Improvement of Yield per se in Sugarcane - Translational Genomics for Crop Breeding: Abiotic Stress, Yield and Quality - page 229

homo(eo)logous haplotypes. Sugarcane does not
appear to have undergone a major reshaping of
its genome as a consequence of polyploidization.
Additional series of homo(eo)logous BACs are
needed to refine these first results.
A recent NGS-based strategy called
genotyping-by-sequence (GBS), associated with
one step aimed at reducing genome complexity
(generating a limited amount of sequence data
to be analyzed), is another new way to discover
SNP (Baird et al. 2008; Elshire et al. 2011).
The concept is based on acquiring the sequence
adjacent to a set of particular restriction enzyme
recognition sites, rather than randomly sequenc-
ing the whole genome. Large amounts of
polymorphism data can be generated by massive
parallel sequencing. This approach increases the
coverage for a given sequenced site, increasing
both the confidence in base identity and the
likelihood that the same sites will be sequenced
in multiple samples. This promising approach,
which could allow simultaneous SNP discovery
and genotyping, is currently under investigation
in sugarcane (Glynn et al. 2011; D'Hont pers.
com.).

phenology, morphogenesis, carbon acquisition,
and allocation among sinks. Yield formation
can thus be described dynamically as a set of
interactive equations using only a small num-
ber of genotypic parameters. These parameters
control plant reaction norms that are at the basis
of plant growth response to the environment
(Dingkuhn et al. 2005). The different param-
eters involved in phenotype expression can be
considered as synthetic component traits, pre-
sumably controlled by fewer genes than the
integrative, complex agronomic trait. In this
sense, if the models represent relevant biolog-
ical processes, the model parameters measured
can be expected to be closer to gene or QTL
effects, in the sense that Genotype x Environ-
ment 'noise' is reduced (Hammer et al. 2002;
Reymond et al. 2003; Yin et al. 2003; Dingkuhn
et al. 2005). In such a heuristic approach, vari-
ation in parameters among genotypes can be
interpreted as the expression of allelic diversity
and analyzed accordingly in QTL or associa-
tion studies. Model parameter values can be esti-
mated by optimizing a relevant criterion based
on deviations of predictions from observations,
using target files containing observations on the
plant (classical phenotyping) and on the envi-
ronment (e.g., weather, soil). Traits resulting
from ecophysiological models can be used for
more detailed investigation of the biological pro-
cesses involved in the development of yield.
Model-assisted phenotyping has already been
used for peach (Quilot et al. 2004; Quilot et al.
2005), barley (Yin et al. 1999), maize (Reymond
et al. 2003), and rice (Dingkuhn et al. 2006;
Luquet et al. 2006; Luquet et al. 2007). Sev-
eral studies allowed ecophysiological model-
ing of sugarcane yield elaboration with mod-
els like Mosicas (Martine et al. 2000; Martine
et al. 2001; Martine 2003; Martine 2007), Apsim
(Keating et al. 1999; Keating et al. 2003), and
Canegro (O'Leary 2000). Some authors are
beginning to evaluate the feasibility of model-
assisted phenotyping in sugarcane (Luquet et al.
2010; Martine

Model-Assisted Phenotyping

For overcoming the problems caused by geno-
type by environment interactions in yield traits
possible solutions include model-assisted phe-
notyping. The “gene-to-phenotype” approach
connects ecophysiological models to statistical
methods for detecting complex traits (Hammer
et al. 2004; Chenu et al. 2009; Prudent et al.
2011). This approach should improve the detec-
tion of QTLs and our understanding of the
genetic architecture of complex biological pro-
cesses. Phenotypic traits for production poten-
tial interact strongly with the environment, and
are the result of multiple processes that are dif-
ficult to measure and tag at the genetic level.
Plant growth modeling can advance our under-
standing of complex biological systems by for-
malizing dynamic interactions among several
biological processes, such as those related to

et al. 2010; Nibouche et al.

2010).