Biology Reference
In-Depth Information
five cM, corresponding to 3,000-7,000 markers
to cover the entire polyploid genome (Raboin
et al. 2008).
Wei and colleagues (2010) recently published
the only association-mapping study related to
cane and sucrose yields using a large panel of 480
clones representative of the modern germplasm
of current breeding programs. The panel was
tested in three different locations using a
replicated design. This panel was genotyped with
a Diverse Arrays Technology (DArT) microar-
ray (Heller-Uszynska et al. 2010). The authors
generated 15,360 DArT score markers and used
them as continuous markers, assuming that this
quantitative scoring was related to the num-
ber of copies of each marker per genotype,
which would be a major advantage in the highly
polyploid sugarcane. Nevertheless these 15,360
markers corresponded to 1,531 discrete polymor-
phic markers within a 0.05-0.95 frequency range.
Several methods can be used to account for
structuration in the panels. A Bayesian cluster-
ing model implemented in STRUCTURE soft-
ware (Pritchard et al. 2000) or kinship analy-
sis inferred from molecular data implemented in
SPAGEDI software (Hardy and Vekemans 2002;
Yu et al. 2006) are currently used in association-
mapping studies in plants but are not suitable
for high autopolyploid species. Principal compo-
nent analysis (Price et al. 2006), genomic control
(Devlin and Roeder 1999), or pedigree matrix
are suitable alternative methods to account for
the structure of the panel in the genomic con-
text of sugarcane. Wei and colleagues (2010)
analyzed marker/trait associations with differ-
ent mixed linear models they adapted from the
conventional basic model of Yu and colleagues
(2006) to minimize the risk of type I and type
II errors. These models combined cofactors rep-
resenting levels of relatedness between individ-
uals, genotype by environment interactions, and
spatial variation within trials. The cofactor used
for levels of relatedness between individuals was
a pedigree matrix that is very efficient, as it dra-
matically reduces the number of significant asso-
ciations. Finally, using the most elaborate model,
with P
0.01, the authors detected 47 discrete
and 352 continuous markers associated with cane
yield, and 42 discrete and 377 continuous mark-
ers associated with sugar content. Depending on
the traits surveyed, these numbers of associations
were 3 to 6-fold lower when considering a thresh-
old
P
value of 10
−
3
. Six associations for sucrose
content were still perceptible at P
<
10
−
4
, while
at this threshold no significant markers would
be expected by random chance. The results of
this first association study of sugarcane yield
components are encouraging, thanks to the large
panel surveyed and the phenotypic data that
were acquired on a multi-loci experimental basis.
However, the repeatability of these marker trait
associations and validation of these markers at
the scale of a breeding program remain the major
concern.
<
Perspectives
In highly polyploid sugarcane, tagging useful
QTLs related to complex traits such as cane or
sucrose yield will always be a challenge, irre-
spective of the strategy used, whether bi-parental
QTL or panel association studies. Regarding
ploidy levels, one would expect a reduction in
the mean size of allele effects within the loci
of interest. Analyses show medium to large pro-
portions of trait variation explained by a swarm
of barely significant QTLs with small individ-
ual effect size (R
2
). Moreover QTLs are often
specific to a single environment and crop cycle.
Validating QTLs across different genetic back-
grounds for breeding purposes is a difficult task
(Piperidis et al. 2008). In sugarcane, as in other
crops, the use of MAS appears to be useful for
traits with simple inheritance, such as disease
resistance (Daugrois et al. 1996; Raboin et al.
2006; Aljanabi et al. 2007). The advantage of
using of MAS for quantitative traits such as
yield is more questionable. Up until now, the
only known example of using markers for selec-
tion in sugarcane is for rust resistance (Costet
et al. 2012). Numerous QTL mapping experi-
ments in many species have been published in
Search WWH ::
Custom Search