Applying Genomics Tools for Breeding Submergence Tolerance in Rice - Translational Genomics for Crop Breeding: Abiotic Stress, Yield and Quality - page 24

the submergence breeding program, association
mapping or family-based QTL approaches could
be effective for exploiting routinely available
breeding populations. The main constraint from
the breeding program's perspective is the cost per
sample of using high-throughput SNP assays as
well as the expense in field trials for phenotyping.
As the ideal number of markers for association
mapping in rice is

ple is still high (around US$40/sample). There-
fore, in the majority of cases, SNP genotyp-
ing for background selection would need to be
combined with SSR genotyping for foreground
and/or recombinant selection. Alternatively, SSR
genotyping could be adopted as a “first round”
of background selection to reduce sample num-
bers for a “second round” of SNP genotyping,
because simulation studies suggest that a small
number of carefully selected markers can be used
for accurate background selection in early back-
cross generations (Hospital et al. 1992; Visscher
1996).
The high SNP multiplex level of these new
platforms could enable recombinant and back-
ground selection steps to be combined, poten-
tially saving a considerable amount of labor,
but this scheme would obviously depend on the
funds available for the MABC program. Other
medium through-put SNP genotyping platforms
(24- to 96-plex) such as the Fluidigm EP1 system
provide lower costs per sample, and may provide
several new options for MABC. For example,
combinations of genotyping with different SNP
platforms could be implemented, or in different
generations (M. Thomson, IRRI, pers. comm.).
In general, computer simulations can provide
useful guidelines regarding the use of combined
low- and high-throughput marker systems (Her-
zog and Frisch 2011).

5,000, based on the level of
linkage disequilibrium, high-throughput marker
assays are essential (Mather et al. 2007). There-
fore, using association-mapping methods with
candidate gene markers and scanning specific
regions delimited by QTLs are probably the most
realistic methods that can be implemented within
the breeding program, at least in the short term.
Genome-wide association studies offer great
potential for identifying new QTLs in germplasm
collections, but this research cannot be per-
formed by public-sector breeding programs.

>

Beyond the SUB1Gene

EfficientDevelopmentofAdditionalSub1
Varieties

Currently several IRRI and other associated
projects are under way to develop addi-
tional enhanced Sub1 varieties for Bangladesh,
India, Nepal, Pakistan, Cambodia, Vietnam, and
Africa. Furthermore, the SUB1 gene has been
incorporated into a wide range of IRRI's elite
breeding material, across several programs, as a
way to develop a broad range of varieties for dif-
ferent ecosystems across many countries, as an
“insurance” against short-term flooding, even in
areas where flooding does not normally occur.
Despite its success, MABC is a labor-
intensive and resource-demanding exercise. To
date, all of the Sub1 varieties have been
developed using SSR markers. The advent
of new high-throughput SNP genotyping plat-
forms promise to improve efficiency in the
future (Thomson et al. 2012), because the
time required for background selection would
decrease markedly. Although the cost per SNP
data point is cheaper than SSRs, the cost per sam-

DevelopmentofSub1 plus Varieties:Increasing
SubmergenceTolerance

Generally, the effect of SUB1 is tolerance of sub-
mergence during flash flooding of up to 14 days.
In some regions of Asia, flash flooding may occur
for longer periods of time - up to 21 days (or even
longer). The identification of additional QTLs
for submergence tolerance is now a major objec-
tive of the IRRI submergence breeding program
so that varieties with submergence tolerance of
> 14 days can be developed. Furthermore, the

incorporation of additional QTLs could increase
the survival and vigor of regeneration after sub-
mergence stress.