Ongoing mapping efforts and increasingly efficient techniques for
detecting QTL in both breeding and targeted mapping populations are providing a
rapid increase in the number markers available for use in breeding. With large
numbers of markers, population sizes required to combine desirable alleles
across all of these loci can be prohibitive. Choice of the most efficient
strategies can reduce populations sizes required to recover a target genotype by
orders of magnitude and extend our ability to combine large numbers of target
alleles.
Currently, most breeding crosses will be varying both for well mapped, often
simply inherited, traits and others for which no marker association has been
established. Even with ongoing discovery of marker-trait associations, genetic
progress will continue to depend on introducing new genetic variation and at
least initially, the value of this new variation will be unknown. In both instances,
there will be a need to balance selection for known desirable alleles already
present in breeding populations with phenotypic selection to retain desirable new
alleles and allelic combinations. Strategies allowing for, or maximizing the
probability of reducing, linkage drag in combination with selection for known
target alleles is likely to be important in effective exploitation of this unmapped
genetic variation.
Strategies to efficiently combine large numbers of target alleles and
balance marker and phenotypic selection will be discussed using examples from our
breeding program. The effects of incomplete linkage of markers with genes, use
of flanking markers and coupling and repulsion phase linkages between target
loci will be considered.