Transcript Bulk Selection Methods

Bulk Selection Methods
The Basic Scheme
• Grow out the F2 generation. Allow natural
selection to occur. Harvest seed from remaining
plants.
• Grow out the F3 generation. Allow natural
selection to occur. Harvest seed from remaining
plants.
• Continue for several generations. Select single
plants, grow progeny rows, harvest seed for yield
testing.
Florell (1929)
•
•
•
•
•
•
19 wheat crosses grown as bulks
First selections in F5 or F6
Selected 9-49 heads per cross
Planted head rows, kept best (looking) 45
Increased seed for yield test
33 of 45 beat the check. Top yielding line was 55
bu/a vs. 37 bu/a for check
A procedure for inbreeding a segregating
population until the desired level of
homozygosity is achieved. The seed used to
grow each inbreeding generation is a sample
of that harvested from plants the previous
generation. The bulk method is used
predominately with inbred species. It is
totally unsuited for fruit crops and most
vegetables. Unlike the mass selection
method, no human imposed selection is made
during successive generations of inbreeding in
bulk.
Bulk selection was developed in Sweden in
the early 20th century. It was trying to
handle segregating generations of a winter
wheat hybrid that attempted to combine
winterhardiness and high yield. The bulk,
grown over several years in Swedish winters,
was influenced by natural selection during the
approach to homozygosity. Natural selection
increased the frequencies of winter-hardy
types in the population.
Because of the importance of natural selection
in this method, the breeder should carefully
choose environments in which natural
selection is likely to favor the desired
genotypes in the population, i.e., a population
segregating for disease resistance should be
grown in the presence of the pathogen in order
to reduce the productivity of susceptible
plants. Because of this concern for the
environment, the population undergoing the
bulk method would rarely be grown in offseason nurseries.
The central issue in bulk population breeding
is the nature of this correlation : “Is the ability
to survive in competition related to
agricultural worth?”
Genetic Considerations
Genotypic frequencies in a population inbred by the
bulk method are determined by four variables
associated with natural selection in a heterogeneous
population.
1) genetic potential of a genotype for seed
productivity
2) competitive ability of a genotype
3) influence of the environment on the
genotype expression
4) sampling of genotypes to propagate the
next generation
There is no way to know if certain F2 plants
have progeny in F3 generation, or later
generations. There is also no way to predict
the genetic variability for any character in any
generation. If variables favor the desired
genotype, the frequency of desired genotypes
will increase; if not, you’re out of luck!
Subdivision/types of Bulk Method
Practical short-term breeding : Up to F2 to F6
generation.
Survival among pure line genotypes
Survival of different plant types in hybrid derived populations
Long-term Evolutionary breeding
Composite cross method
Long-term Evolutionary breeding
Composite cross method
The two methods are separate entities. If one selects
out of F4, F5, or F6, having bulked in F2, the probably
impacts of natural selection will be small in most
cases. This does not mean insignificant, however.
On the other hand, if one selects out of a bulk
propagated for 15 to 20 generations, the impact of
natural selection may be much greater.
Survival among pure line genotypes : Harlan, J., and
Martini, 1938. J. Agric. Res. 57:189-199.
They reported on the effects of natural selection in
cultivated plants. Mixed 11 pure lines of barley and
grew them for 4-12 years at Ag. Experiment Stations
from Davis, CA to Ithaca, NY. The change in
composition was monitored at each location.
% of
pop
Excellent competitors
Interme diate
Poorer of 2 elim inated
in straight line; other
Interme diate
type in hu mped curve
Poor competitors
At least 1 line was
Elim inated rapidly at
Each station.
Results
1 or 2 lines became dominant very quickly
at each location.
Natural selection in barley is a significant force
at all locations and a force of great magnitude
at some locations.
Similar experiments were conducted by
Suneson and Wiebe (1942) with four varieties
of California barley. They mixed equal
amounts of the varieties and a census was
taken annually over a 9 year period of
propagation. They also measured the mean
agronomic performance in adjacent plots.
‘Atlas’ quickly dominated the bulk
population. ‘Hero’ and ‘Vaughn’ were
virtually eliminated. But differences in
productivity of the cultivars was small.
The experiment was not conclusive in establishing a
relationship between yielding ability and survival.
‘Vaughn’, the highest yielder, was the poorest
competitor; so there was not a very strong positive
association between these characteristics. The
marked differences in competitive ability must have
depended on other characteristics of competitive
ability besides yield. Later Josh Lee (1960) provided
information that might explain superiority of ‘Atlas’
over ‘Vaughn’. ‘Atlas’ accounted for 55% of spikes
in mixed stands, but in a pure stand it has 40% fewer
spikes than ‘Vaughn’. He found that ‘Atlas’ also
produced a larger root system then ‘Vaughn’.
An important feature of these experiments, which may not
necessarily relate to the bulk population method, is that these
experiments all involved inbred homozygous lines. But the
predominant use of the bulk method is to “self-in-bulk” until
homozygosity is reached.
Segregation will occur in bulk hybrids with the result that the
competing genotypes are not expected to be constant from
one generation to another. In previous studies, the
populations investigated were very simple in terms of
components – pure lines. It was not possible to assay for the
survival of an allele or allelic combination per se because that
allele or allelic combination was always associated with all
other alleles in the genotype.
Only at the time of homozygosity (F6 to F8) will the situation
begin to represent that of variety mixtures.
Plant height segregated in this cross in a monofactorial basis.
Samples of the segregating generations were from the F2 to
the F6 generation. One sample was grown with zero Nitrogen
and a second with high N.
Obviously, the short plants were at a disadvantage in
competition with tall ones in this bulks in both cultures.
However, the disadvantage was intensified under high N
conditions. In a pure stand the short plants outyielded the tall
by a mean of 35%. In this study, natural selection was
counterproductive to the goal of breeding high yielding rice
cultivars. Tall genotypes had vigorous vegetation which gave
them a decided competitive advantage for light interception.
Such a high differential in reproductive rates caused the
demise of the short, high yielding genotypes very rapidly.
Peta (TT) x Taic hung N ative I (tt) = F1 selfed
(tall , leafy)
(dwarf)
F2
F3
F4
F5
F6
Expected % Dwarf
25
37.5
43.8
46.9
48.5
% Dwarf
Low N
24.9
31.2
33.5
28.3
23.7
segregating F 2
High N
24.9
22.4
18.9
14.2
8.8
Jennings recommends roguing the population
to remove the good competitors in order to
eliminate the bad effects of competition in
segregating populations. But Allard
concludes that agronomically poor types are
generally poor competitors. So differences in
opinion exist.
Composite Cross Method : Suneson, C. 1956.
Agron. J., 48:188-190.
The composite cross method can be called
“evolutionary plant breeding” because it uses
composite crosses and natural selection. The
rationale is survival ability. Survival ability is
sometimes detrimental to productivity in a pure stand
but has special significance for determining
agricultural fitness.
Develop Composite Cross II : crossed together 28
diverse cultivars of barley (378 crosses). Grown in
bulk from F2 to F29 without selection.
Periodically compare yielding ability of CCII
with ‘Atlas’ using remnant seeds from various
generations.
F3 to F4
F7to F8
F11 to F14
F15 to F20
F21 to F23
F24
% Atlas
67.6
85.0
88.0
106.0
113.8
135.7
The composite bulk was conspicuously inferior to ‘Atlas’ in
yielding ability and general agronomic appearance in early
generations. There was a gradual improvement, however, in
both yielding ability and agronomic type until, by F15 the bulk
equaled ‘Atlas’ and then continued to improve. So natural
selection did cause significant increased in yield.
After 15 generations :
1)
Continued natural selections for significant gain in
yield.
2)
Cyclic hybrid recombination alternated with natural
selection.
3)
Conventional selection and testing to identify best
lines.
Evolutionary plant breeding in barley had produced numerous
cultivars. It can be thought of as a “stretched” bulk method,
resting on the dynamics of natural selection. It is very unique
in the length of its development period. It is theoretically and
practically sound, and its greatest value would be as an
adjunct breeding program. Because of the long wait before
selection can begin, it is not feasible as a single method
breeding program. The department and your farmers might
will not be very happy if it takes you over 20 years to develop
a cultivar!! Perhaps a compromise would be to treat all
crosses by the short-term bulk method, forming derived lines
in the F4 – F6 generations for the breeding project. If the
population is judged useful for the future, it could be
continued as an evolutionary breeding entity.
Pros
• Automatic increase in proportion of
homozygous plants with each generation
• May increase mean yield via natural seln.
• Final purification of lines simplified
• Selection among crosses can occur before
selection within crosses begins, thus the
elite crosses are the ones that remain in the
program
Cons
• Plants of one generation not all represented
by progeny the next generation
• Genotypic frequencies and genetic
variability cannot be defined readily
• Bulk method not suited to greenhouses, off
season nurseries
• Environment must be suitable to reinforce
breeding objectives
Early Generation Testing
• Objective: identify those populations that
are likely to contain superior lines
• Strategy: eliminate those populations of
low potential from the inbreeding process
• Goal: maintain and develop lines from
populations with high genetic potential
Jenkins, 1935
• Usual method of estimating combining
ability in maize was to inbreed lines, then
mate them to a common tester
• Jenkins saved seed from S0:1 lines through
many selfing generations, then crossed them
to common tester
• Found that combining ability was already
determined in S0:1 lines
Self-Pollinated Crops
• Determine the generation for testing
• If it is to be the F2, you will have to grow
the F1 in an environment which favors seed
production
• A more common choice would be F2:3 lines
Self-Pollinated Crops
•
•
•
•
Harvest seed from individual F2 plants
Plant seeds in F2:3 progeny rows
Identify the superior rows
Harvest all seed in each selected row in
bulk
• Grow replicated tests of F2:4 lines
• Grow replicated tests of F2:5 lines
Self-Pollinated Crops
• Harvest selected F5 plants individually
• Grow F5:6 lines in headrows
• Test F5 - derived lines extensively
Breeder’s Decisions
• Generation to test
• Number of reps, locations and years - tradeoff
between early and late generation testing
• Separate program for inbreeding or not
• Selected lines can be advanced by pedigree, bulk,
or SSD
• Number of plants chosen from each hetergeneous
line may vary
Genetic Considerations
• Recall that there is all of the additive
variance among F2:3 lines and one-half of
the additive variance within F2:3 lines
• In later generations of F2 derived lines,
there is still all of the additive variance
among lines, and an increasing amount
within lines, as inbreeding progresses
Genetic Considerations
• Therefore, one may need to take a large
number of heads to adequately sample the
variation within the F2 - derived line
• Now one must decide how to allocate
resources
• Should you sample more lines or more
selections within lines?
Pros
• Inferior individuals and crosses are
discarded early in the process
• One hetergeneous line may yield more than
one cultivar
Cons
• When you commit a lot of resources to
early generation testing, you cannot devote
as much to thorough evaluation of more
inbred material
• If you spend a lot of time testing the early
generations, cultivar release may be delayed