CAB Reviews 2014 9, No. 043
Accelerated plant breeding
B.P. Forster1*, B.J. Till1, A.M.A. Ghanim1, H.O.A. Huynh1, H. Burstmayr2,3 and P.D.S. Caligari4,5
Address: 1 Plant Breeding and Genetics Laboratory, Joint FAO/IAEA Division, IAEA Laboratories, A-2444 Seibersdorf, Austria.
2
University of Natural Resources and Life Sciences, Vienna, Austria. 3 Department for Agrobiotechnology (IFA-Tulln),
Institute for Biotechnology in Plant Production, Konrad Lorenz Str. 20, A-3430 Tulln, Austria. 4 BioHybrids International Ltd,
P.O. Box 2411, Earley, Reading RG6 5FY, UK. 5 Instituto de Ciencias Biológicas, Universidad de Talca, 2 Norte 685, Talca, Chile.
*Correspondence: B.P. Forster. Email: [email protected]
Received:
Accepted:
22 April 2014
17 November 2014
doi: 10.1079/PAVSNNR20149043
The electronic version of this article is the definitive one. It is located here: http://www.cabi.org/cabreviews
g
CAB International 2015 (Online ISSN 1749-8848)
Abstract
The time to develop new cultivars and introduce them into cultivation is an issue of major
importance in plant breeding. This is because plant breeders have an urgent need to help provide
solutions to feed a growing world population, while in parallel, time savings are linked to profitability. Plant breeding processes may in general be broken down into the following five key
elements: (1) germplasm variation; (2) crossing; (3) generation of new genetic combinations;
(4) screening and selection (identification and subsequent fixation of desired allelic combinations);
and (5) line/cultivar development. Each of these has implications in relation to the time taken to
breed a new cultivar; a brief introduction is given for each to highlight the obstacles that may be
targeted in accelerating the breeding process. Specific techniques that provide a time advantage
for these elements are then discussed. Some targets for enhancing the efficiency of plant breeding,
e.g., the manipulation of meiotic recombination, have proven to be recalcitrant. However, other
methods that create new genetic variation along with improvements in selection efficiency compensate to a large extent for this limitation. Progress in accelerating the plant breeding process
continues by exploiting new emerging ideas in science and technology.
Keywords: Accelerated breeding, Mutation breeding, GM breeding, Rapid generation cycling, Doubled
haploidy, Marker-assisted selection, High-throughput screening.
Review Methodology: Searched various websites including the http://worldmeters.info/worldpopulation http://www.census.gov/
popclock; FAO/IAEA database of mutant varieties, http://mvgs.iaea.org. FAO Yearbooks and AGROSTAT, BBC News. Used the
following key words in searches: world population, global forecasts, accelerated breeding. Discussed contents with colleagues.
Introduction – Issues and Drivers of Faster
Breeding Methods
The evolution of crop plants can be described as having
three phases: (1) gathering from the wild; (2) domestication and agronomy and (3) plant breeding [1, 2]. The conversion of wild plants to crop plants involves the selection
of suitable types, which developed in conjunction with
suitable agronomic practices. Domestication and associated agronomy therefore played a significant and intimate
role in mankind’s early food security and thereby human
evolution. However, this phase has been eclipsed by subsequent achievements in plant breeding. Throughout its
history, plant breeding has been a test bed for scientific
innovations, especially those in genetics, botany, physiology and biotechnology, thus plant breeding has capitalized
on research in recombination, heritability, polyploidy,
chromosome engineering, tissue culture, heterosis, genetic linkage mapping, molecular genetics, mutagenesis and
transformation. Plant breeding has been a major factor in
increasing crop production [3–5]. Research and development achievements have added much to the toolbox,
but a major constraint and frustration that remains is the
time to generate de novo, a new cultivar for growers.
Plants are grown for a wide range of uses, chiefly for
food, feed, fibre and fuel, but also for specialized products
http://www.cabi.org/cabreviews
2
CAB Reviews
(e.g., medicines), and social amenity such as recreation
and ornamentation. The global human population is expected to grow from current estimates of 7.1–7.2 billion
(http://worldmeters.info/worldpopulation
http://www.
census.gov/popclock) to 9.1 billion in 2040 [6]. Plant
breeders face a daunting [5] task as crop production is not
keeping pace with demands; crop yields are currently
increasing at about 1.3% per annum, about half the rate that
is required [7]. The area of land under cultivation is not
expected to increase; it has remained static over the past
50 years at about 660 million hectares (data from FAO
Yearbooks), and there is even concern that agricultural
land is degrading. Future increases in crop production are
therefore dependent upon greater yields per unit area of
land area and therefore crops with high yield potentials
must be developed and developed quickly. At the same
time, climate change is providing a major uncertainty as to
how it will limit or change agricultural productivity. For
example, in temperate crops significant yield reductions
are expected as temperatures rise by more than 4 C; while
similarly, in tropical regions a rise of 2 C will cause major
yield losses [8]. Temperature is only one aspect of climate
change. Other direct effects include salinity, drought,
waterlogging and storms, all of which bring with them new
pest and disease problems. Some of the effects of climate
change may be mitigated by the introduction of cultivars
adapted to the new conditions, or growing new crops
species, both of which solutions require rapid breeding.
This review does not intend to go into details of plant
breeding methods or comparisons between them. There
are several excellent reference books that discuss theory
and methods in plant breeding in a wide range crops (e.g.,
[9–14]). Here we provide a general overview of techniques that can accelerate the plant breeding process, from
parental choice to cultivar release.
Components of Plant Breeding
Plant breeding processes may be broken down into the
following elements: (1) access to germplasm or creation
of genetic variation, (2) crossing; (3) generation of
new genetic combinations (normally through meiosis);
(4) generation of segregating populations, (5) selection
(fixation of desired alleles); and (6) line development. Each
of these components introduces a time component
and any, or all, of which can be a bottleneck in the plant
breeding process. The relevance and limitation of each
component are described while specific techniques that
overcome these limitations are then discussed.
The terms ‘traditional plant breeding’ and ‘accelerated
plant breeding’ are difficult to define as continual improvements have taken place in deliberate plant breeding
since Mendel first established the genetic principles
involved, principles that provided a scientific basis for
plant breeding. Therefore, to provide some perspective,
we take for our example of ‘traditional plant breeding’ the
pedigree inbreeding method (PIM) as described by Briggs
and Knowles in 1967 in their book: ‘Introduction to plant
breeding’ [9]. This method has, and continues, to serve
plant breeders well. Modern time-saving methods are
then described and may be compared with PIM. This
means that we will tend to concentrate our focus and
examples on inbreeding species, but in reality the principles apply to species with other breeding systems and
the reader can easily apply the possibilities for the
methodologies to those species
Homozygous lines, commonly referred to as genetically
pure lines, or breeding pure lines, are important goals in
many plant breeding programmes and may be developed
through PIM. Pure lines not only form the end product for
self-pollinating species of the breeding programmes (lines
that may be developed into cultivars in many crops,
notably cereals), but are exploited as parental lines in F1
hybrid cultivar production (notably in maize and many
vegetable crops). New homozygous lines are typically developed from crosses between two distinct parental lines
that are complementary for desired traits. This provides
an opportunity for re-assortment and recombination of
genes and alleles through meiosis and thereby the production of segregating populations in subsequent generations, as first described in the experiments of Mendel
(1822–1864 [15]). In crosses between two pure lines,
segregation is observed first in the F2 generation as the F1
individuals are presumed to be genetically identical but
heterozygous for all loci by which the parents differ. The
plant breeder attempts to make improvements either
negatively, by rogueing out inferior plants, or positively, by
selecting the best, starting with the segregating F2 population. The next objective is to develop pure lines from
the selections and in self-pollinating species this is traditionally done by repeated rounds of selfing in the PIM
(Figure 1). Seed from individual selected plants are sown
out as F3 family rows; this is repeated in the F4 and F5.
From F5 to F6 there is a shift from single plant selection to
single family selection. Seed of an individual family are
grown out in small plots in the F6 bulk harvested, and then
grown in larger replicated plots in F7––F10 generations.
At about the F7 stage the term ‘line’ replaces the term
‘family’ as there should be no visible variation at this stage
(plants within a line look the same). Multi-locational trials
often take place from F8 to F10. Larger strip tests take
place in F11 and F12 generations to bulk seed for potential
commercial release, since at this stage the selected lines
must enter national, regulatory trials to achieve official
cultivar status. This normally includes passing DUS tests,
i.e., the line must be distinct, uniform and stable, in
addition it must normally give superior yields compared
with standard checks or possess other important traits
(such as disease resistance, superior quality, etc.). If the
line passes ‘is accepted’ it may be officially released and
grown by farmers. Thus, for an annual self-pollinating
crop, it can take between 10 and 15 years to produce a
cultivar from an initial cross (Figure 1).
http://www.cabi.org/cabreviews
B.P. Forster et al.
Initial cross
Year
F1
Progeny from cross: first generation.
1
F2
The second generation is normally the first
opportunity to observe segregation and to select
individual plants.
2
F3
Single plant selections.
3
F4
Single plant selections, often includes check
cultivars.
4
F5
Single plant selections. Family rows should begin
to appear uniform, if so rows may be selected as
lines.
5
F6
Small plot evaluation and selection. First
opportunity to estimate yield. Plots are bulked
harvested.
6
F7
Large replicated plots; data on yield, quality and
agronomic performance.
7
F8 to F10
Three to 5 years of multi-locational trials.
8–10
F11 to F12
Seed multiplication and official testing and
certification for cultivar release.
11–15
P1 × P2
3
Figure 1 Scheme for PIM, redrawn and amended from Briggs and Knowles [9].
Genetic Variation
Plant breeding is based, ipso facto, on genetic variation.
Plant breeders aim to harness genetic variation to improve crop plants and the speed at which they can succeed depends on whether the required variation exists
(and if so in which gene pool) or needs to be created.
The latter is not necessarily a disadvantage. The simplest
scenario is where the desired genes are present in the
primary gene pool, i.e., available in elite breeding lines.
These can be used in crosses with other elite lines and
the breeding can focus on the trait of interest as there is
minimal disruption of the elite genetic background. If,
however, the desired variation is only available in secondary, tertiary or wild species germplasm then increasing
efforts, such as repeated rounds of backcrossing, are required to recover an elite genetic background which
expresses the trait of interest. Normally, the more distant
the genetic resource used the greater the delay in developing a cultivar. Plant mutation induction and transformation can create novel variation directly in a favoured
genotype and selected lines can be developed into
cultivars relatively quickly. Plant mutation breeding is a
well-established and rapid breeding strategy; generally
taking 7–9 years to produce a cultivar in a self-pollinating
crop (see section on Mutation Breeding). Transformation
to produced genetically modified cultivars (GM crops)
may take a similar amount of time, the rate being dependent upon the efficiency of selection and screening and
cleaning up the genetic background (see section on
Selection and Screening).
Crossing
Crossing or hybridization is a basic tool in plant breeding.
This brings together genes of two or more parental
lines from which, it is hoped and expected, improved
plants will be produced. Crossing involves the application
of pollen from the male parent to the stigmas of the
female parent. It may involve emasculation to prevent
self-pollination and isolation of the female flowers to
prevent uncontrolled pollination. Pollen is collected from
the male parent (and in some species may be stored) and
placed on receptive stigmas; in some case male sterile
lines can be used. Hybridization is not normally a timelimiting step, but may be restricted by incompatibility
problems and the synchronous production of receptive
stigmas and mature pollen, which in turn may be season
dependent. In many countries, particularly developing
countries, crossing can only be done during clement
conditions during seasonal crop production times and this
http://www.cabi.org/cabreviews
4
CAB Reviews
may be restricted to one opportunity per year, weather
permitting. Investment in controlled environment facilities, such as a greenhouse with lighting and temperature
controls allows for additional crossing opportunities
per year.
Plants exhibit doubled fertilization of the egg (which
develops into an embryo) and of the central cell (which
develops into an endosperm). In wide crossing it is
common for the endosperm to fail, necessitating the
rescue and in vitro culture of the hybrid embryo. Embryo
culture can also speed up the time to growing the next
generation (see section on Rapid Generation Cycling). In
extreme cases, the hybrid embryo loses chromosomes
of the male parent resulting in haploid embryos, these
too need to be rescued and cultured and are valuable
in the development of doubled haploids (DHs) which can
shortcut the breeding process (see section on Doubled
Haploidy).
Generation of New Genetic Combinations
Meiosis provides a natural mechanism for re-assortment
and recombination of genetic information [16]. Breeders
have a particular interest in recombinant events that
improve the phenotype, but these are often rare. Large
populations, or repeat crossing, may be necessary to
achieve the desired recombinant. In wide crossing and
especially in inter-species crossing a large piece of an alien
chromosome may be introduced. Although this may carry
a desired gene for a target trait it is likely to be linked
to many other alien genes with negative effects. To
complicate matters further the alien chromosomal segment is often meiotically inert, i.e., it does not recombine
and is always linked to flanking genes leading to linkage
drag. There is therefore interest in manipulating recombination rates, particularly in specific areas of the genome.
In polyploid species such as wheat there are genes
that control chromosome pairing, notably Ph1, which
ensure strict pairing of homologues at meiosis, when the
gene is absent, or recessive, pairing can take place
between non-homologous chromosomes and can break
up alien segments [17]. The processes of meiosis can also
be exploited in the development of chromosome translocation lines, especially between crop and alien chromosomes, thus speeding up introgression (see section on
Induced Recombination), but in general there are not
many examples of stimulating meiosis to increase
the efficiency of plant breeding within a species, and this
remains a bottleneck in plant breeding, an area for future
developments.
Screening and Selection
Screening and selection are the mechanisms by which
breeders fix desired genes and gene combinations.
Screening methods are performed either by phenotyping
or genotyping test materials. Selections are made based
on analysis of this data based on pre-determined criteria.
Phenotyping
Screening by phenotyping has been the standard technique in plant breeding to select improved crops since
plants were first domesticated. In its basic form it entails
growing up plants/populations and seeing how they perform: are they high/low yielding, tall/short, early/late,
disease susceptible/resistant, drought sensitive/tolerance,
is the seed round/wrinkled, etc.? Tests may be performed
in simple growth chambers and greenhouse, or sophisticated phenomics facilities, or most commonly in the field.
The ability to phenotype effectively and efficiently is a
major concern in plant breeding and high-throughput
methods need to be developed [18–20]. The best phenotypes are selected and advance to the next generation. In
the PIM (Figure 1) lines are selected from segregating
families 5–6 years after the initial cross. In terms of time
and cost savings, the sooner the screening the better,
both between generations (early generations are preferred) and within generations (seedling screens are
better than adult plant screens). Phenotypic screening has
a disadvantage in that it may be influenced by environmental factors, this is especially true when screening for
disease resistance. In such cases genotypic screening may
be more accurate. However, at some point, regardless of
the screening method, selected lines must be tested and
advanced in the field.
Genotyping
The majority of the variation utilized by plant breeders
comes in the form of heritable DNA changes that can be
monitored using molecular techniques. A wide range of
techniques now exist. The use of molecular markers to
assay for genetic variation in linkage disequilibrium with
desired traits, known as marker assisted selection (MAS)
is now a common practice [21]. In developing markers for
traits the marker must be associated with the trait and
thus is dependent upon good phenotypic data [22]. The
advent of array hybridization methods and more recently
next generation sequencing have allowed high-throughput
approaches to measure inheritance of genetic variation
with tens of thousands of single nucleotide polymorphisms recoverable in a single assay. In addition to speeding
up traditional selection and removing potentially interfering environmental factors, advanced methods speed up
the process of genetic mapping, cloning and have provided
new approaches to breeding in the form of genomic
selection, e.g., MutMap [23–25]. Advances in tools for the
discovery of nucleotide variation have also enabled the
development of reverse genetic strategies. Reversegenetics allows the identification of genetic lesions in
http://www.cabi.org/cabreviews
B.P. Forster et al.
targeted gene regions hypothesized to be important for
desired traits. One such example known as targeting
induced local lesions in genomes utilizes induced point
mutations to alter gene function [26] for which protocols
have been developed [27–30]. Current issues in genotyping include the cost and ease of DNA extraction and
the large sample sizes that require analysis (typically 700
lines).
Line Development
A ‘line’ is defined as a group of genetically related siblings
which no longer exhibits trait segregation (which is a
characteristic of a family). Once selected, the line is
usually subject to field performance trials that increase
in scale as the line progresses (and as more inferior
lines dropped). Line development in the PIM is given in
Figure 1, and may require 10 years of development before
entry into national official DUS tests and cultivar release.
Line development is costly in time, as well as in space and
official compliance. Breeders must focus their efforts on
lines with a high potential of success and to advance these
as fast as possible.
As can be judged from above that while there is a need
to develop new cultivars as quickly as possible into production, there are numerous stages in the process that
need to work as efficiently as possible and be linked into
an integrated programme of breeding which is time efficient. Below we discuss some of the possibilities to
accelerate the steps in such programmes.
Techniques in Accelerated Breeding
Plant Mutation Breeding
Plant mutation breeding was heralded as providing a universal solution to breeding problems but subsequently
became more restricted to being used in rather specific
circumstances, particularly easily obtained characteristics,
such as dwarf growth that could be readily selected
visually and thus in large numbers. However, it has gained
more recent popularity and indeed provides a successful
and relatively fast breeding method. The Joint FAO/IAEA
Division of Nuclear Techniques in Food and Agriculture
lists over 3000 officially released mutant cultivars in
over 200 crop species ([31]: http://mvgs.iaea.org). Over
60% of these mutant cultivars have been derived from
gamma irradiation. Typically, mutation breeding takes
7–9 years as compared with 10–15 years in a standard
PIM (Figure 1), for an annual crop. This is because the
main objective in plant mutation breeding is to take
an already favoured cultivar and to improve it through
induced mutation with minimal disruption of the elite
genetic background (Figure 2). Thus the breeding effort to
achieve cultivar status, i.e., the development of lines that
5
are distinct (carry new mutant trait), uniform and stable, is
reduced. A major bottleneck in plant mutation breeding
is the ability to screen for the desired mutant trait as
this is a rare event, occurring in 1 in 1000–1 in 100 000
individuals. Traditionally, mutant induction has deployed a
random approach (using physical or chemical mutagens)
and screening has been done at the phenotypic level
by observations of plant traits, such as yield, height,
flowering time, colour, disease resistance, tolerance to
salt and drought etc. [32]. Recently, high-throughput
phenotyping has been facilitated by the deployment
of phenomics facilities [33]. Genotypic mutant screening
can also be deployed and this comes into play where
the target gene is known [34]; Szarejko review [35, 36].
Methods and trends in plant mutation breeding have been
reviewed recently by [37]; classic examples include the
development of semi-dwarf barley in Europe, salt tolerant
rice in Vietnam, self-compatibility in fruit trees, seedless
fruit in citrus.
A recent example is the development of mutant wheat
lines resistant to Ug99, a race of stem rust disease (Puccinia graminis f. sp. tritici) first identified in Uganda in 1999.
Historically this is the most feared disease of wheat as it
causes severe yield losses. It is a highly mobile disease
as its spores are carried by the wind over large distances.
The disease has now spread to neighbouring countries
in Africa: Kenya (2002), Ethiopia (2003), Sudan (2006)
and South Africa (2009). It has also reached Iran (2007),
Yemen (2007) and Iran (2009) and is set to spread further.
Ug99 is of particular concern as there is little resistance in
wheat cultivars, in 2010 it was estimated that 80–90% of
all global wheat cultivars were susceptible [38]. In 2009
Kenyan wheat cultivars were subject to gamma irradiation
in an attempt to induce mutations for disease resistance.
The first mutant generation (M1 population) was grown at
Eldoret, Kenya (a hot spot for the disease). Resistant
mutant lines were selected in subsequent generations in
both field and greenhouse conditions. Since two generations of wheat can be produced per year in Kenya with
the help of irrigation, it was possible to accelerate the
development of resistant mutant lines. Four promising
mutant lines selected in the M4 population were advanced
and multiplied in 2010 and entered into National Performance Trials in Kenya in 2012 and 2013. In 2014, one of
these, named ‘Eldo Ngano I’ was officially released as a
cultivar for farmers, this was achieved within 5 years of
the initial mutation induction.
Plant mutation breeding using random mutation induction is considered to be a conventional breeding method
and is non-regulated.
Induced Recombination
Physical, chemical and biological mutagens can be used
to enhance the efficiency of homologous chromosome
recombination and induce chromosome translocations in
http://www.cabi.org/cabreviews
6
CAB Reviews
Selected elite genotype, an example of a crop plant
with seven chromosomes (M0 generation)
After mutagenic treatment (e.g., gamma or X-ray
irradiation) the population becomes the first
mutant generation, i.e., the M1 generation. This is
made up of individuals with random mutated loci
(an example of just four individuals from a
population of normally several hundred
individuals).
The M2 generation is produced by selfing the M1
population (four examples of individual M2 plants
are given). Mutant events begin to become
homozygous and thus recessive (most) mutant
genes are unmasked and mutant traits observed.
This is the first opportunity to select for desired
mutants. Screening and selection can be carried out
in subsequent M3 family rows which provide
information on segregation within the family.
Selection of (3) different mutant genotypes
(individuals or lines) with a desired mutation
(circled). The aim is to screen for plants or families
with the desired mutation, but with the least
distribution of the elite genetic background, i.e.,
low mutation load. Mutant lines may be developed
directly as cultivars (normally M7/8 generation) or
used in cross-breeding.
Figure 2 Simplified scheme for plant mutation breeding using random mutagenesis in an inbreeding species.
plants [16, 39]. The mode of action is primarily through
double stranded breaks that are then subject to repair
mechanisms. These can be exploited to speed up the introgression of same species of alien genes for crop improvement. Treatments have been applied to both somatic
(mitotic) and gametic (meiotic) cells. The method has
been used most successfully in wheat, notably in the transfer of disease resistance genes from alien species [40].
A classic example is the transfer of rust resistance into
bread wheat, Triticum aestivum from the wild, related
species Triticum umbellulatum by ER Sears in 1956 [41]. In
this work wheat aneuploid plants carrying an additional
T. umbellulatum chromosome were irradiated with
X-rays at the onset of meiosis. The irradiation stimulated
double-strand breaks and the repair mechanism produced
non-homologous chromosome recombination. Unfortunately, the application of physical mutagenic treatments
to pre-meiotic plants or plant parts is difficult due to
sample space issues in conventional closed irradiators.
It is possible to irradiate whole plants in specialized
field or glasshouse facilities, where plants are arranged
around a central mutagen such as a gamma or X-ray
source, which is activated for a period of time to irradiate
the sample, but such facilities, are limited. An alternative
is to use detached floral parts (such as wheat spikes)
or short plants and placed these inside a sample chamber
in a closed irradiator, after treatment these need to
be maintained at least until mature pollen is available (for
crossing) or, preferably, embryos are sufficiently developed for rescue and culture.
Mitotic intra-chromosomal recombination can be
induced by stress treatments that cause double-strand
breaks in DNA, such as physical and chemical mutagens
and heat stress [42, 43]. Lebel et al. found that low dose
X-ray treatments could double the frequency of spontaneous somatic recombination whereas treatment with
the chemical mutagen, mitomycin C increased the rate
nine times in tobacco. Induced mitotic recombination
events may therefore create new genetic variability that
can be harnessed by plant breeders. Induced recombination has obvious benefits as it allows recombination
of chromosomes that would rarely recombine in nature
(normal meiosis). This can be exploited in transferring
desired genes into breeding materials and in breaking up
linkage blocks where desired genes may be freed from
linked to deleterious genes (linkage drag).
Despite its long history, induced recombination
remains a largely academic practice as further developments are needed for practical, large scale application in
accelerated plant breeding.
In addition to enhancing recombination, supressing
meiotic cross-overs is also of interest to plant breeders in
http://www.cabi.org/cabreviews
B.P. Forster et al.
an approach called ‘reverse breeding’ (reviewed by [44]).
In this approach high performing heterozygous genotypes
are selected and their meiosis is genetically engineered
to reduce or eliminate recombination events. The result
of this is that male and female spores carry nonrecombinant parental haploid genomes, which may then
be manipulated to produce DHs (see section below
on Doubled Haploidy) and thus generate parental lines
for F1 hybrid breeding that reconstitute desired heterozygous genotypes.
GM Breeding
GM is normally defined as an alteration of the genotype by
the insertion or alteration of a specific DNA sequence
using ‘recombinant DNA technologies’ involving artificial
delivery systems. Early GM technology focused on the
insertion of DNA from a foreign species, but there has
been a trend away from transgenics (foreign DNA insertion) to cisgenics (same species DNA insertion) and most
recently to targeted mutagenesis (genome editing) of a
favoured genotype. In addition, methods in GM technology are no longer confined to tissue culture methods (and
thereby tissue culture responsive genotypes), thus opening the way for greater application [45]. Like mutation
breeding, the starting point for GM breeding is typically a
successful genotype, which can be improved for a specific
trait (see Figure 2), but unlike random mutation induction
targets a specific gene. Farmers and end-users generally
favour cultivars/products they are accustom to as these
have tailored agronomic, harvesting, shipping, processing
and marketing procedures. New, improved favoured
cultivars have commercial advantages, especially if the
breeder has ownership of the target genotype. The more
sophisticated the GM technology used the less time it will
take to clean up the genetic background, i.e., reduce the
number of backcross generations; the ideal scenario being
targeted gene editing with no alteration of the genetic
background (no off-target effects). Thus GM breeding has
the potential to be very fast. A major disadvantage of GM
breeding is that the target gene must be known and
sequenced. Currently this is a bottleneck, for example in
2012 only 702 functionally characterized genes were
annotated in rice (a well-researched crop), which represents < 2% of the predicted loci [46]. Until our knowledge of genes improves, random mutagenesis remains a
viable option. Other disadvantages of GM breeding are: it
requires specialized laboratories and is expensive, though
cheaper, easier options are being developed. Examples of
new plant breeding technologies (NPBTs) are zinc finger
nucleases, TAL effector nucleases and clustered regularly
interspaced short palindromic repeats technology [45].
NPBTs have yet to be proven, but have the potential to
produce GM cultivars that cannot be distinguished from
those produced by conventional plant breeding, and this
has sparked a debate about whether or not they should be
7
classed as GM and whether they should be subject to
genetically modified organism (GMO) legislation. These
issues will have a significant effect on the uptake of NPBTs.
Classic examples of first generation of GM crops
include: slow ripening tomato, herbicide resistant soya,
insect resistant maize and pest resistant cotton. Future
targets are wide ranging and in addition to include standard plant breeding goals of yield, quality, nutrition, pest
and disease resistance and abiotic stress tolerance,
also include novel ventures such as the production of
pharmaceuticals, biodegradable plastics, pigments and
cosmetics. It also needs to be recognized that there are
now huge hectarages of GM crops growing in a wide
range of countries [47].
Issues in GM breeding, methods, benefits, safety,
regulations and public concerns, have been reviewed by
Halford and Shewry [48].
Rapid Generation Cycling
The production of new generations is a fundamental
component of plant breeding as this allows another round
of meiosis from which new recombinants can be produced. The time taken to obtain new segregating materials
is often a major time constraint. Classic forms of speeding
up generation times are single seed descent (SSD) and
shuttle breeding. SSD is an old technique first proposed
by Goulden [49], later modified by [50]. In this method
only one seed from a population of F2 plants is grown on
to F3; this process is repeated in the next generations up
to F5/6 when plants approach a high level of homozygosity.
Since only one seed is required per plant in early generations these may be grown in small pots or densely in
seed trays, which facilitate early flowering and can be
grown in a small area. Greenhouses and off-season nurseries may be used to produce more than one generation
in a year. Disadvantages of SSD are that it does not allow
selection in early generations (although this means that
more rounds of meiosis are undertaken before selection
is applied which can be an advantage) and it is important
that there is very low mortality/plant loss in each generation or the risk of unconscious disadvantageous selection becomes real. Shuttle breeding was developed by
the famous plant breeder Norman Borlaug who won the
1970 Nobel Peace Prize for his involvement in the Green
Revolution [51]. His idea was to speed up the breeding
process by growing successive generation in the same
year at different sites. His original work involved growing
wheat during the summer at high altitude sites in
Chapingo and Toluca, Mexico and then almost 2000 km
north at Obregon, Mexico at sea level with irrigated
conditions, and then back to the high-altitude sites. This
produced two generations a year and thus cut the
breeding time by half. Shuttle breeding can cross greater
distances, such as UK$New Zealand, but then involves
certain regulations such as phytosanitary certificates and
http://www.cabi.org/cabreviews
8
CAB Reviews
Repeated
generation
cycles
(6–8/year)
Wheat
Sorghum
Barley
Embryo rescue and transplant
of emerging plants into small
pots
Planting in small pots, light watering and continuous
lighting
Figure 3 Scheme for rapid generation cycling using wheat, sorghum and barley as examples, adapted from Zeng et al.
[52]; Ghanim et al. [53]. In the above scheme plants (wheat, sorghum and barley) are grown in small pots in controlled
environmental conditions (greenhouse). The aim is to limit vegetative growth and promote early flowering, but to provide
sufficient seed for the next generation. Tiller production may be completely prevented in small pots, allowing growth of the
main stem only. However, if preferred tiller production may be encouraged by growing in larger pots, this provides stronger
plants and additional/spare spikes for crossing and more embryos, but increases the time to flowering. The time to flowering
may also be promoted by growing under continuous lighting, limited watering and raised temperatures. The milk-ripe stage
of endosperm development (15–20 days after fertilization) provides embryos that are easily detached from their endosperm
in small grain cereals, and are sufficiently developed to germinate immediately (1–2 days) in culture on simple media
containing basic nutrients (such as half-strength Murashige and Skoog (M&S) medium, [54]) and sugar [52, 53]. In vitro
seedlings may be sampled for DNA analysis in MAS. In cases of winter genotypes, vernalization treatments may be applied
to in vitro seedlings.
material transfer agreements. Since shuttle breeding
exposes breeding lines to more than one environment,
which may include a range of biotic and abiotic stresses,
selection may be made for a wide range of traits, but this
may also be hazardous as the breeding lines may succumb
to a new stress.
Recently, a procedure was described that dramatically
shortens the generation times of wheat and barley,
allowing the production of eight generations of wheat and
nine generations of barley per year [52]. The procedure
combines growing plants in greenhouse conditions, in
small pots with limited watering and high light exposure.
These combine to speed up development. After flowering/crossing, embryos are excised and cultured, thus
circumventing the need to grow seeds to maturity and
avoiding any seed dormancy, thus saving several weeks.
The methods are simple and have been extended to a
wider range of cultivars and other crop species (Figure 3).
Table 1 shows examples of the time savings between
generations; these are for spring genotypes that do not
require vernalization treatments. In the scheme presented
in Figure 3, the culture of embryos that germinate readily
into seedlings can act as a convenient stage to apply vernalization (cold) treatments. The procedure has recently
been applied to sorghum with up to six generations being
produced in one year using small pots, continuous lighting,
limited watering and embryo rescue [53]. The in vitro stage
is also ideal for taking tissue samples, e.g., leaf samples for
DNA extraction from which genotypic assays can be
performed (see section on Marker Assisted Selection).
Some in vitro phenotypic screen may also be carried out
such as salt tolerance and osmotic (drought) stress. Thus
only preferred individuals are advanced to the next generation cycle.
High-throughput Phenotyping
The ability to phenotype has lagged behind highthroughput methods in genotyping, so much so that it
has led to the problem of bridging the ‘phenotypic gap’.
One response to this has been the emergence of large
scale phenomics facilities which involve growing potted
plants in greenhouse with carefully regulated environments, These facilities are highly automated and equipped
with sophisticated data-capturing devices linked to largescale computing and bioinformatics capacities, e.g.,
Australian Plant Phenomics Facility and European Plant
Phenotyping Network. Phenomics facilities can measure
and analyse phenotypic data at all stages in development.
However, costs are high and since the growing environment is artificial the trait data must be correlated with
performance in the field, and this is notoriously difficult.
A more pragmatic approach has been the development of
low cost high-throughput field phenotyping systems [20].
Araus and Cairns [20] have reviewed recent advances
in field-based high-throughput phenotyping which include:
(1) field sensing and imaging; and (2) field sampling for
http://www.cabi.org/cabreviews
B.P. Forster et al.
9
Table 1 Normal and accelerated generation times with respective number of generations per year for wheat, barley and
sorghum genotypes (data taken from [53])
Crop
Normal generation
time (month)
Normal number of
generations per year1
Accelerated generation
time (day)2
Generations
per year
Wheat
Barley
Sorghum
4–5
4–5
5–6
1–2
1–2
1–2
45
40
60
7
8
6
1
The number of generations is often limited to one per year (during the cropping season), in a few environments two generations may be
grown per year in the field.
2
These are averages taken from 5 to 10 genotypes tested for each crop species with the aid of embryo culture.
Table 2 Methods in DH production in crop plants
Method
Basics of methods
Usage and crop examples
References
Androgenesis
Culture of male haploid gametic
cells using anthers or isolated
microspores
Maluszynski et al. [58];
Touraev et al. [60]
Gynogenesis
Culture of female haploid gametic
cells using flowers, pistils, ovaries
or ovules
Production of haploid embryos after
pollination with another species or
genus (wide crossing) or treated
pollen. Involves embryo culture
Widely used: apple, asparagus,
aspen, brassicas, bread wheat,
barley, broccoli, citrus, durum wheat,
flax, maize, oak, potato, rapeseed,
rice, rye, ryegrass, swede, timophy,
tobacco, triticale
Specific use: sugar beet, onion
Specialized
crossing
Spontaneous
The recovery of naturally occurring
haploids
Wide crossing is commonly used in
many species: barley, bread wheat,
durum wheat, oat, triticale. Pollination
with treated pollen, e.g., irradiated
pollen is more specialized: pumpkin
Specific use: oil palm
laboratory-based analyses. A method used in both is nearinfrared reflectance spectroscopy (NIRS). NIRS technology
is rapidly advancing and aeronautical and portable
devices are available for field screening. NIRS is also finding
application in grain analysis where it is normally applied to a
few grams of seed per samples [55]. However, single
seed NIRS methods are being developed and this would
allow for early generation screening and accelerate the
breeding process. The new generation of high-throughput
phenotyping screens for plant breeding are fast, cheap,
non-destructive and accumulate large data sets; whereas
in high-throughput genotyping the data analysis requires
good computing and bioinformatics support. A disadvantage of field screening is that it is season dependant.
For some traits simple high-throughput phenotypic
protocols are well established, such as tests for salt tolerance and some disease resistance traits (e.g., [56, 57]).
These are generally performed in greenhouse conditions
and on several hundred plants, usually seedlings and may
or may not be season dependent. Tests are cheap, simple
and quick and correlated to field performance.
Doubled Haploidy
Doubled haploidy is of interest to plant breeders as
it is the fastest means of producing homozygous lines.
Homozygous lines are the final outcomes of breeding in
Bohanec [61]
Maluszynski et al. [58];
Touraev et al. [60];
Košmrlj et al. [62]
Dunwell et al. [63];
Nasution et al. [64]
many crops (e.g., rice, bread and durum wheat and barley)
and satisfy two of the three DUS criteria for cultivar
status, i.e., uniformity and stability. Doubled haploidy
increased the efficiency of selection, especially for recessive traits from F1 or for mutant traits (which are generally recessive) from M1 plants. DHs may also be used as
parental lines in the production of F1 cultivars, e.g., maize,
pepper and rye. Doubled haploid techniques have now
been applied to over 200 plant species and have become a
standard tool in accelerating the breeding of a wide range
of crops [58–60]. DHs are produced from haploid tissues,
organs and plants which may undergo spontaneous or
artificial chromosome doubling. There are four general
methods of haploid/DH production in crop plants; these
are listed in Table 2.
Figure 4 shows a scheme for the production of DH
cultivars from an initial cross in an annual species; this
takes 5 years, about a third of the time for classic pedigree
inbreeding (Figure 1). Furthermore, the DH breeding
system does not suffer from biasing effects of dominance
and competition between individual genotypes in early
generations, thus selections can be made early, which has
particular advantages for quantitatively controlled traits
such as yield. A feature and potential limitation of DH
breeding is that recombination is confined to meiosis in
the F1 generation. This will help preserve linkage
blocks. However, if there is a need to break linkages it
will be necessary to provide greater opportunities for
http://www.cabi.org/cabreviews
10
CAB Reviews
Initial cross
Year
F1
The F1 is grown up and provides material for doubled haploid
production, though this may be done at any generation.
0.5
DH1
The DH1 generation is grown up and seed recovered from
selected lines in glasshouse conditions.
1
DH2
Observation of rows or small plots in the field, selection for
agronomic performance.
2
DH3
Large replicated plots; data on yield, quality and agronomic
performance. Seed production.
3
DH4
Multi-locational trials, data on yield, quality and agronomic
performance. Identify best lines.
4
DH5
Seed multiplication and official testing and certification for
cultivar release.
5
P1 × P2
Figure 4 Scheme for plant selection and rapid advancement using doubled haploidy.
recombination. This may be done by increasing the size of
the DH population or reverting to the SSD method which
has the potential to generate new recombinants at
each generation until homozygosity is reached. Comparisons of the DH, SSD and PIM breeding methods have
been made by several workers with respect to speed,
recombination frequency, population size, practicalities,
inputs and costs (list of references has been compiled
[59]). Some negative features of DH breeding that need
to be considered are: (1) variation is generated at the
beginning of the process and the breeder must therefore
choose suitable parental lines that will generate the
desired variation (these may be generated from a PIM
scheme); (2) not all genotypes are responsive to DH
production methods. Low-cost molecular methods can be
applied to diversity allele selection in newly created DH
lines [65].
DH can also be used in accelerating backcross conversion (BCC). The aim in BCC is to introgress a desired
gene into a favoured cultivar/genotype (see also Mutation
Breeding). The most successful BCC schemes involve
genotyping for both the introduced gene of interest and
the elite genetic background integrated MAS, thus welldefined genetic marker maps and marker–trait associations must be known or readily obtained. A standard BCC
scheme is described in Figure 5.
Alternatives to BCC are mutation and GM breeding
(see Plant Mutation Breeding and GM Breeding sections
above).
Marker Assisted Selection
Plant breeding has been and continues to be based on
phenotypic selection. This is generally done in the field
where single plants, rows and plots are subject to the
‘breeder’s eye’. Individual plants, families and lines are
selected that appear to perform better than control
checks. Target traits have typically focused on yield,
quality and agronomic fitness, but also pest and disease
resistance and tolerance to drought and salinity where
fields are subject to these environmental influences,
or where glasshouse screens are available. This is a slow
processes often confined to seasonal growing periods
and, for many traits with a strong environmental component, often unreliable. MAS offers solutions. Biochemical markers (proteins and isozymes) were among the first
markers used in genetic studies and plant breeding, these
were both practical and cheap, e.g., endopeptidase marker
for eyespot disease in wheat [66] and in some cases monitored the trait directly, e.g., storage proteins involved in
bread-making quality [67]. However, biochemical markers
were limited and have been eclipsed by DNA markers.
Over the past decades the most widely used markers
in plant breeding have been simple sequence repeats
(SSRs, also known as microsatellites [68]). SSRs are simple
and cheap to use, reliable, co-dominant and have a high
level of polymorphism. As a consequence high-density
SSR maps have been produced and SSR markers linked
to genes of interest have been found. Other commonly
used markers include sequence-tagged sites, sequence
characterized amplified regions and single nucleotide
polymorphisms. An historical account of the development
of DNA markers in plant breeding is given by Moose and
Mumm [69]. Perfect markers are those that define the
alleles of the target gene, however, linked markers are
good enough for plant breeding purposes especially if a
pair of tightly linked flanking markers ( < 1–5 cM on either
side) are deployed. Advantages of MAS are: they can
replace unreliable phenotypic screens, can be used at any
http://www.cabi.org/cabreviews
B.P. Forster et al.
11
Initial cross between recipient and donor parents (2n=14
chromosomes).
P1 X P2
F1
X P1
Progeny from cross: first generation (50% DNA from P1).
BC2
X P1
Example of one genotype from the BC2 population (population
has 75% DNA from P1), with selection for a marker/trait.
BC3
X P1
Example of one genotype from the BC3 population (population
has 75% DNA from P1), with selection for a marker/trait.
BC4
X P1
Example of one genotype from the BC4 population (population
has 75% DNA from P1), with selection for a marker/trait.
BCn-1
Selfing
Example of one genotype from and advanced BC population,
with selection for a marker/trait. This is then selfed to produce a
near isogenic line for P1.
Near isogenic line of P1 carrying the marker/trait introgressed
from P2.
BCn
inbred line
Figure 5 Scheme for backcrossing conversion using marker-assisted or trait selection.
time (not season dependent), reduce the number of lines
for promotion to the next generation and can be used in
rapid generation cycling (more generations per year, see,
e.g., Figure 1). Thus selected lines can be fast-tracked with
savings in time and costs.
In addition to MAS for target genes, genetic fingerprinting can be deployed to monitor and select for the
desired genetic background. This is particularly important
in crossing programmes where an elite line is crossed to
one or more donor lines that are genetically diverse. In
such cases gene pyramiding and backcrossing may be
deployed to introgress the desired genes (using markers,
‘offensive breeding’) and genetic fingerprinting used
to select for the desired genetic background (‘defensive
breeding’). Fingerprinting markers should have wide genome coverage. Amplified fragment length polymorphisms
and restriction fragment length polymorphisms were
commonly used fingerprinting methods, but have given
way to hybridization based and more recently sequencebased procedures [70] Rapid developments in genotyping
have promoted the concept of ‘molecular breeding’
and more recently ‘genomics assisted breeding’ whereby
selection is done in a marker laboratory rather than in the
field [71–73].
Despite the advantages of MAS there are relatively
few examples of practical use in plant breeding. Examples
are typically confined to high value traits, such as disease
resistance or publicly sensitive issues, such as ensuring
breeding lines are not contaminated with GMOs. A major
limitation to MAS is the cost of DNA extraction. Costs
have been coming down with time and recently low-cost
methods have been developed [74]. High-throughput
genotyping methods have recently been used to develop
a new approach for breeding natural quantitative alleles.
Called genomic selection, the method employs thorough
phenotyping and high density genotyping of a ‘training
population’. Through an iterative process, a set of markers associated with the trait(s) of interest in the population are defined (‘model training’). These prediction
models are then applied in selecting the most promising
novel breeding lines from the ‘breeding population’ based
on their marker genotypes only [25]. A major advantage
of genomic selection is that it allows increased numbers of
breeding cycles per unit time. As outlined above, conventional breeding requires 10 or more years from crossing
before the best progeny, those which carry an improved
allelic combination, are identified and can be recombined
for the next breeding cycle. Genomic selection, once a
predictive model is well established and validated, allows
the identification of superior individuals or lines much
earlier, even superior F2 plants can be selected based on
their genetic fingerprint and immediately recombined. Such
a procedure is expected to increase the frequency of
desired alleles in a breeding population much faster,
increasing the selection gain per unit time, but experimental results supporting this hypothesis are still pending.
http://www.cabi.org/cabreviews
12
CAB Reviews
Table 3 Historical perspective of innovations providing efficiency gains in plant breeding
Era
Innovation
Landmark papers and recent reviews
1822–1865
Progeny prediction: Mendel studies publishes laws on
inheritance which provide a scientific basis for plant breeding
Mutation breeding: dawn of plant mutation breeding, from
de Vries’ ideas to the first mutant variety. Mutagenesis allowed
the production of new biodiversity directly in elite germplasm
Plant tissue culture: the dawn of plant tissue culture; from the
concept of cell totipotency, to a wide range of cell, tissue
and organ culture methods
Doubled haploidy: dawn of haploid/DH technologies; from
first observations of haploids to applications in plant breeding.
Provided a quick means of producing homozygous lines
Screening: Phenotypic and genotypic screening; from the dawn
of selection by phenotypic associations (seed weight in beans)
to high-throughput phenomic and genomic screening
Rapid generation cycling: speeding up generation times; from
SSD to the incorporation of a range of methods in shortening
the time between successive generations
GM plant breeding: genetic modification: use of recombinant
DNA technologies to provide novel genetic variation.
Techniques are becoming increasingly sophisticated and
include targeted mutagenesis and other NPBTs
Molecular plant breeding: based on developments in
high density genetic maps, genetic fingerprinting, robust
marker–trait associations and bioinformatics
Mendel [15]; Weiss [80]; Ellis et al. [81]
1901–1934
1902–1960
1921–1952
1923–present
1941–present
1980s–present
1980s–present
Other Issues
Preferred cultivars
Many crops have ‘market preferred’ cultivars. This is true
for high-value crops such as coffee, banana and citrus
and local subsistence cultivars or landraces. Such crops
are grown extensively and often become susceptible to
disease or small changes in the environment. Breeders
therefore have a dilemma in attempting to breed a new
improved cultivar. On one hand they need to produce a
new improved cultivar, on the other they do not want to
disrupt a winning genotype, which conventional crossing
will entail. Subtle changes are therefore desirable and this
will promote the use of gene modification methods such
as mutation and GM breeding, which are also faster than
conventional breeding.
Participatory plant breeding
Several countries promote ‘farmer participatory plant
breeding’ and in such cases the farmers have early access
to (pre-release) seed. Participatory breeding has huge
potential particularly in less developed regions of the
world. Breeders working in close cooperation with
farmers can thus include regionally or culturally desired
cultivar traits much better in their selection programme,
such as taste, aroma, shape or other important characters
in addition to typical breeder’s traits such as yield performance and stress tolerance. Participatory breeding
does not necessarily speed up the breeding scheme itself,
but facilitates the adoption of cultivars by growers,
especially those in rural areas. This approach had high
De Vries et al. [82]; Kharkwal [83];
Shu et al. [35, 36]
Haberlandt [84]; Vasil [85]
Blakeslee et al. [86]; Touraev et al. [60]
Sax [87]; Varshney et al. [88]; Araus
and Cairns [20]
Goulden [49]; Brim [50]; Zeng et al.
[52]. This review; [89]; De La Feunte
et al. [90]
Bevan et al. [91]; Fraley et al. [92];
Herrera et al. [93]; Halford and Shewry
[48]; James [47]; Varshney et al. [94]
Paterson et al. [95]; Tanksley [96];
Collard and Mackill [72]; Moose and
Mumm [69]; Varshney et al. [94]
potential for food security, particularly for less developed
regions where, for many crops, only landraces with low
productivity are available.
Patents, IP and legislation
Intellectual property rights and innovations in plant
breeding are often protected by patents [75], for example,
Dunwell [76] lists over 30 granted patents related to plant
haploids [76]. This is not normally a problem for breeding
companies who live in a commercial world and can pay for
their use. GM breeding however is particularly restricted
by legislation and is currently a hot topic for action and
re-action; some GM breeding companies have indicated
they may withdraw from crop production in some
countries and conversely some countries have called for a
reduction in legislation to promote GM crops (https://
www.gov.uk/government/publications/genetic-modificationgm-technologies).
Accessing foreign germplasm
Germplasm needed for breeding may not be available inhouse and may need to be accessed from other breeders,
gene banks or collections. The transfer of materials across
boundaries often requires the use of phytosanitary certificates, Standard Material Transfer Agreements and quarantine. A further problem may arise from restrictions in
germplasm exchange among breeders. In the past exchange of novel breeding lines was common practice and
included the reciprocal agreement to use exchanged lines
in crosses. This has altered recently: breeding lines are
often exchanged with a ‘test only’ restriction. There is
http://www.cabi.org/cabreviews
B.P. Forster et al.
13
also an International Treaty on Plant Genetic Resources
in Food and Agriculture, (http://www.planttreaty.org/).
These compliances come with costs and time delays.
Table 3 provides a list of some significant innovations in
plant breeding from 1822 to the present day.
Pragmatism
Breeders will always use useful material, and have been
effective in developing cultivars with novel traits using
conventional methods. For example, the 1B/1R translocation was bred into many successful wheat cultivars
because it carried desirable disease resistance. The cause
of the resistance, the wheat/rye translocation, was not
discovered until later [77]. Similarly, graphical genotyping
[78] applied to foundation barley cultivars over time has
revealed the introduction of novel markers in successful
cultivars, and once introduced are retained in subsequent
successful cultivars. This will also be true for genes introduced by GM techniques – once they are released, they
can be used in conventional crossing programmes. Today
we know that these are associated with useful traits, such
as disease resistance [79], but for the breeder performance in the field was the main selection criterion.
For the future, genotypic data, and pedigree information
of successful cultivars could provide useful information in
genotypic breeding, especially if it can be linked to target
traits. National list trials provide an ideal opportunity to
combine data on performance, pedigree and genotype
of elite lines. However, since plant breeding is also a
commercial enterprise, these data are now seen as proprietary and not openly available.
Summary
Post-breeding, cultivar release
In order for a selected line to become a cultivar approved
for release to farmers it normally has to pass national
checks and regulations. These require time and sufficient
seed and the processes come with a cost. Prior to release
multi-locational trials are also performed and again this
requires seed and financing. Seed bulking is also required
to satisfy sale demands. These time consuming steps are
post-breeding and in many countries cannot be speeded
up. However, as noted above, several countries promote
‘farmer participatory plant breeding’ and in such cases the
farmers have early access to (pre-release) seed.
Acknowledgements
Techniques that were widely applauded but failed to
deliver widespread results
Some innovations in genetics and biotechnology have
remained largely academic with little impact on practical
plant breeding, e.g., protoplast fusion, QTL analysis
and studies in model species. The amount of R&D and
the number of publications in these are vastly disproportionate to the meagre promised outcomes in
terms of breeding success. Some techniques have also
waxed strong in their day, e.g., cytogenetics but have
tended in recent times to be replaced by simpler (marker)
methods, while others, such as GM techniques and
genomic breeding have yet to reach their full potential.
Food, feed and fuel security, climate change and profitability are major issues which demand rapid plant breeding
responses. Although some issues have remained recalcitrant to improvement (such as the ability to manipulate
meiotic recombination effectively), thus far these have
been compensated for by developments in selection for
desired phenotypes/genotypes, including: rapid generation
cycling, doubled haploidy and MAS. Since the time
of Mendel plant breeders have applied innovations in
science and biotechnology to promote plant breeding.
Plant breeding continues to be a dynamic process and
based more and more firmly on inter-disciplinary science,
yet more techniques are likely to arise and old ones find
new application within these. Mutation breeding perhaps
is a good example of this, of where it was heralded originally as an all-encompassing solution to plant breeding
problems, failed to fulfil its initial promise (in large part
because of the inability to target any changes and
the generation of a plethora of unwanted undesirable
affects), to rise again with the possibility of controlled and
site-directed changes and methodologies for massive
screening.
Much of the work described in plant mutation breeding
has been supported by the Joint FAO/IAEA Division
programme of Nuclear Techniques in Food and
Agriculture.
References
1. Harlan JR. Crops and man. American Society of Agronomy
and Crop Science 1992;284.
2. Forster BP, Shu Q. Plant mutagenesis in crop improvement:
basic terms and applications. In Shu Q, Forster BP, Nakagawa
H, editors. Plant Mutation Breeding and Biotechnology.
CABI; 2012;C01:9–20.
3. Fehr WR. Genetic contribution to yield gains in five major
crop plants. Crop Science Society of America 1984;7:9.
ISBN: 978-0-89118-586-4.
4. Duvick DN. Plant breeding past achievements and
expectations for the future. Economic Botany 1986;40:289–97.
5. Borlaug N. Feeding a world of 10 billion people: the miracle
ahead. In vitro Cellular and Developmental Biology – Plant
2002;38:221–8.
6. Randers J. A Global Forecast for the Next Forty Years 2052.
Chelsea Green Publishing, White River Junction, VT, USA;
2012. p. 62–159.
http://www.cabi.org/cabreviews
14
CAB Reviews
7. Foley JA, Ramankutty N, Brauman KA, Cassidy ES,
Gerber JS, Johnston M, et al. Solutions for a cultivated planet.
Nature 2011;478:337–342.
26. Mc Callum CM, Comai L, Greene EA, Henikoff S. Targeting
induced local lesions IN genomes (TILLING) for plant
functional genomics. Plant Physiology 2000;123:439–42.
8. Easterling W, Aggarwal P, Batima P, Brander K,
Erda L, Howden M, et al. Contribution of Working Group II to
the Fourth Assessment Report of the Intergovernmental Panel
on Climate Change. In: Parry ML, Canziani OF,
van der Linden PJ, Hanson CE, editors. Climate Change 2007:
Impacts, Adaptation and Vulnerability. Cambridge University
Press, Cambridge, UK; 2007. p. 273–313.
27. Till BJ, Zen T, Comai L, Henikoff S. A protocol for TILLING
and Eco-TILLING in plants and animals. Nature Protocols
2006;1(5):2465–77.
9. Briggs FN, Knowles PF. Introduction to Plant Breeding.
Reinhold Publishing Corporation; 1967.
10. Fehr WR, editor. Principles of Cultivar Development:
Theory and Technique 1987: volume 1. Macmillan, New York.
11. Stoskopf NC, Tomes DT, Christie BR. Plant Breeding Theory
and Practice. Westview Press, Boulder, CO; 1993.
12. Poehlman JM, Sleper DA. Breeding Field Crops. Iowa State
University Press, Ames, USA; 1995. pp 494.
13. Allard RW. Principles of Plant Breeding. 2nd edition. Wiley,
New York; 1999.
14. Brown J, Caligari PDS, Campos H. Plant Breeding.
Wiley–Blackwell Press; 2014 (in press).
28. Till BJ, Zen T, Comai L, Henikoff S. A protocol for TILLING
and Eco-TILLING. In: Shu QY, Forster BP, Nakagawa H,
editors. Plant Mutation Breeding and Biotechnology. CABI;
2012;C22:269–86.
29. Kurowska M, Daszkowska-Golec A, Gruszka D, Marzec M,
Szurman M, Szarejko I, et al. TILLING – a shortcut in
functional genomics. Journal of Applied Genetics
2011;52(4):371–90. doi: 10.1007/s13353-011-0061-1.
30. Slade AJ, McGuire C, Loeffler D, Mullenberg J, Skinner W,
Fazio G, et al. Development of high amylose wheat through
TILLING. BMC Plant Biology 2012;12:69–75.
31. FAO/IAEA database of mutant varieties, http://mvgs.iaea.org
32. Forster BP, Franckowiak JD, Lundqvist U, Thomas WTB,
Leader D, Shaw P, et al. Mutant phenotyping and prebreeding in barley. In: Shu Q, Forster BP, Nakagawa H,
editors. Plant Mutation Breeding and Biotechnology.
CABI; 2012;C25:327–46.
15. Mendel G. Experiments in plant hybridization. In: Peters JA,
editor. Translated from German by the Royal Horticultural
Society of London and Reproduced in Classic Papers on
Genetics. Prentice-Hall, Inc., Englewood Cliffs, NJ; 1865.
1959.
33. Schunk CR, Eberius M. Phenomics in plant biological research
and mutation breeding. In: Shu Q, Forster BP, Nakagawa H,
editors. Plant Mutation Breeding and Biotechnology.
CABI; 2012;C41:535–60.
16. Puchta H, Hohn B. From centiMorgans to base pairs:
homologous recombination in plants. Trends in Plant
Science 1996;1:340–8.
34. Jankowicz-Cieslak J, Huynh OA, Bado S, Matijevic M, Till BJ.
Reverse-genetics by TILLING expands through the plant
kingdom. Emirates Journal of Food and Agriculture
2011;23(4):290–300.
17. Griffiths S, Sharp R, Foote TN, Bertin I, Wanous M, Reader S,
et al. Molecular characterization of Ph1 as a major
chromosome pairing locus in polyploid wheat. Nature
2006;439:749–52.
35. Shu QY, Forster BP, Nakagawa H, editors. Plant Mutation
Breeding and Biotechnology. CAB International and FAO;
2012a. pp. 608.
18. Furbank RT, Tester M. Phenomics – technologies to relieve
the phenotyping bottleneck. Trends in Plant Science
2011;16:635–44.
19. Fiorani F, Schurr U. Future scenarios for plant phenotyping.
Annual Review of Plant Biology 2013;64:17.1–17.25.
20. Araus JL, Cairns JE. Field high-throughput phenotyping: the
new crop breeding frontier. Trends in Plant Science
2014;19:52–61.
21. Lörz H, Wenzel G. Molecular marker systems in plant breeding
and crop improvement. In: Nagata T, Lörz H, Wenzel G,
editors. Biotechnology in Agriculture and Forestry. Springer,
Heidelberg, Germany; 2004. p. 55. ISBN 3-540-20689-2.
22. Jannick JL, Lornez AJ, Iwata H. Genomic selection in plant
breeding: from theory to practice. Briefings in Function
Genomics 2010;9:166–77.
23. Abe A, Shunichi K, Yoshida K, Natsume S, Takagi H,
Kanzaki H, et al. Genome sequencing reveals agronomically
important loci in rice using MutMap. Nature Biotechnology
2012;30:174–8.
24. Crossa J, Beyene J, Kassa S. Genomic prediction in maize
breeding populations with genotyping-by-sequencing.
G3-Genes-Genomes-Genetics 2013;1903–26.
25. Heffner EL, Mark E, Sorrells ME, Jannink J-L. Genomic
selection for crop improvement. Crop Science 2009;49:1–12.
36. Shu QY, Shirasawa K, Hoffmann M, Hurlebaus J, Nishio T.
Molecular techniques and methods for mutation detection
and screening in plants. In: Shu Q, Forster BP, Nakagawa H,
editors. Plant Mutation Breeding and Biotechnology. CABI,
Wallingford, UK; 2012b;C20:241–56.
37. Bado S, Forster BP, Till BJ, Nielen S, Ghanim AMA, Lagoda
PJL. Plant mutation breeding: current progress and future
assessment. Plant Breeding Reviews 2015 (accepted).
38. UG99 FACTSHEET, 2012. Available from: URL: http://
www.iaea.org/newscenter/pressreleases/2013/
prn201311.html
39. Lagoda PJL. Effects of radiation on living cells and plants.
In: Shu Q, Forster BP, Nakagawa H, editors. Plant Mutation
Breeding and Biotechnology, CABI; 2012;C11:123–34.
40. Wang HY, Liu ZH, Chen PD, Wang XE. Irradiation-facilitated
chromosomal translocation: wheat as an example. In: Shu Q,
Forster BP, Nakagawa H, editors. Plant Mutation Breeding
and Biotechnology. CABI; 2012;C19:223–40.
41. Sears ER. The transfer of leaf rust resistance from Triticum
umbellulatum to wheat. Brookhaven Symposia in Biology
1956;9:1–21.
42. Lebel EG, Masson J, Bogucki A, Paszkowski J. Stressinduced intrachromosomal recombination in plant somatic
cells. Proceedings of the National Academy of Science, USA
1993;90:422–6.
http://www.cabi.org/cabreviews
B.P. Forster et al.
43. Puchta H, Swoboda P, Hohn B. Induction of
intrachromosomal recombination in whole plants. Plant
Journal 1995;7:203–10.
44. Dirks R, Van Dun K, De Snoo CB, Van Den Berg M,
Lelivelt CLC, Veormans W, et al. Reverse breeding:
a novel breeding approach based on engineered meiosis.
Plant Biotechnology Journal 2009;7:837–45.
45. Belhaj K, Chaparro-Garcia A, Kamoun S, Nekrasov V. Plant
genome editing made easy: targeted mutagenesis in model
and crop plants using the CRISPR/Cas system. Plant Methods
2013;9:9–39.
46. Yamamoto E, Yonemaru J-I, Yamamoto J, Yano M. OGRO:
the overview of functionally characterized genes in rice online
database. Rice 2012;5:26. Available from: URL: http://
www.thericejournal.com/content/5/26
47. James C. Global Status of Commercialized Biotech/GM
Crops. ISAAA Brief, Ithica, NY; 2013.Available from:
URL: http://www.isaaa.org/resources/publications/briefs/46/
default.asp
15
Plants. A manual. 2003, Kluwer Academic Publishers; 2003.
pp. 333–349.
60. Touraev A, Forster BP, Mohan Jain S, editors. Advances in
Haploid Production in Higher Plants. Springer
Science+Business Media B.V.; 2009.
61. Bohanec B. Doubled haploids via gynogenesis. In: Touraev A,
Forster BP, Mohan Jain S, editors. Advances in Haploid
Production in Higher Plants. Springer Science+Business
Media B.V., Heidelberg, Germany; 2009. Chapter 2:35–46.
62. Košmrlj K, Murovec J, Bohanec B. Haploid induction in Hullless seed pumpkin through parthenogenesis induced by x-rayirradiated pollen. Journal of the American Society for
Horticultural Science 2013;138:310–6.
63. Dunwell JM, Wilkinson MJ, Nelson S, Wening S, Sitorus AC,
Mienanti D, et al. Production of haploids and doubled
haploids in oil palm. BMC Plant Biology 2010;10:218–43.
(Available from: URL: http://www.biomedcentral.com/
1471-2229/10/2/18).
48. Halford NG, Shewry PR. Genetically modified crops”
methodology, benefits, regulation and public concerns.
British Medical Bulletin 2000;56:62–73.
64. Nasution O, Sitorus AC, Nelson SPC, Forster BP,
Caligari PDS. A high-throughput flow cytometry method for
ploidy determination in oil palm. Journal of Oil Palm Research
2013;25:265–71.
49. Goulden CH. Problems in plant selection. In: Burnett RC,
editor. Proceedings of the 7th International Genetics
Congress, Edinburgh, UK. Springer, Heidelberg, Germany;
1939. pp.132–3.
65. Hofinger BJ, Owen A, Huynh OA, Jankowicz-Cieslak J,
Müller A, Otto I, et al. Validation of doubled haploid plants by
enzymatic mismatch cleavage. Plant Methods 2013;9:43.
doi:10.1186/1746-4811-9-43.
50. Brim CA. A modified pedigree method of selection in
soybeans. Crop Science 1966;6:220.
66. Koebner RMD, Miller TE, Snape JW, Law CN. Wheat
endopeptidase: genetic control, polymorphism,
intrachromosomal gene location, and alien variation.
Genome 1988;30:186–92.
51. Norman Borlaug – Biographical. Available from: URL:
http://www.nobelprize.org/nobel_prizes/peace/laureates/
1970/borlaug-bio.html
52. Zeng Z, Wang HB, Chen GD, Yan GJ. A procedure allowing up
to eight generations of wheat and nine of barley per annum.
Euphytica 2013;191:311–6.
53. Ghanim AMA, Bado S, Den A, Ahmed MS, Tassoob-Shirazi F,
Amos EK, et al. Techniques in accelerating plant mutation
breeding in wheat and sorghum. IVPCA 2014 (in press).
54. Murashige T, Skoog F. A revised medium for rapid growth
and bioassay with tobacco tissue culture. Physiologia
Plantarum 1962;15:473–97.
55. Jacobsen S, Søndergaard I, Møller B, Desler T, Munck L. A
chemometric evaluation of the underlying physical and
chemical patterns that support near infrared spectroscopy
of barley seeds as a tool for explorative classification of
endosperm genes and gene combinations. Journal of Cereal
Science 2005;42:281–99.
56. Gregorio GB, Senadhira D, Mendoza RT. Screening rice for
salinity tolerance. IRRI Discussion Paper 1997; series 22,
IRRI, Manila. pp.30.
57. Erb WA, Rowe RC. Screening tomato seedlings for multiple
disease resistance. Journal of the American Society of
Horticultural Science 1992;117:622–7.
58. Maluszynski M, Kasha KJ, Forster BP, Szarejko I. Doubled
Haploid Production in Crop Plants. A manual. Kluwer
Academic Publishers, Dordrecht, Boston, London; 2003.
428pp.
59. Thomas WTB, Forster BP, Gertsson B. Doubled haploids
in breeding. In: Maluszynski M, Kasha KJ, Forster BP,
Szarejko I, editors. Doubled Haploid Production in Crop
67. Payne PI. Genetics of wheat storage proteins and the effect of
allelic variation on bread-making quality. Plant Biology
1987;38:141–53.
68. Gupta PK, Varshney RK, Sharma PC, Ramesh B. Molecular
markers and their applications in wheat breeding. Plant
Breeding 1999;118:369–90.
69. Moose SP, Mumm RH. Molecular plant breeding as the
foundation for 21st century crop improvement. Editor’s choice
series on the next generation of biotech crops. Plant
Physiology 2008;147:969–977.
70. Deschamps S, Llaca V, May GD. Genotyping-by-sequencing
in plants. Biology 2012;1:460–483.
71. Koebner RMD, Sommers RW. 21st century wheat breeding:
plot selection or plate detection? Trends in Biotechnology
2003;21:59–63.
72. Collard BCY, Mackill DJ. Marker-assisted selection: an
approach for precision plant breeding in the twenty-first
century. Philosophical Transaction of the Royal Society B
2008;363:557–72.
73. Varshney RK, et al. Achievements and prospects of genomicassisted breeding in three legume crops of the semi-arid
tropics. 2013. Available from: URL: http://dx.doi.org/10.1016/
j,.biotechadv.2013.01.001
74. Huynh OA, Hofinger BJ, Beshir MM, Jankowicz-Cieslak J, Guo
H., Forster BP, Till BJ. Do-it-yourself molecular biology for
plant breeding: low-cost tools for developing countries. Plant
Genetics and Breeding Technologies; Plant Diseases and
Resistance Mechanisms: Proceedings, February 18-20, 2013,
Vienna, Austria. Medimond – Monduzzi Editore international
http://www.cabi.org/cabreviews
16
CAB Reviews
Proceedings Division, Pianoro, Italy, 2013, p. 33–6.
ISBN 978-88-7587-682-1.
75. Blakeney M. Patenting of plant varieties and plant breeding
methods. Journal of Experimental Botany 2012;63:1069–74.
76. Dunwell JM. Patents and haploid plants. In: Touraev A, Forster
BP, Mohan J, editors. Haploid Production in Higher Plant.
Springer; 2009. p. 97–114.
77. Mettin D, Bluthner WD, Schlegel G. Additional evidence on
spontaneous 1B/1R wheat-rye substitutions and
translocations. In: Sears ER, Sears LMS, editors. Proceedings
of the Fourth International Wheat Genetics Symposium. Alien
genetic material; University of Missouri, Columbia, USA; 1973.
p. 179–84.
78. Young ND, Tanksley SD. Restriction fragment length
polymorphisms maps and the concept of graphical genotypes.
Theoretical and Applied Genetics 1989;77(1):95–101.
79. Russell JR, Ellis RP, Thomas WTB, Waugh R, Provan J,
Booth A, et al. A retrospective analysis of spring barley
germplasm development from “foundation genotypes” to
currently successful cultivars. Molecular Breeding
2000;6:553–68.
85. Vasil IK. A history of biotechnology: From the cell theory of
Schleiden and Schwann to biotech crops. Plant Cell Reports
2008;1423–40.
86. Blakeslee AF, Belling J, Farnham ME, Bergner AD. A haploid
mutant in the Jimson weed, Datura stramonium. Science
1922;55:646–7.
87. Sax K. The association of size differences with seed coat
pattern and pigmentation in Phaseolus vulgaris. Genetics
1923;8:552–60.
88. Varshney RK, Graner A, Sorrells ME. Genomics-assisted
breeding for crop improvement. Trends in Plant Science
2005;10:621–30.
89. Lusser M, Parisi C, Plan D, Rodrı́uez-Cerezo E. Development
of new biotechnologies in plant breeding. Nature
Biotechnology 2012;30:231–9.
90. De La Fuente GN, Frei UK, Lübberstedt T. Accelerating plant
breeding. Opinion, Trends in Plant Science 2013;18:667–72.
91. Bevan MW, Flavell RB, Chilton MD. A chimaeric antibiotic
resistance gene as a selectable marker for plant cell
transformation. Nature 1983;304:184–7.
80. Weiss K. Goings on in Mendel’s garden. Evolutionary
Anthropology 2002;11:40–4.
92. Fraley RT, Rogers SG, Horsch RB, Sanders PR, Flick JS,
Adams SP, et al. Expression of bacterial genes in plant cells.
Proceedings of the National Academy of Science, USA
1983;80:4803–7.
81. Ellis THN, Hofer JMI, Timmerman-Vaughan GM, Coyne CJ,
Hellens RP. Mendel, 150 years on. Trends in Plant Science
2011;16:590–6.
93. Herrera-Estrella L, Depicker A, van Montagu M, Schell J.
Expression of chimaeric genes transferred into plant cells
using Ti-plasmid-derived vector. Nature 1983;303:209–13.
82. De Vries H. Die Mutationstheorie. I. Veit & Co. Leipzig,
Germany 1901 (English translation, 1910. The Open Court
Publication, Chicago).
94. Varshney RK, Bansal KC, Aggarwal PK, Datta SK, Craufurd
PQ. Agricultural biotechnology for crop improvement in a
variable climate: hope or hype? Trends in Plant Science
2011;16:363–71.
83. Kharkwal MC. A brief history of plant mutagenesis. In: Shu Q,
Forster BP, Nakagawa H, editors. Plant Mutation Breeding and
Biotechnology. CABI International and FAO; 2012;C02:21–30.
84. Haberlandt G. Kulturversuche mit isolierten Pflanzenzellen.
Sitzungsber. Akad. Wiss. Wien. Mathematischnaturwissenschafliche Klasse Abteilung 1902;111:69–92.
95. Paterson AH, Landers ES, Hewitt JD, Peterson S, Lincoln SE,
Tanksley SD. Resolution of quantitative traits into Mendelian
factors by using a complete linkage map of restriction length
polymorphisms. Nature 1988;335:721–26.
96. Tanksley SD. Mapping polygenes. Annual Review of Genetics
1993;27:205–33.
http://www.cabi.org/cabreviews