Agriculture Reference
In-Depth Information
are but a few of the possible subtle differences which
might be very important to individuals or plots or fields
or farms.
There are a number of ways that plant breeders try to
take genotype
which direction is dominance acting in, and to what
degree.
Let us consider one particular use of the information,
how can we predict the response to selection? Before you
start a selection programme you would obviously like
to know what sort of response you might expect for any
given input, which traits are worth targeting, which
populations or crosses are best to use. Is it a worth-
while venture? These questions involve many aspects
of the biology of the crop, its handling in agriculture,
the availability of other methods to affect crops, for
example, chemicals such as pesticides, herbicides etc.
But one of the main components that will determine
the outcome is the amount and type of variation that is
present. For example, in the extreme case of no genetic
variation the breeder is wasting time trying to select
superior genotypes!
If genetic variation is present, but small compared
with that due to the environment, then progress can be
made but only very slowly, unless very large numbers are
handled - in other words much of the time the breeder
will be selecting phenotypes which are 'superior' but as
this is mostly due to the environment it will not give
a reliable indication of a ' superior ' genotype. If on the
other hand the phenotype is a good reflection of the
genotype, that is most of the variation is genetic, then
progress will be quick, since when the breeder selects a
good phenotype and uses it as a parent it will pass on
the superior attributes ( via its genes) to its offspring.
One obvious question to ask is can we estimate how
much of the variation we observe, in for example the
F 2 we were looking at, could be ascribed to genetic
differences of segregating genes and how much to envi-
ronmental causes? We therefore need to measure the
proportion of the total observed variation, that is genetic
variation, and such a measure is called heritability .
×
×
E) into
account in breeding but it does mean that there a is need
to carry out trials over a range of environments which
might be simply different sites or over years/seasons or
running trials with defined differences such as water
levels etc. Genotype
environment interactions (G
×
environment at its simplest can
be examined by looking at the variance (or standard
error by taking the square root) of the phenotypes over
the range of environments and selecting for the low-
est variance as being the most stable - remembering
that we also need a good mean expression! Or a much
more sophisticated approach is to use the mean of all
the material grown to provide a biological measure of
the environment and compare (usually by regression
slope of the line) the individual lines, families, clones,
etc. against this.
It is not appropriate to go into greater details here
about G
×
E or the various possibilities to take it into
account. Needless to say many breeding programmes
effectively ignore G
×
E in an explicit way but take
it into account to a modest extent by trialling the
more advanced material at different sites, by the fact
that selection is carried out over a number of years,
etc. A more detailed examination of analyses of mul-
tiple year and multiple location trials is presented in
Chapter 7.
Let us now consider prediction protocols associated
with answering which parents will give the best progeny
and which traits will respond most significantly to the
selection we impose?
Genetically based predictions
HERITABILITY
Plant breeders use all the genetic information (qualita-
tive and quantitative) in just the same way as we use
the information from Mendelian Genetics - in other
words to predict the properties of generations or fam-
ilies that have not actually been observed. So from an
analysis of the observed variation, firstly determine how
much of the variation is due to environmental effects
and how much due to genetic effects. Often it is desir-
able to go further and separate the genetic into additive,
dominance, and other genetic effects, and to determine
For a modern plant breeder (or indeed a farmer with no
knowledge of genetics) to make progress in an organized
programme of selective breeding, two conditions are
a must:
There must be some phenotypic variation within the
crop. This would normally be expected, even if it were
due entirely to the effects of a variable environment
 
Search WWH ::




Custom Search