| Mass selection is a major method of genetic improvement for quantitative trait in random-mating populations. Selection in traditional breeding is based on phenotypic value, called phenotypic value selection (PS). Since the advent and development of molecular markers, great attention has been paid to marker-assisted selection (MAS). Marker score selection (MS) and index selection (IS) are two main MAS methods for quantitative traits. The former uses marker score as selection criterion estimated based on the markers showing significant associations with the target quantitative trait screened by stepwise multiple regression. The latter combines marker score and phenotypic value as an index for selection. Obviously, the selection criterions used by the two methods are an approximation of individual genotypic value (breeding value), similar to the traditional phenotypic value.With the development of molecular marker technology, many quantitative trait loci (QTLs) for the important economic traits have been mapped in plants and animals. However, neither of the proposed MAS methods can make use of the known information (effects and locations for QTLs). To utilize the known information for improving the efficiency of MAS, three new MAS methods called genotype-phenotypic value selection (GPS), genotype-marker score selection (GMS) and genotype-marker score-phenotypic value selection (GMPS) are proposed in this study. In GPS, in each generation, mapped (or known) QTLs are firstly selected according to their flanking markers using a large population of seedlings. Then, the selected individuals are further selected according to their adjusted phenotypic values, in which effects of the known QTLs have been removed. After homozygosity of favorable alleles has been reached at the known QTLs, only PS is conducted in later generations. GMS is similar to GPS, except that the second stage selection in each generation is based on the criterion of marker scores estimated with adjusted phenotypic values. After all favorable alleles have been fixed at the known QTLs, only MS is conducted in latergenerations. GMPS is similar to GMS. The only difference is that after the favorable alleles at the known QTLs have been completely fixed, only PS is conducted in later generations.Monte Carlo simulation was conducted by computer to compare the efficiency and reliability of the new MAS methods (GPS, GMS and GMPS) with the traditional PS and the previously proposed MAS methods MS and IS. A software for the MAS simulation was developed with visual C++6.0. The accumulated genetic advance (AGA) was used to evaluate the efficiency of each method, and the variance of AGA (VAGA) estimated by 100 replicates of simulation was used to evaluate the stability of each method. Main results are summarized as follows:(1) When the initial heritability is high, MAS shows no obvious advantages over PS, indicating that MAS is not necessary in this case. But when the initial heritability is low, the superiority of MAS is significant.(2) The progress is much faster if the linkage among QTLs in the initial population is in total coupling phase than in total repulsion phase. And the superiority of MAS over PS is higher if the initial population is in total coupling phase than in total repulsion phase.(3) Population size is the most important factor affecting the efficiency of MS. The larger population size, the higher the relative efficiency of MS to PS. So, all the methods taking marker score as a selection criterion (MS, IS, GMS and GMPS) should be carried out in large populations.(4) When there are few known QTLs, three new methods containing genotype selection (GPS, GMS and GMPS) exhibit little superiority over the old MAS methods (MS and IS). But as the number of known QTLs increases, superiority of three new methods increases rapidly.(5) With the reduction of population size (seedling stage) for genotype selection, the AGAs of three new methods slow down.Overall, when there are many known QTLs, it is suitable to employ thre...
|
PDF Full Text Request