Genetic diversity is indispensable for plant breeding to improve crops

Genetic diversity is indispensable for plant breeding to improve cropsReceived: 4 December 2019 Accepted: 5 October 2020 Published online: 24 February 2021
DOI: 10.1002/csc2.20377
Crop Science R E V I E W A N D I N T E R P R E T AT I O N PA P E R S
Genetic diversity is indispensable for plant breeding to improve crops
Shilpa Swarup1 Edward J. Cargill1 Kate Crosby2 Lex Flagel1 Joel Kniskern3
Kevin C. Glenn1
2 California Dep. of Fish and Wildlife, 1010
Riverside Pkwy, West Sacramento, CA
95605, USA
CA 95076, USA
Chesterfield Pkwy W., Chesterfield, MO
63017, USA.
Eizenga.
online publication.]
Abstract Plant breeders face multiple global challenges that affect food security, productivity,
accessibility, and nutritional quality. One major challenge for plant breeders is devel-
oping environmentally resilient crop cultivars in response to rapid shifts in cultivation
conditions and resources due to climate change. Plant breeders rely on different crop
genetic resources, breeding tools, and methods to incorporate genetic diversity into
commercialized cultivars. Breeders use genetic diversity to develop new cultivars
with improved agronomics, such as higher yield, biotic and abiotic stress tolerance,
and to improve the nutritional quality of foods for a growing world population. Plant
breeders perform the essential task of strategic integration of new genetic diversity
while preserving important economic traits of individual crops such as relative matu-
rity (maize, Zea mays L.), fruit type (tomato, Lycopersicon esculentum Mill.), plant
type (lettuce Lactuca sativa L.), and habitat type (canola, Brassica napus L.) that are
highly specialized for specific consumer preferences or market needs. This review
provides an industry perspective on how genetic diversity is incorporated for crop
improvement by (a) using a real-life example to highlight the vast amount of genetic
diversity that exists in plants, (b) providing a conceptual example to illustrate strate-
gic challenges a breeder faces while incorporating diversity, (c) describing how and
why it can a decade or more to incorporate diversity into commercialized cultivars,
even when advanced tools and technologies are used, and (d) sharing factors that
plant breeders consider when applying various tools, including genome editing, at
different stages of plant breeding.
1 INTRODUCTION
Abbreviations: CGN, Center for Genetic Resources; GS, genomic
selection; MABC, marker-assisted backcrossing; MAS, marker-assisted
selection; PCA, principal component analysis; SNP, single nucleotide
polymorphism.
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original
work is properly cited.
© 2020 The Authors. Crop Science published by Wiley Periodicals LLC on behalf of Crop Science Society of America
lenges that affect food security, sustainability, and adaptation
to climate change. Genetic diversity can be described as the range of genetic characteristics in a crop or species, whereas
genetic variation is the genetic differences among individu-
als for a specific characteristic, where these genetic differ-
ences reside in one or more DNA sequences. Genetic diver-
sity can be assessed by examining differences in the DNA
sequence in a population of individuals (Choudhury, Khan, &
Crop Science. 2021;61:839–852. wileyonlinelibrary.com/journal/csc2 839
14350653, 2021, 2, D ow
nloaded from https://acsess.onlinelibrary.w
iley O nline L
s and C onditions (https://onlinelibrary.w
iley.com /term
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
Dayanandan, 2014; Haun et al., 2011; van de Wouw, van
Hintum, Kik, van Treuren, & Visser, 2010). Genetic varia-
tion gives rise to phenotypic variation, which is differences
in observable traits within a population. Plant genomes are
dynamic and evolutionarily labile in nature, which results in
more frequent genetic and epigenetic changes in plants that
serve as the sources of a large amount of genetic and phe-
notypic diversity, even between cultivars within a species
(Adams & Wendel, 2005; Kejnovsky, Hawkins, & Feschotte,
2012; Kejnovsky, Leitch, & Leitch, 2009; Leitch & Leitch,
2012; Murat, Peer, & Salse, 2012). Greater genetic diversity
in plants gives them a remarkable ability to adapt to sudden
environmental changes (Raza et al., 2019; Wu, Cai, Zhang,
& Zeng, 2017). The success of crop improvement lies in effi-
ciently identifying and incorporating genetic diversity from
various plant genetic sources including currently cultivated
cultivars, newly developed cultivars, landraces, wild and near
relatives of cultivated cultivars, and germplasm collections
with elite and/or mutant plants. Various genomic tools and
breeding methods have improved the efficiency and preci-
sion of incorporating genetic diversity into commercialized
crop cultivars, but plant breeding still remains a time- and
resource-intensive process (Andersen & Lubberstedt, 2003;
Hospital, Chevalet, & Mulsant, 1992; Morgante & Salamini,
2003; Vaughan, Balazs, & Heslop-Harrison, 2007).
Climate change is projected to result in conditions unfa-
vorable for cultivation of many current crop cultivars, sub-
stantially disrupting food security over this century (Deutsch
et al., 2018; FAO, 2017; Stokstad, 2019; https://www.
globalchange.gov/browse/reports/global-climate-change-
are increasingly encountering extreme weather (e.g., both
higher and lower than typical ambient temperatures and more
intense amounts of rainfall as either floods or droughts),
resulting in shifts in crop planting dates and the appearance
of new plant diseases and pests to a region (Laux, Jäckel,
Tingem, & Kunstmann, 2010). In 2019, a wet spring (April
through June) in North America led to delayed planting of
maize (Zea mays L.) and soybean [Glycine max (L.) Merr.],
resulting in only 49% of the maize and 19% of the soybean
being planted on time vs. the 5-yr average “on-time planting”
of 80 and 47% for maize and soybean, respectively (USDA,
2019). In order to meet such atypical on-farm challenges,
plant breeders need to develop climate-resilient crops with
increasing ability to withstand intense and frequent seasonal
variation (Lobell et al., 2008; Renard & Tilman, 2019).
Developing crops and cultivars adapted to grow in extreme
and fluctuating environments is a unique challenge for
plant breeders, compared with improving crops for slow and
predictable seasonal changes. For example, in maize, anthrac-
nose stalk rot (caused by Colletotrichum graminicola Ces.),
gray leaf spot (caused by Cercospora zeae-maydis Tehon
and Daniels), and Goss’s wilt (Corynebacterium michiga-
Core Ideas An industry perspective on how genetic diversity
is incorporated for crop improvement is shared.
Challenges a breeder faces while incorporating
genetic diversity are discussed.
various tools are shared.
are currently three of the most prevalent maize diseases in
the U.S. but were of little or no concern 40 years ago (Burns
& Shurtleff, 1973). Today, nearly all commercial maize
breeding programs develop cultivars with resistance to the
abovementioned diseases due to their significantly increasing
impact over time on crop productivity, ranging from 6.1%
(0.65 t ha−1) yield loss from anthracnose stalk rot to 11.6%
(1.25 t ha−1) yield loss from Goss’s wilt (Boyd, Ridout,
O’Sullivan, Leach, & Leung, 2013). An example of how
exposure to severe weather affects yield is the cultivation
of sweet cherries [Prunus avium (L.) L.; Kistner, Kellner,
Andresen, Todey, & Morton, 2018]. In both 2002 and
2012, below freezing temperatures resulted in the cherry
flower pistils being frozen and consequently there was a
catastrophic loss of the cherry crop. Subsequently, cherry
breeders prevented the crop loss by successfully identifying
and introducing the “late bloom” trait from sour cherry
(Prunus cerasus L.) into sweet cherry (Cai et al., 2018). In
both crops, plant breeders tapped into the genetic diversity
available from cultivars outside of their typical breeding
germplasm collections to meet specific food security-related
challenges of preventing yield loss (in maize) and crop loss
(in cherry).
commercial breeding programs by (a) using a real life exam-
ple to highlight the vast amount of genetic diversity that exists
in plants (e.g., Brassicaceae plant family), (b) providing a con-
ceptual example to illustrate strategic challenges a breeder
faces while incorporating diversity, (c) describing how and
why it can take a decade or more to incorporate diversity into
commercialized cultivars, even when advanced tools and tech-
nologies are used, and (d) sharing factors that plant breeder’s
consider when applying advanced tools and methods, includ-
ing genome editing, in commercial breeding. Specific exam-
ples for several crops are provided throughout the article.
The topic of sources of genetic diversity for plant breeding
has been extensively reviewed by others and is therefore not
covered in this review (Adams & Wendel, 2005; Henderson
& Salt, 2017; Kejnovsky et al., 2009; Oladosu et al., 2016;
14350653, 2021, 2, D ow
nloaded from https://acsess.onlinelibrary.w
iley O nline L
s and C onditions (https://onlinelibrary.w
iley.com /term
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
Mayrose, 2016).
2 THE EXTENT AND USE OF GENETIC DIVERSITY TO DEVELOP NEW CROP CULTIVARS
Over the past several hundreds of years, genetic diversity in
the Brassicaceae family led to dozens of crops commonly
consumed today: field mustard (Brassica campestris L.),
Ethiopian mustard (B. carinata A Braun), mustard greens [B. juncea (L.) Czern.], oilseed rape (B. napus L.), turnips (B. rapa L.), and cabbages (B. oleracea L.) (Figure 1). Using
Brassica oleracea as one specific example, plant breeding
activities over the past 2,500 yr have generated more than a
half dozen horticultural crops, each very different from the
other in shape, color, and nutrients and all distinct from their
wild ancestor, to meet different consumption and/or orna-
mental purposes (Fahey, 2016). In some crops, genetic dif-
ferences can manifest as visible phenotypic differences (e.g.,
cauliflower vs. cabbage), whereas in other cases they are more
cryptic (e.g., the range of maturity groups across maize culti-
vars) (Kato, Lamb, & Birchler, 2004; Sher, Iqbal, Khan, Yasir,
& Hameed, 2012). Continuing with the example of Bras- sica oleracea, a preference for larger leaves led to the devel-
opment of the older members of this family, kale (Brassica oleracea var. acephala), collard greens (Brassica oleracea var. seballica), and Chinese broccoli (Brassica oleracea var.
alboglabra) (Hahn, Müller, Kuhnert, & Albach, 2016). Subse-
quent to the development of kale, several other cultivar groups
(e.g., albograbra, botrytis, capitata, gemmifera, gongylodes,
and italica) were derived from Brassica oleracea, resulting
in nine vegetables currently available in the produce market,
all progeny of the ancestral Brassica oleracea. Each demon-
strates how selection for particular traits resulted in the differ-
ent cultivar groups known today. For example, selection for
tightly bunched leaves from kale produced cabbage, whereas
a different selection process for thicker stems led to kohlrabi.
Brussel sprouts are an outcome of selection for buds from
kohlrabi. These advances occurred sporadically over time.
In the 12th to 13th century in Cyprus, the modern version
of cauliflower was cultivated. Broccoli, the newest member
of the group, was developed in the 1500s from kohlrabi by
selecting for plants that retain immature buds. The process
continues today: in the 1990s, broccolini was developed as a
hybrid between broccoli (Brassica oleracea var. italica) and
Chinese broccoli (Brassica oleracea var. alboglabra) (Bae-
nas, Marhuenda, García-Viguera, Zafrilla, & Moreno, 2019).
This family of related vegetable species is a remarkable exam-
ple of how plant breeding benefits from (a) the vast amount of
genetic diversity that can be used to produce a diverse array of
new crops over time and (b) the ability to grow a population
of plants to select a few individuals with desired characteris-
tic(s). Plant breeders put these advantages into practice during
multiple stages of the breeding process, which are described
below.
3 GENETIC DIVERSITY: A CONCEPTUAL EXAMPLE OF INCORPORATING GENETIC DIVERSITY FOR CROP IMPROVEMENT
One of the most essential tasks plant breeders undertake is
incorporating diversity into their breeding populations. To
illustrate this concept, two simulated tomato (Lycopersicon esculentum Mill.) breeding populations, each with the same
number of individual plants (n = 1,000), are used to describe
the strategic integration of genetic diversity in a plant breeding
program (Figure 2). The phenotypic trait of interest is tomato
fruit weight. Both populations have the same average value
of 100 g fruit weight. Following Fisher’s infinitesimal model,
phenotypic variation is assumed to follow a normal distribu-
tion in this simulation, which is typical for a quantitative trait
such as fruit weight (Fisher, 1918). However, the population
in Figure 2a was simulated to have lower phenotypic variance
than the population in Figure 2b.
First, consider a scenario in which the goal of a breeder is to
increase tomato fruit weight. The range of potential parental
plants is shown in the top 10% of the population indicated by
red bars in the two plant populations, one with “low” pheno-
typic variation (Figure 2a) and the other with “high” pheno-
typic variation (Figure 2b). As noted above, both populations
have the same phenotypic average tomato fruit weight value
of 100 g, but the average phenotypic trait value of the plants
in the top 10% from the low variation population (red bars in
Figure 2a) is considerably less than the average of the top 10%
in the high variation population (red bars in Figure 2b). This is
due to the difference in phenotypic variance between popula-
tions. Thus, given these two populations, a breeder will make
greater progress toward improving the weight by choosing the
population represented in Figure 2b with higher phenotypic
variation (Edwards, 2000).
In a second scenario, the breeder derives a new population to improve the weight in a tomato cultivar. The new popula-
tion is developed from using 100 individual plants producing
tomatoes with the heaviest fruit weight (the top 10%) from
the population distribution shown in Figure 2b. In this new
scenario, the first progeny generation of the new tomato pop-
ulation will contain only the fraction of genetic variation con-
tained within the 100 plants with the heaviest fruit from the
original parental population of 1,000 plants. If the breeder
continues to intercross only individual plants derived from the
100 plants with the heaviest fruit, repeating this for five breed-
ing cycles, as shown in Figure 2c, results in a plant population
14350653, 2021, 2, D ow
nloaded from https://acsess.onlinelibrary.w
iley O nline L
s and C onditions (https://onlinelibrary.w
iley.com /term
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
842 SWARUP ET AL.Crop Science
F I G U R E 1 The cultivar groups of Brassica oleracea are an example of remarkable genetic diversity existing in a plant family that can be used
to develop nine vegetables. All the vegetables were derived by selecting for plants with the desired traits and/or by combining traits in a hybrid. The
common names for the cultivar groups derived from the wild cabbage (B. oleracea) are kale (acephala), collard greens (seballica), Chinese broccoli
(alboglabra), cabbage (capitata), kohlrabi (gongylodes), cauliflower (botrytis), broccoli (italica), Brussels sprouts (gemmifera), and broccolini
(italica × alboglabra)
nloaded from https://acsess.onlinelibrary.w
iley O nline L
s and C onditions (https://onlinelibrary.w
iley.com /term
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
SWARUP ET AL. 843Crop Science
F I G U R E 2 Histograms representing genetic diversity for tomato fruit weight in 1,000 individual plants in simulated breeding populations. The
number of individual plants is shown on the y axis and fruit weight is shown on the x axis. The populations in Panels a and b each have the same
average value of 100 g fruit weight. The population with a low phenotypic variation is shown in Panel a, and a different population with a high
phenotypic variation is shown in Panel b. The red bars highlight the 10% of plants with the largest fruit weight in each population. Panel c shows
selection breeding in which the top 10% individual plants are advanced at each of the successive five generations (1, 2, 3, 4 and 5) from the
population in Panel b
with less genetic variation. This in turn, results in less pheno-
typic variation between individual plants, making a very slen-
der phenotypic trait distribution where the individuals in the
top 10% are hardly differentiated from the population aver-
age. The net result is that progress in improving the pheno-
typic trait stagnates. Figure 2c shows that only about 3% of
the phenotypic variation is left after five generations of selec-
tion, and the potential to continue to improve the fruit weight
over subsequent generations is significantly decreased. Thus,
long-term breeding progress cannot be performed without the
constant availability and accessibility to genetic variation for
different phenotypic traits like fruit weight. To avoid the loss
of genetic diversity shown in the second scenario, the breeder
continues to incorporate genetic diversity for fruit weight in
addition to other traits in tomato as shown by Schouten et al.
(2019).
In a third scenario, consider that genetic diversity is added
by crossing to a wild relative of tomato that provides resis-
tance to a disease but bears small-sized tomatoes. Breeders
must perform such crosses strategically, since wild tomato
cultivars have traits that may be unfavorable to a farmer and/or
consumer, while being a source of novel genetic diversity.
The unfavorable phenotypic traits are manifest as (a) genetic
background effects and (b) linkage drag with the desired trait.
Genetic background effects include any genetic region of the
wild relative crop that is not linked to the trait of interest. For
example, consider the need to introduce a new trait like dis-
ease resistance from a wild tomato cultivar that is not related
to fruit size. Because fruit size is a complex trait controlled
by multiple genes (Tanksley, 2004), crossing this disease-
resistant wild tomato cultivar with small fruit (<15 g) to a
cultivated cultivar that has large fruit (>200 g) will result
in progeny plants with intermediate fruit size, which is an
unfavorable trait in a large-fruited tomato cultivar. Even in
crops with a relatively short generation time such as tomato,
it can take several years to perform sufficient rounds of back-
crossing to remove unwanted genetic background effects after
introducing new alleles from a wild relative (Bai & Lindhout,
2007).
Linkage drag, on the other hand, describes a deleterious
trait that is genetically linked to a favorable trait and can lead
to either introducing a new deleterious trait into an existing
cultivar or negatively affecting a favorable trait that is already
present in the currently cultivated cultivar. Both the addition
of a deleterious trait or the negative impact on a favorable trait
would be undesirable to a grower or consumer. An example
of linkage drag was the introduction of a linked segment of
DNA from a wild relative of onion, Allium roylei Stearn, while
14350653, 2021, 2, D ow
nloaded from https://acsess.onlinelibrary.w
iley O nline L
s and C onditions (https://onlinelibrary.w
iley.com /term
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
introgressing resistance to downy mildew caused by Per- onospora destructor (Berk.) Casp. into a commercial onion
cultivar, A. cepa L. This resulted in the introduction of a
recessive lethal allele, where only heterozygous genotypes
at the locus survive (Scholten et al., 2007). Another exam-
ple is the introduction of Fusarium wilt resistance (Fusarium oxysporum f. sp. Lycopersici) from a wild relative of tomato,
Solanum pennillii Correll, into a cultivated tomato cultivar.
The resulting tomato progeny had increased plant sensitivity
to bacterial spot disease caused by species of Xanthomonas (Li, Chitwood, Menda, Mueller, & Hutton, 2018). Connecting
this to the simulated populations shown in Figure 2, if link-
age drag reduced the value of an important phenotypic trait
like tomato weight by half such that the high genetic diver-
sity population in Figure 2b had an overall average pheno-
typic trait value of 50 g in tomato weight instead of 100 g
(i.e., shifting the whole distribution to the left by 50 g), then
even though this population…