Agriculture Reference
In-Depth Information
Excluding the error associated with their esti-
mates, narrow-sense heritability should be no
greater in magnitude than broad-sense heritabil-
ity for a given trait. Holland et al. (2003) and Iqbal
et al. (2007a,b) provide calculations of narrow-
and broad-sense heritability for some agronomic
traits and for grain protein content.
Selection results in discrimination among
genetically variable individuals, whereby some
are chosen to establish the next generation.
Response to selection (
R
), also called genetic
advance or genetic gain, is the difference between
the mean phenotypic value of the progeny of
selected individuals (a generation subsequent to
selection) and the mean of the entire population
before selection. A signifi cant response to selec-
tion is represented by a signifi cant change in the
population mean between generations. Response
to selection depends on three factors: (i) magni-
tude of variability,
V
(
P
)
; (ii) confounding effects of
environmental and interaction components of
variability on the genetic variation which deter-
mines heritability; and (iii) the proportion of the
population selected, also known as the selection
intensity. Response to selection (
R
) may be
described mathematically as:
in which
H
N2
is the narrow-sense heritability of
the related trait and
r
G
is the genetic correlation
between the primary trait and related trait.
Assuming the same selection intensity, and if
H
N2
r
G
>
H
N
, then indirect selection is more effec-
tive in changing the primary trait mean than
direct selection. A simple and contemporary
example of indirect selection is selection based on
molecular markers (for further reading, see Allard
1960; Falconer 1981; Acquaah 2007).
METHODS OF SELECTING WHILE
INBREEDING TO DEVELOP A CULTIVAR
Two very early methods of plant breeding which
do not necessarily begin with hybridization are
mass selection and pureline selection. Both selec-
tion methods were used when landraces (a mixture
of genetically different inbred lines with a low
level of natural outcrossing) were brought to new
areas. In this case, new phenotypes may appear
due to the lines no longer being in the previous
environment(s). As mentioned previously, selec-
tion does not create genetic variation but only acts
upon genetic variation already in the population,
and selection acts effectively only on heritable
differences.
In mass selection, deleterious plants or off-
types (plants which do not phenotypically appear
as if they belong in the line) are removed from the
population and the remaining plants are harvested
in bulk (or
en mass
e). It is a very good way to
quickly develop more uniform cultivars but rarely
improves the yield of the cultivar as most of its
individual plants are retained. Mass selection is
rarely used today to create new lines but is com-
monly used to purify advanced lines. Roguing to
remove off-types (contaminants) or variants
(inherent variation) in a breeder or foundation
seed fi eld is a form of mass selection.
Pureline selection differs from mass selection
in that individual plants are selected, their progeny
are evaluated, and the best plant-derived progeny
is used to become the cultivar. Pureline selection
attempts to identify and propagate the best indi-
vidual in a line, whereas mass selection intends to
keep the important aspects of the population. As
R
=
k
÷
V
(
p
)
H
N
In this equation
k
is the standardized selection
differential, which takes into consideration the
mean phenotypic value of the selected portion of
the population, the mean phenotypic value of the
entire population, the phenotypic standard devia-
tion, and the proportion of the population selected.
The equation to measure response to selection is
fundamental to plant breeding and deserves
repeated refl ection by aspiring and rejuvenating
breeders.
This equation for
R
estimates the response due
to direct selection for a trait. For some traits, due
to the diffi culty or the time to measure a trait,
indirect selection (selection for a trait using a
related trait) is preferred (Falconer 1952; Ortiz et
al., 2007). The equation, modifi ed for indirect
selection response, becomes:
R
=
k
÷
V
(
p
)
H
N2
r
G
