Selective genotyping, i.e. increasing the size of
the population phenotyped and genotyping only
individuals from the high and low tails of the population,
can considerably improve the efficiency
of experiments aimed at detecting and locating quantitative
trait loci (QTLs) affecting a single
trait. In this paper we study how selective genotyping can
increase the efficiency of multitrait QTL
experiments. By selecting on an index combining the
variables of interest and having the maximum
correlation with each variable, the efficiency of QTL
detection is increased for each trait. The
efficiency of selective genotyping relative to random
selection strongly depends on the correlation
between the index and each variable. The optimum selection
rate that minimizes costs for a given
experimental power depends also on this correlation and on
the genotyping costs relative to
phenotyping costs. When the population segregating for
the quantitative traits and the markers is
not as simple as a backcross or an F2
population, but is composed of several connected or
unconnected families, selective genotyping can be used
to improve the efficiency of the QTL study.
In this case, the extreme individuals should be selected
within each family. A method is provided
to choose the selection rates within each family in order
to optimize the global power of the
experiment when the family sizes are unequal.