The Plant Journal (2018) 95, 487–503 doi: 10.1111/tpj.

13964

Structural variation and rates of genome evolution in the

grass family seen through comparison of sequences of
genomes greatly differing in size
Jan Dvorak1,*, Le Wang1, Tingting Zhu1, Chad M. Jorgensen1, Karin R. Deal1, Xiongtao Dai2, Matthew W. Dawson2,
Hans-Georg Mu € ller2, Ming-Cheng Luo1, Ramesh K. Ramasamy1, Hamid Dehghani1,3, Yong Q. Gu4 , Bikram S. Gill5,
Assaf Distelfeld , Katrien M. Devos , Peng Qi , Frank M. You , Patrick J. Gulick and Patrick E. McGuire1
6 7,8 7,8 9 10

1
Department of Plant Sciences, University of California, Davis, CA, USA,
2
Department of Statistics, University of California, Davis, CA, USA,
3
Department of Plant Breeding, Faculty of Agriculture, Tarbiat Modares University, Tehran, Iran,
4
Crop Improvement & Genetics Research, USDA-ARS, Albany, CA, USA,
5
Department of Plant Pathology, Kansas State University, Manhattan, KS, USA,
6
School of Plant Sciences and Food Security, Tel Aviv University, Tel Aviv, Israel,
7
Institute of Plant Breeding, Genetics and Genomics (Department of Crop & Soil Sciences), University of Georgia, Athens,
GA, USA,
8
Department of Plant Biology, University of Georgia, Athens, GA, USA,
9
Agriculture & Agri-Food Canada, Morden, MB, Canada, and
10
Department of Biology, Concordia University, Montreal, QC, Canada

Received 18 February 2018; revised 4 May 2018; accepted 8 May 2018; published online 16 May 2018.
*For correspondence (e-mail jdvorak@ucdavis.edu).

SUMMARY
Homology was searched with genes annotated in the Aegilops tauschii pseudomolecules against genes
annotated in the pseudomolecules of tetraploid wild emmer wheat, Brachypodium distachyon, sorghum
and rice. Similar searches were performed with genes annotated in the rice pseudomolecules. Matrices of
collinear genes and rearrangements in their order were constructed. Optical BioNano genome maps were
constructed and used to validate rearrangements unique to the wild emmer and Ae. tauschii genomes.
Most common rearrangements were short paracentric inversions and short intrachromosomal transloca-
tions. Intrachromosomal translocations outnumbered segmental intrachromosomal duplications. The densi-
ties of paracentric inversion lengths were approximated by exponential distributions in all six genomes.
Densities of collinear genes along the Ae. tauschii chromosomes were highly correlated with meiotic recom-
bination rates but those of rearrangements were not, suggesting different causes of the erosion of gene
collinearity and evolution of major chromosome rearrangements. Frequent rearrangements sharing break-
points suggested that chromosomes have been rearranged recurrently at some sites. The distal 4 Mb of the
short arms of rice chromosomes Os11 and Os12 and corresponding regions in the sorghum, B. distachyon
and Triticeae genomes contain clusters of interstitial translocations including from 1 to 7 collinear genes.
The rates of acquisition of major rearrangements were greater in the large wild emmer wheat and
Ae. tauschii genomes than in the lineage preceding their divergence or in the B. distachyon, rice and sor-
ghum lineages. It is suggested that synergy between large quantities of dynamic transposable elements
and annual growth habit have been the primary causes of the fast evolution of the Triticeae genomes.

Keywords: synteny, collinearity, annual, translocation, inversion, chromosome rearrangement.

INTRODUCTION
The rates of gene evolution are typically measured in the et al., 1985). Measuring the rate of change at the level of
numbers of nucleotides substituted per unit of time, which the entire nuclear genome is more complicated because
is embodied in the molecular clock concept (Hasegawa the eukaryotic genome is not a homogeneous entity. It

© 2018 The Authors 487

The Plant Journal © 2018 John Wiley & Sons Ltd
488 Jan Dvorak et al.
consists of genes, transposable elements, satellites and
other components, each evolving with its own pace and
mechanism.
Even evolution of gene space may not be homogeneous.
Individual genes are subjected to duplications, deletions,
errors in recombination and transpositions leading to gene
copy number variation (Massa et al., 2011) and loss of
collinearity of individual genes between homoeologous
chromosomes (Luo et al., 2017). Homoeologous chromo-
somes are also subjected to rearrangements of blocks of
collinear genes, such as inversions, translocations, dele-
tions and duplications, resulting in structural variation. We
will consider erosion of collinearity and chromosome rear-
rangements as two discrete forms of genome evolution
throughout this paper.
Initial studies of plant genome evolution were based on
comparative genetic maps. Although the utility of genetic
maps was limited by their low resolution, they neverthe-
less provided many important insights into genome evolu-
tion in plants (Gale and Devos, 1998; Ming et al., 1998;
Brubaker et al., 1999; Paterson et al., 2004; Luo et al., 2009,
2013; Wu and Tanksley, 2010). Today, genetic and physical
maps have been largely replaced by genome sequences
(Bowers et al., 2003; Tang et al., 2008; Paterson et al.,
2009; Town et al., 2011). However, with the sole exception
of conifer genomes (Birol et al., 2013; Nystedt et al., 2013;
Neale et al., 2014, 2017b; Stevens et al., 2016), all
sequenced genomes have until recently been relatively
small. For example, the largest grass genome with a refer-
ence-quality genome sequence has been the paleote-
traploid, 2.3-Gb maize genome (Schnable et al., 2009). The
recent publications of genome sequences of barley
(Mascher et al., 2017), Aegilops tauschii (Luo et al., 2017;
Zhao et al., 2017) and tetraploid wild emmer wheat (Triti-
cum turgidum ssp. dicoccoides, genomes AABB) (Avni
et al., 2017), with genome sizes ranging from 4.3 to 12 Gb,
represent a paradigm shift in genome sequencing. These
genome sequences make possible comparisons of related
genomes greatly differing in size with the hope that such
comparisons may provide insights into plant genome
evolution.
Barley, Ae. tauschii and wild emmer wheat are members
of the tribe Triticeae of the grass family (Poaceae). The lat-
ter two are the closest wild relatives of common (bread)
wheat (T. aestivum, genomes AABBDD). Aegilops tauschii
is the progenitor of the wheat D genome (Kihara, 1944;
McFadden and Sears, 1946; Nesbitt and Samuel, 1996;
Wang et al., 2013) and wild emmer wheat is the progenitor
of domesticated tetraploid wheat (T. turgidum, genomes
AABB). Hybridization of domesticated tetraploid wheat
with Ae. tauschii produced hexaploid bread wheat (Dvorak
et al., 2012). The genomes of wild emmer wheat and
Ae. tauschii together are therefore equivalent to the gen-
ome of bread wheat but are preferable over the bread
The Plant Journal © 2018 John Wiley & Sons Ltd, The Plant Journal, (2018), 95, 487–503
wheat genome for comparative genomics because they
have not been impacted by artificial selection that accom-
panied wheat domestication and improvements.
The tetraploid genome of wild emmer wheat consists of
the A genome, which was contributed by diploid einkorn
wheat T. urartu and the B genome, which was contributed
by an extinct or undiscovered diploid species closely
related to Ae. speltoides (Dvorak and Zhang, 1990; Dvorak
et al., 1993). The hexaploid genome of bread wheat is thus
a product of convergence of three diploid Triticeae lin-
eages, those of the A, B and D genomes. Their divergence
was estimated to have occurred between 1.4 and 4.5 mil-
lion years ago (MYA) (Huang et al., 2002; Dvorak and
Akhunov, 2005; Gornicki et al., 2014; Middleton et al.,
2014; Bernhardt et al., 2017), although a time as ancient as
7 MYA has also been suggested (Marcussen et al., 2014).
An average divergence time of 3 MY will be used through-
out this study. An array of approaches has been used to
reconstruct the phylogeny of the three lineages. Phyloge-
netic trees placing the A-genome divergence basally (Hsiao
et al., 1995; Yamane and Kawahara, 2005; Li et al., 2015;
Bernhardt et al., 2017; Jorgensen et al., 2017) dominate
these efforts.
Comparative genomics has suggested three variables
that may affect the rates of plant genome evolution: (i) the
genome size and the content of dynamic transposable ele-
ments (TEs), (ii) homologous recombination (HR) and (iii)
generation length (life history):
(i) The comparison of a high-resolution genetic map of
Ae. tauschii, a species with a large genome (4.3 Gb),
with the sequences of small genomes of rice (Oryza
sativa) and sorghum (Sorghum bicolor), 0.43 and
0.73 Gb, respectively, (Goff et al., 2002; Paterson et al.,
2009) uncovered 50 inversions and translocations (Luo
et al., 2009). Of these, 2, 8 and 40 occured in the rice,
sorghum and Ae. tauschii lineages, respectively, indi-
cating that the large Ae. tauschii genome has been
acquiring inversions and translocations more quickly
than the smaller rice and sorghum genomes. Compar-
isons of the complete genome sequence of the three
species produced similar results, although many more
rearrangements were discovered (Luo et al., 2017). In
these comparisons, the numbers of rearrangements
mirrored the genome sizes and it was therefore tempt-
ing to attribute the rate of genome evolution to gen-
ome size (Luo et al., 2009). It was recently pointed out
that if this were true the conifer genomes, which are
even larger than the Triticeae genomes, should be
evolving faster than the Triticeae genomes (Luo et al.,
2017). However, conifers are well known for slow gen-
ome evolution (Nystedt et al., 2013; Neale et al.,
2017a). Revealing in this context was the comparison
of unique k-mer plots of the Ae. tauschii genome with
© 2018 The Authors

those of other genomes including that of the much lar-
ger pine genome (Luo et al., 2017). The fast evolving
Ae. tauschii genome had a lower percentage of unique
k-mer sequences than any other genome including the
pine genome, indicating that Ae. tauschii TEs show an
unprecedented homogeneity. This analysis suggested
that it is the large content of homogeneous TEs, i.e.
TEs that have been subjected to recent amplification,
which is critical for the rate of genome evolution (Luo
et al., 2017).
(ii) Meiotic HR is distally located in the Triticeae genomes
and its rate precipitously declines in the proximal
direction (Akhunov et al., 2003; Luo et al., 2013, 2017;
Avni et al., 2017). The distal, high HR regions of Trit-
iceae genomes are enriched for non-collinear genes
that originated by gene duplications and gene dele-
tions (Luo et al., 2013, 2017). Meiotic HR rate and the
content of Ae. tauschii genes collinear with genes in
the B. distachyon, rice and sorghum genomes were
negatively correlated (Luo et al., 2013, 2017). Likewise,
the rates of gene deletions and gene duplications
along the centromere–telomere axes of the chromo-
somes of diploid and polyploid Triticeae species were
correlated with recombination rates (Dvorak and Akhu-
nov, 2005).
(iii) The relationship between generation length and the
rate of the molecular clock is well known (Kohne, 1970;
Martin and Palumbi, 1993; Andreasen and Baldwin,
2001; Tuskan et al., 2006; Smith and Donoghue, 2008;
Luo et al., 2015). The same variable may also affect the
rate of genome evolution. Luo et al. (2015) compared
several criteria of genome divergence in three dicotyle-
donous woody perennials and three dicotyledonous
herbs. The genomes of the woody perennials appeared
to be diverging more slowly than those of the herbs
for each criterion used.
Here, we used 38 775 high-confidence (HC) genes anno-
tated in the Ae. tauschii pseudomolecules (Luo et al., 2017)
in homology searches against genes annotated in the
pseudomolecules of the A and B genomes of the wild
emmer wheat (Avni et al., 2017) in order to assess gene
collinearity and discover rearrangements that have taken
place in these genomes since their divergence. Based on a
wild emmer genome size of 12 Gb (Avni et al., 2017) and
the sizes of individual wheat chromosome arms (Dvorak
et al., 1984), we estimate the size of the wheat A genome
to be about 5.2 Gb and that of the B genome about 6.3 Gb.
The results of these homology searches were combined
with similar searches against the small B. distachyon gen-
ome (0.27 Gb; International Brachypodium Genome Initia-
tive, 2010) and those of rice and sorghum (Luo et al.,
2017), to quantify changes in gene collinearity and chro-
mosome rearrangements among the six genomes. The
© 2018 The Authors
The Plant Journal © 2018 John Wiley & Sons Ltd, The Plant Journal, (2018), 95, 487–503
Rates of genome evolution in grasses 489
genomes represented three grass subfamilies: Panicoideae
(sorghum), Oryzoideae (rice) and Pooideae (B. distachyon,
Ae. tauschii and wild emmer wheat) and the study covered
an evolutionary time frame spanning more than 50 million
years.
When needed, rearrangements were subjected to valida-
tion with optical BioNano genome (BNG) maps (Hastie
et al., 2013). A BNG map consists of contigs assembled
from overlaps among high-molecular-weight DNA mole-
cules digested with a single-strand restriction endonucle-
ase (nickase). The nicks are labeled with fluorescent
nucleotides and distances between them are optically mea-
sured on stretched DNA molecules (Xiao et al., 2007).
Because a BNG map is produced independently of genome
sequencing, it can be used to validate a genome sequence
assembly (Hastie et al., 2013; Luo et al., 2017).
Validated rearrangements were assigned to branches of
a grass phylogenetic tree (International Brachypodium
Genome Initiative, 2010) and rates with which rearrange-
ments were accumulated in individual branches per MY
were computed. To assess the extent to which these rate
estimates were affected by the phylogenetic distances
among compared genomes, a dataset of homology
searches with 42 004 HC genes annotated in the rice pseu-
domolecules against genes annotated in the Ae. tauschii,
B. distachyon and sorghum pseudomolecules produced
earlier (Luo et al., 2017) and augmented here with similar
homology searches against genes annotated in the wild
emmer pseudomolecules (Avni et al., 2017) was analyzed.
The rates of genomic changes in terms of the accumula-
tion of major rearrangements per MY were computed and
compared with those obtained using the Ae. tauschii HC
genes as queries in homology searches. The two sets of
estimates of rates of genomic changes were jointly consid-
ered in relation to genome size, recombination rate and life
history including generation time.
RESULTS
Genome comparisons
In each genome comparison it was determined for each
target gene if it was in a collinear position relative to its
neighbors on the pseudomolecule. If the determination
was affirmative, the gene was considered to be at a colli-
near location in that genome comparison. If a block of col-
linear genes along homoeologous chromosomes was in a
different orientation or place, that genomic change was
considered a chromosome rearrangement and its cause
(inversion, translocation, duplication, etc.) was determined.
Gene collinearity
To quantify gene collinearity, 38 775 HC genes annotated
in the Ae. tauschii pseudomolecules v4.0 (Luo et al., 2017)
were used as queries in homology searches against genes
490 Jan Dvorak et al.

annotated in the pseudomolecules of the A and B genomes in the proximal, low-recombination regions of the
of wild emmer wheat and results were combined with sim- Ae. tauschii pseudomolecules, about 70% of the genes and
ilar searches against genes annotated in the pseudo- declined towards the distal, high-recombination regions to
molecules of B. distachyon, rice and sorghum performed about 20 to 30% of the genes (Figure 1a). The numbers of
earlier (Luo et al., 2017). The top hits in wild emmer wheat, collinear genes along the wild emmer wheat pseudo-
B. distachyon, rice and sorghum were recorded and genes molecules were negatively correlated with meiotic HR rates
collinear with genes along the Ae. tauschii pseudo- computed by Luo et al. (2017). For the wild emmer A and
molecules in these five genomes were color coded in a B genomes, Pearson correlation coefficients were
resulting matrix (Table S1). The highest numbers of genes r = 0.59 and 0.58, respectively (P < 0.0001, N = 181). In
collinear with the Ae. tauschii HC genes were in the wild homoeologous groups 1, 2 and 3 and the long arms of 7A
emmer wheat A and B genomes, respectively, 19 370 and and 7B, the numbers of genes collinear with the
19 415 genes (50.0 and 50.0%) and the lowest number was Ae. tauschii pseudomolecules were significantly greater in
in the sorghum genome, 13 066 (33.7%) genes (Table 1). the B genome than in the A genome (Figure 1a). The differ-
To assess to what extent these numbers were affected ences between the A- and B-genome homoeologues in the
by the choice of the query genome, a dataset based on remaining homoeologous groups were not statistically
homology searches with 42 004 HC genes annotated in the significant.
rice pseudomolecules v7.0 against genes annotated in the Dot plots were constructed to visualize synteny of the
Ae. tauschii, B. distachyon and sorghum pseudomolecules Ae. tauschii pseudomolecules with those of wild emmer
was downloaded from Luo et al. (2017) and combined with wheat (Figure 1b). Except for 4A and the tips of 5AL and
similar searches against genes annotated in the wild 7BS, which harbor reciprocal translocations, dot plots
emmer pseudomolecules performed here. Results aligned the homoeologous pseudomolecules along their
obtained with this dataset (Table S2) and those obtained entire lengths. Paracentric inversions appeared to be the
using genes annotated in the Ae. tauschii pseudo- principal type of chromosome rearrangement shown by
molecules as queries (Table S1) were similar (Table 1). the dot plots. They were apparent in arms 1AS, 1BL, 2AL,
The numbers of collinear genes obtained when the rice 4AL, 4BL, 5BL, 7AL and 7BS. Duplications were apparent
genes were used as queries and the A-, B- and Ae. tauschii between the long arms of the chromosomes of homoeolo-
genome genes were used as targets ranged from 13 520 gous groups 1 and 3 and 2 and 6 and were consistent with
(32.2%) to 12 828 (30.5%). These numbers were close to major self-synteny blocks within the Ae. tauschii genome
13 401 (34.6%) genes found to be collinear in the homol- (Luo et al., 2017) due to a pan-grass whole genome dupli-
ogy searches using the Ae. tauschii genes as queries and cation (WGD) (Paterson et al., 2004).
rice genes as targets (Table 1). In total, 1989 rearrangements were detected in the six
The numbers of genes collinear with the Ae. tauschii genomes. For 1126 of these, the presence of the rearrange-
genes along wild emmer pseudomolecules were quantified ment could be assessed in all genomes, making it possible
in non-overlapping windows of 200 genes. Expressing to assign it to a specific phylogenetic branch (Table S1).
gene collinearity per gene rather than per Mb made the This could not be accomplished for the remaining 863 rear-
collinearity estimates along chromosomes independent of rangements and these were not assigned to a phylogenetic
the variation in gene density that exists in the Ae. tauschii branch. A vast majority of these unassigned rearrange-
genome (Luo et al., 2017). Disregarding the centromeric ments involved two (686) or three (80) genes. Only 46 rear-
regions (gaps in the profiles), the highest collinearity was rangements that could not be assigned to a branch were
major rearrangements, involving four or more genes.
Table 1 Numbers and percentages of genes collinear in the target
Based on this analysis, only rearrangements involving four
genomes with genes annotated in the Ae. tauschii or O. sativa or more genes were considered in the estimation of gen-
pseudomolecules ome evolution rates.
The rearrangements assigned to branches of the phylo-
Ae. tauschii
genetic tree were classified as inversions, intrachromoso-
genes O. sativa genes
mal translocations and interstitial or terminal
Target genome No. % No. % interchromosomal translocations, duplications and nested
chromosome insertions (NCIs) and by the number of colli-
Wild emmer A 19 370 50.0 13 013 31.0
Wild emmer B 19 415 50.0 12 828 30.5 near genes involved (2, 3 and >3). An NCI is an insertion of
Ae. tauschii – – 13 520 32.2 an entire chromosome into another chromosome (Luo
B. distachyon 14 601 37.9 16 863 40.1 et al., 2009). For each inversion and translocation, the rear-
O. sativa 13 401 34.6 – – rangement type, the starting nucleotides of the first and
S. bicolor 13 066 33.7 16 503 39.3
last Ae. tauschii HC gene involved in the rearrangement
Query genes (total) 38 775 42 004
and the branch of the grass phylogenetic tree in which the

© 2018 The Authors

The Plant Journal © 2018 John Wiley & Sons Ltd, The Plant Journal, (2018), 95, 487–503
 Rates of genome evolution in grasses 491

Figure 1. Gene collinearity.

(a) Quantification of collinearity along the Ae. tauschii pseudomolecules (horizontal axes) with genes annotated in the A- and B-genome pseudomolecules of
wild emmer wheat. Each profile presents the number of collinear genes in a series of 200-gene non-overlapping windows (vertical axis). The P-value (paired t-
test) for the entire chromosome is indicated at the lower right for each pair of pseudomolecule profiles. In addition, for the profiles of 7A and 7B, the P-values of
the short and long arms are also shown. If P < 0.05, the B-genome profile was more closely related to that of the Ae. tauschii genome profile than was the A-
genome profile. The gaps in the profiles are the locations of centromeric regions, in which collinearity was not investigated.
(b) Dot plots comparing the 14 wild emmer wheat pseudomolecules with the seven of Ae. tauschii. Each dot consists of a sequence of three or more collinear
genes. The plots are oriented with the tips of the short arms to the left (x-axis) and bottom (y-axis). Large gaps in the profiles are centromeric regions.

rearrangement originated can be found in Table S1. There
were 176, 158 and 125 rearrangements unique to the A and
B genomes of wild emmer wheat and the genome of
Ae. tauschii, respectively.
To determine if the large numbers of rearrangements
unique to the three Triticeae genomes were real or errors
in genome sequence assemblies, unique rearrangements
involving three or more collinear genes in the A, B and
Ae. tauschii genomes were validated with BNG maps (Fig-
ure 2a). Rearrangements involving only two collinear
genes could not be validated because they were usually
below the resolution of the BNG map. They were excluded
from the subsequent analyses. A BNG map for wild emmer
wheat accession Zavitan, used in wild emmer wheat
sequencing, was constructed (Table 2) to validate wild
emmer wheat rearrangements. BNG maps for Ae. tauschii
accessions AS75 and CIae 1 were constructed (Table 2)
and added to the existing BNG maps for Ae. tauschii
accessions AL8/78 and CIae 23 (Luo et al., 2017). Thus, to
date, Ae. tauschii rearrangements have been validated
with four BNG maps.
Of the 421 wild emmer wheat and Ae. tauschii rear-
rangements subjected to validation with BNG maps, 297
were validated (Figure 2b). Rearrangements that failed vali-
dation were left in Table S1, but their failure to be vali-
dated was indicated and they were excluded from further
analyses. Altogether, 938 inversions, 161 translocations, 13
NCIs, three deletions and 11 segmental duplications
remained assigned to individual branches of the grass phy-
logenetic tree (Tables S1 and S3). The length measure-
ments (Table S1) of all these rearrangements among the
five taxa in the phylogenetic tree were gauged on the dis-
tances between genes in the Ae. tauschii pseudo-
molecules.
Inversions
Of the inversions involving ≥3 collinear genes, 167
involved three collinear genes and 460 involved >3

© 2018 The Authors

The Plant Journal © 2018 John Wiley & Sons Ltd, The Plant Journal, (2018), 95, 487–503
492 Jan Dvorak et al.

Figure 2. Validation of inversions with a BioNano genome (BNG) map.

(a) Examples of confirmed (top) and not confirmed (bottom) inversions in the wild emmer wheat A genome. If an inversion deduced by comparison with other
genome sequences is confirmed, the sequence scaffold and BNG map will be identical as the inversion will be present in both. If an inversion is not confirmed,
the scaffold and BNG map will differ, since the inversion is an artefact of assembly and is present only in the sequence scaffold. The green rectangles symbolize
genome fragments and the blue rectangles symbolize BNG contigs. The red lines between the sequence fragments and the BNG contigs connect corresponding
restriction sites and the blue numbers are sequence coordinates in Mb. The four vertical red arrows in each alignment indicate the positions of genes used for
rearrangement validation. The two internal genes are within a putative inversion close to the breakpoints and the two external genes are outside of the inver-
sion. In the top alignment, the order of restriction sites in the region of the 4A pseudomolecule marked by the four genes was collinear with the order of restric-
tion sites in the BNG contig, thus validating the inversion detected in the sequence. The bottom alignment shows a disagreement between the sequence and
BNG map. The order of restriction sites in the region of the pseudomolecule marked by the four genes was inverted relative to their order in the BNG contig,
indicating an error in the assembly of the 3A pseudomolecule.
(b) Percentages of validated inversions in the wild emmer wheat A and B genomes and the Ae. tauschii (Aet) genome. Percentage values sharing the same letter
(above each column) are not significantly different at the 5% significance level (paired t-test with Bonferroni correction).

Table 2 Characteristics of BNG maps of

Parameter Zavitan AS75 CIae 1 wild emmer wheat and Ae. tauschii
Nickase enzyme Nt.BspQI Nt.BspQI Nt.BspQI
No. DNA molecules put into contigs 3 338 157 1 197 206 1 179 029
N50 DNA molecule length (Kb) 340.5 349.8 343.3
Minimum DNA molecule length (Kb) 180 180 180
Total length of DNA molecules (Gb) 1101 403.8 393.2
No. of genome equivalents 110 96 93
No. contigs 7098 3603 5507
Total length of map (Mb) 10 675 4292 4690
N50 of map (Mb) 2.23 1.72 1.12

Zavitan is an accession of wild emmer wheat and AS75 (China) and CIae 1 (Pakistan) are
accessions of Ae. tauschii ssp. tauschii.

collinear genes (Table S1). There were 115, 91, 55, 103, 36
and 50 paracentric inversions involving ≥3 collinear genes
unique to the A and B genomes of wild emmer wheat and
the genomes of Ae. tauschii, B. distachyon, rice and sor-
ghum, respectively (Tables S3 and 3). The mean lengths of
these genome-unique inversions ranged from 8.0 Mb (me-
dian = 1.0 Mb) in the sorghum genome to 1.6 Mb (me-
dian = 0.6 Mb) in the Ae. tauschii genome, indicating that
most were short (Figure 3). These measurements were
based on the Ae. tauschii pseudomolecules and are there-
fore comparable. The median inversion lengths were
The Plant Journal © 2018 John Wiley & Sons Ltd, The Plant Journal, (2018), 95, 487–503
shorter in the large genomes than in the small genomes,
but the sample size was insufficient to test statistically the
hypothesis. The densities of the genome-unique paracen-
tric inversion lengths were closely approximated by expo-
nential distributions in all six genomes (Figure 3). In the
rice genome, this relationship was accounted for by a sin-
gle distribution and in the remaining five genomes two dis-
tributions were necessary (Figure 3).
The following analyses were made to assess the distri-
bution of paracentric inversions in relation to meiotic HR
rates. First, the location of the most distal gene within each
© 2018 The Authors
 Rates of genome evolution in grasses 493

Table 3 Numbers and rates of inversions and intrachromosomal translocations for branches of the grass phylogenetic tree

Intrachr. transloc.
Branch Time (MY) Inv. 2 genes Inv. 3 genes Inv. > 3 genes >3 genes All minor rearrang. All major rearrang.

A 3 36 (12.00b,c) 25 (8.33b,c) 91 (30.33c) 13 (4.33b) 69 (23.00b) 107 (35.67c)

B 3 32 (10.67a,b) 25 (8.33b) 65 (21.67c) 11 (3.67b) 76 (25.67b) 81 (27.00c)
Aet 3 49 (16.33c) 11 (3.67a,b) 44 (14.67b) 7 (2.33a,b) 64 (21.33b) 61 (20.33b)
AB ? 3 2 6 0 5 8
AetA ? 3 2 4 0 5 4
AetB ? 1 0 10 1 1 12
AetBA 32 28 (0.88a) 32 (1.00a) 76 (2.38a) 9 (0.28a) 69 (2.16a) 98 (3.06a)
AetABBd 12 12 (1.00a) 15 (1.25a) 29 (2.42a) 4 (0.33a) 28 (2.33a) 36 (3.00a)
Bd 35 61 (1.73a,b) 21 (0.60a) 82 (2.34a) 15 (0.43a) 91 (2.60a) 105 (3.00a)
Os 47 28 (0.60a) 16 (0.34a) 20 (0.43a) 4 (0.06a) 46 (0.98a) 28 (0.60a)
Sb 53 58 (1.09a) 17 (0.32a) 33 (0.62a) 16 (0.30a) 77 (1.45a) 54 (1.02a)
Total 309 158 444 77 520 592

inv., inversion; intrachr., intrachromosomal; transloc., translocation; rearrang., rearrangements.

Rates are per million years (in parentheses) for each type of structural change. Rates sharing the same letter suffix are not significantly dif-
ferent at the 5% significance level. See Figure 4 for identity of phylogenetic tree branches. Time estimates (column 2) are means of ranges
reported in (International Brachypodium Genome Initiative, 2010). For the divergence time of the A, B and Ae. tauschii genomes, see
Introduction.

inversion involving >3 collinear genes in the Ae. tauschii
and the wild emmer wheat A and B genomes was deter-
mined on the Ae. tauschii pseudomolecules. The number
of inversion starting points in the three genomes per 200-
gene window was counted and used as a variable in com-
puting the Pearson correlation coefficient. The other vari-
able was the average HR rate in the 200-gene window (Luo
et al., 2017). No correlation between HR rate and the distri-
bution of the inversions in the three genomes was
detected (r = 0.03, P = 0.64, N = 181).
The problem with this analysis is that there were 863
rearrangements, mostly inversions, which could not be
assigned to a lineage. The failure to include these inver-
sions into the correlation analysis could have biased the
data and have caused the absence of correlation. To assess
that, the unassigned inversions that appeared to be located
in the B. distachyon, rice and sorghum genomes were
excluded and the rest was subdivided to those involving
two collinear genes and >2 collinear genes, both groups
were combined with the original dataset and Pearson cor-
relation coefficients were re-computed. For the dataset
including unassigned inversions involving two collinear
genes r = 0.32 (P < 0.0001, N = 181) and for the dataset
including inversions involving >2 collinear genes r = 0.15
(P = 0.04, N = 181). The increase in the size of correlation
coefficients indicated that long inversions (4 or more
genes) do not correlate with HR but short inversions, prin-
cipally those involving two genes, do correlate with HR.
The rearrangement dataset in Table S1 showed cases in
which two different inversions shared a common break-
point, defined as an interval between neighboring collinear
genes, one external and one internal to an inversion break-
point. Because the data were based on comparison of
diverged genomes it was impossible to refine the location
of a breakpoint to the level of nucleotide sequences. To
study this phenomenon, we scored double inversions con-
sisting of two inversions, one nested within the other, while
sharing a breakpoint. For instance, a region of wild emmer
pseudomolecule 1A involving markers AET1Gv20576000
and AET1Gv20583900 is inverted and markers that are
ascending on the 1D pseudomolecule are descending on the
1A pseudomolecule. We name this inversion InvA
(AET1Gv20576000-83900) using a naming convention
described in Experimental procedures. The progression of 3
and 4 collinear markers at each end of this inversion is
ascending indicating the presence of two additional inver-
sions. Each of these inversions, InvA (AET1Gv20576000-200)
and InvA (AET1Gv20583000-900), share one breakpoint with
inversion InvA(AET1Gv20576000-83900) suggesting that the
breakpoints that produced the major inversion also pro-
duced the two additional inversions. We analyzed 363 inver-
sions involving >3 collinear genes for the existence of these
double inversions with the restriction that the nested inver-
sion involved ≥3 collinear genes. We found 22 (6%) of such
cases (Table 4).
Translocations
In total, 79 interstitial intrachromosomal translocations, 19
interstitial interchromosomal translocations, 9 terminal
interchromosomal translocations and 13 NCIs were allo-
cated to branches of the phylogenetic tree (Table S3). Most
of the intrachromosomal translocations involving 3 or
more collinear genes were short. In the A and B genomes
of wild emmer wheat and the Ae. tauschii genome, they
averaged about 7.6 Mb (median = 0.6 Mb), 2.1 Mb (me-
dian = 0.5 Mb) and 1.7 Mb (median = 0.1 Mb), respectively
(too few of them were detected in the remaining genomes
to make their statistical treatment meaningful).

© 2018 The Authors

The Plant Journal © 2018 John Wiley & Sons Ltd, The Plant Journal, (2018), 95, 487–503
494 Jan Dvorak et al.

A B Aet
125
μ1 = 0.93 Mb μ1= 0.93 Mb μ1 = 0.80 Mb
μ2 = 8.87 Mb μ2 =11.91 Mb μ2= 8.44 Mb
α = 0.71 α = 0.87 α = 0.89
100

75

3.22 Mb b 2.48 Mb ab 1.61 Mb a

50 1.05 Mb 0.88 Mb 0.63 Mb

25

0
Count

Bd Os Sb
125
μ1 = 0.21 Mb μ1 = 0.96 Mb
μ2 =24.49 Mb μ2 =36.01 Mb
α = 0.98 α = 0.80
100

75

2.60 Mb b 2.37 Mb ab 8.01 Mb b

50
1.25 Mb 1.28 Mb 1.03 Mb

25

0
0 50 100 150 0 50 100 150 0 50 100 150
Inversion length in Mb

Figure 3. Paracentric inversion lengths.

Densities of the lengths of inversions involving 3 or more collinear genes measured in Mb in the wild emmer wheat A and B genomes and the genomes of Aegi-
lops tauschii (Aet), Brachypodium distachyon (Bd), rice (Os) and sorghum (Sb). The inversion lengths were scaled and gauged on the distances between genes
in the Ae. tauschii pseudomolecules. The curves in red show mixtures of two exponential distributions fitted to the data. A single exponential distribution fit the
rice data. The mean inversion length (top number) and the median (bottom number) are given for each graph. A Kolmogorov–Smirnov test indicated that the
data deviated significantly from normality (P ≤ 0.01). They were therefore transformed (see Experimental procedures) and analyzed with ANOVA for a completely
randomized design which showed a significant genome effect (P ≤ 0.10). Means followed by the same letter do not differ at the 5% significance level (Duncan’s
test). Also given in each graph are the means (l1 and l2) for the mixture components and the mixture proportion (a).

An intrachromosomal translocation of a chromosome TAetBA (AET4Gv20018000-9500; AET5Gv20867400-600), was

segment could either leave the original segment in place
or not, depending on the nature of the translocation pro-
cess and subsequent recombination of the chromosome.
The former results in a segmental intrachromosomal dupli-
cation whereas the latter is a true translocation. Resolution
was sufficient only for the analysis of translocations
unique to the wild emmer wheat and the Ae. tauschii gen-
omes. Of the 26 characterized, 7 (28%) were intrachromo-
somal duplications whereas the rest were true
translocations.
There were three reciprocal translocations 4DS/5DL, 4AL/
5AL and 4AL/7BS, all terminal, in the wild emmer and
Ae. tauschii genomes. The 4DS/5DL reciprocal translocation,
shared by the three Triticeae genomes. The 4AL/5AL and
4AL/7BS reciprocal translocations, TA(AET4Gv20754000-
5100; AET5Gv21126000-200) and TAB(AET5Gv21238100-
9800; AET7Gv20264500-800), respectively, were unique to
the wild emmer genome.
Clusters of translocations were observed between rice
chromosomes Os11 and Os12, sorghum chromosomes
Sb5 and Sb8, Triticeae chromosomes 4 and 5 and within
B. distachyon chromosome Bd4. The translocations
involved regions corresponding to the distal 4 Mb of rice
chromosomes Os11 and Os12. The numbers of the translo-
cations ranged from 84 in sorghum to four in the three Trit-
iceae genomes (Table 5). The translocated segments were

© 2018 The Authors

The Plant Journal © 2018 John Wiley & Sons Ltd, The Plant Journal, (2018), 95, 487–503
 Rates of genome evolution in grasses 495

Table 4 Double inversions indicating recurrent chromosome and the A genome, the Ks values should not differ. This
breaking in individual branches of the phylogenetic tree hypothesis was tested using two AetA inversions involving
Branch N Recurrent Percent eight genes and seven AetB inversions involving 107 genes
(Table S5). No evidence was found for differences between
A 77 7 9.1 the Ks values in the AetA inversions. For the AetB inver-
B 59 2 3.4
sions, Ks values between the Ae. tauschii and the B-gen-
Aet 39 1 2.6
AB 4 0 0.0 ome haplotypes were significantly smaller than the Ks
AetA 1 0 0.0 values between the Ae. tauschii haplotypes and the corre-
AetB 9 1 11.1 sponding A-genome alleles (Table 6). The Ks values were
AetBA 57 6 10.5 therefore consistent with the divergence of the A, B and
AetABBd 22 1 4.5
Ae. tauschii genomes by bifurcations as shown in
Bd 56 2 3.6
Os 13 0 0.0 Figure 4.
Sb 26 2 7.9
Rates of genome evolution
Total 341 22
Inversions, translocations and duplications involving >3
collinear genes and NCIs were combined into a class of
uniformly small, involving from 1 to 7 collinear genes major rearrangements (Tables 3 and S3). The greatest num-
(Table S4). The density of the number of genes per translo- ber of them was present in the A genome (107) and the
cation was approximated by Weibull distribution smallest number was in the rice genome (28) (Table 3). To
(P < 0.01). Distribution fit for individual chromosomes are estimate the rate of change in each branch of the phyloge-
in Table S4. The number of genes per translocated seg- netic tree, the number of major rearrangements was
ment decreased in the proximal direction. Pearson correla- divided by the length of a branch in MY (Table 3). The high-
tion coefficients of the number of collinear genes per est rates were in the A-, B- and Ae. tauschii genome
translocated segment and the distance from the terminus branches, 35.7, 27.0 and 20.3 rearrangements per MY,
ranged from 0.38 to 0.17 (Table S4). These transloca- respectively and the lowest rates were in the rice and sor-
tions were considered exceptional and were not included ghum branches, 0.6 and 1.0 major rearrangement per MY,
into rearrangement quantifications. respectively (Figure 4). The rates in the subfamily Pooideae
were higher than those in Panicoideae and Oryzoideae (Fig-
Phylogeny of the A, B and Ae. tauschii genomes
ure 4 and Table 3). The rate in the B. distachyon branch
We employed 460 inversions involving >3 collinear genes was 3.0 major rearrangements per MY, which was similar
in the reconstruction of the phylogeny of the three Trit- to the rate in the Pooideae branches preceding the diver-
iceae genomes. The resulting tree had a topology similar gence of Brachypoideae and Triticeae (3.0 major rearrange-
to the tree reported earlier (International Brachypodium ments per MY) and preceding the radiation of the A, B and
Genome Initiative, 2010) with the difference that this one Ae. tauschii genomes (3.1 major rearrangements per MY).
included three Triticeae genomes. The branching indicated The slowest rates of genomic change were in the
that the A genome diverged prior to the divergence of the branches most distant from Ae. tauschii (Figure 4). This
B and Ae. tauschii genomes (Figure 4). fact raised a legitimate concern whether the ability to
This pattern of dichotomous genome divergence pre- assign a rearrangement to a phylogenetic branch did not
dicted that the numbers of synonymous nucleotide substi- decline with the phylogenetic distance and was not the
tutions (Ks) of the Ae. tauschii alleles within inversions cause of the variation in the rates among the branches of
shared with the B-genome (category AetB) will be smaller the tree. We therefore reversed the direction of the analysis
than Ks between the same Ae. tauschii alleles and the A- and analyzed collinearity of Ae. tauschii, wild emmer,
genome alleles. For inversions shared by the Ae. tauschii B. distachyon and sorghum genes using the 42 004 rice

Table 5 Numbers of translocations and translocated genes in chromosome segments corresponding to the distal ends of short arms of rice
chromosomes Os11 and Os12 in the three Triticeae genomes and the genomes of B. distachyon, rice and sorghum

Translocations (no.) Genes (no.)

Genome Os12 homoeologue Os11 homoeologue Os12 homoeologue Os11 homoeologue

Triticeae 0 4 0 4
B. distachyum 26 11 35 13
Rice 28 47 54 104
Sorghum 35 49 59 103

© 2018 The Authors

The Plant Journal © 2018 John Wiley & Sons Ltd, The Plant Journal, (2018), 95, 487–503
496 Jan Dvorak et al.

rearrangements assigned to the branches of the tree using

the Ae. tauschii genes as queries (P = 0.001, two-tailed
paired t-test, N= 11). The rates were the most different in
the three Triticeae branches (Figure 4). In the wild emmer
A-genome branch and the Ae. tauschii branch the reduc-
tions were statistically significant (P = 0.008 and P = 0.01,
respectively, two-tailed paired t-test, N = 7). In the rest of
the tree, the differences were minor and not statistically
significant except for the branch preceding the divergence
of Triticeae and Brachypodieae (branch AetABBd)
(P = 0.01, two-tailed paired t-test, N = 7). Despite these dif-
ferences, the analyses were consistent in showing more
than an order of magnitude faster rate of genomic change
in the Triticeae genomes compared to the rice and sor-
ghum genomes and faster rate in the Pooideae branches
than in the Panicoideae and Oryzoideae branches.

Figure 4. Rates of genome evolution.
Rates of genomic change in terms of the number of major chromosome rear-
rangements per MY during the grass family radiation. The phylogenetic tree
is from International Brachypodium Genome Initiative (2010) and rebuilt to
include the wild emmer and Ae. tauschii genomes. A and B represent the
wild emmer wheat genomes and Aet, Bd, Os and Sb are the Ae. tauschii,
B. distachyon, rice and sorghum genomes, respectively. Bootstrap confi-
dence is specified by bold black numbers to the right of branches. The num-
bers in red are the rates of acquisition of major rearrangements by individual
branches. The top red numbers are rates based on using genes annotated in
the Ae. tauschii genome as queries in homology searches and the bottom
red numbers are rates based on using genes annotated in the rice genome
as queries in homology searches. Divergence time in MY is given for each
node. The grass subfamily is indicated for each branch.
genes as queries in homology searches, as described in
the section Gene collinearity (Table S2). In total, 415 major
rearrangements were assigned to the phylogenetic tree
branches, which was significantly fewer than the 594 major
DISCUSSION
Rates of genome evolution
Searches for homology using the Ae. tauschii HC genes as
queries uncovered 594 major rearrangements that could
be allocated into specific branches of the grass phyloge-
netic tree. Of them, 247 (41.7%) were unique to one of the
three Triticeae genomes and have evolved in the past 3
MY. Similar results were obtained using rice genes as
queries in searches for homology. They uncovered 415
major rearrangements that could be allocated into the
branches of the phylogenetic tree and 139 (33.5%) of them
were unique to one of the three Triticeae genomes.
Because major rearrangements unique to the Triticeae
genomes were validated with BNG maps in both datasets,
it is unlikely that this excess of rearrangements in the Trit-
iceae genomes was caused by errors in sequence

Table 6 Average Ks values for genes within shared inversions

Ks

Inversion Chrom. Start (bp) End (bp) Genes (no.) A versus B D versus A D versus B

AetA 1D 4 200 706 4 378 346 5 0.165 0.105 0.110

AetA 1D 60 675 934 60 758 258 3 0.088 0.080 0.083
AetB 1D 54 678 481 55 665 632 8 0.091 0.083 0.088
AetB 2D 77 874 727 78 733 204 4 0.082 0.068 0.062
AetB 2D 629 018 334 629 304 845 4 0.084 0.087 0.084
AetB 3D 96 990 346 109 275 480 46 0.128 0.116 0.089
AetB 3D 109 641 478 118 437 345 41 0.115 0.100 0.067
AetB 4D 17 690 222 17 830 447 5 0.135 0.125 0.095
AetB 7D 73 724 397 73 980 729 2 0.161 0.124 0.082
AetA Mean 0.137a 0.096a 0.100a
AetB Mean 0.118a 0.105a 0.073b

The inversions are identified by chromosome location and starting and ending coordinates. The Ks values are for collinear A-, B- and Aet-
genome genes located within the inversions shared by the Ae. tauschii branch and the A-genome branch (AetA) and by the Ae. tauschii
branch and the B-genome branch (AetB). Means with the same letter suffix in a row are not statistically significant at the 5% probability
level.

© 2018 The Authors

The Plant Journal © 2018 John Wiley & Sons Ltd, The Plant Journal, (2018), 95, 487–503

assembly. Furthermore, the same excess of major rear-
rangements in the Ae. tauschii lineage was evident from
genetic maps (Luo et al., 2009, 2013).
In a previous study (Luo et al., 2017), the rearrange-
ments unique to the Ae. tauschii genome were allocated to
the 35-MY time span, covering the time since the diver-
gence of Triticeae and Brachypodieae to the present. That
produced a rate of 5.1 major rearrangements per MY. By
inserting the wild emmer wheat into the analysis, the 35
MY-timeline was subdivided into 32 MY prior to and 3 MY
after the radiation of the A, B and Ae. tauschii genomes.
The rate prior to the radiation was 3.1 major rearrange-
ments per MY (2.5 using rice genes as queries). A similar
rate, 3.0 major rearrangements per MY (1.9 using rice
genes as queries), was estimated in the branch preceding
the divergence of the Triticeae and Brachypodieae lin-
eages. However, the rate since the radiation was 20.3
major rearrangements per MY (15.0 using rice genes as
queries) in the Ae. tauschii lineage. Even if the time of radi-
ation of the three genomes is pushed back to 7 MYA, using
a conservative time estimate (Marcussen et al., 2014), the
rate (9.4 major rearrangements per MY) would still be
greater than the rate preceding radiation of the three gen-
omes. Equally fast rates, 35.7 and 27.0 major rearrange-
ments per MY (16.3 and 15.0, respectively, using rice
genes as queries), were observed in the wild emmer wheat
A and B genomes, respectively.
The sizes of the three Triticeae genomes ranged from
4.3 Gb for Ae. tauschii to 6.3 Gb for the wild emmer B gen-
ome. The Ae. tauschii and wild emmer genomes were esti-
mated to contain about 84.4 and 82.2% of TEs (Avni et al.,
2017; Luo et al., 2017). Yet, it seems unlikely that the gen-
ome sizes and the content of TEs were the sole causes of
the great increase in the rates of genome evolution after
their radiation 3 MYA. All Triticeae genomes are large
(Dvorak, 2009) and likely contain similar percentages of
TEs. Fast genome evolution would therefore be expected
for the entire tribe if the rates were exclusively a function
of genome size and TE content. If the rate recorded for the
Ae. tauschii branch, 20.3 major rearrangements per MY,
were applicable to the entire tribe, the Ae. tauschii lineage
would be expected to accumulate 203 major rearrange-
ments over the 10-MY period, the estimated divergence
time of the wheat and barley lineages and the age of Trit-
iceae (Dvorak and Akhunov, 2005; Bernhardt et al., 2017).
This number exceeds the total number of 175 major rear-
rangements that the Ae. tauschii lineage accumulated
since the divergence of Triticeae and Brachypodieae 35
MYA. This simple reasoning suggests that the rate of gen-
ome evolution observed in the Ae. tauschii genome accel-
erated recently, as the rate estimates suggest.
What factor(s) in addition to genome size could have
caused this rate acceleration? Aegilops tauschii, wild
emmer wheat and the donors of its A and B genomes are
© 2018 The Authors
The Plant Journal © 2018 John Wiley & Sons Ltd, The Plant Journal, (2018), 95, 487–503
Rates of genome evolution in grasses 497
annuals whereas their progenitors were perennial (Hsiao
et al., 1995). In annual grasses, the generation time is typi-
cally 1–2 years (Wilson et al., 1990). Estimates of the gen-
eration time in perennial grasses are hard to obtain and
undoubtedly vary among species and among populations.
An anecdotal number for a population of Brachypodium
sylvaticum, a perennial grass, suggests that the generation
time is at least an order of magnitude greater than that of
its annual relatives (Haeggstrom and Skyten, 1996). Long
generation time is accompanied by a slow molecular clock
rate (Wilson et al., 1990; Tuskan et al., 2006; Smith and
Donoghue, 2008; Luo et al., 2015) and a slow rate of gen-
ome evolution (Luo et al., 2015). A dependence of genome
evolution on generation length was also observed in mam-
mals. The quickest rate of the accumulation of chromo-
some breaks was observed in the mouse lineage, i.e., in a
lineage involving species with a short generation time
(Murphy et al., 2005). Quantification of meiotic chromo-
some pairing in Triticeae interspecific hybrids indicated
that genomic change has been slow in perennial Triticeae
compared to that in the annual Triticeae (Dvorak and
Zhang, 1992). Another factor that could have affected the
rate of genome evolution in the three Triticeae genomes is
self-pollination, although no relevant data on this relation-
ship are available. Based on currently available evidence
we suggest that the rapid genome evolution in Ae. tauschii
and wild emmer was caused by synergy between large
quantities of active TEs (Luo et al., 2017) and annual repro-
ductive cycle.
This hypothesis may also account for the relatively quick
rate of evolution of the small B. distachyon genome. The
preceding and present analyses of rearrangements in the
B. distachyon genome showed that the rate of change of
the B. distachyon genome was intermediate between the
quick rate in the Ae. tauschii lineage and the slow rate in
the rice and sorghum lineages (International Brachy-
podium Genome Initiative, 2010, Luo et al., 2013). Evidence
was provided that the small B. distachyon genome was
derived from a larger genome (Luo et al., 2013). Brachy-
podium distachyon is a self-pollinating annual (Gordon
et al., 2015) and the intermediate rate of evolution of its
genome is thus consistent with this hypothesis. However,
rice species having the A genome, which includes
O. sativa, are annual (Kellogg, 2009). So are many sor-
ghum species (Price et al., 2005). Yet, rice and sorghum
genomes are evolving slowly. This shows that annual
reproductive cycle alone leads to rates that are slower than
those observed in the A-, B- and D-genome Triticeae.
Synthesis of these observations and those reviewed in
Introduction suggests that genome size matters only if it is
accompanied by large quantities of active TEs. Likewise,
generation length (annual versus perennial life cycle)
seems to play a role if it is accompanied by a large and
dynamic pool of TEs.
498 Jan Dvorak et al.
Causes of genome evolution
Of the 592 major rearrangements detected using
Ae. tauschii genes as queries and assigned to the phyloge-
netic tree, 444 (75%) were inversions, almost all paracentric
and the rest were translocations, duplications and NCIs. Of
translocations, 77 (13%) were intrachromosomal. Since
paracentric inversions are by definition intrachromosomal
rearrangements, 88% of all rearrangements were intrachro-
mosomal events. A mechanism by which rearrangements
originate must be consistent with this reality. Intrachromo-
somal rearrangements may hypothetically originate by: (i)
two double-strand DNA breaks (DSBs) joined by the non-
homologous end-joining (NHEJ) mechanism in an inverted
orientation; (ii) repair of single DSB by intrachromosomal
HR; or (iii) alternative transposition.
If the DSBs in mechanism (i) were independent of each
other and homogeneously distributed along a chromo-
some arm, the average length of a paracentric inversion
and intrachromosomal translocation would be about a
third of the chromosome arm, since two random uniform
breaks divide a chromosome arm into three interchange-
able segments. This expected length is much longer than
what we found. Thus (i) seems an unlikely mechanism to
account for the lengths of most of the inversions and
translocations.
An attractive feature of mechanism (ii) is that it could
account for the origin of both paracentric inversions and
intrachromosomal translocations. Intrachromosomal HR of
repeats in the opposite orientation is expected to result in
an inversion and intrachromosomal HR of repeats in the
same orientation is expected to result in excision of a cir-
cular intermediate that can be inserted somewhere else.
Thus, if a DSB would occur in a TE and if the members of
the TE family were distributed homogeneously along a
chromosome arm and targeted for intrachromosomal HR
sequentially (first the closest, next the second closest, etc.),
the likelihood of a TE being chosen for recombination
would be multiplicative and exponentially decline with the
distance along the chromosome arm. The fact that mix-
tures of two exponential distributions provided signifi-
cantly better fits to the inversion length data than a single
distribution in five of the six genomes would be consistent
with this hypothesis only if inversions were generated by
mechanism (ii) and TEs were non-homogeneously dis-
tributed along the chromosome arm, which has been
shown (Luo et al., 2017). A difficulty with this hypothesis is
that it requires HR taking place within the same chromatid.
However, DSB repair by HR appear to be limited to S and
G2 phases in the cell cycle and uses sister chromatids;
DSB in G1 are repaired by NHEJ (Rothkamm et al., 2003;
Ira et al., 2004; Knoll et al., 2014; Orthwein et al., 2014; Shi-
bata, 2017). HR of inverted duplicated sequences were
shown to produce dicentric chromosomes and complex
The Plant Journal © 2018 John Wiley & Sons Ltd, The Plant Journal, (2018), 95, 487–503
rearrangements rather than inversions in yeast (VanHulle
et al., 2007).
Mechanism (iii), alternative transposition, takes place
between termini of two DNA transposons (Zhang et al.,
2006; Huang and Dooner, 2008) located in the same chro-
matid. If one TE is defective and the termini of the two
transposons face in opposite orientation, transposition
results in translocations and inversions (Zhang et al.,
2009). A ring intermediate (Zhang et al., 2009) can be rein-
serted resulting in intrachromosomal duplications and
translocations. Another important aspect of alternative
transposition as the prime mechanism for the origin of
major chromosome rearrangements is its independence
from meiotic HR. In contrast, erosion of collinearity corre-
lated highly with meiotic HR in this and a previous study
(Luo et al., 2017). We therefore suggest that meiotic HR
errors are the primary cause of collinearity erosion. Intra-
chromosomal HR and alternative transposition are the pri-
mary causes of major chromosome rearrangements. The
recurrent breaking of the chromosome due to alternative
transposition (Zhang et al., 2009) is consistent with our
finding inversions with common breakpoints.
These considerations do not apply to clusters of translo-
cations present in the distal 4 Mb of rice chromosomes
Os11 and Os12, since they evolved by a different process.
The two chromosomes are homoeologous and originated
by the pan-grass WGD (Paterson et al., 2004). In contrast
with the rest of the chromosomes duplicated by the WGD,
this homoeologous region recurrently recombined produc-
ing clusters of interchromosomal translocations. Similar
clusters of translocations are detectable between sorghum
chromosomes Sb5 and Sb8, which correspond to Os11
and Os12, respectively and within B. distachyon chromo-
some Bd4, which is a compound chromosome partly corre-
sponding to Os11 and Os12 (International Brachypodium
Genome Initiative, 2010). It was originally suggested that
the translocations originated by recurrent gene conver-
sions (Wang et al., 2011). Most of these translocations con-
tain a single gene, which is consistent with gene
conversions, but some contain from two up to seven colli-
near genes. Translocations involving several genes are too
long to be produced by gene conversion, which has led to
suggesting crossovers as the mechanism of their origin
(Wicker et al., 2015). Since the densities of translocations
with 1–7 genes can be approximated by either Logistic or
Weibull functions it is possible that translocations involv-
ing several genes originated by multiple independent gene
conversions. A decline in the number of genes per translo-
cation with distance from the chromosome terminus is
apparent in the rice, sorghum and B. distachyon genomes.
This pattern is more consistent with gene conversions
rather than crossovers as the cause, since lower crossover
frequencies in the proximal regions are expected to pro-
duce larger translocations, not smaller as was observed.
© 2018 The Authors

Rearrangement dataset and its practical utility
Most inversions are unique evolutionary events that
rarely revert to the ancestral state. That makes inversions
an exceptional and currently largely unexploited tool for
phylogenetic reconstructions. We show here that the
boundaries of many inversions are clearly distinguishable
even after 50 MY of evolution. In inversion heterozygotes,
meiotic recombination by crossovers is suppressed
within the inverted haplotype because odd-number cross-
over products are inviable. Even number crossover
recombination produces viable products only if it
involves two of the four chromatids. Since multiple
crossovers are largely precluded by crossover interfer-
ence in short genetic intervals, recombination between
the inverted and structurally wild-type haplotypes is
effectively suppressed in short inversions. The haplotypes
of the same inversion shared by different phylogenetic
lineages must therefore be of the same age and all dif-
ferences between them must have evolved since the
inversion origin. These properties of inversions are
important for timing recent speciation events and were
employed here to validate the phylogeny of the three
Triticeae genomes.
To date, DNA phylogenetic analyses have primarily relied
on analyses of nucleotide sequences, with its set of assets
and drawbacks. The rapidly increasing number of refer-
ence-quality genome sequences opens the door to the use
of structural variation in phylogenetic studies. The dataset
of structural rearrangements provided here (Tables S1 and
S2) will be useful for other types of analyses, such as phylo-
genetic reconstructions in the grass family, comparative
functional genomics (by indicating putative orthologues
across the grass family) and comparative gene mapping.
Absence of structural variation is a critical prerequisite
for success in projects involving recombination between
homoeologous chromosomes. Tribe Triticeae involves over
300 species (Lo€ ve, 1984). Most of them can be hybridized
with wheat and represent therefore a valuable genetic
resource for wheat genetics and improvement. Structural
variation between genomes is the greatest obstacle to gene
introgression in the tribe. Inspection of the dataset of struc-
tural rearrangements provided in Tables S1 and S2 and
results of their validation with BNG maps may assist target-
ing specific genes and specific regions for introgression to
wheat.
EXPERIMENTAL PROCEDURES
Gene collinearity and structural chromosome analyses
The following genome sequences were analyzed: Ae. tauschii v4.0
(Luo et al., 2017), wild emmer wheat (Avni et al., 2017), B. dis-
tachyon, v3.1 (International Brachypodium Genome Initiative,
2010), rice (Oryza sativa) v7.0 (Chantret et al., 2005) and
© 2018 The Authors
The Plant Journal © 2018 John Wiley & Sons Ltd, The Plant Journal, (2018), 95, 487–503
Rates of genome evolution in grasses 499
sorghum (Sorghum bicolor), v3.1 (Paterson et al., 2009). The
methodology used for gene collinearity analysis has been
described earlier (Luo et al., 2017) and only essential informa-
tion will be provided here. The amino acid sequences of 38 775
HC genes located on the Ae. tauschii pseudomolecules (Luo
et al., 2017) were downloaded from (http://aegilops.wheat.ucda
vis.edu/ATGSP/annotation/) and the amino acid sequences of
65 012 HC genes of wild emmer wheat were downloaded from
http://wewseq.wixsite.com/consortium (Avni et al., 2017). Those
of B. distachyon v3.1, rice v7.0 and sorghum v3.1 were down-
loaded from Phytozome (https://phytozome.jgi.doe.gov/pz/portal.
html). BLASTP homology searches using amino acid sequences
of the Ae. tauschii HC genes as queries and amino acid
sequences of wild emmer wheat, B. distachyon, rice and sor-
ghum used as targets were performed. A default BLASTP
parameter setting was used. The top alignment score was
recorded.
Collinearity of top hits in the wild emmer wheat A and B
genomes and the genomes of B. distachyon, rice and sorghum,
ordered according to the order of HC genes along a
Ae. tauschii pseudomolecule, was analyzed as follows. Three or
more genes were considered collinear if the starting
nucleotides of the top hits follow an ascending or descending
order and distances between them were <0.5 Mb on the
B. distachyon, rice and sorghum pseudomolecules or <5 Mb on
the Ae. tauschii and wild emmer wheat pseudomolecules.
Noncolinear genes interrupting a sequence of collinear
genes were allowed. If an Ae. tauschii gene was homologous
to tandem duplicated genes on a target pseudomolecule, only
one of the duplicated genes was recorded as collinear,
provided that it was in a collinear position on the pseudo-
molecule.
A spreadsheet dataset showing collinearity of each of the
38 775 Ae. tauschii HC genes along the wild emmer wheat A-
and B-genome pseudomolecules and B. distachyon, rice,
sorghum pseudomolecules was constructed (Table S1). Cells
containing collinear genes were colored whereas those that were
not collinear were left colorless. Changes in gene order due to
inversions or translocations were thus indicated by changes in
cell color. For each rearrangement, three things were recorded:
its starting gene in bp and ending gene on the Ae. tauschii pseu-
domolecule, the branch of the grass phylogenetic tree (Figure 4)
in which the rearrangement had taken place and the type of rear-
rangement. Rearrangements were coded as follows: A – inversion
of two genes, B – inversion of three genes, C – inversion of >3
genes, D – translocation of two genes within a chromosome, E –
translocation of three genes within a chromosome, F – transloca-
tion of >3 genes within a chromosome, iT – interstitial transloca-
tion between chromosomes, T – terminal translocation, Dup –
duplication of a segment, Del – deletion of a segment.
A similar analysis was performed using 42 004 HC genes anno-
tated on the rice pseudomolecules v7.0 as queries in searches
against genes annotated on the Ae. tauschii, wild emmer wheat,
B. distachyon and sorghum pseudomolecules (Table S2).
To quantify collinearity, Ae. tauschii genes collinear with a
specific genome were counted and expressed as percent of all
Ae. tauschii genes. Collinearity along a pseudomolecule was
expressed as the number of Ae. tauschii genes collinear with
genes on a target pseudomolecule per non-overlapping window
of 200 Ae. tauschii genes. Paired t-tests with Bonferroni correction
using the percentages of collinear genes in the same 200-gene
intervals as variables were used to test statistical significance of
differences between genomes (Figure 1a).
500 Jan Dvorak et al.
Inversion and translocation lengths
As the compared genomes differed in their sizes, the length of
each rearrangement involving 3 or more collinear genes was
expressed in terms of the length of the rearrangement on the
Ae. tauschii pseudomolecule. The lengths were measured from
the position of the first to the last Ae. tauschii collinear gene
located within the rearrangement.
Mean paracentric inversion lengths were computed for each of
the six genomes and statistical significance among the means was
analyzed as follows. Most parametric statistical methods include
the assumption that the sample is drawn from a population where
the variables are normally distributed. A Kolmogorov–Smirnov test
to check the normality of the distribution using SPSS (ver. 22; IBM
Corp, 2013, https://www.ibm.com/analytics/data-science/predic
tive-analytics/spss-statistical-software) indicated that the data devi-
ated significantly from normality (P ≤ 0.01). This was visualized by
histograms, which all showed positive skewness (skewed to the
right). The data were transformed by the Boxcox transformation
(Box and Cox, 1964). The transformed data were normally dis-
tributed.
Analysis of variance (ANOVA) of transformed paracentric inver-
sion lengths in a completely randomized design (CRD) showed
significant genome effects (P ≤ 0.10). Duncan’s test of means com-
parison was used to group the mean paracentric inversion length
with a < 0.05.
An exponential function was fitted to each histogram (counts
of paracentric inversions on y-axis and their length on x-axis).
Exponential functions were estimated in two steps. In the first
step, an exponential distribution was fitted to the paracentric
inversion length data by maximum likelihood estimation (MLE).
In the second step, the estimated exponential distribution was
multiplied by the total number of paracentric inversions. A likeli-
hood ratio test was conducted to test the null hypothesis that
inversion lengths follow an exponential distribution against the
alternative hypothesis that the lengths follow a mixture of expo-
nential distributions. This null hypothesis was rejected for all spe-
cies but rice (P < 0.0004). For the species where the null
hypothesis was rejected, a mixture of two exponential distribu-
tions was fitted by MLE, which produced estimates for the means
of the component distributions l1 and l2, as well as the mixture
proportion a (Figure 3).
Ks analysis
Seven Ae. tauschii paracentric inversions shared with the wild
emmer wheat B genome (category AetB) involving 107 genes and
two Ae. tauschii paracentric inversions shared with the wild
emmer wheat A genome (category AetA) involving eight genes
were analyzed (Table S5). Amino acid sequences of compared
genes were aligned using ClustalW (Thompson et al., 1994), which
aligned and delimited the CDS. Amino acid sequences were then
replaced by nucleotide sequence and the number of synonymous
nucleotide substitutions were counted for each compared gene
pair. Ks values were calculated using Bioperl module Bio::Align::
DNAStatistics.
For a statistical analysis, mixed model ANOVA was performed
to assess significance of variation in Ks values among the A, B
and Ae. tauschii genomes, using Ks values between genes
within an inversion as replications and different inversions as
independent sampling units. The null hypothesis of zero mean
difference was tested. The difference in mean divergence was
then tested by two-sided conditional t-test, Bonferroni cor-
rected.
The Plant Journal © 2018 John Wiley & Sons Ltd, The Plant Journal, (2018), 95, 487–503
Phylogenetic tree and bootstrap confidence on its nodes
The presence/absence of each paracentric inversion involving >3
collinear genes (category C) among the six genomes was con-
verted to binary format, 1 for the presence and 0 for the absence
of an inversion in a genome. Dice dissimilarity index was imple-
mented in the DARwin v6.0.15 program (Perrier and Jacquemoud-
Collet, 2006) to calculate the pairwise genetic distance between
genomes. Dice dissimilarity index measures were used to draw a
Weighted Neighbour-Joining (NJ) tree (Figure 4) with DARwin
v6.0.15. Bootstrap analysis (10 000 replicates) was performed
using DARwin v6.0.15 on the NJ tree. The FigTree v1.4.3 (http://
tree.bio.ed.ac.uk/software/figtree/) program was used to edit the
NJ tree.
Dot-plot
The annotated primary transcripts and corresponding protein
sequences in the emmer wheat genome assembly for accession
Zavitan and the Ae. tauschii AL8/78 assembly v4.0 were down-
loaded. Only the first transcript for each gene was retrieved. A
BLASTP search was conducted using the Ae. tauschii proteins as
queries and the emmer wheat proteins as targets. The top two hits
with an E-value <1e-5 were recorded. Homologous gene pairs
identified by BLASTP were used to detect syntenic blocks using
the software MCscanX (Wang et al., 2012). Collinear segments for
all possible pairs of chromosomes were detected using a match
score of 50, a gap penalty of 1, an E-value threshold of 1e-05 and
a minimum of three genes and maximum gap size between two
consecutive proteins of 25 to declare a collinear block. The pair-
wise comparative dot-plot using the MCscanX output was drawn
using R (Figure 1b).
De novo BioNano genome map assembly
High-molecular-weight (HMW) DNA was isolated from young
leaves of wild emmer wheat accession Zavitan, Aegilops tauschii
ssp. tauschii accession AS75 and Ae. tauschii ssp. tauschii acc.
CIae 1 by Amplicon Express (Pullman, WA, USA). The nicking
endonuclease Nt.BspQI (New England BioLabs, Ipswich, MA,
USA) was used to nick HMW DNA molecules. The nicked DNA
molecules were labeled and stained according to the instructions
of the IrysPrep Reagent Kit (BioNano Genomics), as described in
detail in Luo et al. (2017). The DNA sample was loaded onto the
nanochannel array of an IrysChip (BioNano Genomics) and was
automatically imaged by the Irys system (BioNano Genomics).
Raw DNA molecules >20 kb were collected and converted into.
bnx files by AutoDetect software to obtain basic labelling and
DNA length information. Molecules >180 kb that passed quality
control were aligned, clustered and assembled into BNG contigs
using the BioNano Genomics assembly pipeline as described in
previous publications (Lam et al., 2012; Cao et al., 2014). The P-
value thresholds used for pairwise assembly, extension/refine-
ment and final refinement stages were adjusted according to the
genome sizes (1e 5/genome size in Mb). The initial BNG maps
were then visually checked for chimeric contigs, which were dis-
sociated.
Sequence alignments on a BNG map and inversion
validation
To compare a nucleotide sequence with a BNG map, the nucleo-
tide sequences were digested in silico with the restriction endonu-
clease Nt.BspQ1 by using Knickers (BioNano Genomics). The
alignment of a nucleotide sequence with the BNG map or
© 2018 The Authors

alignments between BNG maps were computed with RefAligner
(BioNano Genomics). The alignments were visualized in IrysView
(BioNano Genomics) and those of interest were saved. Software
packages for these operations were obtained from BioNano Geno-
mics (https://bionanogenomics.com/support/software-downloads/).
Each paracentric inversion unique to the Ae. tauschii (Luo et al.,
2017) or wild emmer wheat A or B genomes (Avni et al., 2017)
and involving three collinear genes and >3 collinear genes (cate-
gories B and C, respectively, in Tables S1 and S2) was validated
(Figure 2). Coordinates (in bp) on the pseudomolecule of collinear
genes flanking a breakpoint were used to identify the breakpoint
region on the BNG contig and a pseudomolecule.
Naming structural variants
We used the following abbreviations for rearrangements: inversion
(Inv), translocation (T), duplication (Dup), deletion (Del) and nested
chromosome insertion (NCI). We employed the HC gene set v2.0
annotated in the Ae. tauschii genome v4.0 as a reference for the
grass family and the branches of the grass phylogenetic tree as indi-
cators of rearrangement location and origin. Thus, InvAetA
(AET1Gv20186400-900) is an inversion shared by the Ae. tauschii
(Aet) and wheat A genomes starting with HC Ae. tauschii gene
AET1Gv20186400 and ending with NC gene AET1Gv20186900. For
translocations, Ae. tauschii genes flanking a breakpoint are given,
e.g. TAetBA (AET4Gv20018000-9500; AET5Gv20867400-600) is a
reciprocal translocation with breakpoints on 4D between genes
AET4Gv20018000 and AET4Gv20019500 and 5D between genes
AET5Gv20867400 AET5Gv20867600. Once defined, the transloca-
tion can be abbreviated as, e.g., T(4AL;5AL).
Availability of data and materials
All data generated in this study are included in this published arti-
cle and Supporting Information files. Published Aegilops tauschii
sequences on which these analyses are based are available from
this link http://aegilops.wheat.ucdavis.edu/ATGSP/data.php.
ACKNOWLEDGEMENT
This publication is based upon work supported by the US National
Science Foundation (NSF) under grant number IOS-1238231.
CONFLICT OF INTEREST
The authors declare that they have no competing interests.
AUTHORS’ CONTRIBUTIONS
J.D, M.-C.L., K.R.D., C.M.J., T.Z., L.W., H-G.M, P.E.M, B.S.G.,
K.M.D., Y.Q.G and A.D. planned the work. M.-C.L., K.R.D and
T.Z. performed BNG mapping and analyses; J.D., L.W., T.Z.,
H.-G.M., X.D., K.M.D., P.Q., P.E.M., Y.Q.G., R.K.R., H.D.,
F.M.Y. and P.J.G. performed the analyses of genome struc-
ture and evolution. H-G.M., X.D., M.W.D., R.K.R. and H.D.
performed statistical analyses. J.D. organized and managed
the contributions to this publication and was primary author.
All authors read and approved the final manuscript.
SUPPORTING INFORMATION
Additional Supporting Information may be found in the online ver-
sion of this article.
Table S1. Collinearity with Aegilops tauschii pseudomolecules of
best blast hit targets in the wild emmer A- and B-genome
© 2018 The Authors
The Plant Journal © 2018 John Wiley & Sons Ltd, The Plant Journal, (2018), 95, 487–503
Rates of genome evolution in grasses 501
pseudomolecules and those of Brachypodium distachyon, Oryza
sativa and Sorghum bicolor.
Table S2. Collinearity with Oryza sativa pseudomolecules of best
blast hit targets in the wild emmer A- and B-genome pseudo-
molecules and those of Aegilops tauschii, Brachypodium dis-
tachyon and Sorghum bicolor.
Table S3. Synopsis of rearrangements, identified using Aegilops
tauschii genes as blast queries, allocated to individual branches of
the phylogenetic tree (Figure 4).
Table S4. Analysis of translocations between Oryza sativa chro-
mosomes Os11 and Os12, within Brachypodium distachyon chro-
mosomes Bd4 and between Sorghum bicolor chromosomes Sb5
and Sb8, wild emmer wheat chromosomes 4B and 5B.
Table S5. Ka and Ks values in inversions shared between Aegilops
tauschii and either the A or B genomes of wild emmer wheat.
REFERENCES
Akhunov, E.D., Goodyear, J.A., Geng, S. et al. (2003) The organization and
rate of evolution of the wheat genomes are correlated with recombination
rates along chromosome arms. Genome Res. 13, 753–763.
Andreasen, K. and Baldwin, B.G. (2001) Unequal evolutionary rates between
annual and perennial lineages of checker mallows (Sidalcea, Malvaceae):
evidence from 18S-26S rDNA internal and external transcribed spacers.
Mol. Biol. Evol. 18, 936–944.
Avni, R., Nave, M., Barad, O. et al. (2017) Wild emmer genome architecture
and diversity elucidate wheat evolution and domestication. Science, 357,
93–97.
Bernhardt, N., Brassac, J., Kilian, B. and Blattner, F.R. (2017) Dated tribe-
wide whole chloroplast genome phylogeny indicates recurrent hybridiza-
tions within Triticeae. BMC Evol. Biol. 17, 141.
Birol, I., Raymond, A., Jackman, S.D. et al. (2013) Assembling the 20 Gb
white spruce (Picea glauca) genome from whole-genome shotgun
sequencing data. Bioinformatics, 29, 1492–1497.
Bowers, J.E., Chapman, B.A., Rong, J.K. and Paterson, A.H. (2003) Unravel-
ling angiosperm genome evolution by phylogenetic analysis of chromo-
somal duplication events. Nature, 422, 433–438.
Box, G.E.P. and Cox, D.R. (1964) An analysis of transformations. J. Roy.
Stat. Soc. Series B, 26, 211–234.
Brubaker, C.L., Paterson, A.H. and Wendel, J.F. (1999) Comparative genetic
mapping of allotetraploid cotton and its diploid progenitors. Genome,
42, 184–203.
Cao, H.Z., Hastie, A.R., Cao, D.D. et al. (2014) Rapid detection of structural
variation in a human genome using nanochannel-based genome map-
ping technology. Gigascience, 3, 34.
Chantret, N., Salse, J., Sabot, F. et al. (2005) Molecular basis of evolution-
ary events that shaped the hardness locus in diploid and polyploid wheat
species (Triticum and Aegilops). Plant Cell, 17, 1033–1045.
Dvorak, J. (2009) Triticeae genome structure and evolution. Plant Genet.
Genomics, 7, 685–711.
Dvorak, J. and Akhunov, E.D. (2005) Tempos of deletions and duplica-
tions of gene loci in relation to recombination rate during diploid and
polyploid evolution in the Aegilops-Triticum alliance. Genetics, 171,
323–332.
Dvorak, J. and Zhang, H.B. (1990) Variation in repeated nucleotide
sequences sheds light on the phylogeny of the wheat B and G genomes.
Proc. Natl Acad. Sci. USA, 87, 9640–9644.
Dvorak, J. and Zhang, H.B. (1992) Molecular tools for study of the phy-
logeny of diploid and polyploid species of Triticeae. In 1st Int. Symp. on
Triticeae. Helsingborg: Hereditas, pp. 37–42.
Dvorak, J., McGuire, P.E. and Mendlinger, S. (1984) Inferred chromosome
morphology of the ancestral genome of Triticum. Plant Syst. Evol. 144,
209–220.
Dvorak, J., di Terlizzi, P., Zhang, H.B. and Resta, P. (1993) The evolution of
polyploid wheats: identification of the A genome donor species. Gen-
ome, 36, 21–31.
Dvorak, J., Deal, K.R., Luo, M.C., You, F.M., von Borstel, K. and Dehghani,
H. (2012) The origin of spelt and free-threshing hexaploid wheat. J.
Hered. 103, 426–441.
502 Jan Dvorak et al.
Gale, M.D. and Devos, K.M. (1998) Plant comparative genetics after
10 years. Science, 282, 656–659.
Goff, S.A., Ricke, D., Lan, T.H. et al. (2002) A draft sequence of the rice gen-
ome (Oryza sativa L. ssp japonica). Science, 296, 92–100.
Gordon, S.P., Liu, L. and Vogel, J.P. (2015) The genus Brachypodium as a
model for perenniality and polyploidy. In Plant Genetics and Genomics:
Crop Models (Vogel, J.P., ed). Switzerland: 2015 Springer International
Publishing, pp. 313–325.
Gornicki, P., Zhu, H.L., Wang, J.W., Challa, G.S., Zhang, Z.Z., Gill, B.S. and
Li, W.L. (2014) The chloroplast view of the evolution of polyploid wheat.
New Phytol. 204, 704–714.
Haeggstrom, C.A. and Skyten, R. (1996) Flowering and individual survival of
a population of the grass Brachypodium sylvaticum in Nato, Aland
islands, SW Finland. Ann. Bot. Fenn. 33, 1–10.
Hasegawa, M., Kishino, H. and Yano, T.A. (1985) Dating of the human ape
splitting by a molecular clock of mitochondrial-DNA. J. Mol. Evol. 22,
160–174.
Hastie, A.R., Dong, L.L., Smith, A. et al. (2013) Rapid genome mapping in
nanochannel arrays for highly complete and accurate de novo sequence
assembly of the complex Aegilops tauschii genome. PLoS ONE, 8,
e55864.
Hsiao, C., Chatterton, N.J., Asay, K.H. and Jensen, K.B. (1995) Phylogenetic-
relationships of the monogenomic species of the wheat tribe, Triticeae
(Poaceae), inferred from nuclear rDNA (internal transcribed spacer)
sequences. Genome, 38, 211–223.
Huang, J.T. and Dooner, H.K. (2008) Macrotransposition and other complex
chromosomal restructuring in maize by closely linked transposons in
direct orientation. Plant Cell, 20, 2019–2032.
Huang, S., Sirikhachornkit, A., Su, X., Faris, J., Gill, B.S., Haselkorn, R. and
Gornicki, P. (2002) Genes encoding plastid acetyl-CoA carboxylase and 3-
phopshoglycerate kinase of the Triticum/Aegilops complex and the evo-
lutionary history of polyploid wheat. Proc. Natl Acad. Sci. USA, 99, 8133–
8138.
International Brachypodium Genome Initiative (2010) Genome sequencing
and analysis of the model grass Brachypodium distachyon. Nature, 463,
763–768.
Ira, G., Pellicioli, A., Balijja, A. et al. (2004) DNA end resection, homologous
recombination and DNA damage checkpoint activation require CDK1.
Nature, 431, 1011–1017.
Jorgensen, C., Luo, M.-C., Ramasamy, R., Dawson, M., Gill, B.S., Korol, A.B.,
Distelfeld, A. and Dvorak, J. (2017) A high-density genetic map of wild
emmer wheat from the Karaca Dag  region provides new evidence on the
structure and evolution of wheat chromosomes. Front. Plant Sci. 8, 1798.
Kellogg, E.A. (2009) The evolutionary history of ehrhartoideae, oryzeae and
oryza. Rice, 2, 1–14.
Kihara, H. (1944) Discovery of the DD-analyser, one of the ancestors of Triti-
cum vulgare (Japanese). Agric. Hortic. 19, 13–14.
Knoll, A., Fauser, F. and Puchta, H. (2014) DNA recombination in somatic
plant cells: mechanisms and evolutionary consequences. Chromosome
Res. 22, 191–201.
Kohne, D. (1970) Evolution of higher-organism DNA. Q. Rev. Biophys. 3,
327–375.
Lam, E.T., Hastie, A., Lin, C. et al. (2012) Genome mapping on nanochannel
arrays for structural variation analysis and sequence assembly. Nat. Bio-
tech. 30, 771–776.
Li, L.F., Liu, B., Olsen, K.M. and Wendel, J.F. (2015) A re-evaluation of the
homoploid hybrid origin of Aegilops tauschii, the donor of the wheat D-
subgenome. New Phytol. 208, 4–8.
Lo€ ve, A. (1984) Conspectus of the Triticeae. Feddes Rep. 95, 425–521.
Luo, M.C., Deal, K.R., Akhunov, E.D. et al. (2009) Genome comparisons
reveal a dominant mechanism of chromosome number reduction in
grasses and accelerated genome evolution in Triticeae. Proc. Natl Acad.
Sci. USA, 106, 15780–15785.
Luo, M.C., Gu, Y.Q., You, F.M. et al. (2013) A 4-gigabase physical map
unlocks the structure and evolution of the complex genome of Aegilops
tauschii, the wheat D-genome progenitor. Proc. Natl Acad. Sci. USA, 110,
7940–7945.
Luo, M.C., You, F.M., Li, P.C., Wang, J.R., Zhu, T.T., Dandekar, A.M., Leslie,
C.A., Aradhya, M., McGuire, P.E. and Dvorak, J. (2015) Synteny analysis
in Rosids with a walnut physical map reveals slow genome evolution in
long-lived woody perennials. BMC Genomics, 16, 707.
The Plant Journal © 2018 John Wiley & Sons Ltd, The Plant Journal, (2018), 95, 487–503
Luo, M.C., Gu, Y.Q., Puiu, D. et al. (2017) Genome sequence of the progeni-
tor of the wheat D genome Aegilops tauschii. Nature, 551, 498–502.
Marcussen, T., Sandve, S.R., Heier, L. et al. (2014) Ancient hybridizations
among the ancestral genomes of bread wheat. Science, 345, 1250092.
Martin, A.P. and Palumbi, S.R. (1993) Body size, metabolic-rate, genera-
tion time and the molecular clock. Proc. Natl Acad. Sci. USA, 90, 4087–
4091.
Mascher, M., Gundlach, H., Himmelbach, A. et al. (2017) A chromosome
conformation capture ordered sequence of the barley genome. Nature,
544, 427–433.
Massa, A.N., Wanjugi, H., Deal, K.R. et al. (2011) Gene space dynamics
during the evolution of Aegilops tauschii, Brachypodium distachyon,
Oryza sativa and Sorghum bicolor genomes. Mol. Biol. Evol. 28, 2537–
2547.
McFadden, E.S. and Sears, E.R. (1946) The origin of Triticum spelta and its
free-threshing hexaploid relatives. J. Hered. 37, 81–89, 107–116.
Middleton, C.P., Senerchia, N., Stein, N., Akhunov, E.D., Keller, B., Wicker,
T. and Kilian, B. (2014) Sequencing of chloroplast genomes from wheat,
barley, rye and their relatives provides a detailed insight into the evolu-
tion of the Triticeae tribe. PLoS ONE, 9, e85761.
Ming, R., Liu, S.C., Lin, Y.R. et al. (1998) Detailed alignment of Saccharum
and Sorghum chromosomes: comparative organization of closely related
diploid and polyploid genomes. Genetics, 150, 1663–1682.
Murphy, W.J., Larkin, D.M., Everts-van der Wind, A. et al. (2005) Dynamics
of mammalian chromosome evolution inferred from multispecies com-
parative maps. Science, 309, 613–617.
Neale, D.B., Wegrzyn, J.L., Stevens, K.A. et al. (2014) Decoding the massive
genome of loblolly pine using haploid DNA and novel assembly strate-
gies. Genome Biol. 15, R59.
Neale, D.B., Martınez-Garcıa, P.J., De La Torre, A.R., Montanari, S. and Wei,
X.X. (2017a) Tree genome sequencing: novel insights into plant biology.
Annu. Rev. Plant Biol. 68, 457–483.
Neale, D.B., McGuire, P.E., Wheeler, N.C. et al. (2017b) The Douglas-fir gen-
ome sequence reveals specialization of the photosynthetic apparatus in
Pinaceae. G3, 7, 3157–3167.
Nesbitt, M. and Samuel, D. (1996) From staple crop to extinction? The
archaeology and history of hulled wheats. In Hulled Wheats. Promoting
the Conservation and Use of Underutilized and Neglected Crops. 4. Proc.
1st Internatl. Workshop on Hulled Wheats (Padulosi, S., Hammer, K. and
Heller, J., eds). Castelvecchio Pacoli, Tuscany, Italy: International Plant
Genetic Resources Institute, Rome, Italy, pp. 41–100.
Nystedt, B., Street, N.R., Wetterbom, A. et al. (2013) The Norway spruce
genome sequence and conifer genome evolution. Nature, 497, 579–584.
Orthwein, A., Fradet-Turcotte, A., Noordermeer, S.M., Canny, M.D., Brun,
C.M., Strecker, J., Escribano-Diaz, C. and Durocher, D. (2014) Mitosis
inhibits DNA double-strand break repair to guard against telomere
fusions. Science, 344, 189–193.
Paterson, A.H., Bowers, J.E. and Chapman, B.A. (2004) Ancient polyploidiza-
tion predating divergence of the cereals and its consequences for com-
parative genomics. Proc. Natl Acad. Sci. USA, 101, 9903–9908.
Paterson, A.H., Bowers, J.E., Bruggmann, R. et al. (2009) The Sorghum
bicolor genome and the diversification of grasses. Nature, 457, 551–556.
Perrier, X. and Jacquemoud-Collet, J.P. (2006) DARwin software http://dar
win.cirad.fr/darwin.
Price, H.J., Dillon, S.L., Hodnett, G., Rooney, W.L., Ross, L. and Johnston,
J.S. (2005) Genome evolution in the genus Sorghum (Poaceae). Ann.
Bot. 95, 219–227.
Rothkamm, K., Kruger, I., Thompson, L.H. and Lobrich, M. (2003) Pathways
of DNA double-strand break repair during the mammalian cell cycle.
Mol. Cell. Biol. 23, 5706–5715.
Schnable, P.S., Ware, D., Fulton, R.S. et al. (2009) The B73 Maize Genome:
complexity, diversity and dynamics. Science, 326, 1112–1115.
Shibata, A. (2017) Regulation of repair pathway choice at two-ended DNA
double-strand breaks. Mutat. Res. 803, 51–55.
Smith, S.A. and Donoghue, M.J. (2008) Rates of molecular evolution are
linked to life history in flowering plants. Science, 322, 86–89.
Stevens, K.A., Wegrzyn, J.L., Zimin, A. et al. (2016) Sequence of the sugar
pine megagenome. Genetics, 204, 1613–1626.
Tang, H.B., Bowers, J.E., Wang, X.Y., Ming, R., Alam, M. and Paterson, A.H.
(2008) Perspective – Synteny and collinearity in plant genomes. Science,
320, 486–488.
© 2018 The Authors

Thompson, J.D., Higgins, D.G. and Gibson, T.J. (1994) CLUSTAL W: improv-
ing the sensitivity of progressive multiple sequence alignment through
sequence weighting, position-specific gap penalties and weight matrix
choice. Nucleic Acids Res. 22, 4673–4680.
Town, C., Schmidt, R. and Bancroft, I. (2011) Comparative genome analysis
at the sequence level in the Brassicaceae. Genet. Genomics Brassicaceae,
9, 171–194.
Tuskan, G.A., DiFazio, S., Jansson, S. et al. (2006) The genome of black cot-
tonwood, Populus trichocarpa (Torr. & Gray). Science, 313, 1596–1604.
VanHulle, K., Lemoine, F.J., Narayanan, V., Downing, B., Hull, K., McCul-
lough, C., Bellinger, M., Lobachev, K., Petes, T.D. and Malkova, A. (2007)
Inverted DNA repeats channel repair of distant double-strand breaks into
chromatid fusions and chromosomal rearrangements. Mol. Cell. Biol. 27,
2601–2614.
Wang, X.Y., Tang, H.B. and Paterson, A.H. (2011) Seventy million years of
concerted evolution of a homoeologous chromosome pair, in parallel, in
major Poaceae lineages. Plant Cell, 23, 27–37.
Wang, Y.P., Tang, H.B., DeBarry, J.D. et al. (2012) MCScanX: a toolkit for
detection and evolutionary analysis of gene synteny and collinearity.
Nucleic Acids Res. 40, e49.
Wang, J.R., Luo, M.C., Chen, Z.X., You, F.M., Wei, Y.M., Zheng, Y.L. and
Dvorak, J. (2013) Aegilops tauschii single nucleotide polymorphisms
shed light on the origins of wheat D-genome genetic diversity and
© 2018 The Authors
The Plant Journal © 2018 John Wiley & Sons Ltd, The Plant Journal, (2018), 95, 487–503
Rates of genome evolution in grasses 503
pinpoint the geographic origin of hexaploid wheat. New Phytol. 198,
925–937.
Wicker, T., Wing, R.A. and Schubert, I. (2015) Recurrent sequence exchange
between homeologous grass chromosomes. Plant J. 84, 747–759.
Wilson, M.A., Gaut, B. and Clegg, M.T. (1990) Chloroplast DNA evolves
slowly in the palm family (Arecaceae). Mol. Biol. Evol. 7, 303–314.
Wu, F.N. and Tanksley, S.D. (2010) Chromosomal evolution in the plant
family Solanaceae. BMC Genomics, 11, 182.
Xiao, M., Phong, A., Ha, C. et al. (2007) Rapid DNA mapping by fluorescent
single molecule detection. Nucleic Acids Res. 35, e15.
Yamane, K. and Kawahara, T. (2005) Intra- and interspecific phylogenetic
relationships among diploid Triticum-Aegilops species (Poaceae) based
on base-pair substitutions, indels and microsatellites in chloroplast non-
coding sequences. Am. J. Bot. 92, 1887–1898.
Zhang, J.B., Zhang, F. and Peterson, T. (2006) Transposition of reversed Ac
element ends generates novel chimeric genes in maize. PLoS Genet. 2,
1535–1540.
Zhang, J.B., Yu, C.H., Pulletikurti, V., Lamb, J., Danilova, T., Weber, D.F.,
Birchler, J. and Peterson, T. (2009) Alternative Ac/Ds transposition
induces major chromosomal rearrangements in maize. Gene Dev. 23,
755–765.
Zhao, G.Y., Zou, C., Li, K. et al. (2017) The Aegilops tauschii genome reveals
multiple impacts of transposons. Nat. Plants, 3, 946–955.