Agriculture Reference
In-Depth Information
America (Hyten et al.
2006
). Before choosing the appropriate number of genetic
markers (usually SNPs) for a genome scan, it is necessary to have some under-
standing of the LD in the population selected for the study. In general, LD decreases
with distance between marker loci, more slowly in inbreds (soybean), faster in
outbred species (maize), although breeding practices have a large impact (Flint-
Garcia et al.
2003
). LD is, however, very non-uniform across the genome, with both
general trends (more LD in centromeric regions) and pronounced local fluctuation
(Rafalski
2010
). For instance, LD in Arabidopsis extends for roughly 10 kb, which
is a nearly ideal distance for mapping since it extends up to the gene level
(Bergelson and Roux
2010
). Genetic resolution of any mapping methodology
ultimately depends on the amount of recombination available in the experimental
population, as measured by the rate of decay of LD (Rafalski
2010
). In collections
of distantly related individuals many generations have passed and much recombi-
nation occurred since the last common ancestor, therefore resolution of association
mapping will, in general, be considerably higher than in simple biparental
populations. The power of association mapping is strongly dependent upon the
quality of phenotypic data. It is important to stress that in most cases it is necessary
to use well-controlled environmental conditions, including, when possible, use of
growth chambers, especially for the collection of samples for metabolomic or
biochemical phenotypes. Relevance of such phenotypes for field performance
will have to be separately established (Rafalski
2010
). High throughput methods
to precision phenotyping, frequently referred to as phenomics, are developing
rapidly and automated facilities for high precision phenotyping are being
established (E. Finkel
2009
). The use of such a facility has been recently presented
as a powerful tool to investigate plant response to drought stress by Tisn
´
et al. (
2013
), allowing the precise control of watering condition of more than
700 plants. In this way, the environmental variance was strongly reduced allowing
the identification of QTLs for complex traits related to drought response.
Validation and Applications
Validation of the hypotheses generated by association mapping constitutes an
integral part of the experiment (Rafalski
2010
). In one approach, NILs differing
in the alleles at the candidate locus are constructed by repeated backcrossing into a
reference genetic background (Vlad et al.
2010
). The resulting NILs are then
phenotyped side by side, and the amount of phenotypic variation ascribed to the
presence of introgressed segment is estimated. Biparental populations segregating
for the relevant alleles at the associated locus may also be used (Bel´ et al.
2008
).
Alternatively, the association experiment could be expanded by the inclusion of
additional individuals in the expectation that the strength of the association should
improve if the association hypothesis is correct. Association mapping is usually
performed with the objective of applying the results for genotype-based selection of
superior individuals in plant breeding, or as a step toward positional cloning
