o r i g i n a l r es e a r c h

Fruit Quality Traits Have Played Critical Roles
in Domestication of the Apple
M. Awais Khan, Kenneth M. Olsen, Valpuri Sovero,
Mosbah M. Kushad, and Schuyler S. Korban*

Abstract
With its long history of cultivation, the domesticated apple, Malus
× domestica Borkh., is an excellent model for studying domestication in long-lived perennial plants. As apples have been
transported from their center of origin in the Tian Shan region in
central Asia and moved along the famous Silk Road trading path,
they have undergone selection for distinct phenotypic traits. In this
study, a high-throughput single nucleotide polymorphism (SNP)
genotyping assay is used to investigate relationships of Malus
species, draw inferences on the domestication history, identify
traits critical for domestication, and assess potential genetic loci
under selection. A total of 160 Malus accessions, including wild
species, old and new apple cultivars, and advanced selections,
were genotyped. Of 1536 SNPs from GoldenGate oligonucleotide pool assays (OPAs), 901 SNPs fulfilled filtering criteria. A
principal component analysis (PCA) revealed that most M. × domestica accessions grouped together. A total of 67 loci, including 13 genomic SNPs and 54 genic SNPs, have been identified
as potential targets for selection during evolution of the apple
genome from its wild progenitor M. sieversii. Genes introgressed
from local wild species into M. × domestica are associated with
adaptation to local environments, while genes for fruit quality
traits are derived from M. sieversii.

S

ince ancient times , plants have been undergoing
selection for desirable traits through the process of
domestication. During domestication, wild species of
plants and animals undergo intensive artificial selection
based on desirable morphological characters that eventually fixes beneficial alleles derived from wild species or
from new mutations (Paran and Van der Knaap, 2007),
collectively referred to as the “domestication syndrome”
(Hammer, 1984). Over many centuries, selection for
traits of interest has gone on both intentionally, with targeted desired traits, and unintentionally, by propagating
favorable phenotypes of each generation (Gepts, 2004).
Domestication of species provides not only insights into
cultures and lifestyles of human civilizations, but also
serves as a well-designed natural experiment to gain new
knowledge about evolution, and allows us to pursue studies on the influence of selection on modifying functions
of genes controlling desirable traits (Purugganan and
Fuller, 2009; Olsen and Wendel, 2013). To date, several
studies have focused on the genetics and genomics of the

M.A. Khan, and S.S. Korban, Dep. of Natural Resources &
Environmental Sciences, and V. Sovero and M. M. Kushad, Dep.
of Crop Sciences, Univ. of Illinois, Urbana, IL 61801; K.M. Olsen,
Dep. of Biology, Washington Univ., St. Louis, MO 63130-4899;
S.S. Korban, Dep. of Biology, Univ. of Massachusetts-Boston,
Boston 02125. M.A. Khan, current address, International Potato
Center (CIP), La Molina, Lima, Peru. Received 25 Apr. 2014.
*Corresponding author (korban@illinois.edu).
Published in The Plant Genome 7
doi: 10.3835/plantgenome2014.04.0018
© Crop Science Society of America
5585 Guilford Rd., Madison, WI 53711 USA
An open-access publication
All rights reserved. No part of this periodical may be reproduced or
transmitted in any form or by any means, electronic or mechanical,
including photocopying, recording, or any information storage and
retrieval system, without permission in writing from the publisher.
Permission for printing and for reprinting the material contained herein
has been obtained by the publisher.
the pl ant genome



november 2014



vol . 7, no . 3 	

Abbreviations: EST, expressed sequence tag; Fst, subpopulation
relative to the total variance; GWAS, Genome-Wide Association
Study in Genome Association and Prediction Integrated Tool, an R
package; IBS, identity-by-state matrix; K, number of subpopulations;
LD, linkage disequilibrium; LG, linkage group; NADP, nicotinamide
adenine dinucleotide phosphate; NCBI, National Center for
Biotechnology Information; OPA, oligonucleotide pool assay; PC,
principal component; PCA, principal component analysis; QTL,
quantitative trait locus; SNP, single nucleotide polymorphism.

1

of

18

evolution of animals (e.g., dogs (Pollinger et al., 2010),
horses (Outram et al., 2009), and cattle (Götherström et
al., 2005), and plants, such as maize (Zea mays L.; Wang
et al., 1999; Clark et al., 2004; Hufford et al., 2012) and
rice (Oryza sativa L.; Konishi et al., 2006; Fuller et al.,
2009; Huang et al., 2012). Evaluation and analysis of
genome-wide genetic diversity and linkage disequilibrium (LD) in crops and their wild progenitors can be
used in “selective sweep” mapping and in other population genetic studies to identify genomic regions targeted
for selection during domestication (Tian et al., 2009).
Two of the best-known examples of extended LD are the
maize tb1 gene that controls plant architecture (Wang et
al., 1999; Clark et al., 2004), and the Y1 gene that controls
kernel color (Palaisa et al., 2004).
Unlike self-compatible annual crop species, wherein
lifecycles are short and several generations can fix favorable genetic combinations over a relatively short time,
many fruit trees have long lifecycles, and are self-incompatible (Hummer and Janick, 2009). Therefore, it will be
very useful to investigate whether or not selection during
domestication of fruit trees has undergone a different
mechanism to rapidly shape their genomes in relatively
few generations during cultivation (Miller and Gross,
2011). Thus far, there are few studies on domestication of
fruit crops. In one recent study (Myles et al., 2011), archeological findings have confirmed that wine grape (Vitis
vinifera subsp. vinifera), a vegetatively-propagated vine
crop, originated in the Near East. However, following
domestication and as grape cultivation moved westward
towards Europe, wild V. vinifera subsp. sylvestris from
Western Europe contributed to V. vinifera subsp. vinifera
cultivars, resulting in significant introgression into modern Western European cultivars.
The cultivated apple, M. × domestica Borkh., can
serve as an excellent model to study domestication in
long-lived perennial fruit trees. The Malus genus includes
about 25 to 33 species. Progenitors of M. × domestica
have been proposed to have arisen ~8 to 12 million years
ago in forests close to the borders of Kazakhstan and
China (Kellerhals, 2009). Incidence of high genetic diversity in both wild and cultivated apples of the Tian Shan
region in central Asia, availability of chloroplast and
nuclear DNA phylogenies, as well as common fruit and
tree morphologies strongly suggest that the domesticated
apple is more similar to M. sieversii (abundant in Tian
Shan forests) rather than to other wild apples of Western
Europe and North America (Juniper and Mabberley,
2006). Neolithic agricultural techniques were pioneered
in the Fergana Valley in Uzbekistan, initiating early
efforts of human selection and propagation of desirable
traits, and sweet apples, similar to extant apples, first
appeared there. Subsequently, apples have been transported from their center of origin, along the Silk Road
trading path, to Europe and then to the Americas while
undergoing further selection for distinct phenotypic
traits based on new environmental conditions and different organoleptic preferences. During this path of
2

of

18	

domestication, secondary introgression into the genome
of domesticated apples must have occurred from additional wild Malus species (Harris et al., 2002).
Large fruit size and low-fruit acidity (sweetness) have
likely played important roles in selection during domestication of the apple (Juniper and Mabberley, 2006). Among
Malus species and hybrids, there are large differences in
fruit size, from ~4 cm or less in diameter for the wild progenitor M. sieversii (Harris et al., 2002) up to 10 cm for the
modern cultivar Spokane Beauty, but the molecular basis
is not conclusively known. Differences in fruit size among
Malus species have been attributed to differences in cell
number and/or cell size (Harada et al., 2005). It has also
been reported that ‘Golden Delicious’, a large-sized apple,
differs from small-sized apple cultivars both in early cell
production (proliferation) as well as in the duration of cell
proliferation, suggesting that both factors influence final cell
number and fruit size of apple (Malladi and Hirst, 2010).
Although it is widely accepted that the origin of the
apple is from M. sieversii in the Tian Shan region and its
domestication occurred along the commercial Silk Road
path, the extent of contribution, as well as the number of
other Malus species and traits that might have been critical in apple domestication remain in question. In this
study, a high-throughput SNP genotyping assay (Khan
et al., 2012a) is used to investigate relationships of wild
Malus species with the cultivated apple, M. × domestica,
draw inferences on the domestication history of apples,
identify traits critical for domestication, and assess
potential genetic loci under selection.

Materials and Methods
Plant Material

In this study, 160 accessions from 30 worldwide Malus
species, including accessions from wild species, old and
new cultivars, as well as advanced selections from a germplasm collection located at the University of Illinois at
Urbana-Champaign, were used (Table 1, Supplementary
Table S1). Out of 30 Malus species and a single hybrid
group (advanced selections from breeding programs and
cultivars originating from crosses among different Malus
species), 20 species had only one accession and five species,
including M. × domestica, M. orientalis, M. prunifolia, M.
sieversii, and M. sylvestris had 52, 11, 5, 21, and 5 accessions, respectively. The hybrid group consisted of 31 accessions, and each of M. baccata and M. micromalus had four
accessions. Species with fewer than four accessions were
excluded from some analysis, such as analysis of relationships of species, unless stated otherwise.
All accessions have been propagated (by grafting)
onto Bud-9 apple rootstocks, and were 6 yr in age. These
were grown and maintained in an experimental orchard
(in a completely randomized block with at least four
replications per accession) at the University of Illinois
at Urbana-Champaign. The origin of this germplasm
collection has been described previously by Potts et al.
(2012). Briefly, these accessions are a subset of the entire
the pl ant genome



november 2014



vol . 7, no . 3

Table 1. Accessions from 30 Malus species, including accessions from wild Malus species, M. × domestica, and
advanced selections from the germplasm collection at University of Illinois, Urbana-Champaign, used in the study.
Species
M. angustifolia
M. arnoldiana
M. asiatica
M. baccata
M. bhutanica
M. coronaria
M. dawsoniana
M. × domestica

No.

Species

No.

1
1
2
4
1
1
1
52

M. halliana
M. hartwigii
M. honanensis
Malus hybrid spp.
M. kansuensis
M. kirghisorum
M. microcarpa
M. orientalis

1
2
1
31
1
1
4
11

Malus germplasm maintained at the clonal repository
of the Plant Genetic Resources Unit in Geneva, NY.
This collection has been selected by the US Apple Crop
Germplasm Committee (CGC) for its wide diversity for
various horticultural characters, such as fruit size, fruit
ripening, cold hardiness, and resistance to diseases and
insects, among others, and botanical traits, such as leaf
shape and leaf color (Korban and Tartarini, 2009).

DNA Extraction and Genotyping
Young leaves were collected from trees of each of 160
Malus accessions (Table 1), growing in the orchard, and
were quickly transported to the lab on ice. Leaves were
freeze-dried in liquid N, and crushed into fine powder
for DNA extraction.
Genomic DNA extraction was done using the
Cetyltrimethyl Ammonium Bromide (CTAB) extraction
method, but with slight modifications to remove polysaccharides, polyphenols, and RNA, followed by a phenolchloroform extraction method as previously described
(Khan et al., 2012a). DNA was quantified using NanoDrop
ND-1000 spectrophotometer (NanoDrop Technologies
Inc., Wilmington, DE), and normalized to a total of 250 ng
per sample. DNA from each sample was genotyped using
the Illumina 1536 GoldenGate assay on a BeadStation
system (Illumina Inc., San Diego, CA) at the W.M. Keck
Center for Functional Genomics (University of Illinois at
Urbana-Champaign), as previously described (Khan et al.,
2012a). The GoldenGate OPAs used for genotyping consisted of 1536 SNPs, consisting of 1411 genic SNPs along
with an additional 125 genomic SNPs (Khan et al., 2012a).
Scoring of SNPs after removal of outliers, and scaling of raw hybridization intensity data of SNPs were performed using the genotyping function in the BeadStudio
package (Illumina, San Diego, CA) as recommended by
Illumina and as previously described (Khan et al., 2012a).
If needed, normalized intensity values were manually
inspected and corrected, and SNPs with a GenCall (GC)
score ³ 0.35, based on an average GC scores for genotypes, and showing errors in segregation were removed.
Clean data were used for further analysis.

Population Genetic Structure
PLINK (Purcell et al., 2007) was used to check for missing values at the level of individual SNPs as well as at the
k ahn et al .: domesti cation of apples

	

Species
M. prattii
M. prunifolia
M. robusta
M. rockii
M. sieversii
M. sikkimensis
M. soulardii
M. spectabilis

No.
1
5
2
1
21
1
1
1

Species

No.

M. sublobata
M. sylvestris
M. toringo
M. transitoria
M. yunnanensis
M. zhaojiaoensis
M. zumi
Total

1
5
2
1
1
1
1
160

accession level, and for minor allele frequency for genotyping. Accessions with ³15% missing data and SNPs
with ³10% missing data were excluded, while SNPs with
allele frequencies ³0.05 were kept for further analysis.
Filtered SNP data were used for assessing population
groupings and relationships among Malus accessions.
A PCA was performed on genotypic data from 901 SNPs
of 160 Malus accessions from 30 species and one hybrid
group (includes advanced selections and cultivars resulting
from crosses of two species) in R. Relationships among accessions (Table 1) were inferred based on PCA1 and PCA2. All
data were graphically displayed with XLSTAT. Population
structure was also assessed for these same accessions using
901 SNPs in STRUCTURE (Pritchard et al., 2000) using the
admixture model. The parameters used in STRUCTURE
were 1,000,000 burnin period followed by 10,000,000 MCMC
runs, along with five replications each for K (number of subpopulations) values of 1 to 15.
A nearest neighbor approach was also employed to
detect outliers, as implemented in PLINK (Purcell et al.,
2007). This consisted of closely related Malus accessions
from the first nearest neighbor to the fifth nearest neighbor, and these were subjected to network analysis and
graphical display in Pajek (Batagelj and Mrvar, 2004).
A pairwise identity-by-state (IBS) similarity matrix was
also calculated using PLINK (Purcell et al., 2007). This
was also used to obtain a distance matrix (1 – IBS) which
was then used in SplitsTree 4 (Huson and Bryant, 2006)
to calculate and display the NeighborNet network.

Linkage Disequilibrium Decay
The chromosomal locations of SNPs were determined
using BLAST (Basic Local Alignment Search Tool)
searches against the apple draft genome sequence
(Velasco et al., 2010) posted on the Genome Database
for Rosaceae (http://www.rosaceae.org/, verified 24 July
2014). The physical positions of 882 SNPs on all 17 linkage
groups (LGs) of the apple was used to calculate genomewide LD patterns in all 160 Malus accessions from 31 species and M. × domestica accessions using PLINK (Purcell
et al., 2007). Pairwise r2 values were visually displayed
by plotting physical distance against r2 values using
XLSTAT. The LD decay threshold was inferred by drawing a trend line based on nonlinear logarithmic regression
curve of r2 as a function of physical distance.
3

of

18

Loci under Selection

Two approaches were used to scan SNPs for signatures
of selection: a subpopulation relative to the total variance (Fst)–based outlier detection method implemented
in LOSITAN (Antao et al., 2008) and a Bayesian genome
scan approach as implemented in BayeScan (Foll and
Gaggiotti, 2008). Loci showing signatures of selection
detected by each method were compared, and overlapping loci were identified and considered candidate SNPs
under selection. Allele frequencies for the candidate
SNPs were calculated across the different species, and
the identities of genes containing SNPs were assessed
through blastn BLAST searches against the NCBI
(National Center for Biotechnology Information) database for M. × domestica expressed sequence tags (ESTs)
and unigenes. Corresponding apple EST sequences with
highest similarities were recorded and BLAST searched
using blastx against nonredundant protein sequences.

Trait Evaluation
Accessions were evaluated for various phenotypic traits
over 2 yr (2010 and 2011) as follows:
Leaf traits. Three mature (fully-expanded) leaves
from each genotype were collected, and scanned using
a scanner (Epson GT-15000). Scanned images were
imported into LAMINA software (Bylesjö et al., 2008)
to quantify several leaf parameters including leaf area,
leaf perimeter, leaf indentation, and leaf circularity.
LAMINA was calibrated using 30 × 30 mm bar, and
default settings were used for the analysis as manual
threshold value (1 to 255) = 100, minimum object size (%
of total image size) = 0.05, minimum object density (% of
total image size) = 10, serration detection pixel threshold
= 10, and boundary coordinates = 50.
Flowering. Twice a week, and for a period of 4 wk
during the months of April to May in 2010 and 2011, data
were recorded on flowering dates for each accession. Date
was recorded for each genotype when most of the flowers
of an accession were in full bloom.
Fruit quality traits. A minimum of three mature
fruits were collected from each of 160 Malus accessions.
Fruits were kept in cold storage and evaluated within 2
wk of collection for 10 fruit quality traits as described
below. For many wild accessions, fruit size was small,
and 20 to 50 fruits per genotype were collected instead.
Fruit dimension parameters. Measurements of fruit
circumference, length, and diameter at midpoint were
recorded in centimeters using calipers, while fresh weight
was recorded in grams of individual fruits. Subsequently,
data collected from each tree of an accession were averaged
over all trees to obtain a single measurement per genotype.
Peel (skin) and flesh coloration. Fruit skin pigmentation was measured using a Minolta CR-200 (Minolta
Co. Ltd., Japan) reflectance colorimeter for each of three
fruits per accession from three different locations along
the surface of a fruit. The Minolta CR-200 records colors in CIE L*a*b* color space coordinates, wherein L*
corresponds to lightness, and a* and b* correspond to
4

of

18	

chromaticity indices. For each genotype, the average
of measurements per genotype was used to calculate
chroma (C*) and hue angle (H) as follows: C* = (a*2 +
b*2)1/2 and H = tan–1 (b*/a*) (Lipka et al., 2012). Fruits
were then cut in half, and fruit flesh color was also measured using the colorimeter at two different places on the
flesh; C*and H were calculated as described above.
Sugar content and acidity: Peeled fruit flesh from
each genotype was crushed, and extracted juice was used
to determine sugar content and acidity. For wild Malus
species with small fruit sizes, at least 20 fruits were
pooled and crushed to obtain sufficient juice for acidity
and sugar content analyses. A drop of apple juice from
each accession was placed on a hand-held refractometer
(Leica Inc., Buffalo, NY). Brix values were recorded as
total soluble solids and used as equivalent to the percentage of sugar. Total titratable acid was measured by
titrating 1.0 mL of apple juice of each accession to an
endpoint pH of 8.2 using 0.01 M sodium hydroxide. The
concentration of malic acid (MA) was calculated according to the Organization for Economic Cooperation and
Development (OECD) guidelines as follows:
%MA = titer ´ AMF ´ 100/10 (mL juice);
an acid multiplication factor (AMF) of 0.0067 was used.

Distribution and Correlation of Phenotypic
Traits across Malus Species

Best linear unbiased predictions (BLUPs) were extracted
for four traits related to fruit dimension parameters, four
leaf attributes, skin and flesh color, sugar and malic acid
content, and flowering date using the linear mixed model
in R-package lme4. Pairwise Pearson correlation coefficients between traits were calculated using the data for
all accessions, and PCA was performed based on all traits,
for all species together and for different combinations of
species, using R packages gplots and FactoMineR.

Genome-Wide Association Study
Phenotypic data for 14 traits and genotypic data for
901 SNPs collected from 158 accessions were used for
Genome-Wide Association Study (GWAS). This was conducted in GAPIT (Genome Association and Prediction
Integrated Tool), an R package available at http://www.
maizegenetics.net/gapit (verified 25 July 2014). GAPIT
uses the Efficient Mixed Model Association (EMMA)
method (Lipka et al., 2012). Sequences containing SNPs
with significant associations were BLAST searched against
the NCBI database for M. × domestica ESTs using blastn
to retrieve full length sequences of ESTs. Then, retrieved
EST sequences were searched against the unigene dataset. Sequences of apple ESTs with highest similarities to
sequences of significant SNPs were determined through
BLAST searches using blastx against nonredundant protein sequences to confirm sequence identities.

the pl ant genome



november 2014



vol . 7, no . 3

Results
Relationship of the Domesticated Apple
with Wild Malus Species

A total of 160 Malus accessions, part of a core collection
of 200 accessions maintained at five different sites around
the United States, including Urbana, IL, and Geneva, NY,
previously collected from different parts of the world by
the U.S. National Plant Germplasm System, including wild
species, old and new apple cultivars, and advanced selections, were genotyped (Table 1). Of 1536 SNPs from the
GoldenGate OPAs (Khan et al., 2012a), 901 SNPs fulfilled
the filtering criteria of £15% missing values for accessions,
£10% missing values for SNPs, and with allele frequencies of ³0.05. A PCA based on these 901 SNPs revealed
that most M. × domestica (domesticated apple) accessions
grouped together (Fig. 1A). Accessions of hybrid genotypes also showed high similarities with M. × domestica,
as most hybrids were crosses between M. × domestica and
wild species or with progenies of such crosses. Overall,
most accessions of domesticated apples and hybrids clustered together. However, a few accessions of M. sieversii,
two accessions of M. orientalis, one accession of M. prunifolia, and two accessions, PI 588757 and PI 589421, identified as M. × hartwigii Koehne (M. baccata × M. halliana)
and Malus spp. (catch-all record for unidentified species)
by the USDA GRIN database, from the “others” group of
accessions, also showed stronger similarities with M. ×
domestica. Furthermore, M. sieversii and the majority of
M. orientalis accessions were grouped together, while four
out of five accessions of M. sylvestris used in this study
clustered together. Similarly, most M. micromalus accessions were also grouped together.
Bayesian analysis of genetic structure using
STRUCTURE software (Pritchard et al., 2000) has indicated that there are five clusters (K = 5) in the population
used in the study (Fig. 1B). Delta-K values (Evanno et al.,
2005) and log-likelihood (Ln p[D]) values estimated for all
160 Malus accessions as well as for 102 accessions of seven
Malus species confirmed that K = 5 is the optimal number of clusters. Of seven Malus species, only M. sylvestris
is represented by a single cluster, while four are in two
clusters and another two show admixtures from all five
clusters. Although the orange color component is clearly
attributed to the genome of M. sylvestris, three accessions
of M. sylvestris show that some portion of their genomes
belong to other clusters, mainly of green, red, and blue
color components (Fig. 1). The clusters represented by red
and green color components can be equally attributed
to genomes of either M. micromalus and M. baccata or
M. sieversii and M. orientalis, respectively. Thus, these
latter genomes cannot be readily assigned to a single
cluster. Moreover, a single accession of M. micromalus is
an admixture of the genome represented by green color,
while a single accession of M. baccata has a completely
different genome composition represented by green and
blue components of the genome. Two additional accessions of M. baccata have predominantly red components
k ahn et al .: domesti cation of apples

	

along with some contributions from orange and green
clusters. For genomes of two M. orientalis accessions, it
is revealed that they predominantly belong to the blue
cluster. The genome of M. prunifolia shows an admixture
of green, blue, and red components, wherein the proportions of blue and green components are larger than that of
the red component. As to be expected, the genome of M.
× domestica has contributions from genomes represented
by all five clusters. However, neither blue nor yellow components of the M. × domestica genome are substantially
present in other species.
Relationship assessments were conducted for 102
Malus accessions based on identity-by-state (IBS) distance matrix (1 – IBS) using 901 SNPs, and these results
were similar to those obtained using PCA and Bayesianbased clustering. Most accessions of M. × domestica
were closely related, and some accessions showed strong
genetic similarities with multiple accessions (Fig. 1C).
For example, apple cultivars Rosemary Russet, Smith
Jonathan, Jonathan, and Golden Delicious exhibited
strong genetic relationships with multiple accessions or cultivars in the collection used in this study.
Interestingly, five accessions of M. sylvestris did not have
direct relationships with M. × domestica accessions;
whereas, M. sieversii, M. orientalis, M. prunifolia, M. baccata, and M. micromalus all exhibited multiple network
connections to cultivated apples (Fig. 1C).
The NeighbourNet network has further clarified
relationships between accessions and their species (Fig.
1D). Similar to above findings, most accessions of M. ×
domestica have grouped together, and M. sieversii and
M. orientalis have also predominantly grouped together.
Furthermore, ‘Nagano’ (M. prunifolia) along with two
accessions (99TU-20-01 [PI 4539] and RUS 98 05-05 [PI
633822]) of M. orientalis have strong genetic similarities with accessions of the domesticated apple, similar
to findings obtained by PCA (Fig. 1A). The accession
‘Nagano’ may have been mislabeled, while the two M.
orientalis accessions are open-pollinated, according to
the description in GRIN, suggesting an uncertain pedigree. Four M. sieversii accessions, including KAZ 96
07-07 (PI 613958), KAZ 95 08-06 (PI 613976), KAZ 96
07-06 (PI 613994), and KAZ 96 08-16 (PI 613998) show
genetic similarities to two groups of old apple cultivars,
represented by ‘Petrel’ and ‘Irish Peach’, and are grouped
similar to our earlier observations (Fig. 1A). Another
important finding of the NeighbourNet network is the
observed genetic similarities of domesticated apples to
both M. sieversii and M. orientalis, as well as to other
wild Malus species including M. sylvestris. Among popular domesticated cultivars close to M. sieversii and M.
orientalis are Cox’s Orange Pippin, Anna, Empire, Fuji,
Nova Easygro, and McIntosh. Other popular domesticated apple cultivars that have genetic similarities with
other wild Malus species (M. sylvestris, M. baccata, M.
prunifolia, and M. micromalus) include Irish Peach,
Granny Smith, Gala, Golden Delicious, and Honey Crisp.
Three apple cultivars, Crimson Beauty (PI 589024),
5

of

18

Figure 1. (A) Principal component (PC) analysis based on genotypic data from 901 single nucleotide polymorphisms (SNPs) showing population genetic structure in 160 Malus accessions from 30 species and one hybrid group (advanced selections from breeding programs
and cultivars originating from crosses of different Malus species). Relationship of accessions is shown based on PC1 and PC2. Accessions
from seven species, as well as the hybrid group, are colored. Note: Group named “others” includes 58 accessions from rest of the species
(Table 1). (B) Population genetic structure (K = 2 to 5) in seven Malus species (M. × domestica, M. orientalis, M. sylvestris, M. sieversii,
M. micromalus, M. prunifolia, and M. baccata). Genetic structure was estimated using a Markov chain Monte Carlo (MCMC) algorithm
implemented in STRUCTURE 2.2 using a total of 901 SNPs. (C) Network relationships for 102 Malus accessions from seven species
inferred from an identity-by-state distance matrix. Species other than M. × domestica are collapsed into one circle each to emphasize the
relationship of domesticated cultivars with different wild species. The relationship is based on nearest neighbor approach implemented in
PLINK to detect outliers using 901 SNPs in 102 Malus accessions from the first nearest neighbor to the fifth nearest neighbor. (D) Network
built with Neighbor–Net option in SplitsTree 4 using 901 SNPs based on identity-by-state (IBS) distance matrix (1 – IBS).

6

of

18	

the pl ant genome



november 2014



vol . 7, no . 3

Figure 2. Genome-wide linkage disequilibrium (LD) decay patterns for (A) 160 Malus accessions and (B) M. × domestica accessions
only. The LD decay was estimated using 882 single nucleotide polymorphisms (SNPs) along 17 chromosomes in PLINK and plotted in
XLSTAT. Plots illustrate physical distances between pairs of SNPs with their corresponding r2 values; the red line indicates the trend line
based on a nonlinear logarithmic regression curve of r2 on physical distance.

Antonovka 172670-B (PI 589956), and Antonovka 1.5
Pounds, show strong similarities to the M. sieversii and
M. orientalis group as well as to M. sylvestris.

Linkage Disequilibrium Decay
A BLAST search against the apple draft genome sequence
(Velasco et al., 2010) allowed for assigning physical positions of 882 SNPs out of 901 SNPs. Based on the trend
line, an r2–value of 0.1 was determined to be the threshold
for significant LD decay. Pairwise LD patterns indicated
that for the population used in this study, all 160 Malus
k ahn et al .: domesti cation of apples

	

accessions, LD decayed rapidly to an average r2–value of
0.2 or less within 2.8 Mb of physical distance for most
SNPs (Fig. 2A); while LD decay for M. × domestica accessions was more rapid, 0.2 r2 for about 2.5 Mb (Fig. 2B).

Supporting Evidence for Selection
The Fst–based outlier detection method implemented
in LOSITAN (Antao et al., 2008) identified 51 SNPs as
potentially showing signatures of selection. A total of
16 candidates were identified following the Bayesian
genome scan in BayeScan (Foll and Gaggiotti, 2008), and
7

of

18

among these, 14 were shared with LOSITAN candidates.
Over both detection methods, a total of 67 detected
candidate SNPs have been identified. Of these, 13 were
genomic SNPs, identified using the genome sequence of
apple (Velasco et al., 2010), 54 were genic SNPs (Khan et
al., 2012a), using EST sequences, and a single genic SNP
(MdSNPui04662) whose chromosomal position could
not be established (Table 2). The remaining loci were
distributed across all 17 Malus chromosomes. The chromosome with the highest number of loci under selection

was Chromosome 7, with nine candidate loci, and this
was followed by Chromosomes 12, 5, and 17 with 8, 7,
and 6 loci, respectively (Table 2).
Allele frequencies differ significantly among species
for SNPs detected as potential candidates for selection.
BLAST searches of sequences containing candidate SNPs
against the NCBI EST database for M. × domestica using
blastn identified those corresponding Malus ESTs. This
was followed by blastx searches of EST sequences against
nonredundant protein sequences, and this provided a

Table 2. Inferred functions and gene ontology classifications of loci showing signatures of selection between
wild and domesticated apples. Loci under selection were identified using Fst-based outlier detection method
implemented in LOSITAN and a Bayesian method for genome scan implemented in BayeScan.
Single nucleotide Chromopolymorphism
some

Position (bp)

Protein ID

Unigene/blastx†

Id (%)

M.
M. ´
Len (aa) sieversii domestica

M.
hybrid

MdSNPui04662B

0

34966610

XP_003537933.1

predicted: salt tolerance protein-like isoform 1,
COL domain class transcription factor

84.5

237

1

1

1

GDsnp01916L
MdSNPui07528L
MdSNPui08653L

1
1
2

13185230
29037653
19130120

XP_002277103.1
XP_003531078.1

94.3
92.1

581
266

0.9
1
1

0.47
0.9
0.8

0.43
0.88
0.83

MdSNPui09323L
GDsnp00588L
MdSNPui08596L
MdSNPui11939L
MdSNPui01585L

2
3
3
4
4

29798888
2107141
4382784
12449449
19100858

XP_002263156.1
NP_181757.2

asparagine synthetase 1
predicted: 26S proteasome non-ATPase
regulatory subunit RPN12A-like
predicted: dnaJ protein homolog, heat shock protein
transcription factor bHLH130

96.4
77

415
122

94.8

756

0.86
0.31
0.54
0.4
0.9

0.88
0.53
0.76
0.48
0.93

MdSNPui06825L
MdSNPui06330L
MdSNPui08785L

4
5
5

22399444
291398
680454

XP_002336919.1
XP_003610595.1
XP_003607876.1

53.1
77
46.5

245
260
202

0.8
1
1

0.29
0.78
0.8

0.24
0.79
0.9

MdSNPui08770L

5

5292467

XP_002266308.2

68.3

221

1

0.8

0.66

MdSNPui04326L

5

6362466

XP_002316737.1

98.9

263

1

0.85

0.67

MdSNPui07414L
MdSNPui03960L
MdSNPui08927L
MdSNPui07915L

5
5
5
6

17952130
18984890
22399567
20851825

XP_002266481.2
NP_178171.1
NP_188284.1
XP_002275117.1

89.5
85.4
82.8
52.1

114
402
612
241

1
1
0.73
1

0.89
0.86
0.18
0.81

0.9
0.74
0.22
0.79

MdSNPui11553L
MdSNPui09322BL
MdSNPui01415L
GDsnp00699L
GDsnp02291L
MdSNPui09399L

7
7
7
7
7
7

20876
5479896
11215023
20880172
21051042
21683503

transcribed
XP_002263156.1
NP_188258.1
XP_003550092.1

acidity, vacuolar H+-translocating
inorganic pyrophosphatase
putative CC-NBS-LRR protein
Proline-rich protein
F-box protein PP2-B10, Phloem protein 2
(pp2–1 gene), cultivar Florina,
predicted: oryzain alpha chain-like,
cysteine protease, putative
chloroplast LHCII type I chlorophyll a-b
binding precursor protein
sugar, predicted: dCTP pyrophosphatase 1-like
acidity, 3-isopropylmalate dehydrogenase 2
predicted: translocase of chloroplast 132, chloroplastic-like
predicted: putative pectinesterase/
pectinesterase inhibitor 26-like
heme binding protein
dnaJ protein homolog, heat shock protein
GDSL esterase/lipase APG, zinc finger protein
transmembrane protein 70 homolog, mitochondrial-like

1
0.83
0.93
0.88
1

96.4
78.9
82.2

415
350
240

79.5

209

1
1
0.93
1
0.93
0.98

0.78
0.73
0.53
0.93
0.49
0.56

0.9
0.72
0.48
0.91
0.52
0.69

MdSNPui09396BL

7

21698261

XP_002280495.1

79.5

209

1

0.7

0.79

MdSNPui06769BL

7

22658462

XP_001699012.1

66

103

1

0.61

0.76

MdSNPui03340L

7

24570799

NP_194094.1

67.3

268

0.85

0.3

0.41

XP_002318956.1

XP_002280495.1

predicted: putative pterin-4-alphacarbinolamine dehydratase 1-like
predicted: putative pterin-4-alphacarbinolamine dehydratase
peptidyl-prolyl cis-trans isomerase, FKBP-type,
FK506-binding protein, putative
reticulon-like protein B2-like isoform 1

(cont’d)
8

of

18	

the pl ant genome



november 2014



vol . 7, no . 3

Table 2. Continued.
Single nucleotide Chromopolymorphism
some

Position (bp)

Protein ID

M.
M. ´
Len (aa) sieversii domestica

M.
hybrid

Unigene/blastx†

Id (%)

Alpha-helical ferredoxin, NADH-quinone
oxidoreductase subunit I
predicted: LOB domain-containing protein 38
predicted: extensin-3-like
elongation factor 1 alpha
predicted: probable WRKY transcription factor 23, WRKY1
fruit size, putative axial regulator YABBY 2,
YABBY domain class transcription factor
beta-ureidopropionase-like
adenine nucleotide alpha hydrolases-like protein
lipid transporter
aquaporin, MIP family, PIP subfamily, aquaporin PIP2
aquaporin, MIP family, PIP subfamily, aquaporin PIP2
protein FAF-like, chloroplastic-like
putative senescence-associated protein SAG102
Glucan endo-1,3-beta-glucosidase precursor, putative
NADH-Cytochrome b5 reductase, putative
membrane steroid-binding protein 2-like
ubiquitin-conjugating protein-like

95.5

221

0.9

0.45

0.52

81.4
53.6
98.2
78
79.8

70
166
166
195
182

0.98
1
0.93
1
1

0.54
0.84
0.53
0.87
0.87

0.43
0.72
0.45
0.93
0.86

89.7
73
62.7
91.5
91.5
56.2
58.3
83.5
92.1
70.4
100

397
241
118
283
283
104
229
478
277
186
147

58.1

362

1
1
0.65
0.9
0.88
0.98
1
0.93
1
0.85
0.85
1
1
0.95
0.98
1

0.81
0.79
0.15
0.46
0.35
0.59
0.87
0.49
0.72
0.38
0.34
0.86
0.7
0.57
0.49
0.95

0.9
0.78
0.22
0.59
0.5
0.72
0.97
0.45
0.74
0.41
0.36
0.88
0.79
0.66
0.62
0.9

86.8

121

1

0.59

0.74

AP2-associated kinase A. thaliana, serine/
threonine protein kinase, putative
rhodanese-like domain-containing protein

62.5

517

0.68
1

0.14
0.87

0.16
0.83

78.5

232

XP_002310744.1
XP_002281376.1

disease resistance protein
40S ribosomal protein S30

58.3
98.4

156
61

NP_176247.2
XP_002270485.1
XP_002273994.1
XP_003594789.1

predicted: pantothenate kinase 2-like, partial
alpha-glucan water dikinase, chloroplastic
alanine aminotransferase 2 isoform 1
Ras-related small GTP-binding protein

79.2
87.9
93.3
97.4

173
495
480
192

XP_003537933.1

predicted: salt tolerance protein-like isoform 1,
COL domain class transcription factor
putative medium chain acyl-CoA oxidase
acidity, predicted: NADP-dependent malic enzyme-like
predicted: salt tolerance protein-like isoform 1,
COL domain class transcription factor
Plastidic glucose transporter 4, putative sugar transporter

81.9

237

0.88
1
0.98
1
1
0.95
1
1
1
0.6
1

0.44
0.81
0.42
0.67
0.69
0.44
0.66
0.75
0.81
0.1
0.73

0.48
0.84
0.41
0.67
0.67
0.36
0.57
0.64
0.83
0.12
0.72

79.1
84.5
81.9

569
128
237

1
0.93
1

0.85
0.46
0.79

0.88
0.52
0.81

85.5

543

1

0.96

0.88

MdSNPui03173L

8

3443435

NP_173114.1

MdSNPui08601BL
MdSNPui06951L
MdSNPui00311L
MdSNPui11107L
MdSNPui09716L

8
8
9
9
10

12201364
21839764
15310382
31173087
7720319

XP_003545336.1
XP_002264360.2
XP_002284964.1
XP_002277882.1
XP_002277937.1

MdSNPui00408L
GDsnp00258L
MdSNPui11654L
MdSNPui04695L
MdSNPui04696L
GDsnp01640L
MdSNPui05178L
GDsnp00173L
MdSNPui01058L
MdSNPui10357L
MdSNPui01007BL
GDsnp01874L
GDsnp01647BL
GDsnp01523L
GDsnp00296BL
MdSNPui10109L

10
10
11
11
11
11
11
12
12
12
12
12
12
12
12
13

8218445
30458919
3990834
11357665
11357905
15273189
32270624
306576
1336099
1496415
26392506
30306267
30409507
31227148
31342412
25537095

XP_003563110.1
NP_001154325.1
NP_196380.4
XP_002313510.1
XP_002313510.1
XP_002279648.2
XP_003590616.1
XP_002277003.1
XP_003520929.1
XP_003553837.1
XP_002284203.1

MdSNPui05417BL

13

32069849

NP_194351.1

GDsnp01888L
MdSNPui09720L

14
14

232566
13545797

NP_850199.1

MdSNPui08089L
MdSNPui00937L
MdSNPui06548BL
MdSNPui08800BL
GDsnp01179BL
MdSNPui08640BL
MdSNPui07532BL
MdSNPui00129B
MdSNPui07063L
MdSNPui11018L
MdSNPui04666BL

14
15
15
15
16
16
16
16
16
17
17

23936419
15668840
22681289
46732617
4301368
4483457
4736021
5097313
8964268
458698
5786868

MdSNPui00867L
MdSNPui02268L
MdSNPui04667L

17
17
17

19195304
19660061
20057880

NP_172119.1
XP_003551741.1
XP_003537933.1

MdSNPui11052L

17

20831486

NP_568328.1

XP_002284638.1

NP_565969.1

predicted: 1-aminocyclopropane-1-carboxylate
oxidase homolog 1
ATP synthase subunit G protein, hydrogen-transporting
ATP synthase, rotational mechanism, putative

ATP, adenosine triphosphate; CC, coiled-coil domain; dCTP, deoxycytidine triphosphate ; dnaJ, deoxyribonucleic acid chaperon protein; FAF, Finegoldia magna adhesion factor; FKBP, FK-binding protein; GDSL, GlyAsp-Ser-(Leu) [GDS(L)] motif protein; LOB, lateral organ boundaries; LRR, leucine-rich repeat; NADH, nicotinamide adenine dinucleotide hydrogenated; NADP, nicotinamide adenine dinucleotide phosphate; NBS,
nucleotide-binding site;; PIP, prolactin-inducible protein; WRKY, conserved WRKYGQK peptide sequence and a zinc finger motif (CX4-7CX22-23HXH/C; YABBY, Adaxial–abaxial polarity transcription factor.
†

k ahn et al .: domesti cation of apples

	

9

of

18

Figure 3. Pearson correlation analysis for all fruit quality and leaf traits for 131 apple accessions. Traits are grouped based on correlations among them, lighter colors correspond to stronger correlations.

listing of predicted possible functional associations for 55
SNP loci. A total of 12 SNPs did not show similarities with
any database entry, including four genic SNPs (Khan et al.,
2012a) and eight genomic SNPs (Velasco et al., 2010).
Among those 55 loci with functional associations, five
loci (MdSNPui01585, MdSNPui07414, MdSNPui03960,
MdSNPui09716, and MdSNPui02268) showed sequence
similarities (80–95%) with transcripts predicted to be
associated with sugar metabolism, acidity, and fruit size.
Moreover, a total of seven loci on Chromosome 5 exhibited selection signatures. Of these, the SNP sequence of
MdSNPui07414 showed similarity with dCTP pyrophosphatase 1-like, a sugar related protein. A second SNP of
Chromosome 5, MdSNPui03960, showed similarity to
an acidity related protein, 3-isopropylmalate dehydrogenase 2. In addition, two other SNPs of particular interest,
MdSNPui09716 and MdSNPui02268, were located on
Chromosomes 10 and 17, respectively. An EST sequence
from NCBI (CV631928.1) harboring SNP MdSNPui09716
showed high sequence similarity (E-value = 2e–111) to
YABBY-2, a transcription factor previously reported to
be associated with tomato fruit development (Bartley
and Ishida, 2003) and more recently identified in apple
(Costa et al., 2010), exhibiting the highest levels of change
in mRNA abundance during fruit development. The
sequence of the second SNP of interest, MdSNPui02268,
was similar to that of an NADP (nicotinamide adenine
dinucleotide phosphate)-dependent malic enzyme-like
element. Overall, allele frequencies of these five SNPs were
similar in both M. × domestica and the hybrid group, but
differed significantly from allele frequencies in M. sieversii
10

of

18	

and (Table 2), which was consistent for selection of alleles
during domestication.

Distribution, Correlation, and Clustering
of Phenotypic Traits

Phenotypic data for 131 accessions were used for assessing trait distributions, correlations among traits, and
percentage contribution of traits to phenotypic diversity
using PCA. Overall, phenotypic traits were normally
distributed, with the exception of fruit size parameters,
as well as sugar content and acidity. As expected, fruits
of wild Malus species were small in size, with high
acidity and low sugar content. Pairwise Pearson correlation coefficients between these traits revealed that
fruit size, including fruit weight, diameter, length, and
circumference, yielded high positive correlations (Fig.
3). Moreover, leaf character parameters showed strong
positive correlations amongst themselves as well as with
fruit size parameters. Both fruit size and leaf parameters
had strong negative correlations with sugar content,
acidity, as well as with both peel (or skin) and flesh coloration (Fig. 3). Overall, fruit size, leaf attributes, and
flowering time were clustered together. Similarly, sugar
content, acidity, peel and flesh coloration were clustered
together (Fig. 3). Malic acid content and peel coloration
covaried in the same direction, while flesh coloration and
sugar content also covaried in the same direction (Fig.
4). The first principal component (PC1), predominantly
of fruit size parameters and leaf attributes, explained
~43% of the observed variation in Malus accessions
used in this study. The PC2, predominantly of peel, flesh
the pl ant genome



november 2014



vol . 7, no . 3

Figure 4. Principal component analysis (PCA) showing Dimension Parameters (Dim) 1 and 2 for all fruit quality traits and leaf attributes,
skin, and flesh color in 131 Malus accessions, demonstrating grouping and direction of the traits.

coloration, sugar content, and acidity, explained ~13%
of the observed variation; whereas, PC3, primarily of
flesh coloration and sugar content, explained ~9% of the
observed variation (Fig. 5).
The PCA for 131 Malus accessions, M. × domestica
vs. M. sieversii, and M. × domestica vs. M. sylvestris
did not reveal any large differences between the PC
and variation explained. For those major contributors
to PC2, including malic acid content and leaf traits,
they were generally responsible for higher variation
between M. × domestica vs. M. sieversii than for either
131 Malus accessions or M. × domestica vs. M. sylvestris
(Fig. 5). Moreover, the amount of variation explained
by fruit flesh color and sugar content, key contributors
to PC3, between M. × domestica and M. sieversii, was
k ahn et al .: domesti cation of apples

	

significantly higher compared with their contributions
toward variation among 131 Malus accessions as well
as between M. × domestica and M. sylvestris (Fig. 5).
Overall, this may point to the important role of acidity
and color traits in the divergence of M. × domestica from
M. sieversii as both parameters were negatively correlated
with fruit traits, which contributed minimally to the
observed variation between the two species.

Genome-Wide Association Study
Following Bonferroni’s correction for multiple comparisons, no significant SNPs associated with observed
phenotypic variations were detected. However, on visual
inspection, 36 SNPs with the highest p-values were
deemed to be likely candidates associated with one or
11

of

18

Figure 5. Principal component (PC) analysis showing overall contribution and importance of fruit quality attributes, leaf traits, and
skin and flesh color for the first three PCs (i.e., PC1, PC2, and PC3) for 131 accessions, between M. × domestica and M. sieversii and
between M. × domestica and M. sylvestris.

more traits from all 14 phenotypic traits evaluated (Table
3). Among these SNPs, three were from the draft apple
genome sequence (Velasco et al., 2010), while the remainder (33) were EST-derived SNPs (Khan et al., 2012a). All
these SNPs were distributed across 15 chromosomes
with no associations identified on Chromosomes 10 and
13. Each of five SNPs showed potential associations on
Chromosomes 3, 4, and 15, while four potential associations were found on Chromosome 12, and three associations were found on each of Chromosomes 5, 6, and 16.
Allele frequencies for these SNPs ranged from 0.1 to 0.5.
The EST sequences for all genic SNPs with putative
associations were obtained using blastn searches against
the apple (M. × domestica) EST NCBI database. Unigenes
of corresponding apple EST sequences with highest
similarities and blastx searches against nonredundant
protein sequences aided in assigning functions to genes
containing these SNPs. It was found that two SNPs
(MdSNPui10632, MdSNPui10633) from Chromosome 15
were deemed good candidates as they revealed associations with malic acid content and with fruit size parameters. In fact, the sequence of the MdSNPui10632 SNP,
associated with malic acid content, was predicted to be
similar to an NADP-dependent malic enzyme-like (Table
3, Fig. 6); while the sequence of the MdSNPui10633 SNP,
associated with fruit size parameters, was similar to
malate dehydrogenase (oxaloacetate-decarboxylating;
NADP+). In addition, these two SNPs were very closely
located to each other. Yet another SNP of potential interest, MdSNPui03155, located on Chromosome 11, was
associated with peel color intensity, and was present in
a gene with sequence similarity to a sucrose synthase
2-like protein (Table 3).

12

of

18	

Discussion
Population Genetic Characteristics
of Domesticated Apples

Many of the apple cultivars (M. × domestica) used in this
study are predominantly derived from and/or linked to a
few founder cultivars, thus affirming their close relationships to one another and suggesting a shared common
ancestry. For example, cultivars Rosemary Russet, Smith
Jonathan, Jonathan, Golden Delicious, and Macintosh
have strong genetic relationships with many cultivars
as demonstrated by their links to multiple accessions or
cultivars. These findings are consistent with an earlier
report (Noiton and Alspach, 1996), wherein Cox’s Orange
Pippin, Golden Delicious, Red Delicious, Jonathan, and
McIntosh have been deemed as the most frequent founder
cultivars. However, Red Delicious was not included in
our study sample. In addition, Cox’s Orange Pippin
was not detected as a founder and this might be due to
the composition of the different sets of cultivars used
in this study. This may also account for the detection of
Rosemary Russet as a key founder in this study, but not in
a previous report (Noiton and Alspach, 1996).
Overall, a high level of diversity is observed within the
M. × domestica collection used in this study. Considering
the global genetic diversity in domesticated apples, it is
reasonable to assume that a certain level of genetic diversity for various economic horticultural traits is present
within M. × domestica. It is known that most current
apple cultivars have been identified either as chance
seedlings or bud sports of a single cultivar or developed
from breeding programs (Noiton and Alspach, 1996).
According to Evans et al. (2011), the majority of apple
breeding programs around the world, and particularly
European apple breeding programs, rely on improved
the pl ant genome



november 2014



vol . 7, no . 3

Table 3. Single nucleotide polymorphisms (SNPs) significantly associated with 14 phenotypic traits identified using Genome-Wide Association Study (GWAS) in GAPIT (Genome Association and Prediction Integrated Tool), an R
package. GAPIT uses Efficient Mixed Model Association (EMMA) method55. Sequences of SNPs showing significant
associations were BLAST searched against the National Center for Biotechnology Information database for Malus
× domestica ESTs using blastn and the unigenes for the corresponding apple EST sequence with highest similarity
were recorded. The EST sequence was also BLAST searched using blastx against nonredundant protein sequences
for homology. Chromosome position (bp), minor allele frequency (MAF), protein id from sequence blast results for
top hit against nonredundant protein sequences, and GWAS significance level (p-value) for the trait for each SNP is
also provided.
SNPs

Chrom

Position
(bp)

Protein
ID

Unigene/blastx†

MAF

MdSNPui04946

1

9582771

XP_003633964.1

reticulon-like protein B1 isoform 2 V. vinifera 79.0 214

0.3

MdSNPui07247

2

8794039

XP_002276608.1

MLO9 protein V. vinifera 79.2 557

0.1

MdSNPui11355

3

1020516

NP_001031195.4

SC35-like splicing factor 33 A. thaliana 75.2 214

0.4

MdSNPui01364

3

7039674

NP_189621.1

general transcription factor 2-related zinc finger protein A. thaliana 47.0 230

0.2

MdSNPui00312

3

9737978

XP_003547694.1

elongation factor 1-alpha-like isoform 1 G. max 98.7 446

0.3

MdSNPui07830

3

16053768

XP_003634629.1

probable LRR receptor-like serine/threonine-protein
kinase At4g08850-like V. vinifera 75.7 470

0.4

MdSNPui05042

3

32301103

XP_003520230.1

glutathione S-transferase-like G. max 77.4 212

0.3

MdSNPui11046

4

4403759

XP_003525411.1

PR-5 protein G. max 85.6 221

0.2

MdSNPui02526

4

4548762

NP_176500.1

rhomboid-like 2 A. thaliana 78.2 243

0.4

MdSNPui07106

4

4692027

NP_567380.2

Ara4-interacting protein A. thaliana 72.3 158

0.3

MdSNPui11858

4

6878878

NP_196891.1

xyloglucan:xyloglucosyl transferase A. thaliana 83.3 286

0.3

MdSNPui11898

4

11733909

NP_190930.1

pyrophosphorylase 4 A. thaliana 89.3 215

0.3

MdSNPui07242

5

12313108

NP_565719.1

actin depolymerizing factor 6 A. thaliana 83.6 145

0.3

MdSNPui02859

5

17835394

NP_193544.1

ribosomal protein L32–1 A. thaliana 88.7 132

0.3

GDsnp02729

5

27341501

MdSNPui07718

6

4990256

NP_568645.1

duplicated SANT DNA-binding domain-containing protein A. thaliana 45.5 121

0.3

MdSNPui07719

6

4990499

NP_568645.1

duplicated SANT DNA-binding domain-containing protein A. thaliana 45.5 121

0.3

0.003

0.004
0.002

0.1

MdSNPui01624

6

5961708

NP_567522.5

peptidyl-prolyl cis-trans isomerase A. thaliana 71.2 333

0.5

MdSNPui02362

7

21049993

XP_003541082.1

PREDICTED: 4-hydroxy-3-methylbut-2-enyl diphosphate
reductase-like G. max 91.8 379

0.5

MdSNPui11906

8

13931723

XP_002274966.1

profilin-1 isoform 1 V. vinifera 87.0 130

0.4

MdSNPui04262

9

1793397

XP_003553833.1

carbonic anhydrase, chloroplastic-like isoform 1 G. max 80.3 329

0.4

MdSNPui03155

11

34496174

XP_003521575.1

sucrose synthase 2-like G. max 87.0 59

0.2

MdSNPui07111

12 21458032

XP_003537829.1

V-type proton ATPase 16 kDa proteolipid subunit-like isoform 1 G. max 100.0 163

0.1

MdSNPui07110

12 21458804

XP_003537829.1

V-type proton ATPase 16 kDa proteolipid subunit-like isoform 1 G. max 100.0 163

0.3

MdSNPui06752

12 30130898

NP_001236007.1

putative alto/keto reductase G. max 83.4 319

0.2

GDsnp02228

12 30184029

MdSNPui10600

14 17709952

0.1
XP_002269986.1

putative Holliday junction resolvase V. vinifera 83.8 167

0.4

NADP-dependent malic enzyme-like G. max 89.2 582

0.3

malate dehydrogenase (oxaloacetate-decarboxylating)(NADP+) A. thaliana 86.7 578

0.3

MdSNPui10632

15

3202683

XP_003522719.1

MdSNPui10633

15

3203185

NP_196728.1

MdSNPui05276

15

12489150

XP_003524915.1

70 kDa peptidyl-prolyl isomerase-like isoform 1 G. max 85.4 547

0.3

MdSNPui02219

15

14789358

XP_002285072.1

DEAD-box ATP-dependent RNA helicase 56-like V. vinifera 94.2 427

0.2

MdSNPui04592

15

25096687

XP_003540182.1

probable serine/threonine-protein kinase DDB_G0279405-like G. max 80.5 169

0.3

MdSNPui04087

16

2650965

NP_001189504.1

uvrB/uvrC motif-containing protein A. thaliana 78.5 219

GDsnp01179

16 4301368

MdSNPui08271

16

15864142

NP_188935.1

MdSNPui05184

17

11061114

XP_003580532.1

†

FlowerFruit
ing circumference

0.001

0.4
0.2

ADP-ribosylation factor C1 A. thaliana 92.8 181

0.2

36.4 kDa proline-rich protein-like B. distachyon 60.4 105

0.4

0.001

ADP, adenosine triphosphate; NADP, nicotinamide adenine dinucleotide.

k ahn et al .: domesti cation of apples

	

13

of

18

Figure 6. Genome-wide association study (GWAS) for leaf area, total soluble solids, and fruit weight for 158 apple accessions showing
single nucleotide polymorphisms (SNPs) with significant associations compared with background. Colors represent SNPs located on
alternating chromosomes.

traditional cultivars that have been either bred or discovered as chance seedlings several decades or centuries
ago. Thus, only a few cultivars have served as the germplasm pool of most modern cultivars. Concerns over the
use of a narrow pool of cultivars of domesticated apples
in breeding programs have also been raised previously
(Noiton and Alspach 1996), as this would highly influence
future sustainability of apple production by decreasing
horticultural and fruit quality attributes, rendering them
prone to diseases and abiotic stresses. Similar to Noiton
and Alspach (1996), a thoughtful approach for increasing the genetic basis of our apple improvement programs
is needed. Patzak et al. (2012) used 10 SSRs in 273 M. ×
domestica accessions, including 130 reference world cultivars and 143 old and local genotypes. Their results suggest
that many recent breeding programs are exploiting wider
diversity in the Malus germplasm, thus contributing to
enhanced sustainability of the domesticated apple.

14

of

18	

Relationships of Wild Malus Species
with M. × domestica
The observed high numbers of relationships between
M. sieversii and domesticated apples in the network
analysis (Fig. 1C) suggest that M. sieversii is the primary
progenitor of M. ´ domestica, as previously reported in
earlier studies (Velasco et al., 2010; Harrison and Harrison, 2011; Cornille et al., 2012). The elevated LD in the
pooled dataset compared to M. × domestica likely corresponds to genome-wide genetic differentiation among
the different Malus species. There are also secondary
introgressions from multiple Malus species into the M.
× domestica genome following its origination from M.
sieversii. Bayesian-based population genetic structure
analysis (Fig. 1B) identified five clusters (K = 5). In addition to M. sylvestris, a total of seven species could not be
separated into distinct clusters. M. micromalus and M.
baccata showed similar compositions of their genomes
while M. sieversii and M. orientalis showed similar patterns in their genomes. This might be due to the fact
that markers used for assessing these relationships were
not sufficiently informative to distinguish beyond two
the pl ant genome



november 2014



vol . 7, no . 3

clusters for these four species. However, patterns shared
by these species indicated that M. micromalus and M.
baccata were as close to each other as were M. sieversii
and M. orientalis, and these relationships were further
confirmed using additional analyses (Fig. 1A, C, D). As
for accessions showing admixture, this could be due to
likely coancestry or secondary introgression from cocultivation or original location of the plant material.
In this study, the genome composition of M. ×
domestica shows admixture for all five clusters, but it has
a major contribution from the blue component and some
contribution from the yellow component (Fig. 1B). It is
notable that both blue and yellow population components are hardly represented in the wild species used in
the study. Based on these findings, it is proposed that the
closest wild progenitor populations are either unsampled
in this study or they are extinct. However, as presented in
previous studies (Velasco et al., 2010; Cornille et al., 2012),
substantial contributions from M. sieversii, the presumed
progenitor, and M. sylvestris, the major secondary contributor to the domesticated apple genome, are observed.
As mentioned above, these findings are in agreement
with previous studies (Velasco et al., 2010; Harrison
and Harrison, 2011; Cornille et al., 2012), reporting
that M. × domestica must have originated from M.
sieversii. However, these results do not offer sufficient
confidence in the claim that M. sylvestris, the European
wild apple, has made only minor contributions to M.
× domestica when compared with other wild species of
Malus. Whether or not M. sylvestris harbors traits different from all other wild apple species that will render it
uniquely capable of introgressing into M. × domestica is
yet to be determined. Moreover, it is important to assess
the extent to which domesticated apples from specific
regions carry hallmarks of wild apples from their neighborhoods. Depending on the set of accessions and their
geographic regions of origin, these factors can lead to different determinations of such relationships. In this study,
several accessions of both M. orientalis and M. prunifolia
have been included, and findings have suggested that
both species have made considerable contributions to the
domesticated apple along the Silk Road. For example, it
has been previously reported that ‘Mandshurica 2330’ (PI
322713) is a separate species (‘Mandshurica’; Velasco et
al., 2010). However, this is in contrast to previous USDA
reports as well as to findings obtained in this study with
twice as many M. baccata included, which demonstrate
that the accession PI 322713 is indeed similar to other M.
baccata accessions. Thus, a wider sampling of accessions
along the Silk Road will provide a more comprehensive
overview of the relationships among various Malus species, as well as their relationships and contributions to
the domesticated apple, M. × domestica, genome.
In a recent study (Cornille et al., 2012), a comprehensive set of 839 accessions consisting of 368, 168, 215,
40, and 48 accessions of M. × domestica, M. sieversii, M.
orientalis, M. sylvestris, and M. baccata, is used to assess
genetic relationships among Malus using 26 SSR markers.
k ahn et al .: domesti cation of apples

	

Based on their findings, M. sieversii has been confirmed
as the progenitor of the domesticated apple. Moreover,
it has been suggested there is a bidirectional gene flow
between domesticated apples and their wild species, and
M. sylvestris is deemed as the main secondary contributor to the M. × domestica genome (Cornille et al., 2012).
Nevertheless, these findings remain inconclusive as to
the extent of contribution as well as the number of introgressed wild species into the M. × domestica genome.

Signatures of Selection across the Malus
Genome and Corresponding Distribution
of Phenotypic Traits in M. × domestica vs.
Wild Species

During genome evolution of domesticates, genetic regions
as well as genes undergoing selection are predicted to
show significant reductions in genetic diversity when
compared with neutral genes, as well as to higher levels of differentiation between the progenitor and the
domesticated population (Tanksley and McCouch, 1997;
Yamasaki et al., 2005). In this study, a total of 67 loci,
including 13 genomic SNPs (Velasco et al., 2010) and 54
genic SNPs (Khan et al., 2012a) have been identified as
potential targets for selection during the evolution of the
M. × domestica genome from its wild progenitor M. sieversii. It is likely that there are some false positives among
these detected loci. Nevertheless, loci identified by both
Fst and Bayesian-based outlier detection methods, serving as potential targets for selection, as well as exhibiting
significant differences in allele frequencies among M. ×
domestica and M. sieversii, are likely to be strong candidates. Regardless of the possible role of these loci in Malus
domestication, these are proposed as potential candidates
for selection, and deserve further investigation.
Except for a single SNP (MdSNPui04662), chromosomal positions of 66 SNPs were established along all 17
LGs of the Malus genome. It is presumed that genes associated with key traits for apple domestication are present
in the regions with most loci under selection (Table 2).
Besides clusters of disease resistance, quantitative trait
loci (QTLs) for fire blight disease (Calenge et al., 2005;
Durel et al., 2009; Khan et al., 2006; Khan et al., 2012b),
and QTLs for plant vigor and fruit quality traits of apple
have also been reported to be present along these LGs
(King et al., 2001; Liebhard et al., 2003; Longhi et al.,
2012). Quantitative trait loci with moderate effects for
variation in fruit texture (sensory traits, crispiness, and
firmness, among others) on LGs 7, 5, and 12 have been
reported (King et al., 2001). In addition, a strong QTL
(LOD = 7) for flowering time on LG 17 has also been
identified (Liebhard et al., 2003). Furthermore, QTLs
for each of fruit flesh firmness and fruit weight, located
on LG 12, along with two QTLs for flowering time with
moderate effects, located on LG 7, as well as two QTLs
for increased plant length, located on LG 17, have also
been reported (Liebhard et al., 2003).
15

of

18

In this study, marked differences in allele frequencies for loci between domesticated and wild apples have
been observed, suggesting that selection against one
of the alleles might have occurred. It is interesting to
note that allele frequencies in M. × domestica are in
agreement with those of the hybrid group used in this
study, thus suggesting that similar forces of selection
are active in genomes of both groups. Based on BLAST
searches of sequences containing candidate SNPs, five
loci (MdSNPui01585, MdSNPui07414, MdSNPui03960,
MdSNPui09716, and MdSNPui02268) show >80%
sequence similarities to genes known to control sugar
content, acidity, and fruit size. These are key traits during
early domestication events of apple genome as reported
previously by Juniper and Mabberley (2006).
On LG 5, a single sequence containing SNP
MdSNPui07414 shows similarity to dCTP pyrophosphatase 1-like a sugar related protein, while the sequence of
another SNP, MdSNPui03960, from LG 5 shows similarity to an acidity related protein, 3-isopropylmalate dehydrogenase 2. Two additional SNPs of major interest are
MdSNPui09716 and MdSNPui02268, located on LGs 10
and 17, respectively. The sequence of SNP MdSNPui09716
shows similarity to a previously predicted fruit qualityrelated locus (Costa et al., 2010), while the sequence of
SNP MdSNPui02268 is similar to NADP-dependent
malic enzyme-like. Altogether, these five loci are potentially the best candidates for future studies.
In this study, distribution and correlation of traits
in PCA analyses also point towards the potential role
of specific phenotypic traits in apple domestication. In
general, fruits of wild Malus species are characterized by
small size, high acidity, and low sugar content. This is also
accompanied with strong positive correlations among
fruit dimension parameters. Furthermore, these parameters have strong negative correlations with sugar content,
acidity (malic acid content), as well as with both peel and
flesh coloration (Fig. 3). These phenotypic covariance
findings suggest that when larger fruit size is favored
during domestication, selection for low acidity and high
sugar content is observed. The positive correlations of
malic acid content and peel color, as well as of flesh color
and sugar content, suggest that skin coloration serves as
an indicator of fruit ripening and exposure to sun would
contribute to accumulation of high sugar content. The
PC1 of phenotypic traits, which predominantly involves
fruit dimension parameters and leaf attributes, explains
~43% of the variation, but does not seem to account for
observed variation between M. × domestica and M. sieversii. The malic acid content, as well as flesh color and sugar
content, main contributors to variations explained by PC2
and PC3, respectively, accounted for more of the observed
variation between M. × domestica and M. sieversii than
among all 131 accessions or between M. × domestica and
M. sylvestris (Fig. 5), and are likely to be key contributing
traits from M. sieversii during domestication of apples.
In contrast to previous studies (Juniper and Mabberley,
2006), fruit dimension parameters are significantly
16

of

18	

different between M. × domestica and M. sieversii; however, they do not seem to be the main traits under selection between these two species. Instead, our findings indicate that other sensory attributes, such as color and taste,
in the form of low acidity, are more important during the
domestication process.

Genome-Wide Association Study
Following Bonferroni’s correction of multiple comparisons
in significance tests, none of the SNPs above the significance threshold could be identified for trait associations in
the GWAS. This could be due to the following two reasons:
(i) rapid LD decay, and (ii) the small size of the association mapping population, which can both influence the
power of association mapping (Khan and Korban, 2012). It
is expected that, with a larger size population and a more
dense SNP coverage across the genome, signals of SNPs
and genomic regions would be above the threshold. To
assess whether trends toward significant associations were
present in our dataset, every LG was visually inspected
to identify trends for elevated associations. A total of 36
SNPs with the highest p-values were identified using visual
inspection (Table 3), and as valuable candidates for further
research, are reported here. The distribution of SNPs on
different chromosomes provided further support of the
level of genetic complexity of the traits. Although these
were not significant, elevated associations with traits and
sequence similarities with transcripts relevant to key traits
emphasized the potential roles of the SNPs in trait expression. For example, SNPs MdSNPui10632 and MdSNPui10633 on LG 15 showing associations with malic acid
and fruit dimension parameters along with sequence similarities to malic acid metabolism related enzymes, as well
as SNP MdSNPui03155 on LG 11 with association to skin
color pigmentation (or hue) and sequence similarity with
sucrose synthase 2-like protein, are all good candidates for
further studies (Table 3).

Supplemental Information Available
Supplemental information is included with this article.
Acknowledgments

This work was funded by a grant received from USDA-NIFA-SCRI
grant AG 2009-51181-06023. Partial funding was also provided by the
University of Illinois Office of Research Project 875-325, and University of
Illinois Office of Research Project 875-922.

References

Antao, T., A. Lopes, R. Lopes, A. Beja-Pereira, and G. Luikart. 2008.
LOSITAN: A workbench to detect molecular adaptation based on a Fstoutlier method. BMC Bioinform. 9:323. doi:10.1186/1471-2105-9-323
Bartley, G.E., and B.K. Ishida. 2003. Developmental gene regulation during
tomato fruit ripening and in-vitro sepal morphogenesis. BMC Plant
Biol. 3:4. doi:10.1186/1471-2229-3-4
Batagelj, V., and A. Mrvar. 2004. Pajek—Analysis and visualization of large
networks. Springer, Berlin. Available at http://link.springer.com/content/pdf/10.1007/978-3-642-18638-7_4 (verified 18 July 2014).
Bylesjö, M., V. Segura, R.Y. Soolanayakanahally, A.M. Rae, J. Trygg, P.
Gustafsson, S. Jansson, and N.R. Street. 2008. LAMINA: A tool for
rapid quantification of leaf size and shape parameters. BMC Plant Biol.
8:82. doi:10.1186/1471-2229-8-82
the pl ant genome



november 2014



vol . 7, no . 3

Calenge, F., D. Drouet, C. Denancé, W.E. Van de Weg, M.-N. Brisset, J.-P.
Paulin, and C.-E. Durel. 2005. Identification of a major QTL together
with several minor additive or epistatic QTLs for resistance to fire blight
in apple in two related progenies. Theor. Appl. Genet. 111:128–135.
doi:10.1007/s00122-005-2002-z
Clark, R.M., E. Linton, J. Messing, and J.F. Doebley. 2004. Pattern of diversity in the genomic region near the maize domestication gene tb1. Proc.
Natl. Acad. Sci. USA 101:700–707. doi:10.1073/pnas.2237049100
Cornille, A., P. Gladieux, M.J. Smulders, I. Roldán-Ruiz, F. Laurens, B. Le
Cam, A. Nersesyan, J. Clavel, M. Olonova, L. Feugey, I. Gabrielyan, X.G.
Zhang, M.I. Tenaillon, and T. Giraud. 2012. New insight into the history
of domesticated apple: Secondary contribution of the European wild
apple to the genome of cultivated varieties. PLoS Genet. 8:e1002703.
doi:10.1371/journal.pgen.1002703
Costa, F., C.P. Peace, S. Stella, S. Serra, S. Musacchi, M. Bazzani, S.
Sansavini, and W.E. Van de Weg. 2010. QTL dynamics for fruit firmness and softening around an ethylene-dependent polygalacturonase
gene in apple (Malus × domestica Borkh.). J. Exp. Bot. 61:3029–3039.
doi:10.1093/jxb/erq130
Durel, C.-E., C. Denancé, and M.-N. Brisset. 2009. Two distinct major QTL
for resistance to fire blight co-localize on linkage group 12 in apple
genotypes Evereste and Malus floribunda clone 821. Genome 52:139–
147. doi:10.1139/G08-111
Evanno, G., S. Regnaut, and J. Goudet. 2005. Detecting the number of
clusters of individuals using the software STRUCTURE: A simulation
study. Mol. Ecol. 14:2611–2620. doi:10.1111/j.1365-294X.2005.02553.x
Evans, K.M., A. Patocchi, F. Rezzonico, F. Mathis, C.E. Durel, F. FernándezFernández, A. Boudichevskaia, F. Dunemann, M. Stankiewicz-Kosyl, L.
Gianfranceschi, M. Komjanc, M. Lateur, M. Madduri, Y. Noordijk, and
W.E. van de Weg. 2011. Genotyping of pedigreed apple breeding material
with a genome-covering set of SSRs: Trueness-to-type of cultivars and
their parentages. Mol. Breed. 28:535–547. doi:10.1007/s11032-010-9502-5
Foll, M., and O. Gaggiotti. 2008. A genome-scan method to identify
selected loci appropriate for both dominant and codominant markers:
A Bayesian perspective. Genetics 180:977–993. doi:10.1534/genetics.108.092221
Fuller, D.Q., L. Qin, Y. Zheng, Z. Zhao, X. Chen, L.A. Hosoya, and G.-P.
Sun. 2009. The domestication process and domestication rate in
rice: Spikelet bases from the Lower Yangtze. Science 323:1607–1610.
doi:10.1126/science.1166605
Gepts, P. 2004. Crop domestication as a long-term selection experiment.
Plant Breed. Rev. 24:1–44.
Götherström, A., C. Anderung, L. Hellborg, R. Elburg, C. Smith, D.G.
Bradley, and H. Ellegren. 2005. Cattle domestication in the Near East
was followed by hybridization with aurochs bulls in Europe. Proc. Royal
Soc. B Biol. Sci. 272:2345–2351.
Hammer, K. 1984. Das domestikationssyndrom. Kulturpflanze 32:11–34.
doi:10.1007/BF02098682
Harada, T., W. Kurahashi, M. Yanai, Y. Wakasa, and T. Satoh. 2005.
Involvement of cell proliferation and cell enlargement in increasing
the fruit size of Malus species. Sci. Hortic. (Amsterdam) 105:447–456.
doi:10.1016/j.scienta.2005.02.006
Harris, S.A., J.P. Robinson, and B.E. Juniper. 2002. Genetic clues to the
origin of the apple. Trends Genet. 18:426–430. doi:10.1016/S01689525(02)02689-6
Harrison, N., and R.J. Harrison. 2011. On the evolutionary history of the
domesticated apple. Nat. Genet. 43:1043–1044. doi:10.1038/ng.935
Huang, X., N. Kurata, X. Wei, Z.X. Wang, A. Wang, Q. Zhao, Y. Zhao,
K. Liu, H. Lu, W. Li, Y. Guo, Y. Lu, C. Zhou, D. Fan, Q. Weng, C.
Zhu, T. Huang, L. Zhang, Y. Wang, L. Feng, H. Furuumi, T. Kubo, T.
Miyabayashi, X. Yuan, Q. Xu, G. Dong, Q. Zhan, C. Li, A. Fujiyama, A.
Toyoda, T. Lu, Q. Feng, Q. Qian, J. Li, and B. Han. 2012. A map of rice
genome variation reveals the origin of cultivated rice. Nature 490:497–
503. doi:10.1038/nature11532
Hufford, M.B., X. Xu, J. van Heerwaarden, T. Pyhäjärvi, J.M. Chia, R.A.
Cartwright, R.J. Elshire, J.C. Glaubitz, K.E. Guill, S.M. Kaeppler, J. Lai,
P.L. Morrell, L.M. Shannon, C. Song, N.M. Springer, R.A. SwansonWagner, P. Tiffin, J. Wang, G. Zhang, J. Doebley, M.D. McMullen, D.
Ware, E.S. Buckler, S. Yang, and J. Ross-Ibarra. 2012. Comparative
k ahn et al .: domesti cation of apples

	

population genomics of maize domestication and improvement. Nat.
Genet. 44:808–813. doi:10.1038/ng.2309
Hummer, K.E., and J. Janick. 2009. Rosaceae: Taxonomy, economic importance, genomics. In: K.M. Folta and S.E. Gardiner, editors, Genetics and
genomics of Rosaceae. Springer, New York. p. 1–17.
Huson, D.H., and D. Bryant. 2006. Application of phylogenetic networks
in evolutionary studies. Mol. Biol. Evol. 23:254–267. doi:10.1093/molbev/msj030
Juniper, B.E., and D.J. Mabberley. 2006. The story of the apple. Timber
Press, Portland, OR.
Kellerhals, M. 2009. Introduction to apple (Malus × domestica). In: K.M.
Folta and S.E. Gardiner, editors, Genetics and genomics of Rosaceae.
Springer, New York. p. 73–84.
Khan, M.A., B. Duffy, C. Gessler, and A. Patocchi. 2006. QTL mapping
of fire blight resistance in apple. Mol. Breed. 17:299–306. doi:10.1007/
s11032-006-9000-y
Khan, M.A., Y. Han, Y.F. Zhao, and S.S. Korban. 2012a. A high-throughput
apple SNP genotyping platform using the GoldenGateTM assay. Gene
494:196–201. doi:10.1016/j.gene.2011.12.001
Khan, M.A., and S.S. Korban. 2012. Association mapping in forest trees and
fruit crops. J. Exp. Bot. 63:4045–4060. doi:10.1093/jxb/ers105
Khan, M.A., Y. Zhao, and S.S. Korban. 2012b. Molecular mechanisms of
pathogenesis and resistance to the bacterial pathogen Erwinia amylovora, causal agent of fire blight disease in Rosaceae. Plant Mol. Biol. Rep.
30:247–260. doi:10.1007/s11105-011-0334-1
King, G.J., J.R. Lynn, C.J. Dover, K.M. Evans, and G.B. Seymour. 2001.
Resolution of quantitative trait loci for mechanical measures accounting for genetic variation in fruit texture of apple (Malus pumila Mill.).
Theor. Appl. Genet. 102:1227–1235. doi:10.1007/s001220000530
Konishi, S., T. Izawa, S.Y. Lin, K. Ebana, Y. Fukuta, T. Sasaki, and M. Yano.
2006. An SNP caused loss of seed shattering during rice domestication.
Science 312:1392–1396. doi:10.1126/science.1126410
Korban, S.S., and S. Tartarini. 2009. Apple structural genomics. In: K.M.
Folta and S.E. Gardiner, editors, Genetics and genomics of Rosaceae.
Springer, New York. p. 85–119.
Liebhard, R., M. Kellerhals, W. Pfammatter, M. Jertmini, and C.
Gessler. 2003. Mapping quantitative physiological traits in
apple (Malus × domestica Borkh.). Plant Mol. Biol. 52:511–526.
doi:10.1023/A:1024886500979
Lipka, A.E., F. Tian, Q. Wang, J. Peiffer, M. Li, P.J. Bradbury, M.A. Gore,
E.S. Buckler, and Z. Zhang. 2012. GAPIT: Genome association and
prediction integrated tool. Bioinformatics 28:2397–2399. doi:10.1093/
bioinformatics/bts444
Longhi, S., M. Moretto, R. Viola, R. Velasco, and F. Costa. 2012.
Comprehensive QTL mapping survey dissects the complex fruit texture
physiology in apple (Malus × domestica Borkh.). J. Exp. Bot. 63:1107–
1121. doi:10.1093/jxb/err326
Malladi, A., and P.M. Hirst. 2010. Increase in fruit size of a spontaneous
mutant of ‘Gala’ apple (Malus × domestica Borkh.) is facilitated by
altered cell production and enhanced cell size. J. Exp. Bot. 61:3003–
3013. doi:10.1093/jxb/erq134
Miller, A.J., and B.L. Gross. 2011. From forest to field: Perennial fruit crop
domestication. Am. J. Bot. 98:1389–1414. doi:10.3732/ajb.1000522
Myles, S., A.R. Boyko, C.L. Owens, P.J. Brown, F. Grassi, M.K. Aradhya,
B. Prins, A. Reynolds, J.M. Chia, D. Ware, C.D. Bustamante, and E.S.
Buckler. 2011. Genetic structure and domestication history of the grape.
Proc. Natl. Acad. Sci. USA 108:3530–3535. doi:10.1073/pnas.1009363108
Noiton, D.A., and P.A. Alspach. 1996. Founding clones, inbreeding, coancestry, and status number of modern apple cultivars. J. Am. Soc. Hortic.
Sci. 121:773–782.
Olsen, K.M., and J.F. Wendel. 2013. A Bountiful Harvest: Genomic insights
into crop domestication phenotypes. Annu. Rev. Plant Biol. 64:47–70.
doi:10.1146/annurev-arplant-050312-120048
Outram, A.K., N.A. Stear, R. Bendrey, S. Olsen, A. Kasparov, V. Zaibert,
N. Thorpe, and R.P. Evershed. 2009. The earliest horse harnessing and
milking. Science 323:1332–1335. doi:10.1126/science.1168594
Palaisa, K., M. Morgante, S. Tingey, and A. Rafalski. 2004. Long-range patterns of diversity and linkage disequilibrium surrounding the maize Y1
gene are indicative of an asymmetric selective sweep. Proc. Natl. Acad.
Sci. USA 101:9885–9890. doi:10.1073/pnas.0307839101

17

of

18

Paran, I., and E. Van der Knaap. 2007. Genetic and molecular regulation of
fruit and plant domestication traits in tomato and pepper. J. Exp. Bot.
58:3841–3852. doi:10.1093/jxb/erm257
Patzak, J., F. Paprstein, A. Henychová, J. Sedlák, and D. Somers. 2012.
Comparison of genetic diversity structure analyses of SSR molecular marker data within apple (Malus × domestica) genetic resources.
Genome 55:647–665. doi:10.1139/g2012-054
Pollinger, J.P., K.E. Lohmueller, E. Han, H.G. Parker, P. Quignon, J.D.
Degenhardt, A.R. Boyko, D.A. Earl, A. Auton, A. Reynolds, K. Bryc, A.
Brisbin, J.C. Knowles, D.S. Mosher, T.C. Spady, A. Elkahloun, E. Geffen,
M. Pilot, W. Jedrzejewski, C. Greco, E. Randi, D. Bannasch, A. Wilton,
J. Shearman, M. Musiani, M. Cargill, P.G. Jones, Z. Qian, W. Huang,
Z.L. Ding, Y.P. Zhang, C.D. Bustamante, E.A. Ostrander, J. Novembre,
and R.K. Wayne. 2010. Genome-wide SNP and haplotype analyses
reveal a rich history underlying dog domestication. Nature 464:898–
902. doi:10.1038/nature08837
Potts, S.M., Y. Han, M.A. Khan, M.M. Kushad, A.L. Rayburn, and S.S.
Korban. 2012. Genetic diversity and characterization of a core collection of Malus germplasm using simple sequence repeats (SSRs). Plant
Mol. Biol. Rep. 30:827–837. doi:10.1007/s11105-011-0399-x
Pritchard, J.K., M. Stephens, and P. Donnelly. 2000. Inference of population
structure using multilocus genotype data. Genetics 155:945–959.
Purcell, S., B. Neale, K. Todd-Brown, L. Thomas, M.A. Ferreira, D. Bender,
J. Maller, P. Sklar, P.I. de Bakker, M.J. Daly, and P.C. Sham. 2007.
PLINK: A tool set for whole-genome association and population-based
linkage analyses. Am. J. Hum. Genet. 81:559–575. doi:10.1086/519795
Purugganan, M.D., and D.Q. Fuller. 2009. The nature of selection during
plant domestication. Nature 457:843–848. doi:10.1038/nature07895
Tanksley, S.D., and S.R. McCouch. 1997. Seed banks and molecular maps:
Unlocking genetic potential from the wild. Science 277:1063–1066.
doi:10.1126/science.277.5329.1063

18

of

18	

Tian, F., N.M. Stevens, and E.S. Buckler. 2009. Tracking footprints of maize
domestication and evidence for a massive selective sweep on chromosome 10. Proc. Natl. Acad. Sci. USA 106:9979–9986. doi:10.1073/
pnas.0901122106
Velasco, R., A. Zharkikh, J. Affourtit, A. Dhingra, A. Cestaro, A.
Kalyanaraman, P. Fontana, S.K. Bhatnagar, M. Troggio, D. Pruss, S.
Salvi, M. Pindo, P. Baldi, S. Castelletti, M. Cavaiuolo, G. Coppola,
F. Costa, V. Cova, A. Dal Ri, V. Goremykin, M. Komjanc, S. Longhi,
P. Magnago, G. Malacarne, M. Malnoy, D. Micheletti, M. Moretto,
M. Perazzolli, A. Si-Ammour, S. Vezzulli, E. Zini, G. Eldredge, L.M.
Fitzgerald, N. Gutin, J. Lanchbury, T. Macalma, J.T. Mitchell, J. Reid, B.
Wardell, C. Kodira, Z. Chen, B. Desany, F. Niazi, M. Palmer, T. Koepke,
D. Jiwan, S. Schaeffer, V. Krishnan, C. Wu, V.T. Chu, S.T. King, J. Vick,
Q. Tao, A. Mraz, A. Stormo, K. Stormo, R. Bogden, D. Ederle, A. Stella,
A. Vecchietti, M.M. Kater, S. Masiero, P. Lasserre, Y. Lespinasse, A.C.
Allan, V. Bus, D. Chagné, R.N. Crowhurst, A.P. Gleave, E. Lavezzo, J.A.
Fawcett, S. Proost, P. Rouzé, L. Sterck, S. Toppo, B. Lazzari, R.P. Hellens,
C.E. Durel, A. Gutin, R.E. Bumgarner, S.E. Gardiner, M. Skolnick, M.
Egholm, Y. Van de Peer, F. Salamini, and R. Viola. 2010. The genome
of the domesticated apple (Malus × domestica Borkh.). Nat. Genet.
42:833–839. doi:10.1038/ng.654
Wang, R.-L., A. Stec, J. Hey, L. Lukens, and J. Doebley. 1999. The limits of selection during maize domestication. Nature 398:236–239.
doi:10.1038/18435
Yamasaki, M., M.I. Tenaillon, I.V. Bi, S.G. Schroeder, H. Sanchez-Villeda,
J.F. Doebley, B.S. Gaut, and M.D. McMullen. 2005. A large-scale screen
for artificial selection in maize identifies candidate agronomic loci
for domestication and crop improvement. Plant Cell 17:2859–2872.
doi:10.1105/tpc.105.037242

the pl ant genome



november 2014



vol . 7, no . 3