Rapid Cycle Breeding in Crops with Genomic Selection.

In contrast to animal breeding programs, plant breeders typically use some variation of a cross-inbreed-select breeding scheme.  This breeding methodology has two fold utility; 1) to develop pure-line cultivars (or single cross-hybrids) and 2) to distribute genetic variance between lines and provide homogenous material for replicated testing.  The development of genomic selection (GS), however, provides a method for single plant selection based on prediction models developed from dense genome-wide molecular markers.  Using GS the breeding cycle length can be drastically reduced by selecting parents prior to phenotyping.  To further shorten the breeding cycle and increase the rate of genetic gain, we propose a FastGS breeding scheme where individual F₁s are continually selected and intermated in a recombinational nursery and the breeding cycle progresses without any inbreeding.  To evaluate the effectiveness of FastGS we simulated a maize-like genome with 100 effect loci and 3,000 SNP markers.  Breeding programs were evaluated for a 10 year period of 1) a typical cross-inbreed-select breeding program using both phenotypic and genotypic selection and 2) a FastGS program where S0 plants were continually selected and intermated and inbred lines (for advanced testing) were derived from this recombinational nursery. To minimize the time for each breeding cycle and maximize genetic gain overtime, FastGS is a rapid cycle breeding methods that provides further improvement over inbreeding prior to selection.  However, with fast cycle breeding, there is greater risk for loss of favorable allelic diversity.  Therefore, methods should be put in place to conserve low frequency favorable alleles in the breeding program.