Premium
Efficiency of quantitative trait loci‐assisted selection in correlations between identified and residual genotypes
Author(s) -
LIN Ching Y.
Publication year - 2008
Publication title -
animal science journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.606
H-Index - 38
eISSN - 1740-0929
pISSN - 1344-3941
DOI - 10.1111/j.1740-0929.2007.00493.x
Subject(s) - quantitative trait locus , heritability , selection (genetic algorithm) , biology , genotype , genetics , trait , genetic correlation , residual , genetic variation , mathematics , gene , computer science , algorithm , artificial intelligence , programming language
This study quantified the efficiency of quantitative traits loci (QTL)‐assisted selection in the presence of correlations ( ρ qr ) between identified ( q ) and residual ( r ) genotypes. Two levels of heritability (h 2 = 0.1 or 0.3), two levels of correlation ( ρ qr = −0.3 or 0.3) and five proportions of genetic variance explained by QTL detected (= 0.1, 0.2, 0.4, 0.6 or 0.8) were combined to give 20 scenarios in all. QTL‐assisted selection placed a larger index weight on the QTL genotype than on the phenotype in 17 of 20 scenarios, yielding a greater response in the QTL genotype than in residual genotype. Although QTL‐assisted selection was superior to phenotypic selection in all 20 scenarios, QTL‐assisted selection showed a greater advantage over phenotypic selection when ρ qr was positive than when ρ qr was negative. Doubling the proportion of detected QTL variance to genetic variance does not result in a twofold increase in the genetic response to QTL‐assisted selection, suggesting that economic returns diminish for each additional cost of detecting extra QTL. The correlation between q and r would make the interpretation (or prediction) of QTL effects difficult and QTL‐assisted selection strategy must consider the joint effect of q and r . When q and r are not independent, a failure to account for ρ qr in QTL‐assisted selection would underestimate the genetic responses when ρ qr is positive, but overestimate the genetic responses when ρ qr is negative. Estimation bias is more serious at high heritability than at low heritability. Accounting for ρ qr would improve the efficiency of QTL‐assisted selection and the accuracy of QTL detection. The generalized procedure developed in this study allows for quantifying the efficiency of QTL‐assisted selection and assessing estimation bias for ignoring the correlation between q and r for all possible combinations of h 2 , ρ qr , and .