SIMULATIONS AND PRACTICAL PROBLEMS OF APPLYING MULTIPLE MARKER ASSISTED SELECTION AND DOUBLED HAPLOIDS TO WHEAT BREEDING PROGRAMS

Marker assisted selection (MAS) and wheat doubled haploids (DH) are re latively new technologies, recently applied to wheat breeding programs . Simulations demonstrate that DHs increase the efficiency of MAS, and offer faster strategies for combining large numbers of genes with a m inimum number of marker tests. When small numbers of marked loci (1-3) are selected simultaneously, selection of DH progeny is 5-6 times mor e efficient than selecting F-4 derived families. Combining 4-8 marked loci, screening of F-2 plants and using only those plants homozygous o r segregating for all of the marked loci as parents for DH production (10-31% of F-2 plants) is 3-10 times as efficient as using F-1 plants. A number of protocols have been proposed involving sib-matings and se lection to fix some genes, with further selection in the second genera tion to improve the proportion of useful DH lines. In one scheme (reco mbinant F-2 selection) all F-2 plants, either homozygous or heterozygo us for the marked alleles, are intercrossed at random and the recurren t F-2 plants still having these alleles are used for DH production. An alternative strategy (recurrent DH selection) is to select from an in itial DH population and intercross those lines having most favourable marked loci with a second cycle of DHs to fix all favourable marked lo ci. Combining more than 12 marked gene loci does not seem feasible, du e to the very large numbers of F(2)s (>2000) required. This has implic ations when using MAS for quantitative trait loci, where many minor ge ne loci would have to be combined. Direct selection for some multi-gen ic quantitative traits amongst the DH lines may be more efficient than using MAS where recurrent selection is used. At the Cereal Research C entre, the practical problems of using these protocols as part of the spring wheat breeding program are being evaluated.