Application of Index Selection in Upland Cotton Breeding Program Using QuLine Simulation

Comparative Genomics and Bioinformatics

N/A

Li, Weihua; Ye, Guoyou; Li, Weihua; Ye, Guoyou

Li, Weihua; Ye, Guoyou; Li, Weihua; Ye, Guoyou

Li, Weihua; Li, Weihua

When selection is applied to the improvement of the economic value of plants, it is generally applied to several characters simultaneously and not just to one, because economic value depends on more than one character. How, then, should selection be applied to the component traits in order to achieve the maximum improvement of economic value? There are several possible procedures. One might select in turn for each trait singly in successive generations (tandem selection); or one might select for all the traits at the same time but independently, rejecting all individuals that fail to come up to a certain standard for each traits regardless of their values for any other of the traits (independent culling levels). In this study we compared four index selection methods, which were applied simultaneously to all the component traits together, with tandem selection to investigate the effectiveness of these methods on giving rapid improvement of economic value in Upland cotton breeding program. Lint yield, pre-frost harvest ratio, fiber length, uniformity ratio, fiber strength, fiber fineness, Fusarium wilt, Verticilium wilt and days to boll opening were chosen as component traits of index selection, appropriate weight being given to each trait according to its relative economic importance based on the genetic information gained so far. One environment type, four inheritance models (additive, dominance, over-dominance, and epistasis), two levels of linkage strength between lint yield and fiber strength, formed a total of eight GE systems. Then computer simulation approach was used to generate the starting population of genotypes by the QU-GENE engine, followed by the manipulation of the starting population according to the breeding and selection strategy to be tested with application module QuLine, which was developed in a collaborative project between The University of Queensland and CIMMYT. Results from the analysis of variance showed that inheritance model (IM), linkage strength (LS), among or within family selection involved in index selection strategies (AFS) and two-way interaction between inheritance model and linkage strength (IM*LS) were the four main contributors to the overall variation of lint yield gain; whereas IM, AFS, IM*LS contributed most to the variation from addition of simple traits (AST) and improvement of optimum index (OI) was mainly attributed to AFS, IM, IM*AFS. Among the four genetic models highest genetic gain of lint yield, AST and OI were all obtained from epistasis model after one breeding cycle, however, no significant difference was observed from dominance and over-dominance models for optimum index selection (OIS). 2.93 percent higher genetic gain of lint yield was achieved when genes were tightly linked, recombination frequency of which being 0.001, than that when genes were independently inherited; in contrast, 3.02 percent less genetic gain of OIS was obtained from high linkage strength of genes between lint yield and fiber strength than from low linkage strength. More than double genetic gain of OI was produced using OIS as compared with tandem selection for component traits of selection index. Ten percent more genetic gain was achieved for OI when within family selection of OI was implemented than that when within family selection of lint yield was conducted from F2 through F7.