Copy Number Variation in Domestication – Cell

 

 

Review
Copy

Number

Variation

in

Domestication

Zoe N. Lye1 and Michael D. Purugganan1,2,*

Domesticated plants have long served as excellent models for studying evo-
lution. Many genes and mutations underlying important domestication traits
have been identified, and most causal mutations appear to be SNPs. Copy
number variation (CNV) is an important source of genetic variation that has been
largely neglected in studies of domestication. Ongoing work demonstrates the
importance of CNVs as a source of genetic variation during domestication, and
during the diversification of domesticated taxa. Here, we review how CNVs
contribute to evolutionary processes underlying domestication, and review
examples of domestication traits caused by CNVs. We draw from examples
in plant species, but also highlight cases in animal systems that could illuminate
the roles of CNVs in the domestication process.

Domestication is a Coevolutionary Process
Domestication is an evolutionary process that arises from coevolutionary interactions where
one species controls the reproduction and dispersal of another species for the benefit of the
former. Human-associated domestication as an evolutionary process began in the Paleolithic
and continued into the Neolithic, with the shift of hunter-gatherers to pastoralists and farmers
(cid:1)12 000 years ago, leading to the evolution of hundreds of crop plant species [1].
beginning
Moreover, domestication also occurred in animals, and there are dozens of known domesti-
cated
is now generally thought that domestication was a
protracted process that unfolded over thousands of years [3,4] and, it was during this period,
that genetic changes led to adaptation to agricultural environments and differentiation from wild
ancestors.

livestock and pet species [2]. It

(ii) diversification and/or

The early evolution of domesticated species occurs in two distinct phases: (i) initial domesti-
cation, where control over reproduction and dispersal is established, resulting in the origin of
the new domesticated species; and
improvement, where the
domesticated species develops local or population-specific adaptations to different environ-
its center of origin [3–5]. Many of the
ments or cultural preferences as
adaptive traits arising during this process may have evolved under the process termed
‘unconscious selection’, which acts similar to natural selection because
incipient domes-
ticates adapt to
in human-associated environments [1,2]. Nevertheless, many key
traits, particularly those associated with diversification, may have evolved under more
intense selection.

it spreads from

living

Most studies on the evolutionary genetics of domestication have used SNPs to examine
population relationships and to identify causal genetic variants often through genetic mapping
and genome-wide association studies. The role of CNVs
in the evolution of domesticated
species is not as well appreciated. In recent years, as whole-genome sequencing methods
have allowed the genome-wide characterization of CNVs, they have become the subject of
increased interest, broadening our understanding of the genetic basis of evolution. Here, we
review the role of CNVs in domestication, focusing primarily on plant species, but also providing

352

Trends in Plant Science, April 2019, Vol. 24, No. 4
© 2019 Elsevier Ltd. All rights reserved.

https://doi.org/10.1016/j.tplants.2019.01.003

Highlights
Whole-genome
resequencing, pan-
genomics, and developing computa-
tional methods have allowed charac-
terization of CNVs in diverse species.

Loss-of-function CNVs can cause
some of
the critical domestication
traits
in plants, whereas other CNVs
are associated with postdomestication
diversification traits, such as environ-
mental adaptation, disease resistance,
fruit size, and cultural preferences.

An exhaustive table of characterized
CNVs associated with domestication
phenotypes in a plant and animal sys-
tems is included.

1Center for Genomics and Systems
Biology, 12 Waverly Place, New York
University, New York, NY 10003, USA
2Center for Genomics and Systems
Biology, New York University Abu
Dhabi, Saadiyat Island, Abu Dhabi,
United Arab Emirates

*Correspondence:
mp132@nyu.edu (M.D. Purugganan).

field

Glossary
Amplification: the same sequence
of DNA is duplicated multiple times,
typically in tandem.
Chimeric gene: a gene comprising
coding sequences derived from two
or more other genes.
Experimental evolution: the use of
laboratory or controlled
experiments to investigate the
processes of evolution. Typically,
organisms with short generations
times are used to simulate processes
that would take longer in larger
organisms.
Fixation: increase in frequency of a
genetic variant, eventually resulting in
all members of a population sharing
the same variant at a locus.
Fixation index (FST): a measure of
genetic differentiation between two
populations.
Microhomology: identical short
DNA sequences, 1–4 bp in length.
Pan-genome: the entire gene set
contained within a species, taking
into account PAV between
individuals in a species. Not all
individuals carry all of the genes in
the pan-genome.
Photoperiod: day length; many
plants use day length as a signal to
enter various stages of the life cycle.
Purifying selection: selection
against disadvantageous alleles.
Tandem array: cluster of genes
created by repeated duplications.
flowering
Vernalization: induction of
by prolonged exposure to cold (i.e.,
winter).

examples from domesticated animal species that could point to contrasting patterns between
these two groups.

Copy Number Variation
CNVs are polymorphisms within species in which sections of a genome differ in copy number
between individuals, and include deletions, duplications, or amplifications (see Glossary) of
DNA sequence. Originally, CNVs were only thought of in terms of copy changes in functional
genetic features. Today, many researchers adopt a more expansive definition in light of the
ability to discover gains and losses of genomic material in an unbiased genome-wide manner
(Box 1). This can the include transposable elements and noncoding sequences. The definition
‘segmental
of a CNV continues to be somewhat arbitrary and is often conflated with the terms
duplication’
is typically 1 kb,
or
although many studies
(bps)
[6,7].
interest to
Nevertheless, CNVs that
researchers.

include smaller variants of as few as 50 base pairs
include functional sequences continue to be of most

‘structural variant’. The defined minimum

length of a CNV

Box 1. CNV Detection Methods

For reviews on major methods for CNV detection applied to domesticated species, see [7]. For reviews of CNV detection
from next-generation sequencing data, see [103,104].

fluorescence signals of a test and
Array comparative genome hybridization (aCGH) is based on the comparison of
reference sample hybridized to a microarray of tiled probes covering an entire genome. The use of smaller probes
increases the specificity of CNV detection in this method; however, aCGH (is more accurate in detecting deletions than
duplications [7,105]. SNP microarrays are also applied to CNV detection by comparing probe
intensities across
samples. They are also able to distinguish CNV alleles because they can use allele-specific probes [7].

Next-generation sequencing (NGS)-based methods fall into three major categories: read-depth (RD), read-pair (RP),
and split-read (SR) methods [7,103]. RD methods detect CNVs by comparing normalized read depth from short-read
sequence data aligned to a reference genome. Low or zero RD
is
interpreted as an increase in copy number. RP methods are based on the idea that read pairs should map to a reference
separated by approximately the same distances as the insert size. If read-pairs map farther away from each other than
expected, a deletion is detected; if they are too close together, an insertion is detected. SR methods use paired-end
reads and detect CNVs by aberrant mapping to a reference genome. For example, when only half of a read-pair maps to
a genome, a CNV breakpoint is identified. Whole-exome sequencing data are also applied to CNV discovery using a RD
identification of CNV
approach to
breakpoints and infer structure of CNVs [7].

local realignments are also used to refine the

identify CNVs [104]. Additionally,

interpreted as a deletion and

increased RD

is

Each method comes with a different set of biases. RP methods are less effective in repetitive regions and their accuracy
is dependent on the size of the insert [103,104]. SR methods are biased to detect smaller CNVs [103]. RD methods
typically have higher false positive rates and are biased towards detecting large variants [7]. The effectiveness of these
methods is also dependent on sample read depth. Due to these shortcomings, CNV studies using NGS data typically
combine multiple computational approaches to minimize false positives [7].

field of CNV discovery. Most, if not all, CNV discovery
There does not appear to be clear methodological standards in the
methods were developed for use in humans, and can be benchmarked against gold standard sets of known human
variants. In domesticated species, gold standard CNV sets do not exist to evaluate the efficacy of different meth-
odologies. Rather, researchers rely on simulations to benchmark methods or simply take existing methods at face value.
There are more than 50 published methods for detecting CNVs from NGS data. Selecting an appropriate method for a
given data set and species is a challenge to anyone designing a CNV study. As multiple high-quality reference genomes
are created for domesticated species and third-generation long-read sequencing becomes available, we expect to see
an increase in CNV studies and the development of more novel methodologies. Long read sequences have already
been used to resolve CNV
is critical that new
methodologies developed are accurately compared to existing methods to ensure that research is comparable across
platforms.

in tandem repeats where traditional methods are

limited [106]. It

Trends in Plant Science, April 2019, Vol. 24, No. 4 353

(A)

NAHR

(B)

SSA

(C)

Transposon excision

Retrogene

(D)

mRNA

DNA

Retrotransposase

Figure 1. Mechanisms of Copy Number Variation (CNV) Formation. (A) Nonallelic homologous recombination
(DSB) repair, a direct repeat,
(NAHR; unequal crossing over): during a recombination-based double-strand break
represented in green, is used as homology and incorrectly pairs during crossing over, this causes a reciprocal deletion
flanked by
and duplication of sequence between the repeats (purple). In this scenario, the resulting CNV break point is
tracts of homologous sequence. (B) Single-strand annealing (SSA). During double-strand break repair, the 50 stands are
resected to expose complimentary sequences either side of the break (green). Although this
is similar to the micro-
>30 base pairs (bp). This
homology-mediated end joining repair pathway, SSA requires longer tracts of homology, typically
flank
can result in significant deletions of intervening sequence (purple). (D) Transposon excision. Transposons (pink ovals)
a unique sequence (purple). Both transposons excise simultaneously, removing the unique sequence with them, and can
result in a deletion. (D) Retro-gene formation. Retrotransposon activity causes insertion of a coding sequence into the
genome (gene is shown in green with white boxes representing introns). mRNA (red) from the gene is reverse transcribed to
DNA. This DNA can be occasionally inserted into the genome and become a retrogene, a copy of the original gene lacking
introns (green box). These genes can be inserted into another gene, creating a chimeric gene, or become under control of
different promoter sequences and take on a new expression regime.

CNVs are formed through a variety of genetic mechanisms (reviewed in [6]). A key mechanism is
nonallelic homologous recombination (NAHR) or unequal crossing over, which results from
aberrant homology recognition during homology-based DNA repair or meiosis [6] (Figure 1A).
CNVs formed by this mechanism are characterized by tracts of homology on either side of the
CNV. NAHR is common in repetitive regions and an important source of tandem duplications
and deletions [8]. Another mechanism
is a double-
>30 bp
strand break repair process where broken ends are joined by annealing at homologies
in length, which can result in significant deletions [6] (Figure 1B).

is single-strand annealing (SSA), which

Transposable elements are also a source of CNVs; they can result in copy number change by
capturing DNA segments during excision and moving or deleting DNA segments [6] (Figure 1C).
Retrotransposon activity can also create CNVs through retrogenes; these are DNA insertions
into the genome resulting from reverse-transcribed mRNA that might take on a new function or
form a chimeric gene [9] (Figure 1D). For example, the sun locus in tomato is a retrotrans-
poson-mediated gene duplication that places the SUN gene under a different regulatory
element, altering fruit development to result in an oval fruit [10].

CNVs can also arise following polyploidization, when a genome doubles all genes are dupli-
cated; subsequent deletions in either of the subgenomes lead to a change in copy number.
Fractionalization of the maize genome has contributed to high intraspecific variation in copy
number and presence–absence variation (PAV) [11,12].

Other proposed mechanisms include microhomology-mediated break-induced replication
(MMBIR), whereby, during meiosis, a replication fork stalls and the lagging strand anneals to a
different replication in the vicinity as a result of microhomology, which can lead to complex

354

Trends in Plant Science, April 2019, Vol. 24, No. 4

rearrangements and duplications [13]. Undoubtedly, there are other stochastic and poorly
understood causes of CNV formation, especially ones that involve larger CNVs (for a compre-
hensive review of these mechanisms, see [6]).

CNVs Are Generally Deleterious . . .
CNVs have become the subject of increased interest, broadening our understanding of the
genetic basis of evolution. CNVs are thought to be generally deleterious and subject to
purifying selection and, thus, affect coding sequences
less frequently than noncoding
sequences [14–16]. Deletion CNVs can lead to loss of function (LOF), whereas duplication
CNVs affecting entire protein-coding genes can be deleterious if they affect dosage-sensi-
tive genes [14,17]. Simulation of the effects of genic CNVs in regulatory networks demon-
strated that increases in gene copy number by one or two copies can have large effects on
overall expression patterns due to regulatory feedbacks [18]. CNVs have been identified as
in altering gene
expression quantitative trait
expression [19,20].

loci (eQTLs), further demonstrating their role

fitness. Thus, most genic CNVs are predicted to occur

Given the generally deleterious effects of CNVs, it would be expected that CNVs that do affect
gene expression should be restricted to functional classes that can tolerate expression
changes without costs to
lowly
expressed genes at the periphery of gene regulatory and gene interaction networks, where
change in copy number is less impactful. Dopman and Hartl measured this in Drosophila and
found significantly
lower representation of deletion CNV genes among genes with known
interactions [21]. They also measured the ratio between nonsynonymous (Ka) and synonymous
site (Ks) mutations in open reading frames of CNV genes and found that CNV genes had a
higher ratio than did non-CNV genes, suggesting that CNV genes are under relaxed selective
constraint [21]. Keel et al. extended the investigation of CNVs in interaction networks to cows by
quantifying the number of interactions of CNV genes in a protein–protein interaction network
[22]. They demonstrated that CNV genes were likely to have fewer network connections than
were non-CNV genes, supporting the prediction that CNV genes are functionally constrained
and tend to occur at the periphery of interaction networks.

in

CNVs May Have a Role in Rapid Adaptation under Strong Selective
Pressure
While CNVs are generally deleterious, they also appear to be a key mechanism that can
enable adaptation during a period of strong selection. This phenomenon
in
experimental evolution of microbes under nutrient limitation, where spontaneous duplica-
tion of nutrient transporters repeatably occurs, conferring adaptation to the nutrient-limited
environment. In yeast under glucose-limited conditions, for example, amplifications of the
HXT6 and HXT7 genes, encoding high-affinity glucose transporters, were observed; under
limitation, there was amplification of SUL1, which encodes a high-affinity sulfate
sulfate
transporter [23].

is observed

This effect has also been observed in multicellular systems. An experimental evolution study of
Arabidopsis thaliana grown under stress conditions of high heat and salicylic acid showed
increased CNV formation [24]. Similar results were found in an experimental evolution study in
Caenorhabditis elegans that selected for recovered fecundity following inbreeding and muta-
gen application, where an increase in the frequency of copy number change was observed
during the adaptive recovery stage [25]. Interestingly, CNVs
in replicate populations were
identified at the same genome regions but had different breakpoints, suggesting recurrent
adaptation [25].

Trends in Plant Science, April 2019, Vol. 24, No. 4 355

in response to strong selective
In the natural environment, adaptive CNVs are also found
pressures. For example, amplifications of P540 genes have conferred insecticide resistance in
[26–28]. Given that many species
aphids and multiple disease vector mosquito species
undergo a period of strong selective pressure during domestication, typically in the postdo-
mestication diversification stage, CNVs could have been an
important source of genetic
diversity underlying adaptation during domestication.

in detection methodologies

CNVs Are Widespread in Domesticated Species
The results of experimental evolution experiments suggest that CNVs contribute to the rapid
adaptation associated with domestication and during population expansion of the domesti-
cated species. Advances
in Box 1), reduced
sequencing costs, and proliferation of sequencing data have expanded CNV studies, and
CNVs have been described in most major crop plant species, including rice, maize, potato,
soybean, barley, cucumber, melon, apple, and grapevine (Table 1, Key Table) [29–37]. Not
surprisingly, they have also been examined in domesticated animal species, such as silkworm,
sheep, goat, pig, chicken, cow, horse, and dog (Table 1) [22,38–43]. These studies demon-
strate that CNVs are a pervasive source of genetic variation
in domesticated taxa, and
examples from both plants and animals serve to highlight both common and contrasting
features of CNVs in both groups.

(summarized

Early studies of CNVs in domesticated species used few samples, although they nevertheless
provided key insights. In rice (Oryza sativa), for example, whole-genome comparisons of two
cultivars found 641 CNVs ranging in size from 1.1 kb to 180.7 kb [44]. An analysis of two inbred
(cid:1)400 genomic regions exhibiting duplications and pervasive
lines in maize (Zea mays) found
PAV affecting more than 700 genes [45].

increasing sample sizes resulted

in more thorough catalogs of CNVs and other
However,
structural variants. Later analysis of 11 maize and 14 wild relative teosinte
individuals, for
example, found 3889 CNVs, most of which were segregating in both maize and teosinte [46]. In
the case of rice, a recent study of 3010 rice varieties identified thousands of deletions and
hundreds of duplications affecting between 100 bp and 1 Mb [47]. Indeed, there are often
significant inconsistencies in the results of CNV analyses from different studies within the same
species due to differences in sample sizes, breeds used, and methodologies of CNV detection
(Box 1) [22,48,49]. This was highlighted in a recent analysis, albeit in a domesticated animal
species, which characterized CNVs in European cattle (Bos taurus) populations and compared
the results to 18 previous studies [39]. Prior studies had identified between 27 and 3438 CNVRs
and, of those data sets, 6–46% of CNVRs overlapped with CNVRs discovered in the present
study [39]. Altogether, this analysis
indicated that CNVs may affect as much as 63 Mb of
the genome, and that cattle have a higher level of CNV diversity than a single study would
predict [39].

The role that CNVs have in the evolution of domesticated taxa is becoming clear. Over past
three decades, considerable effort has been applied to identify the causal mutations and genes
associated with domestication and diversification traits [1,3]. We compiled an exhaustive list of
genes from the literature and found 39 examples where CNVs appear to have a role in trait
evolution in plant and animal domesticated species (Table 1).

The size of these CNVs associated with domestication and diversification ranged from
(cid:1)1 kb to
(cid:1)1 Mb. Plant domestication CNVs affected both domestication and diversification traits,
whereas animal CNVs were all associated with postdomestication diversification traits. Plant

356

Trends in Plant Science, April 2019, Vol. 24, No. 4

Key Table
Table 1. Examples of CNVs Affecting Domestication Traitsa

Locus

Ppd-B1

Ppd-D1a

Vrn-A1

Fr-A2

Rht-D1c

Sub1a

GL7

qSW5/
GSE5

SNORKEL1
SNORKEL2

Pup1

sh1

Sh1

Species

Type

Wheat

mCNV: (cid:1)25 kb

Deletion: 2 kb

mCNV: (cid:1)30 kb

mCNVb

Duplication: >1 mb

Insertionb

Duplication: 17.1 kb

Phenotype

early flowering

Photoperiod insensitivity
(short day growth)

Increased vernalization
requirement

Description

Pseudo-response regulator (Ppd-B1)

Pseudo-response regulator (Ppd-D1a)

Trait type

Diversification

Diversification

Refs

[62]

[79]

MADS-box transcription factor

Diversification

[62]

Frost resistance

Transcription factor, C-repeat Binding Factor (CBF-A14)

Dwarf phenotype; increased
yield

DELLA protein, gibberellic acid insensitive

Submergence tolerance

Ethylene receptor

Grain length

Uncharacterized gene function, homologous to LONGIFOLIA
in Arabidopsis

Domestication/
Diversification

Grain width

GSE5, plasma membrane-associated protein

Deletion: 2 alleles: 950
and 1212 bp

Rice

Insertion: 20.9 kb

q-AG-9-2

Insertionb

Submergence tolerance

Ethylene response factor; transcription factor

Diversification

[84]

Anaerobic germination
tolerance

Trehalose-6-phosphate phosphatase (OsTPP7), sugar
signaling and metabolism

Insertion: 90 kb

Low phosphorous tolerance

PSTOL receptor-like cytoplasmic- kinase

African Rice

Deletion: 30 kb

Sorghum

Deletion: 2.2 kb

Shattering

YABBY transcription factor

Seed shattering

YABBY-like transcription factor

Soybean

mCNV: 31 kb

Rhg1-b

Resistance to cyst nematode
disease

Multiple genes: alpha-SNAP involved in snare membrane
traffic, wound-inducible protein 12 (WI12), a
predicted amino acid transporter

Resistance to leaf rust
disease

Resistance to head smut
disease

mCNVb

Rp1

Cluster of leucine-rich repeat high CN haplotypes

Diversification

[88]

Insertion: 147 kb

ZmWAK

Multiple receptor-like kinase alleles

Diversification

[89]

mCNV: 30 kb

MATE1

Aluminum toxicity resistance

Multidrug and toxic compound extrusion 1 (MATE1)

Duplication: (cid:1)1.5 kb

Tunicate (TU)

Pod corn

ZMM19MADS-box gene

mCNVb

mCNVb: (cid:1)6 kb

Duplication: 22 kb

Bot1

HvFT1

VRN-H1

Boron toxicity resistance

Boron efflux transporter (Bot1)

Flowering time

Mobile florigen signaling protein

Freezing tolerance

C-repeat binding factors (CBF2A-CBF4B)

Diversification

Diversification

Diversification

Diversification

Diversification

[67]

[69]

[66]

[90]

[91]

T
r
e
n
d
s

i

n
P
a
n
t

l

i

S
c
e
n
c
e
,

A
p
r
i
l

2
0
1
9
,

V
o

l
.
2
4
,

N
o

.
4

3
5
7

Maize

Barley

Diversification

Diversification

Diversification

[80]

[81]

[74]

[68]

[82]

Domestication/
Diversification

Diversification

[83]

Diversification

Domestication

Domestication

Diversification

[85]

[86]

[52]

[87]

3
5
8

T
r
e
n
d
s

i

n
P
a
n
t

l

i

S
c
e
n
c
e
,

A
p
r
i
l

2
0
1
9
,

V
o

l
.
2
4
,

N
o

.
4

Goat

Pig

Dog

Table 1. (continued)

Species

Type

Cucumber

Duplication: 30.2 kb

Female (F) locus

Gynoecy

Locus

Phenotype

Description

Retrogene insertion: 24.7 kb

SUN

Elongated fruit shape

IQ67 domain-containing family, function uncharacterized

Tomato

Deletion: 14 kb

CSR-D

Fruit weight

Truncated cell size regulator (CSR-D), uncharacterized protein

Multiple genes, all likely flowering regulatory:
aminocyclopropane-1-carboxylic acid synthase gene (ACS1);
ethylene synthesis; truncated myb transcription factor
(Csa6G496960); branched-chain amino acid
aminotransferase (Csa6G496970)

Transcription factor controls switch from vegetative to
flowering state

Agouti signaling protein (ASIP), S-adenosylhomocysteine
(AHCY), and itchy homolog E3 ubiquitin protein ligase
promoter (ITCH)

Mast/stem cell growth factor receptor linked to tyrosine kinase
receptor genes

Mast/stem cell growth factor receptor

Major lauric acid (medium-chain fatty acid) omega
hydroxylase, lipogenesis

Common bean

Deletion allele: 5840 bp
Insertion allele: 4171 bp

PvTFL1y

Determinate growth

Sheep

Duplication > 100 kb

Ovine ASIP, AHCY

White coat

Duplication 190 kbOvine kb

ASIP, AHCY

ASIP, AHCY

Duplication: 450 kb

Duplication: (cid:1)480 kb

KIT

KIT

Cattle

mCNV: (cid:1)1562 bp

CYP4A11

White coat

Coat Color

White coat

Body fat

Duplication: (cid:1)133 kb

FGF3, FGF4, FGF19,
ORAOV1

Ridgeback

FGF, embryonic development; oral cancer overexpressed
(ORAOV1;) function uncharacterized

Retrogene insertion: 5 kb

fgf4

Chondrodysplasia (short
legs)

FGF4 retrogene

mCNVb

AMYB2

Starch diet

AMYB2, pancreatic amylase

Duplication: 98 kb

Intergenic region

Blue eyes

Amplification: 3.2 kb

SOX5

Peacomb (cold tolerance)

Duplication: 176 kb

PRLR and SPEF2

Late feathering

Intergenic region adjacent to Hox gene ALX4, which has role in
eye development

Intron 1 of SOX5 is a SRY-related HMG box family of
transcription factors

Prolactin receptor (PRLR); inhibits follicle activation; sperm
flagellar protein 2 (SPEF2), thought to be involved in signal
transmission

Chicken

Duplication: 130 kb

EDN3

Fibromelanosis
(pigmentation)

Endothelin 3 gene (EDN3), receptor, melanoblast/melanocyte
mitogen

Duplication: 20 kb

Duplex-comb

Comb shape

Silkworm

mCNVb

CBP

Cocoon color

Duplication upstream of eomesodermin (EOMES), a t-box
transcription factor

Carotenoid-binding protein (CBP); cystolic transporter of
carotenoid pigments

aInsertion and deletion alleles are distinguished by the state that is associated with the phenotype.
bSize of CNV varies, is not precisely known, or amplification units are of variable size.

Trait type

Diversification

Refs

[29]

Diversification

Domestication/
Diversification

[10]

[92]

Domestication

[53]

Diversification

[54]

Diversification

Diversification

Diversification

Diversification

[55]

[93]

[94]

[95]

Diversification

[56]

Diversification

[96]

Diversification

Diversification

[77]

[97]

Diversification

[98]

Diversification

[99]

Diversification

[100]

Diversification

[101]

Diversification

[41]

CNVs found in 14 genes were associated with duplications or amplifications, while 11 genes
had CNV insertions or deletions; by contrast, of the 14 animal genes identified, all but one were
sequence duplications and/or amplifications. Of the duplications and/or amplifications, 22 were
tandem duplications. The prevalence of tandem duplications
in known crop CNVs may be
find tandem rather than dis-
because previous QTL-mapping techniques made
it easier to
persed duplications. Plants also tend to have higher genetic redundancy and, thus, may be
more robust to deletion and/or PAV mutations; thus, they may be able to better tolerate the
deleterious nature of most CNVs.

There were also some differences in the types of domestication and/or diversification genes
affected by CNVs in plants versus animals. Crop CNVs had a more diverse array of functions,
with CNV mutations found in transcription factors related to photoperiod signaling, develop-
ment, stress tolerance, and resistance genes (R-genes). By contrast, animal CNVs were largely
found in genes that encode growth factors and receptors, and genes related to development.

include seed nonshattering or suppression of seed dormancy and,

CNVs and Domestication Traits and Genes
Domestication traits are those that distinguish a domesticated species from its wild progen-
itors, and are the requisite traits for cohabitation with human societies. In crops, common
domestication traits
in
animals, critical changes occur in behavioral traits [1–3]. The genes underlying domestication
‘domestication genes’, are thought to have arisen early during the evolution of crop
traits, or
and livestock species (or may even be present in the wild ancestor at low frequencies) [3,4].
Therefore, examining genetic differentiation between wild ancestors and domesticates has long
been a strategy to discern the underlying genetic causes of domestication. These approaches
are also SNP-centric, and CNVs between wild species and domesticates could further illumi-
nate the genetic basis of domestication.

Domestication traits are common to all members of a domesticated species