Estimation of genetic parameters and breeding values in shrimp F enneropenaeus chinensis using the REML/BLUP procedure

An analysis of a selection experiment was used to assess the impact of various animal model struc- tures on REML estimates of variance components. The analyses were carried out based on 162 d body mass （BM） of 1 287 animals from 21 paternal half-sib groups of Fenneropenaeus chinensis. Estimated breeding values （EBV） of BM of all individuals were estimated using eight statistical models （A, AB, ABC, ABDC, ABMFC, ABMDC, ABFDC and ABMFDC） and BLUP （best linear unbiased prediction）. These models were designed involving factors such as sex, spawn date as fixed effects, maternal genetic effects, full-sib family effects as random effects, mean BM of families at tagging and age at recording （covariate）. The results demonstrate the importance of correct interpretation of effects in the data set, particularly those that can influence resemblance between relatives. The data structure and the particular model that was applied markedly influenced the magnitude of variance component estimates. Models based on few effects obtained upward biased estimates of additive genetic variance. The accuracy of genetic parameters and breeding value es- timated by ABFDC model was higher than other models. The results imply that additive genetic direct value, full-sib family effects, and covariance effects besides sex and spawn date as fixed effects were very important for estimating genetic parameters and breeding value of body mass. This model had a heritability estimate of 162 d BM of 0.44. The comparison of the efficiency of selection based on breeding values or phenotypic value revealed great difference： average breeding value of the best 24 families selected by the 162 d BM breeding value and phenotype were 0.577 g and 0.366 g, respectively, representing a 36.57% higher efficiency in the former. In conclusion, selection based on breeding value was more effective than selection based on phenotypic value. Our results indicate that effects influencing the magnitude of estimates should be taken into account when estimating heritability and breeding values for BM.

ASP-Based Programs of Best Linear Unbiased Prediction-Estimated Breeding Values in Breeding Stock

In order to improve the breeding effect of livestock, the data were read from an Excel file with Active Server Page （ASP） programs, and the breeding values of breeding stock were calculated by best linear unbiased prediction （BLUP） method.

Estimation of genetic parameters for growth trait of turbot using Bayesian and REML approaches

Bayesian and restricted maximum likelihood （REML） approaches were used to estimate the genetic parameters in a cultured turbot Scophthalmus maximus stock. The data set consisted of harvest body weight from 2 462 progenies （17 months old） from 28 families that were produced through artificial insemination using 39 parent fish. An animal model was applied to partition each weight value into a fixed effect, an additive genetic effect, and a residual effect. The average body weight of each family, which was measured at 110 days post-hatching, was considered as a covariate. For Bayesian analysis, heritability and breeding values were estimated using both the posterior mean and mode from the joint posterior conditional distribution. The results revealed that for additive genetic variance, the posterior mean estimate （σa^2 =9 320） was highest but with the smallest residual variance, REML estimates （σa^28 088） came second and the posterior mode estimate （σa^2=7 849） was lowest. The corresponding three heritability estimates followed the same trend as additive genetic variance and they were all high. The Pearson correlations between each pair of the three estimates of breeding values were all high, particularly that between the posterior mean and REML estimates （0.996 9）. These results reveal that the differences between Bayesian and REML methods in terms of estimation of heritability and breeding values were small. This study provides another feasible method of genetic parameter estimation in selective breeding programs of turbot.

Importance of “connectedness” between herds for effective across herd genetic evaluation

“Connectedness” is an essential component of genetic evaluations. The degree of connectedness affects the accuracy of comparing estimated breeding values (EBVs) from one herd or contemporary group to the other. It can be measured through Connectedness Rating (CR) which is based on variances and covariance among the estimates of contemporary group effects. A computing algorithm and a computer program for estimating CR is available. The minimum required level of connectedness depends upon the size of the contemporary groups, the level of accuracy and the residual variance. About 48% CR is required to detect differences between EBVs that are greater than 20% of the standard deviation in the trait, for group sizes of about 100 animals. Higher levels are necessary for smaller group sizes and for more accurate comparisons. Breeders participating in a common genetic evaluation program should therefore exchange their superior genetics and possibly use some common testing facilities for meaningful estimates of breeding values. Maintaining a good connectedness level will make the genetic evaluation program more useful for selection of superior breeding animals and achieving faster rate of genetic progress.

A Method for Accuracy of Genetic Evaluation by Utilization of Canadian Genetic Evaluation Information to Improve Heilongjiang Holstein Herds

The objectives of this study were to set up a new genetic evaluation procedure to predict the breeding values of Holstein herds in Heilongjiang Province of China for milk and fat production by utilizing Canadian pedigree and genetic evaluation information and to compare the breeding values of the sires from different countries. The data used for evaluating young sires for the Chinese Holstein population consisted of records selected from 21 herds in Heilongjiang Province. The first lactation records of 2 496 daughters collected in 1989 and 2000 were analyzed. A single-trait animal model including a fixed herd-year effect, random animal and residual effects was used by utilizing Canadian pedigree and genetic evaluation information of 5 126 sires released from the Canadian Dairy Network in August 2000. The BLUP procedure was used to evaluate all cattle in this study and the Estimated Breeding Values (EBV)for milk and fat production of 6 697 cattle (including 673 sires and 6 024 cows) were predicted. The genetic levels of the top 100 sires originated from different countries were compared. Unlike the BLUP procedure that is being used in conjunction with the single-trait sire model in Heilongjiang Province of China now, the genetic evaluation procedure used in this study not only can be used simultaneously to evaluate sires and cows but also increase the accuracy of evaluation due to using the relationships and genetic values of the Canadian evaluated sires with more daughters. The results showed that the new procedure was useful for genetic evaluation of dairy herds and the comparison of the breeding values of these sires imported from different countries showed that a significant genetic improvement has been achieved for milk production of the Heilongjiang Holstein dairy population by importing sires from foreign countries, especially from the United States due to the higher breeding values.

Influence of Allele Frequency on Predicting Animal Phenotype Using Back-Propagation Artificial Neural Networks

To overcome the obstacle of the fascinating relation in predicting animal phenotype value, we have developed a neural network model to detect the complex non-linear relationships between the genotypes and phenotypes and the possible interactions that cannot be expressed with equations. In this paper, back-propagation neural network is used to discuss the influences of different allele frequencies on estimating the polygenic phenotype value. To ensure the precision of prediction, normalization was needed to train the prediction model. The results show that back-propagation artificial neural networks can be used to predict the phenotype value and perform very well in allele frequency from 0.2 to 0.8, when the allele frequency is very small （less than 0.2） or big （more than 0.8）; however, the prediction model was not reliable and the predicted value should be carefully tested.

Statistical considerations for genomic selection

Genomic selection is becoming increasingly important in animal and plant breeding, and is attracting greater attention for human disease risk prediction. This review covers the most commonly used statistical methods and some extensions of them, i.e., ridge regression and genomic best linear unbiased prediction, Bayesian alphabet, and least absolute shrinkage and selection operator.Then it discusses the measurement of the performance of genomic selection and factors affecting the prediction of performance. Among the measurements of prediction performance, the most important and commonly used measurement is prediction accuracy. In simulation studies where true breeding values are available, accuracy of genomic estimated breeding value can be calculated directly. In real or industrial data studies, either trainingtesting approach or k-fold cross-validation is commonly employed to validate methods. Factors influencing the accuracy of genomic selection include linkage disequilibrium between markers and quantitative trait loci, genetic architecture of the trait, and size and composition of the training population. Genomic selection has been implemented in the breeding programs of dairy cattle, beef cattle, pigs and poultry. Genomic selection in other species has also been intensively researched, and is likely to be implemented in the near future.