Agriculture Reference
In-Depth Information
derived from earlier inbreeding generations, espe-
cially the F
2
(Table 13.1); hence they tend to be
produced from narrow crosses where there are
fewer important genes segregating.
The major difference between single-seed
descent and doubled haploid breeding is in cumu-
lative recombination frequency, as the single-seed
descent process involves more segregating gener-
ations of selfi ng where effective recombination
can occur (Riggs and Snape 1977). Because most
authors have found few or small differences
between the three methods, one can conclude for
the populations in question that the additional
recombination was not benefi cial. This result may
be to due to the relatively narrow crosses com-
monly used in single-seed descent and doubled
haploid plant breeding. If past history repeats
itself, the methods to create doubled haploids will
become less expensive and will feature fewer
culture-induced variants.
Two additional aspects of doubled haploid
methods are important. First, recovery of mutants
is very effi cient at the haploid level because each
locus is hemizygous and expressed (especially
important for recessive alleles). Second, while
most programs create haploid plants and then
double all of them, selection with molecular
markers at the haploid level is possible for many
traits, thus allowing only selected plants to
undergo the chromosome doubling process. If it
is not possible to select at the haploid state, selec-
tion using only homozygous diploid lines is very
effi cient for studying mutants, because both target
and nontarget mutants will be expressed and not
hidden by heterozygosity. Similarly, using molec-
ular markers in selection becomes more effi -
cient because the heterozygous lines have been
removed.
Finally, with the rapid speed that doubled
haploid lines theoretically can be made, consider-
able theory has developed on how to use doubled
haploid lines to estimate population means and
among-line standard deviations (Choo et al., 1979;
Choo and Reinbergs 1979; Snape et al., 1984;
Caligari et al., 1985; Choo 1988; Baenziger 1996;
Baenziger et al., 2001). While much of the research
attempts to answer theoretical concerns with
plant breeding methodology and understanding a
given crop, a very practical use of doubled haploid
breeding can be to estimate the value of a cross.
Reinbergs et al. (1976) estimated that as few as 20
doubled haploid lines in barley (
Hordeum vulgare
L.) will accurately estimate among-progeny line
mean and standard deviation. Hence even if
doubled haploid technology is expensive and few
lines can be made, it could allow breeders to iden-
tify the most important crosses with which to
work.
Doubled haploid cultivars have been released
in a number of countries and some have become
dominant cultivars. For example in Canada in
2007, three of the fi ve most widely grown culti-
vars in the market class Canada Western Red
Spring (CWRS) were doubled haploid cultivars.
'Lillian' (DePauw et al., 2005) accounted for
about 15% of the CWRS acreage, 'Superb'
accounted for 12% (T.F. Townley-Smith, pers.
comm.), and 'McKenzie' (Graf et al., 2003)
accounted for 7%. 'Snowbird' (Humphreys et al.,
2007b) and 'Kanata' (Humphreys et al., 2007a)
accounted for all of the Canada Western Hard
White acreage. 'Andrew' (Sadasivaiah et al., 2004)
accounted for 99% of the Canada Western Soft
White Spring acreage.
Backcrossing
Backcrossing is a method of recurrent hybridiza-
tion by which a desirable allele for a specifi c trait
is substituted for the alternative or undesirable
allele, as initially proposed by Briggs (1938). The
parental source of the desirable allele is desig-
nated the donor parent, and the parent used
repeatedly in hybridization is the recurrent parent.
The recurrent parent is reconstituted at the rate
of 1 − (1/2)
n
,
in which
n
is the number of recur-
rent crosses or backcrosses. The effectiveness of
the backcross method depends upon (i) the heri-
tability of the trait, (ii) the degree to which the
expression of the trait is independent of back-
ground genes for its expressivity, (iii) expression
of the trait or its detection phenotypically (or with
markers) in the F
1
for subsequent backcrossing,
(iv) linkage of the desirable gene with undesirable
genes from the donor, and (v) the number of
backcrosses necessary to recover a desirable level
