results in this research was provided quantitative genetics information for growth and flesh trait using crossbreeding and family selection method, especially providing data for flesh trait technical route decision’s making.The main results of this research were as follows:1 Growth trait and flesh trait analysis with complete diallel cross breeding from three selective strains of C. gigas1) Growth traitThe results show that growth performance is significant difference in different mating groups on 180 d,330 d and 450 d (P< 0.05, n= 950). Growth performance, general combining ability (GCA), special combining ability (SCA) and heterosis potence are not identical in different growth stages. The best growth performance groups are C (?)K♀, J (?)C♀ and K (?)J♀ on 180 d,330 d and 450 d, respectively. But K (?)K♀ shows growth disadvantage in each stage. The GCA is non-negative in both sire and dam of the C strain, while GCA of the K strain is non-positive. SCA is different in each stage, but is consistent with heterosis potence approximately.2) Flesh traitThe results show that growth performance of pure strain of C (?)C♀ is excellent to other group on GW and DP, meanwhile K (?)K♀ is inferior to others (except DP). General combining ability (GCA) value of MC, GC, TPC, TFC, ZC, SC and AC are approximately to zero in dam, while the situation is more complex in sire. Special combining ability value in most indexes of flesh trait is negative. Most of the (27/54) hybrid offsprings showed no heterosis (-1.0< hp< 1.0), and even some hybrid groups (13/54) showed hybrid weakness (hp<-1.0).2 Genetic parameters analysis of growth trait on C. gigas1) Comparison of average phenotype valueFamilies reared at RS area grow well at both of 188d and 338d, while inversely at 474d. C. gigas at KTD area is higher than RS area in SH, SL and WW, but SW is similar with each other in both area. SH, SL, SW, WW differs greatly among families at 188d,338d,474d and 560d, respectively. And some of growth trait index show significant difference at some period.2) HeritabilityThe results show that heritability differs between locations and varies among times. All of the index of growth trait show moderate to high heritability in RS area, while KTD area shows high heritability. Parental model was used to calculate total heritability (time and location was treated as fixed factors), the heritability of SH, SL, SW and WW is 0.25±0.08,0.29±0.09,0.14±0.05 and 0.26±0.09, respectively.3) Genetic and phenotypic correlationBoth genetic and phenotypic correlations are different in RS and KTD among 118d,338d,474d and 560d, but both of them are positively correlated. Generally speaking, the value of genetic correlation is higher than phenotypic correlation.4) Early to later correlationIn KTD area, rT of SH and WW are showed high early to later correlation, but SL and SW varies greatly between different period. In RS area, the early to later correlation of growth trait is much lower, the results indicate early selection on growth trait in RS area may not obtain good ideal genetic process.5) G by E interactionFor SL, the genetic correlation between RS and KTD is 0.51-0.73, indicate that the G by E between the two locations reaches slight to moderate level. For WW, rg is around 0.50 (0.48-0.73), almost reaches significant level. The rg value of other index of growth trait is very low.6) Evaluation of breeding valueIndividual model was used to evaluate individual breeding values of growth trait of C. gigas. Efficiency analysis of individual selection of SH and WW at different growth stages. Based on phenotypic value and the individual breeding values (IBV), we found that selection efficiency which was using IBV is significantly higher than the selection according to phenotypic value. Pedigree breeding value (PBV) was obtained by parental model, the result is similar to IBV, i.e., selection efficiency which was using PBV is significantly higher than the selection according to phenotypic value on SH and WW.3 Establishment of near infrared (NIR) model of numerous biochemical compositions on C. gigasThe advantages of the NIR model are their efficiency, low cost and friendly to environment. In this study, we collected representative samples of C. gigas and used Fourier transform near infrared (FT-NIR) spectrum technology to establish two models:fresh tissue (homogenate) and frozen-dried (powder). Chemical real value was tested and NIR spectrum was collected.1) Models on fresh tissue samplesThe results indicated that NIR models of MC, GC and TPC have good predictive veracity. Correlation coefficient in calibration (RC) was 0.96-0.99, correlation coefficient in cross validation (RCV) was 0.93-0.99, and correlation coefficient in external prediction (RCV) was 0.97-0.99. But the parameters indicated that NIR models of TFC, ZC, SC, TC and AC were not acceptable.2) Models on frozen-dried tissue samplesIn this part, powder form’s NIR models were established, the models were better than fresh sample models, the main reason are:most of the chemical contents (GC, TPC, TFC, ZC, SC, TC and AC) had accurately predictive ability. Correlation coefficient in calibration (RC) was 0.92-0.99, correlation coefficient in cross validation (RCV) was 0.78-0.99, and correlation coefficient in external prediction (RCV) was 0.91-0.99. NIR model of TC was unsatisfied, the parameters indicated that it was not accurate enough (RC=0.83, RCV=0.75, REV=0.80).4 Genetic parameters analysis of flesh trait on C. gigas1) Phenotype valueThe average value of DP, GC, TFC and SC of all families reared at RS area grew better than KTD area, while GW, MC, TPC, ZC and AC were inversed. Significant comparison show that, GC is not reaching significant level, TPC reaches significant level, and other index of flesh trait are reaching extremely significant level.2) HeritabilityIndividual model was used to calculate heritability of RS and KTD area. The results show that heritability differs between locations. The results are GW (0.17±0.08, 0.63±0.14), DP (0.53±0.13,0.35±0.12), MC (0.45±0.13,0.40±0.13), GC (0.58±0.13, 0.61±0.14), TPC (0.65±0.14,0.46±0.14), TFC (0.94±0.11,0.32＋0.12), ZC (0.46±0.14, 0.21±0.10), SC (0.80±0.13,0.91±0.12) and AC (0.71±0.13,0.56±0.14) [Note, former is RS’s heritability and latter is the heritability of KTD]. Parental model was used to calculate total heritability (location was treated as fixed factor), the heritability is showed as follows:GW (0.43±0.15), DP (0.27±0.10), MC (0.20±0.08), GC (0.30±0.10), TPC (0.26±0.10), TFC (0.47±0.16), ZC (0.23±0.09), SC (0.47±0.14) and AC(0.30±0.11).3) Genetic and phenotypic correlation(1) Genetic correlation between growth and flesh trait:the relationship between growth trait and flesh trait is very complex, some of them are positive correlation, and some of negative correlation. Also, genetic correlation varies from different areas. ① RS:using SH as direct selection index, we could positively select SC and AC, and negatively select GC and DP. Using SL as direct selection index, we could negatively select DP. Using SW as direct selection index, we could positively select SC. And using WW as direct selection index, we could negatively select GW. ② KTD:using SH as direct selection index, we could negatively select TFC. Using SL as direct selection index, we could positively select GW, negatively select TPC. Using SW as direct selection index, we could positively select GC, negatively select ZC, SC and AC. Using the genetic correlation between the growth trait and flesh trait, we could realize the breeding goals for flesh trait form the growth trait according to different area. (2) Correlations between index of flesh trait. The results show that selection for some of the index may be influence other flesh trait index, if they are negatively genetic correlated, we should rearrange the breeding schedule according to their relationship. (3) phenotypic correlation:Absolute value of phenotypic correlation between growth trait and flesh trait generally less than 0.5. Tendency between genetic and phenotypic correlation is almost consistent, but phenotypic correlation is always less than genetic correlation.4) G by E interactionAll of the index of flesh trait are less than 0.5, G by E interaction is reaching serious level. Among them, the genetic correlation of TPC, SC and AC between locations is approximately 0. However, in this experiment the standard error is high.5) Evaluation of breeding valueBecause flesh trait could not be measured lively, family selection method should be choosen on flesh trait. Efficiency comparative analysis was taken between phenotypic value and individual breeding values (IBV), we found that selection using IBV is significantly higher than the selection according to phenotypic value. Families were taken out according to breeding value of each index. There were few common families for the same index between RS and KTD culture area.