Agriculture Reference
In-Depth Information
genotyped, say for dairy cows or thoroughbred horses, with the conse-
quent need for a prediction of breeding value (Calus et al. 2008).
Compared with animals and for success in crops, a reduction in cost
of marker genotyping as well as a change in mindset has been critical for
each program that starts GS on an experimental basis for crop improve-
ment. The attentive help of statisticians is also needed as well as very
complete databases for phenotypic data on the lines produced by a
breeding program for GS to be successful. The approach of GS starts off
with a calibration set of individuals and then a validation population
with the aim of predicting phenotypes based solely on genotyping
without the need for further expensive
field or greenhouse con
rmation
of traits.
Multiple rounds of calibration and validation within a breeding
population are normally needed to assess the ef
ciency of GS methods
and make predictions in crop breeding programs. Among the
rst to
apply GS modeling to crop plants was a group of small grain breeders
from Cornell University (Heffner et al. 2009; Zhong et al. 2009) and a
group from CIMMYT working on international maize and wheat collec-
tions (Crossa et al. 2010). Progress with two diverse maize populations in
Europe has also been noted (Rincent et al. 2012).
GS holds great promises for long-lived species that have breeding
cycles beyond the limits of typical cross-and-select programs, but which
produce lots of offsprings, like fruit trees and forestry species. One
example of this is the use of GS for selection in Loblolly Pine (Resende
et al. 2012). Although GS is a
it merits attention as an
extension and alternative in the realm of association genetics for plant
breeding, as evidenced by an entire recent issue in the journals
G3/
Genetics
devoted to this topic of interest.
“
work in progress,
”
VIII. OUTLOOK
Association genetics in plants is a
fieldwith over a decade of research but
that is still under development, with challenges of appropriate pheno-
typing methods, accessibility to genotyping, and development of new
statistical tools or software. Most of these
fields are improving to close
the gaps in association studies, and enhance its power as a feasible tool
for crop genetics and for breeding progress. The use of AM in QTL
dissection requires evolutionary aspects of the population to be
decoded, and a clear idea of the type of organism used in order to detect
the amount of markers needed. Since breeders have by intuition known
for a long time that evolutionary aspects of a population can have a
