Agriculture Reference
In-Depth Information
(UCGM) was developed by CIAT (Orjuela et al. 2010). With this aim,
we derived a set of 165 anchors, representing clusters of three SSR
markers arranged into nonrecombining groups. Each anchor consists of
at least three closely linked SSRs, located within a distance below the
genetic resolution provided by common segregating populations (
200
individuals). We chose anchors that were evenly distributed across the
rice chromosomes, with spacing between 2 and 3.5Mbp (except in the
telomeric regions, where spacing was 1.5Mbp). Anchor selection was
performed using
in silico
tools and data: the
O. sativa
cv. Nipponbare
rice genome sequence, the CHARM (Comprehensive High-Throughput
Arrays for Relative Methylation) tool and information from the Gra-
mene database and the OrygenesDB database. Sixteen AA-genome
accessions of genus
Oryza
were used to evaluate polymorphisms for
the selected markers, including accessions from
Oryza sativa
,
Oryza
glaberrima
,
Oryza barthii
,
Oryza ru
<
pogon
,
Oryza Glumaepatula
,and
Oryza meridionalis
. High polymorphism was found for the tested
Oryza sativa
wild rice combina-
tions. We developed Paddy Map, a simple database that is helpful in
selecting optimal sets of polymorphic SSRs for any cross that involves
the previously mentioned species. We validated the UCGM by using it
to develop three interspeci
×
Oryza glaberrima
or
Oryza sativa
×
c genetic maps and then comparing genetic
SSR locations with their physical positions on the rice pseudomole-
cules. We demonstrated that the UCGM is a useful tool for the rice
genetics and breeding community, especially in strategies based on
interspeci
c hybridization. The Paddy Map Database is available
5. Genomic Selection.
GS is the most recent MAS method. It considers
markers on the entire genome, rather than only those signi
cantly asso-
ciated with the traits of interest, as in the case of other marker-assisted
selection. In GS, individuals of a training population (TP) are both
phenotyped and genotyped to develop a prediction model. GEBV are
calculated on a genotyped breeding population (BP) as the sum of effects
of genetic markers across the entire genome. GS then potentially cap-
tures all the QTLs that contribute to the trait variation (Meuwissen et al.
2001). Many simulation and empirical results have shown the effective-
ness of GS for modeling the sum of effects of unknown QTLs across the
genome (see references in Bernardo 2013). In a CIRAD-CIAT project,
aTPof
tropical japonica
rice was designed to build a prediction model
for GS. The TP encompassed 355 lines from RS populations, which
well represent our breeders
working collection. All entries were gen-
otyped at high density through GBS. The TP was phenotyped for yield
'
