Strategies of marker-aided recurrent selection

Abstract

The choice of a marker-aided recurrent selection (MARS) strategy depends on a number of factors. An experimental comparison of these factors can be costly or infeasible, but a theoretical comparison can provide some guidelines. This study was conducted to evaluate the genetic aspects of expected responses to various MARS methods and to compare the efficiencies of MARS relative to traditional phenotypic recurrent selection (PRS) and to marker-aided mass selection (MS). The MARS methods examined here are MS, half-sib family (HSF), full-sib family (FSF), S1 family (S1F), testcross progeny (TC), combined FSF selection with within-family MS (CSFS), and combined HSF selection with within-family MS (CSHS). If a large fraction of the additive genetic variance (p) in a character can be explained by marker loci, the efficiencies of MARS relative to corresponding PRS can be great. This is most pronounced when family size (n) is small and decreases as n increases. An exception to this is for CSHS, where MARS is superior to PRS across all family sizes. Similarly, CSFS performs better than FSF selection under a wider range of family sizes (1 < n < 50). With a constant rate selected and a relatively large family size, S1F, CSFS and CSHS selection are superior to MS, whereas HSF selection is inferior to MS. With constraints of the same effective size selected for each method, MS, which takes advantage of the greatest selection intensity, is the most efficient method. Nevertheless, MARS among small families or individual plants can produce substantial increases in selection responses.