Agriculture Reference
In-Depth Information
continuing to inbreed. Alternatively it is prefera-
ble to discard genotypes which are low yielding
or fail to have the desirable alleles (e.g., they are
susceptible to diseases), which allows the breeder
to allocate resources to select within more promis-
ing early generation lines.
As tools are developed to enable whole-genome
scans, it may become possible to know which F
2
plants have the highest proportion of desirable
alleles. For example, with codominant markers
for all 20 desirable genes, one could potentially
multiplex all 20 markers to identify the genotype
with all desirable alleles, followed by subsequent
inbreeding and continued selection among the
remaining heterozygous loci to identify the geno-
type which has all 20 desirable alleles in a homo-
zygous condition. While this example is for
progeny from a biparental cross, whole-genome
scans can similarly identify the best F
1
plants for
selection in more complex crosses (three-way and
double crosses).
Early-generation yield testing of individual
lines or families has been assessed by various
researchers with diverse results, but generally the
consensus is that high-yielding materials can
be distinguished from low-yielding materials.
Lupton and Whitehouse (1957) reported a signifi -
cant correlation coeffi cient of 0.56 for grain yields
of F
4
versus F
5
lines of winter wheat grown in
successive years. Shebeski (1967), Briggs and
Shebeski (1971), and DePauw and Shebeski
(1973) reported signifi cant correlations for grain
yield between generations of F
3
, F
4
,
and F
5
. These
studies used frequently repeated systematic con-
trols to address spatial variability. To overcome
the cost of using repeated controls, statistical
techniques such as moving-mean (Townley-
Smith and Hurd 1973) and nearest-neighbor
(Wilkinson et al., 1983) analyses have been used
to partition genetic variation from environmental
variation.
In Australia, Fischer et al. (1989) and Gras and
O'Brien (1992) reported signifi cant early inter-
generation correlations for end-use quality traits.
Early-generation testing for quality has been the
basis for improvement in spring bread wheat
(Lukow 1991) and in durum wheat in Canada
(Clarke 2005).
Some contemporary plant breeding programs
exploit these early-generation relationships to
select simultaneously for qualitative as well as
quantitative traits (DePauw et al., 2007). The F
2
plants are selected for resistance to diseases in
nurseries with high pressure from diseases such
as rusts and for other simply inherited traits such
as plant height. The F
3
generation is used to mul-
tiply seed for subsequent multilocation yield trials
and disease nurseries (Color Plate 27). These F
3
nurseries can be used to screen for simply inher-
ited traits such as diseases, time to maturity, and
plant architecture. Nonreplicated F
4
yield-trial
nurseries are grown at multiple locations, and
specialized disease nurseries may be established if
these diseases are not reliably expressed in the
yield-trial nurseries (Color Plate 28).
The harvested grain may subsequently be used
for end-use quality testing. Prior to harvest, heads
are selected to establish families of those lines
which express the highest concentration of desir-
able alleles for all quantitative and qualitative
traits. This procedure is very expensive and is
suitable for application to populations which have
a very high midparent value and are expected to
have complementary genes for all traits to make a
“fi eld-ready” cultivar (Color Plate 29).
Mechanization and computerization have
enabled breeders to handle very large populations
to enhance their opportunities to identify the rare
recombinants with all the desired traits. Mecha-
nized seeders have replaced human-powered
seeders in the past 60 years (Fig. 13.1). Many
wheat breeding programs assess fi vefold more
genetic materials today than 30 years ago. Most
breeding programs are able to score more traits
on more lines each generation, analyze the data,
and interpret it than 30 years ago. Molecular
markers are being integrated as another tool to
enhance effi ciency and achieve genetic gain
through selection.
Impact of molecular markers on
wheat breeding
From a breeder's perspective, identifying QTLs
will be very useful in understanding how impor-
tant agronomic traits are inherited (Campbell
