W39
Station d'Amelioration des Arbres Forestiers, INRA, 45160 ARDON, FRANCE
As tree improvement is limited by delayed evaluation of performance for
quantitative traits, the identification of implicated QTL could provide a valuable
tool for early selection. We are developing a method for QTL detection by using
connected families from a mating design. The plant material analysed for this study
is a factorial mating design involving 12 European larches and 12 Japanese larches.
The hybrid families have been evaluated for 11 years for growth and morphological
characters and for wood quality (wood gravity assessed by pilodyn and
microdensitometry). Genotypes of parental trees are determined at molecular marker
loci. Family frequencies for a marker are deduced from parental genotypes for this
marker. Statistical relationships between marker frequency and quantitative
performance are then investigated. Significant relationships can be validated at an
individual level in families showing segregation for the marker. The establishment
of genetic maps for each parent (68 markers for the European parent and 81 markers
for the Japanese parent) provides spaced markers for further QTL
detection. This method for QTL identification presents advantages over the
traditional QTL detection which is usually achieved by examining the relationships
between marker genotypes and performance of individuals segregating within one
family: (i) genotyping is restricted to parents for the whole set of markers and
to hybrid families for markers potentially linked to QTL, (ii) the multiplicity of
parents increases the probability for at least one family to cosegregate at the QTL
and at the associated marker locus, (iii) QTL detected within many families are
more suitable for marker-assisted-selection at the level of the mating design.