Journal of Experimental Botany, Vol.

Page63, No.
1 of 1611, pp.695709,
2, pp. 40454060,20122012
doi:10.1093/jxb/ers105 Advance
doi:10.1093/jxb/err313
doi:10.1093/jxb/ers105 AdvanceAccess
Accesspublication
publication 417November,
April, 20122011
This paper is available online free of all access charges (see http://jxb.oxfordjournals.org/open_access.html for further details)

RESEARCH
REVIEW PAPER
PAPER

Association
In Posidonia mapping
oceanicaincadmium
forest trees
induces
and fruit
changes
cropsin DNA
methylation and chromatin patterning
M. Awais Khan* and Schuyler S. Korban*
Department of Natural
Maria Greco, AdrianaResources & Environmental
Chiappetta, LeonardoSciences,
BrunoUniversity
and Mariaof Illinois, Urbana,
Beatrice IL 61801 USA
Bitonti*
*Department
To whom correspondence shouldofbe
of Ecology, University addressed.
Calabria, E-mail: of
Laboratory awais@illinois.edu; korban@illinois.edu
Plant Cyto-physiology, Ponte Pietro Bucci, I-87036 Arcavacata di Rende,
Cosenza, Italy
* To whom correspondence should be addressed. E-mail: b.bitonti@unical.it
Abstract
Received 29 May 2011; Revised 8 July 2011; Accepted 18 August 2011
Association mapping (AM), also known as linkage disequilibrium (LD) mapping, is a viable approach to overcome
limitations of pedigree-based quantitative trait loci (QTL) mapping. In AM, genotypic and phenotypic correlations are
investigated in unrelated individuals. Unlike QTL mapping, AM takes advantage of both LD and historical
Abstract
recombination present within the gene pool of an organism, thus utilizing a broader reference population. In plants,
In
AMmammals,
has beencadmium is widely
used in model considered
species as a non-genotoxic
with available carcinogen
genomic resources. acting through
Pursuing AM in treea methylation-dependent
species requires both
epigenetic mechanism. Here, the effects of Cd treatment on the DNA methylation
genotyping and phenotyping of large populations with unique architectures. Recently, genome patten are examined together with
sequences and
its
genomic resources for forest and fruit crops have become available. Due to abundance of single analysed
effect on chromatin reconfiguration in Posidonia oceanica. DNA methylation level and pattern were nucleotidein
actively growing organs, under short- (6 h) and long- (2 d or 4 d) term and low (10 mM) and
polymorphisms (SNPs) within a genome, along with availability of high-throughput resequencing methods, SNPs canhigh (50 mM) doses of Cd,
through a Methylation-Sensitive
be effectively used for genotyping Amplification
trees. In addition Polymorphism technique and
to DNA polymorphisms, copy annumber
immunocytological approach,
variations (CNVs) in the
respectively. The expression of one member of the CHROMOMETHYLASE (CMT) family, a DNA
form of deletions, duplications, and insertions also play major roles in control of expression of phenotypic traits. methyltransferase,
was
Thus,also
CNVs assessed by qRT-PCR.
could provide Nuclear
yet another chromatin
valuable resource,ultrastructure
beyond those wasofinvestigated
microsatellitebyand
transmission electron
SNP variations, for
microscopy. Cd treatment induced a DNA hypermethylation, as well as an up-regulation
pursuing genomic studies. As genome-wide SNP data are generated from high-throughput sequencing efforts, of CMT, indicating that de
these
novo methylation did indeed occur. Moreover, a high dose of Cd led to a progressive heterochromatinization
could be readily reanalysed to identify CNVs, and subsequently used for AM studies. However, forest and fruit crops of
interphase nuclei and apoptotic figures were also observed after long-term treatment. The
possess unique architectural and biological features that ought to be taken into consideration when collecting data demonstrate that Cd
perturbs theand
genotyping DNA methylation
phenotyping status
data, through
as these willthe involvement
also dictate whichof aAM
specific methyltransferase.
strategies SuchThese
should be pursued. changes are
unique
linked toas
features nuclear
well aschromatin reconfiguration
their impact on undertaking likely
AM to establish
studies a new balance
are outlined of expressed/repressed chromatin.
and discussed.
Overall, the data show an epigenetic basis to the mechanism underlying Cd toxicity in plants.
Key words: Association mapping, linkage disequilibrium, perennial plants, quantitative trait loci.
Key words: 5-Methylcytosine-antibody, cadmium-stress condition, chromatin reconfiguration, CHROMOMETHYLASE,
DNA-methylation, Methylation- Sensitive Amplification Polymorphism (MSAP), Posidonia oceanica (L.) Delile.

Introduction
Increasing the efficiency of selection by maximizing the use et al., 1988). Since then, QTL mapping has been widely
Introduction
of desired genetic variation is one of the critical objectives used in plants for genetic dissection of biomass, yield, and
In any
of thebreeding
Mediterranean
programme. coastal
Mostecosystem,
agronomically theimportant
endemic Although
disease not essential
resistance for plant
traits. Although there growth,
are many in published
terrestrial
seagrass
traits arePosidonia
complex, oceanica (L.) Delile
are controlled plays agenes,
by multiple relevant
androle
are plants,
reports Cd onisQTLs,
readily only
absorbeda few by QTLs
roots and have translocated
been usedinto in
by ensuring inprimary
quantitative nature. production,
The principlewater oxygenation
of mapping and
a quantita- aerial
breedingorgans while, in acquatic
programmes (Bernardo, plants, it is and
2008), directly
thesetaken
are upin
provides
tive niches(QTL)
trait locus for some animals,
was first besides
described counteracting
in the early 20th by
factleaves.
majorIn plants,
genes ratherCd absorption
than QTLs.inducesUsually, complex changes
QTL intervals
coastal erosion
century throughwherein
by Sax (1923) its widespread
seed sizemeadows
in bean, a(Ott, 1980;
quantita- at
arethe
quite long, ;510
genetic, biochemical and physiological
cM (wherein 1 cM is equallevels to 3000 which
kb
Piazzi
tive trait, was 1999;
et al., Alcoverro
associated with seed coat
et al., 2001). There
colour, is also
a morpho- ultimately
of DNA), and account
containfor its
manytoxicity
genes.(Valle and Ulmer,
Therefore, 1972;
transferring
considerable
logical marker. evidence
However, thatQTL oceanica plants
P. identification didarenot able to
receive Sanitz
multiplediminor
ToppiQTLsand Gabrielli, 1999; Benavides
across genotypes et al., 2005;
via marker-assisted
absorb attention
serious and accumulate
until the metals from of
introduction sediments
polymerase (Sanchiz
chain Weber et(MAB)
breeding al., 2006;
can leadLiu to al., 2008).
et genetic dragThe due to most obvious
presence of
et al., 1990;
reaction Pergent-Martini,
(PCR)-based 1998;markers
molecular Masertiinetthe 2005)
al., late thus
1980s. symptom of Cd
undesirable toxicity
traits is a reduction
in these regions. in plant prior
Often, growthtodue to
their
influencing metal bioavailability
Molecular markers in the marine
offer great opportunities forecosystem.
dissecting an inhibition
utility of photosynthesis,
in positional cloning or MAS respiration, and nitrogen
in crop improvement
For
complexthis traits
reason, this QTL
using seagrass is widely
mapping. considered
Systematic to be
identifica- metabolism,
efforts, as well
individual genesas ain reduction
QTL regions in water
must be andidentified
mineral
ation
metalof bioindicator
QTLs in plants specieswas(Maserti et al., 1988;
first described by Pergent
Steven uptake
via (Ouzonidouaettime-consuming,
fine-mapping, al., 1997; Perfus-Barbeoch
laborious, etand 2000;
al., costly
et al., 1995;
Tanksleys Lafabrie
group while et al., 2007).a Cd
constructing is one fragment
restriction of most Shukla This
effort. et al.,is2003;
due to Sobkowiak
the necessityandof Deckert,
making2003). large numbers
widespread
length heavy metals ingenetic
polymorphism-based both terrestrial and marine
map for tomato fruit of At the genetic
crosses to elicitlevel, in both
sufficient animals
numbers of and
meioticplants, Cd
events.
environments.
quality traits using an interspecific backcross (Paterson can induce chromosomal
Additionally, QTL identification aberrations,
is basedabnormalities
on bi-parental in

The
2011Author [2012]. Published by Oxford University Press [on behalf
The Author(s). behalf of
of the
the Society
Society for
for Experimental
Experimental Biology].
Biology]. All
All rights
rights reserved.
reserved.
For
ThisPermissions, please article
is an Open Access e-mail:distributed
journals.permissions@oup.com
under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-
nc/3.0), which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
2 of 16| Khan
4046 and
| Khan Korban
and Korban
crosses, and quite often QTLs are specific to the bi-parental
population used in identifying these QTLs. Thus far, most
QTLs identified in either plant or mammalian populations
are not useful in a wide range of genetic backgrounds
(Sorkheh et al., 2008).
Association mapping (AM), also known as linkage
disequilibrium (LD) mapping, has been proposed as an
alternative approach to overcome limitations of pedigree-
based QTL mapping. In AM, genotype and phenotype
correlations are investigated in unrelated individuals.
Unlike QTL mapping, AM takes advantage of LD as well
as historical recombinations present within the gene pool of
an organism, thus utilizing a broader reference population
(Breseghello and Sorrels, 2006a; Ersoz and Buckler, 2007;
Myles et al., 2009). If two alleles from separate loci occur
together more often than otherwise predicted, on the basis
of their individual frequencies, i.e. non-random association
of alleles at separate loci, they are deemed to be in LD.
Only those molecular markers that are tightly linked to the
trait and located within the extent of LD decay will
demonstrate significant marker-trait association. If markers
are not tightly linked to a trait, they will be separated by
recombination during meiosis throughout the evolutionary
history of the crop. Accumulating meiotic events in
a population will increase the statistical power and mapping
resolution for detecting associations. However, it should be
noted that the rate of LD decay should be sufficient enough
to statistically identify associations, but not too high as it
will make it difficult to narrow down the target genomic
region.
AM requires availability of large numbers of polymor-
phic markers and is more complex than QTL mapping, as
historical factors such as population admixture, selection,
and genetic drift can bias the detected association. More-
over, the population genetic structure as well as effects due
to non-random mating (relatedness) must be accounted for
in the analysis to avoid false positive (spurious) associations
(Fig. 1; Pritchard et al., 2000; Bradbury et al., 2007; Zhu
et al., 2008; Zhang et al., 2010). Population structure
influences both the power and precision of detecting
associations. However, it can be overcome with good
sampling and by using appropriate algorithms to detect
groupings in a population and accounting for these in an
association mapping analysis (Zhu et al., 2008). Early on,
LD mapping had been used in human studies to understand
the genetic control of disease. Nowadays, it has rapidly
gained interest among plant scientists for studying biomass
traits, yield, and disease resistance, among others. Reviews
on the concept, methodology, prospects, and status of LD
and AM studies in plants have already been published
(Neale and Savolainen, 2004; Gupta et al., 2005; Brese-
ghello and Sorrels, 2006a; Ersoz and Buckler, 2007;
Abdurakhmonov and Abdukarimov, 2008; Sorkheh et al.,
2008; Zhu et al., 2008; Myles et al., 2009; Neale and
Kremer, 2011).
AM in plants was initially undertaken in model species
with available genomic resources, such as those of Arabi-
dopsis and maize. The first candidate gene-based AM in
plants was conducted to study flowering time in maize
inbred lines, wherein 92 maize recombinant inbred lines
(RILs) and the candidate gene Dwarf8, influencing quanti-
tative variation of flowering time and plant height, were
used (Thornsberry et al., 2001). Multiple polymorphisms
associated with flowering time were identified, including
a deletion in a key domain of the coding region of Dwarf8.
As for the genetic control of flowering in maize, extensive
studies have been conducted in Arabidopsis using AM. The
first study in Arabidopsis to elucidate the genetic control of
flowering using LD mapping identified two haplotype
groups in the genomic region of the photoperiod receptor
CRYPTOCHROME2 associated with flowering time varia-
tion (Olsen et al., 2004). Later, Aranzana et al. (2005)
demonstrated that identification of known major genes
involved in flowering (FRI) and pathogen resistance (Rpm1,
Rps2, and Rps5) could have been accomplished using
genome-wide AM, even in small and structured samples.
Recently, Su et al. (2011) pursued association analysis of
drought tolerance in maize with genes coding for key
enzymes in the abscisic acid biosynthesis pathway, nced and
rab28.
To further facilitate identification of loci controlling
quantitative traits with increased power and mapping
resolution through joint linkage-association analysis in
maize and Arabidopsis, a nested association mapping
(NAM) population and multiparent advanced generation
inter-cross (MAGIC) lines, respectively, were recently con-
structed (Kover et al., 2009; McMullen et al., 2009). NAM
was developed by crossing 25 diverse maize inbred lines to
line B73, a reference inbred line previously used for
constructing a physical map and for genome sequencing.
These 25 diverse inbred lines maximized the genetic di-
versity of NAM as resulting RIL families represented
diverse origins and types of maize. Whereas, MAGIC lines
corresponded to a set of 527 RILs generated by intercross-
ing a heterogeneous stock of 19 accessions of Arabidopsis
thaliana. These represented genetically diverse and highly
recombinant inbred lines of A. thaliana that were suitable
for high-resolution mapping of segregating traits. Buckler
et al. (2009) studied variation in flowering time using NAM
population, a set of 5000 maize RILs. Although they did
not identify any major gene effects, they noted additive
QTLs of numerous small genetic effects that were shared
among families, but with few genetic or environmental
interactions. In contrast to rice and Arabidopsis, wherein
flowering time was controlled by a few genes with large
effects, epistasis, and environmental interactions, it was
proposed that flowering time in maize was controlled by
a simple additive model (Buckler et al., 2009). Using
a candidate gene-based AM approach and utilizing 275
accessions of A. thaliana, Ehrenreich et al. (2009) identified
210 genes that were significantly associated with flowering
time, with CO arguably possessing the most promising
associations. Later, MAGIC lines were used to test a subset
of these genes to identify associations linked to flowering
time. It was confirmed that approximately one-third of
significant polymorphisms that had been associated with
 Association
Association mapping
mapping in forest
in forest trees
trees andand fruit
fruit | 3| of
crops
crops 4047
16

Fig. 1. A schematic representation of a protocol to conduct an association mapping study. This protocol can be used for candidate
gene-, genome-wide-, and copy number variation-based association mapping, with some modifications. The overall process of
association mapping involves genotyping, phenotyping, identification, and interpretation of results.

flowering time in these accessions were now replicated in
both populations, including polymorphisms at the CO,
FLC, VIN3, PHYD, GA1, and FRI loci. Both NAM and
MAGIC lines demonstrated the power and prospects of
a combined linkage and association mapping strategy for
identifying polymorphism-trait associations in plants. In
addition to maize and Arabidopsis, candidate-gene AM has
also been used in wheat, rice, barley, and sorghum to study
yield, flowering, and disease resistance traits (Breseghello
and Sorrels, 2006b; Ravel et al., 2006; Agrama et al., 2007;
Tommasini et al., 2007; Murray et al., 2009; Stracke et al.,
2009; Cockram et al., 2010; Maccaferri et al., 2011). These
studies demonstrated feasibility of conducting genome-wide
AM in plants and emphasized the need of having appropri-
ate controls to detect false positives as well as availability of
appropriate sample size.
So far, LD-based studies have primarily focused on
plants with short lifecycles, but with only few examples in
plants having long juvenility periods such as those of forest
trees and grapes. AM is particularly suited for forest trees
and perennial horticultural crops to overcome their charac-
teristic pedigree-based mapping limitations. This is due to
their long generation time, laborious and time-consuming
trait evaluation, slow maturation, and, at times, their
polyploidy. In this review, we will focus on the importance,
current status, and potential of AM and LD studies in
forest trees and fruit crops. In addition, we will provide
insights into the role of these studies in the genetic
improvement of these long-lived perennial plants.
Superiority and significance of association
mapping in forest trees and fruit crops
Most economic traits of horticultural and forest trees, such
as wood properties, disease resistance, fruit quality, and
biomass traits, are controlled by multiple genes exhibiting
quantitative distribution of phenotypes. Thus, quantitative
genetic strategies have been used to identify genes control-
ling these traits (Neale, 2007). However, the most widely
used approach to study quantitative traits in plants, i.e.
QTL mapping, has not been promising in pursuing genetic
studies in tree species. Fruit and forest trees are character-
ized by long juvenility periods; moreover, establishing and
maintaining bi-parental crosses and progenies of trees in
multiple locations for QTL mapping are often difficult and
costly (Rikkerink et al., 2007). For example, apple trees
require periods of 57 years to overcome juvenility, as well
as considerable economic input before seedlings are re-
productive and amenable for phenotypic evaluation. Com-
monly, an F1 progeny resulting from full-sib crosses
between two outbred parents is used for QTL mapping in
perennial forest trees and fruit crops. Recombination in an
F1 cross is attributed to a single meiotic event, resulting in
low resolution of QTLs. Often, the region of the QTL spans
;510 cM, with the likelihood of presence of hundreds of
genes within this region. For example in conifers, a QTL
interval spans ;15 cM (Brown et al., 2003; Gonzalez-
Martnez et al., 2007). This suggests that the majority of
identified QTLs, particularly those with minor to moderate
4 of 16| Khan
4048 and
| Khan Korban
and Korban
effects, are of limited utility in tree genetic improvement
efforts. Utilizing QTLs in MAB without fine-mapping is not
advisable due to genetic drag effects and/or rapid decay of
LD in most fruit and forest trees. However, fine-mapping of
the QTL region to locate a target gene(s) is both time-
consuming and laborious (Neale and Savolainen, 2004).
Moreover, it should also be noted that most QTLs are
limited to either a single or a few genetic backgrounds, and
will not be useful across a wide gene pool. For instance, it is
likely that disease resistance could break down before genes
conferring resistance are identified and fine-mapped via
QTL mapping. For all these reasons, QTL mapping seems
laborious and time-consuming in trees and fruit crops,
particularly when the trait is controlled by multiple genes
with small effect. Therefore, it is highly desirable to have
access to genomic tools and genetic approaches that could
accelerate detection of functionally important genetic
regions in trees (Savolainen and Pyhajarvi, 2007).
AM serves as a viable alternative approach that can
overcome the limitations of pedigree-based mapping in
perennial plants. In contrast to QTL mapping, AM requires
large numbers of molecular markers and utilizes recombi-
nation events present within the gene pool of an organism.
It is only those markers that are tightly linked to the locus
of interest and located within the extent of the LD decay of
the genome that will show significant associations. AM
could also serve as a powerful tool in dissecting quantitative
traits into their individual gene components in fruit and
forest trees due to their large architectural tree frames,
random mating, unstructured populations, and rapid decay
of LD (Komulainen et al., 2003; Neale and Savolainen,
2004; Rikkerink et al., 2007).
Current status of LD studies in forest trees
and fruit crops
Forest trees
Forest trees are not only important due to their economic
value for wood biomass and fuel, but they are also critical
for maintaining plant and animal biodiversity and temper-
ing environmental effects, as well as their natural resource
and aesthetic values. Threats to forest trees from biotic and
abiotic factors are rapidly increasing (Neale, 2007).
Although there is wide variation among different forest
trees, generally they are out-crossing, long-lived, and at
early stages of domestication (Savolainen and Pyhajarvi,
2007). Despite similarities among forest trees, there are
large differences between genetic and lifecycle character-
istics of forest trees that render some better suited for
genetic studies than others. For instance, two of the most
important forest trees, poplar and pine, are different in
many aspects. Poplar is a dioecious angiosperm with 29
species, 38 chromosomes (2n), and a genome size of 450
Mb; moreover, LD decays at a rate of <500 bp, and has
a nucleotide diversity of 0.51% (Neale and Ingvarsson,
2008). In contrast, pine is a monoecious gymnosperm with
;100 species, 24 chromosomes (2n), and a genome size of
1040 Gb; moreover, pine LD decays at a rate of <500 to
2000 bp, and has 12% nucleotide diversity (Neale and
Ingvarsson, 2008). Since the introduction of QTL mapping
methods, forest geneticists have been investigating quantita-
tive traits of both economic (growth, wood properties, and
disease resistance) and adaptive (cold tolerance and phenol-
ogy) values. According to Neale and Ingvarsson (2008),
AM approaches in forest trees offer advantages over
pedigree-based genetic tests due to the availability of large
size random-mating populations with minimal population
structure, presence of adequate levels of nucleotide di-
versity, rapid decay of LD, direct determination of haploids
by sequencing of haploid seed megagametophyte tissues,
and reliable phenotypic evaluations in clonally propagated
plants.
Availability of genomic resources is one of the key
limitations for conducting an AM study in a crop. Genome
sizes of forest trees vary considerably, impacting availability
of genomic resources, and influencing AM strategies. For
example, sequencing of the genome of Populus trichocarpa,
which is four times the size of the Arabidopsis genome, has
been long completed (Tuskan et al., 2006). Sequencing of
the pine genome is still ongoing as the size of this genome is
;100-fold larger than that of the Arabidopsis genome (Pavy
et al., 2005). On average, there is a slightly lower nucleotide
diversity in poplar compared to that of pine, but LD decay
is almost equal (Ingvarsson, 2005). LD declines rapidly in
forest trees, within ;1 kb, compared to that of self-
pollinated plant species, thus requiring availability of large
numbers of molecular markers for AM. In addition to the
genome sequence of poplar, large numbers of expressed
sequence tag (EST) sequences are also available for multiple
forest trees. Thus, there are considerable genomic resources
and wide genetic diversity in forest trees that would serve
well in pursuing AM strategies. Therefore, due to expanding
genomic resources and limitations of QTL methods in forest
trees, AM studies have been carried out in perennial forest
trees.
As for other crops, conducting AM in tree species
requires both genotyping and phenotyping of a large
population (Fig. 1). SNPs are currently the markers of
choice for genotyping trees due to their abundance within
a genome and the availability of high-throughput rese-
quencing and genotyping methods. For those conifers with
limited EST databases, SNPs are identified by resequencing
candidate genes in a small (10100) panel of individual
trees, using haploid seed megagametophytes, known as the
SNP discovery panel (Neale, 2007). For those forest trees
with abundant EST data, SNPs are discovered in silico
using bioinformatics tools. Identified SNPs are then geno-
typed on a larger population using high-throughput geno-
typing methods, such as the Illumina GoldenGate assay
(Eckert et al., 2009). For accurate phenotypic evaluation,
individuals of a population are grown in replicated trials by
vegetative propagation or subjected to family-based testing.
Moreover, to accurately estimate the effects of the SNP
genotype by minimizing the effects of population genetic
 Association
Association
structure and genetic relatedness, a large population is
usually used. So far, candidate gene-based AM studies and
population genetic neutrality tests have frequently been
used to study wood-related economic and adaptive traits,
and to investigate gene behaviour under natural selection
conditions in forest trees (Neale, 2007; Eckert et al., 2009).
Conifers (pine and spruce species)
As mentioned above, SNPs are generally identified by
resequencing ESTs using a small SNP discovery panel and
are then genotyped on large samples of unrelated trees by
high-throughput genotyping assays (Pavy et al., 2008;
Eckert et al., 2009). Sufficient numbers of EST sequences
are available for Picea and Pinus (Pavy et al., 2005; Rungis
et al., 2005) to pursue SNP discovery and conduct LD
studies in conifers. The NCBI database, as of 6 February
2012, has 328791, 247119, 36384, 34524, 341325, and
296253 EST sequences for Pinus taeda (loblolly pine), Pinus
contorta (lodgepole pine), Pinus banksiana (jack pine), Pinus
pinaster (meritime pine), Picea glauca (white spruce), and
Picea sitchensis (sitka spruce), respectively.
Gonzalez-Martnez et al. (2007) conducted the first multi-
gene association study in forest trees. A population of 422
435 unrelated loblolly pine trees (P. taeda L.) trees, in
a clonally replicated trial, was used to conduct an associa-
tion analysis study of 58 SNPs, from 20 wood- and
drought-related candidate genes and wood property traits.
These traits included earlywood and latewood specific
gravity, percentage latewood, earlywood microfibril angle,
and wood chemistry (lignin and cellulose contents). They
used mixed linear models to perform AM analysis, where
population structure and relatedness was accounted for.
The strongest association was observed between allelic
variations in tubulin, a gene involved in the formation of
cortical microtubules, and the earlywood microfibril angle.
It was suggested that due to rapid LD decay in conifers,
SNPs revealing genetic associations were likely to be located
in close proximity to causative polymorphisms (Gonzalez-
Martnez et al., 2007). Furthermore, a strong association
between a SNP within the candidate gene 4-coumarate CoA
ligase (4cl) and percentage latewood was also detected in
this study, thus confirming previous findings based on co-
location of a QTL for percentage latewood. In another
study, Gonzalez-Martnez et al. (2008) took advantage of
both pedigreed crosses and genetic diversity to develop the
first family-based AM approach in plants, known as
a quantitative transmission disequilibrium test. A popula-
tion of loblolly pine, consisting of 961 clones from 61
families, was evaluated at two sites to test for genetic
associations between 46 SNPs (identified from 41 biotic and
abiotic stress-inducible genes) and carbon isotope discrimi-
nation, a measure of water use efficiency. Several candidate
gene associations were detected together with two very
promising associations, a cell structure stabilizing dehydrin
gene (dhn-1) and a cell wall reinforcement protein gene (lp5)
(Gonzalez-Martnez et al., 2008). This study demonstrated
mapping
mapping in forest
in forest trees
trees andand fruit
fruit | 5| of
crops
crops 4049
16
successful candidate gene association mapping in trees using
a complex family structure.
The inherent broad phenotypic adaptation to diverse
environments that is present in natural populations of forest
trees is key to successful establishment of any tree species
either as domesticated or undomesticated (Petit and
Hampe, 2006). Therefore, understanding the genetic control
of adaptability is critical in dealing with current problems in
forests. Eckert et al. (2010) have used a genome-wide
dataset of SNPs genotyped across 3059 functional genes to
study patterns of population structure and to identify
genetic loci linked to response to aridity across a natural
range of loblolly pine. After accounting for confounding
effects of shared ancestry with correlations between genetic
and environmental variations, five loci correlating with
aridity have been identified. These loci are primarily in-
volved in abiotic stress responses to temperature and
drought.
Successful establishment of forest trees depend not only
on their adaptability to constantly changing environments,
but also to their responses to emerging pathogens. Pitch
canker, a disease caused by the necrotrophic pathogen
Fusarium circinatum, is an important fungal disease of
loblolly pine. Quesada et al. (2010) have identified genes
that are associated with resistance to pitch canker in
loblolly pine using a set of 498 largely unrelated clonally
propagated genotypes. Upon inoculation of these genotypes
with F. circinatum and measuring lengths of necrotic tissues
after 4, 8, and 12 weeks, significant associations have been
detected in 10 out of 3938 SNPs tested. These 10 SNPs have
exhibited small effects, thus suggesting that resistance is
quantitative, and that whole-genome scans must be con-
ducted to identify remaining variations for this trait.
Recently, Dillon et al. (2010) have evaluated the utility of
LD mapping to detect associations between SNPs and wood
quality in a natural population of Pinus radiata. After
accounting for both population structure and experimental
error, a total of 10 significant associations (P < 0.05, q <
0.1) have been detected with one or more traits, out of 149
loci investigated. Significant associations have been re-
examined in an Australian land race, and those associations
previously observed in the discovery population have
further validated associations of two genes with wood
density. Decreased wood density is associated with a minor
allele, thus suggesting that these SNPs may be under weak
negative purifying selection for wood density. These find-
ings clearly demonstrate the utility of LD mapping in
detecting associations, even when the power of detecting
SNPs with small effects is anticipated to be low (Dillon
et al., 2010).
Various neutrality tests have also been used to identify
genes and genomic regions that have been targets of natural
selection in forest trees. Due to long generation times,
molecular evolution at neutral sites in forest trees is very
slow, and, in contrast to their slow rate of neutral evolution,
large tree populations respond quickly to natural selection.
Therefore, detecting traces of selection may be easier in trees
than in many other species due to low demographic effects
6 of 16| Khan
4050 and
| Khan Korban
and Korban
in large randomly mating tree populations (Savolainen and
Pyhajarvi, 2007). According to Neale and Ingvarsson
(2008), of ;290 genes of trees subjected to neutrality tests,
20% have exhibited departures from neutrality. In one of
the early studies describing nucleotide diversity, Heuertz
et al. (2006) have conducted LD and tests of neutrality in
spruce (Picea) and have surveyed DNA polymorphisms at
22 loci in ;47 haplotypes from seven populations. Their
results have revealed that the overall nucleotide variation is
limited, being lower than that observed in most plant
species. LD is also restricted and does not extend beyond
a few hundred base pairs. Moreover, neutrality tests have
revealed presence of an excess of both rare and high-
frequency-derived variants and pointed to a severe bottle-
neck. It has been concluded that demographic departures
from equilibrium expectations and population structure
must be accounted for when detecting selection at candidate
genes and in association mapping studies, respectively.
Recently, neutrality tests have been used to study the effects
of natural selection on 41 candidate genes from loblolly
pine; these genes have been selected primarily from host
pathogen interactions together with 15 drought-tolerance
and 13 wood-quality genes identified in previous studies
(Ersoz et al., 2010). Patterns of both directional selection
and selective sweep consistent with the arms-race model of
disease response evolution have been detected, as well as
patterns consistent with diversifying selection.
To further elucidate the genetic control of adaptive traits
in Douglas fir, Krutovsky and Neale (2005) studied LD and
haplotype and nucleotide frequencies, and performed neu-
trality tests in cold-hardiness and wood quality-related
candidate genes. Coastal Douglas fir (Pseudotsuga menziesii
Franco), an important tree of western North American
forests, has evolved complex adaptive mechanisms. In the
study, genes were selected on the basis of their functions in
other plants and their co-location with cold-hardiness-
related QTLs. On average, the frequency of SNPs was one
SNP per 46 bp across coding and noncoding regions. LD
within genes decayed relatively slowly, but steadily, while
a neutrality test suggested directional selection (Krutovsky
and Neale, 2005).
Poplar
Thus far, poplar (Populus trichocarpa) is the only forest tree
with a complete genome sequence, thus allowing resequenc-
ing of different genotypes to identify SNPs for genetic
studies. Ingvarsson et al. (2006) first reported on genetic
dissection of complex adaptive traits using candidate genes
in poplar. They identified SNPs in the phytochrome gene
(phyB2) that co-located to a previously reported QTL for
timing of bud set, and this was genotyped in 16 Populus
tremula populations collected along a latitudinal gradient in
Sweden. It was found that there was a significant clinal 3
latitude variation for bud set. A sliding-window scan of
phyB2 identified four SNPs with significant clinal varia-
tions, suggesting an adaptive response of phyB2 to local
photoperiodic conditions. Additionally, Hall et al. (2007)
also found several phenological traits of P. tremula with
strong genetic differentiation and clinal variation across the
latitudinal gradient. It was suggested that genetic differenti-
ation at candidate loci was better described by FST at
neutral loci rather than by QST at quantitative traits. Later,
Ingvarsson et al. (2008) reported that polymorphism varied
substantially across the phyB2 region when surveyed within
an 80-kb region surrounding the phytochrome locus, but
there were no signs of deviations from neutral expectations.
Moreover, using 41 SNPs in a mapping population, they
identified two non-synonymous SNPs in the phyB2 gene
associated with variations in timing of bud set. These SNPs
explained between 1.5 and 5% of the observed phenotypic
variation in bud set. It was proposed that due to low LD in
this region, these SNPs were strong candidates that were
causally linked to variation in bud set.
Eucalyptus
Eucalyptus (Eucalyptus nitens) is a widely adapted forest
tree, and has been the focus of studies on adaptation as
a key element in genetic conservation of forests. Thumma
et al. (2005) have used a candidate-gene-based LD mapping
approach to identify alleles associated with microfibril
angle, a wood quality trait affecting stiffness and strength
of wood in eucalyptus. SNPs detected in the cinnamoyl CoA
reductase gene, a key lignin gene, have been used to
genotype 290 unrelated trees from a natural population of
E. nitens. Two haplotypes that are significantly associated
with microfibril angle, as confirmed in two full-sib families
of E. nitens and Eucalyptus globules, have been identified. In
a recent study, Thumma et al. (2009) have demonstrated the
potential of revealing functional polymorphisms underlying
quantitative traits by integrating both QTL and association
mapping methods. First, a marker from the COBRA-like
gene, whose Arabidopsis homologue has been implicated in
cellulose deposition, was found to be strongly associated
with a QTL for cellulose content in a full-sib family. By
genotyping SNPs and a simple sequence repeat (SSR)
marker in an association population, LD analysis has
revealed that LD declines within the length of the COBRA-
like gene. Subsequent association mapping analysis has
contributed to fine-resolution mapping of the effect of this
gene to a SNP marker.
Grapes
The cultivated grapevine (Vitis vinifera L.) is one of the first
economically important fruit crops whose genome was
sequenced. Having access to the genome sequence has
facilitated the identification and development of genetic
markers and has also enabled LD studies in grapes. The
cultivated grapevine is primarily a self-pollinated perennial
fruit crop. Most of todays cultivated grapes have been
derived from controlled crosses among a few select culti-
vars, and the elite selections have been maintained by
vegetative propagation. There are only a few studies
 Association
Association
describing LD patterns across the grapevine genome, as well
as AM studies undertaken (Myles et al., 2011).
Characterization of LD in the wild French grapevine,
V. vinifera L. subsp. sylvestris, was first reported by
Barnaud et al. (2010). LD patterns and extent were assessed
using un-phased SSRs and reconstructed haplotypic data in
a sample of 85 plants from southern France by performing
independence tests and multiallelic r2 analysis. It was found
that the LD decayed rapidly, with r2 values decreasing to
0.1 within 2.7 cM for genotypic data and within 1.4 cM for
haplotypic data. However, when LD was compared to
previous findings from a study on cultivated grapevine
subsp. sativa, LD was extended further, by 12-fold, in
cultivated compared with wild grapevine. In a recent study,
Myles et al. (2011) also reported that there was rapid decay
of LD in V. vinifera that appeared unchanged between the
wild ancestor and the domesticated grape. These differences
in LD patterns between cultivated and wild grapevine could
primarily be due to domestication bottlenecks and vegeta-
tive propagation. It was suggested that the LD in wild
grapevine seemed to extend slightly further than in wild
relatives of other crops, thus rendering it amenable for
future AM studies. Integrating both QTL and AM strate-
gies has proven highly effective in various studies. For
example, Fournier-Level et al. (2010) combined QTL and
AM to study the genetic patterns of anthocyanin content,
a determinant of berry colour, in grapes. A strong QTL,
accounting for 62% of the variation in anthocyanin content,
was mapped in an F1 pseudo-testcross. Markers were
identified in four Myb-type genes, selected based on
metabolic profiles, within the QTL interval, and these were
tested in a core collection of natural grape germplasm
consisting of 141 individuals. Using Bayesian reconstruction
of effective population size dynamics, the presence of
extended LD in these genes was revealed. Furthermore,
a multivariate regression analysis identified five polymor-
phisms in VvMybA genes that accounted for 84% of the
observed variation. It was concluded that the observed
variation in anthocyanin content in grape was mainly
accounted for by a single cluster of three VvMybA genes.
Fruit development is an important economic trait of
grapevine, but there is little known about the genetic and
molecular controls of fleshy fruit development. Houel et al.
(2010) have identified markers from a gene (flb) proposed to
be linked to a fleshless berry mutation by resequencing
genomic regions of two grape cultivars. They have also
sequenced these gene fragments in a highly diverse set of
both cultivated and wild V. vinifera genotypes to identify
possible signatures of domestication in the cultivated
V. vinifera. In this study, SNPs significantly associated to
berry weight variation have been identified in the flb region.
Moreover, eight gene fragments with significant deviations
from neutrality of the Tajimas D parameter have been
identified in the cultivated pool along with putative
signatures of selection. Sweet- and floral-flavoured Muscat
cultivars are highly regarded as table grapes and for wine
making. Using QTL analysis, Muscat flavour determination
has been investigated and a major QTL on chromosome
mapping
mapping in forest
in forest trees
trees andand fruit
fruit | 7| of
crops
crops 4051
16
5 linked to the Muscat flavour has been found to co-localize
with a VvDXS gene encoding 1-deoxy-D-xylulose 5-phosphate
synthase (Emanuelli et al., 2010). Upon resequencing of the
VvDXS gene in an ad hoc association population of 148
grape cultivars, consisting of muscat-flavoured, aromatic,
and neutral accessions, as well as muscat-like aromatic
mutants and non-aromatic offsprings of Muscats, three
SNPs in moderate LD that are significantly associated with
muscat-flavoured cultivars have been identified, after ac-
counting for population structure. They have also identified
a putative causal SNP responsible for a predicted non-
neutral substitution (Emanuelli et al., 2010). Although LD
rapidly decays in grapevine, LD could be extended in those
genomic regions selected during domestication. Therefore,
conducting a candidate gene-based study for traits impor-
tant during domestication is feasible, although the availabil-
ity of high densities of markers across the genome is
required for genome-wide AM (Myles et al., 2011).
Rosaceous fruit crops
The Rosaceae is the third most economically important
plant family (Dirlewanger et al., 2002; Shulaev et al., 2008).
It includes such economically important crops as apple
(Malus), pear (Pyrus), stone fruit (Prunus), strawberry
(Fragaria), and rose (Rosa). Polyploidy is common in
rosaceous plants; moreover, genome sizes vary from moder-
ate to small as in apple (742 Mb) and strawberry (240 Mb),
respectively (Velasco et al., 2010; Shulaev et al., 2011).
Rosaceous plants are often vegetatively propagated to
maintain both additive and non-additive genetic effects in
phenotypes of superior genotypes. According to Rikkerink
et al. (2007), trait evaluation in most fruit breeding
programmes depends on a two-step strategy. In the first
step, a large number of individuals in non-replicated trials
are evaluated to select a small number of individuals with
desirable traits. In the next step, these selected plants are
asexually propagated in a replicated trial for reliable trait
evaluation. For most fruit crops, as in forest trees, genetic
analysis is complicated due to presence of high levels of
heterozygosity, long generation time, long juvenility period,
time-consuming trait evaluation, slow physiological matu-
ration, and polyploidy, with few exceptions (Shulaev et al.,
2008). Costs of growing and maintenance of perennial
plants until they reach maturity to collect phenotypic data
are high, thus efforts to conduct early selection at the
seedling stage are highly desirable. In addition, fruit
evaluation requires quick assessment and special care due
to perishability (Rikkerink et al., 2007). In general, there are
many similarities between forest trees and fruit crops, and
some of the genomic tools developed and knowledge gained
for one group could be transferred to the other group.
However, before implementing such approaches across
these groups, we must acknowledge and understand the
fundamental differences. Fruit crops are mainly maintained
through vegetative propagation in a domesticated state
while forest trees are found in both partially undomesti-
cated and domesticated states (Neale and Ingvarsson, 2008).
8 of 16| Khan
4052 and
| Khan Korban
and Korban
Therefore, LD patterns between forest trees and fruit crops
could be quite different. For vegetatively propagated
domesticated trees, incidence of recombination is unlikely,
thus this can result in extended LD compared to un-
domesticated trees.
There are LD data for only a few plants and only for
specific loci in Rosaceae; therefore, it is rather difficult to
propose general conclusions about LD decay. Recently,
Aranzana et al. (2010) have studied extent, distribution, and
structure of genetic variation in North American and
European commercial peach cultivars, as well as in some
old peach cultivars. They have genotyped 50 SSRs that are
evenly distributed across all linkage groups in 224 peach
cultivars. The population structure analysis has divided the
sample into three main groups, based mainly on their fruit
characteristics, including melting-flesh peach, melting-flesh
nectarine, and non-melting subpopulations. LD analysis has
revealed that LD extends up to 1315 cM in peach. The
extended LD could be attributed to the self-pollination
habit of peach cultivars used in this study and a bottleneck
that must have occurred early on in modern breeding
practices. It is concluded that a high LD suggests that
whole-genome scanning approaches are well suited for
genetic studies of important traits in peach.
The mechanism of inbreeding control could also have
significant influence on LD, as it is assumed that self-
incompatible rosaceous fruits will have rapid LD decay
compared to self-compatible fruits. Furthermore, deliberate
breeding can also influence LD; however, long generation
times and recent breeding efforts in fruit trees may have
reduced these effects (Myles et al., 2011). Compared to
pedigree-based approaches, AM is more suited for rosa-
ceous plants due to their above-described biological fea-
tures. Depending on LD, different AM strategies could be
implemented. Genome size and LD patterns are critical in
determining the number of markers required for conducting
whole-genome scans, as well as economic feasibility and
overall success of AM studies (Myles et al., 2009). De-
veloping sufficient numbers of markers will greatly depend
on available genomic resources for a given crop. Candidate
gene AM may be the method of choice compared to
genome-wide mapping for crops characterized by fast LD
decay and with limited available genomic resources. As to
which strategy is preferable for identifying genes of interest
will also depend on the economic value of the target crop.
AM is preferred for those horticultural fruit crops of high
economic value, such as apple, peach, pear, and strawberry.
For apples, Cevik et al. (2009) have used a pedigree-based
QTL mapping approach and have identified markers
(MdMADS2.1, MdMADS2.2, and MdMADS14) from two
candidate orthologous FRUITFULL-like genes linked to
fruit flesh firmness. Subsequently, they have used an
association analysis to collect further evidence for associa-
tion between MdMADS2.1 and flesh firmness in a popula-
tion of 168 apple accessions. The association mapping
population has allowed for confirmation that MdMADS2.1
and fruit flesh are significantly associated. Additionally,
recently developed high-throughput SNP genotyping assays,
a 1536 SNP GoldenGate assay for apple (Khan et al., 2012)
and a 8000 SNP Infinium assay for apple, peach, and cherry
(Chagne et al., 2012), will provide more opportunities for
candidate gene-based association mapping. Although the
Infinium has 8000 SNPs, this is not sufficient for genome-
wide AM due to rapid LD decay in apple and in other
rosaceous fruit crops. But, this can be used for whole-
genome scanning of marker-trait associations utilizing
a joint linkage-association approach that makes use of both
pedigree and diversity analyses.
Presence of co-linearity between some rosaceous species
should also be taken into consideration when drawing
conclusions based on the economic value of a crop. Co-
linearity among species has been deemed acceptable by the
Rosaceae research community as a unitary system for
genetic analysis to overcome economic costs (Rikkerink
et al., 2007). However, transferability of AM results across
species may not always be feasible as there are also major
biological differences, even within a single species, thus
rendering synteny across species as a minor factor. How-
ever, it should also be noted that, unlike forest trees, where
population structure is minimal, domesticated fruit trees
can have substantial population structure that may con-
found AM inferences. Therefore, a large set of accessions/
genotypes are required for reliable identification of associa-
tions by minimizing spurious associations due to population
structure.
Future of association mapping in forest trees
and fruit crops in the era of next-generation
sequencing
In perennial trees, traditional full-sib crosses, i.e. F1 crosses
between two outbred plants, are used to identify associa-
tions between traits and genetic loci. A maximum of four
alleles would segregate in such a cross. However, in AM,
a large number of alleles present within the gene pool of
a species are tested against the phenotype to detect
significant associations. As discussed earlier, patterns of
LD and availability of genomic resources for a species will
dictate which AM strategy is more appropriate for a given
crop. Although the utility of AM in crop improvement is
well established, its active use is rather slow, mainly due to
limited numbers of genomic resources available for most
agriculturally important crops, with only a few exceptions.
Development of genetic markers is crucial for AM studies.
A whole genome or selected segments of a genome of
a panel of organisms are sequenced to identify differences
across the genome(s) of the panel(s). Subsequently, identi-
fied polymorphisms are genotyped across a larger and more
diverse yet unrelated population.
Due to high costs of sequencing and genotyping, marker
development is considered very costly and thus only justifi-
able in crops of high commercial value. Recently, there has
been a new revolution in technological advances that impact
economic aspects of whole-genome sequencing efforts.
Next-generation sequencing technologies are highly efficient
 Association
Association mapping
mapping in forest
in forest trees
trees andand fruit
fruit | 9| of
crops
crops 4053
16

and costs of sequencing have sharply declined. Currently,
there are many ambitious ongoing whole-genome sequenc-
ing projects for a wide range of organisms. The Genome
10K Consortium of Scientists (G10KCOS) has undertaken
a sequencing project (G10K) supported by zoos, museums,
universities, and research centres to assemble a genomic zoo
the genomes of 10,000 vertebrate species (for mammals,
birds, amphibians, reptiles, and fishes), approximately one
for every vertebrate genus. In November 2010, the Beijing
Genomics Institute (BGI) and the G10KCOS have together
planned to sequence genomes of the first 101 vertebrate
species within 2 years. The 1000 Plants & Animals reference
genome project, launched in January 2010, aims to sequence
genomes of 1000 economically and scientifically important
plant/animal species within 2 years. Moreover, the possibility
of sequencing pooled DNA will also change how AM is
pursued. In organisms where a reference genome sequence
exists (Table 1), genotyping-by-resequencing (GBS, Fig. 2) an
association population will be pursued. Resequencing of
whole genomes using next-generation sequencing will allow
rapid and efficient identification of SNPs and insertions
deletions between genomes. Recently, resequencing of the rice
genome identified a total of 132,462 SNPs, 16,448 insertions,
and 19,318 deletions between the Omachi (landrace of
japonica rice) and Nipponbare genomes (Arai-Kichise et al.,
2011). Validation of a set of selected SNPs has yielded
a success rate of 95 and 88% for the Omachi and Nipponbare
genomes, respectively. Huang et al. (2009) have utilized
a whole-genome resequencing approach to genotype recombi-
nant inbred lines in rice and have reported that this is a faster
approach for collecting data. More importantly, it provides
accurate determination of recombination breakpoints com-
pared to those revealed in a genetic map constructed using
PCR-based markers.
Recently, it has been demonstrated that it is not necessary
to resequence a genome at very high coverage to identify
SNPs and copy number variations (CNVs), as resequenced
samples will be aligned using the already sequenced genome
as a reference (Alkan et al., 2009). Improvements in
sequencing throughput and pair-end sequencing, together
with reduced genome coverage will enable multiplexing of
barcoded samples within the same sequencing lane, effec-
tively reducing sequencing costs (Fig. 2). DNA of two or
more genotypes with different indices could be mixed in
equal molar concentrations and loaded onto a single
sequencing lane to perform paired-end sequencing. After
trimming barcodes, sequences could be aligned to the
reference genome to identify polymorphisms across the

Table 1. Association mapping-relevant biological characteristics of forest trees and fruit crops whose genomes have been sequenced

Sequenced Family Ploidy Chromosomes Pollination Genome LD Protein Reference

genomes (n) size (Mb) extent coding
genes
Strawberry Rosaceae F. vesca is a diploid, 7 Self- 240 34,809 Shulaev et al.
(Fragaria vesca) but others pollination (2011)
species include
tetraploid,
pentaploid,
hexaploid, octoploid
(F. 3 ananassa)
and mixoploid
Cacao Sterculiaceae Diploid 10 Cross- 430 28,798 Argout et al.
(Theobroma pollinated (2011)
cacao)
Apple Rosaceae Predominantly 17 Cross- 742 57,386 Velasco et al.
(Malus 3 diploid, along pollination (2010)
domestica) with some triploids
and tetraploids
Peach Rosaceae Diploid 8 Self- 227 1315 cM 28,702 International
(Prunus persica) pollination Peach
Genome
Initiative
(IPGI)
Papaya Caricaceae Diploid 9 Polygamous 372 28,629 Ming et al.
(Carica papaya) (2008)
Grapevine Vitaceae Predominantly 19 Self- 487 2.7 cM 26,346 Velasco et al.
(Vitis vinifera) diploid but tri pollination (2007)
and tetra-ploids
also exist
Poplar Salicaceae Diploid 19 Cross- 485 <500 bp 45,555 Tuskan et al.
(Populus pollination (2006)
trichocarpa)
4054 | Khan
10 of 16 andand
| Khan Korban
Korban

Fig. 2. Genotyping strategies for candidate gene and genome-wide association mapping (GWAS). The strategy for candidate gene
genotyping is based on Sanger sequencing using ABI technology, while that for GWAS is the genotyping-by-sequencing (GBS) adapted
from Elshire et al. (2011) for Illumina GA sequencing technology. The GBS strategy could be used for both single-nucleotide
polymorphism (SNP) genotyping and copy number variation (CNV) identification. LD, linkage disequilibrium.

population. For example, the HiSeq2000 platform is
capable of generating approximately 120 Gbp per flowcell
(;30 Gb per sequencing lane), and this throughput is
expected to increase further in the near future. An estimated
cost of resequencing an apple population, with a total
genome size of ;742 Mb (Velasco et al., 2010), would add
up to ;$25,000 for 14 individuals per flow cell, with an
average cost of ;$1800 per genotype, assuming that ;203
genome coverage would be sufficient to identify SNPs and
CNVs by resequencing, which can be achieved by multi-
plexing of two genotypes per lane.
Genome-wide association mapping
Success of whole-genome AM depends on density and
robustness of available markers as well as accuracy of
phenotypic evaluation. To achieve appropriate marker
density for a target plant species, it is important to measure
genome-wide LD and have access to high-throughput
genotyping methods. For those forest and fruit crops where
whole-genome sequences are available (Table 1), resequenc-
ing of the AM panel is doable using next-generation high-
throughput sequencing technologies, thus enabling rapid
and cost-effective genome-wide identification of polymor-
phisms. A genome-wide association mapping analysis has
been undertaken for 14 agronomic traits in a population of
rice landraces using a genotyping-by-resequencing approach
(Huang et al., 2010). Resequencing of 517 rice landraces has
contributed to the identification of ;3.6 million SNPs and
aiding in the construction of a high-density haplotype map.
On average, identified loci have accounted for ;36% of the
phenotypic variance, and six loci, closely located to pre-
viously identified genes, have been identified (Huang et al.,
2010). This has demonstrated that integrating high-through-
put genome resequencing and genome-wide AM can serve
as a powerful complementary strategy to bi-parental QTL
mapping for genetic dissection of complex traits in plants.
The next-generation high-throughput sequencing technolo-
gies combined with multiplexing of DNA samples will
significantly contribute to reduced costs (Fig. 2). As GBS
data for large numbers of germplasm become available,
these would be useful in studying various traits over many
years. However, the panel of genotypes for sequencing
should be carefully selected, keeping in mind their diversity
and their future utility in breeding efforts. For example in
apples, an appropriate germplasm set for AM studies would
be the USDA Malus core collection. This Malus core
collection includes diverse apple germplasm along with
a large number of accessions from the cultivated apple,
M. x domestica, which have been distributed to multiple sites
throughout the USA, and has been evaluated for multiple
traits, and continues to undergo evaluation.
 Association
Association
Sequencing of reduced representation genomic libraries
(Altshuler et al., 2000; Baird et al., 2008; Elshire et al., 2011)
could be used for identification and genotyping of SNPs for
genome-wide AM in crops with extended LD and complex
genomes (Fig. 2). The GBS method described by Elshire
et al. (2011) is a robust, simple, quick, and highly specific
approach for species with high diversity. Reduced represen-
tation genomic libraries reduce genome complexity and
allow for multiplexing of large numbers of indexed samples
within a single sequencing lane. Genome complexity re-
duction strategies are usually applied to complex genomes
such as maize, wheat, barley, and soybean to generate
simpler genomic DNA libraries for high-throughput se-
quencing. These reduced representation genomic libraries
are constructed using methylation-sensitive restriction
enzymes that avoid repetitive regions of genomes and can
target lower copy regions with 2- to 3-fold higher efficiency
(Elshire et al., 2011). Reduction of genome complexity by
avoiding repetitive regions markedly simplifies computa-
tionally challenging alignment problems in species of high
levels of genetic diversity. Different restriction enzymes
could be used to digest genomic DNA of target species and
to investigate polymorphism patterns within a species.
Selection of a single or multiple restriction enzymes for
a reduced representation library relies on genomic features
of a given organism and its intended applications (Baird
et al., 2008; Wu et al., 2010). For example, a high-
throughput sequencing approach of pooled DNA fragments
of a reduced representation library of two parental lines of
a soybean mapping population has been used for identify-
ing SNPs (Wu et al., 2010). As the frequency of predicted
putative SNPs of short sequence reads has been detected at
a low sequencing depth, a total of 39,022 putative SNPs
have been identified, and validation of these SNPs have
been predicted at low and high stringencies of 72% and
85%, respectively. It has been suggested that this approach
is more efficient for targeting multiple QTL regions in the
same genetic population, and it can be used for fine-
mapping multiple QTL regions in other crops. Previously,
Hyten et al. (2010) have discovered a total of 710825,047
of predicted SNPs using high-throughput resequencing of
a reduced representation library in soybean. These SNPs
have been subsequently used to develop a high-resolution
genetic map using RILs for assembly of whole-genome
shotgun sequences of soybean.
Candidate gene-based association mapping
Genome-wide AM has been useful in detecting candidate
genes underlying Mendelian traits in maize, rice, and
Arabidopsis, either due to relatively simple genetics and
strong imposed selection to the traits besides availability of
large amount of genomic resources. However, for species
with rapid LD decay and fewer genomic resources, genome-
wide AM is generally not suitable due to large numbers of
markers required to cover the entire genome. Thus, a candi-
date gene-based AM strategy may be the best option.
Roughly, ;140,000, ; 2 million, and ;1015 million
mapping
mapping in forest
in forest treestrees
andand
fruit fruit
crops | 11| of
crops 4055
16
markers are needed to have a reasonable coverage of
Arabidopsis, grapevine, and maize genomes, respectively
(Kim et al., 2007; Myles et al., 2009). AM studies based on
candidate genes require markers only for regions of interest.
The candidate gene-based AM has a higher likelihood of
success when the target trait is well-characterized at the
biochemical and/or physiological level, with known meta-
bolic pathways, to permit selection of candidate genes
(Pflieger et al., 2001). Candidate regions to be included in
AM analysis are also selected based on results from previous
QTL mapping and by assessing comparative studies in
related crops. Selecting candidate genes from closely related
species by synteny requires prior knowledge about the degree
of conservation of gene families. For example, there is a high
synteny between the genomes of apple and pear, as revealed
by the apple genome sequence (Velasco et al., 2010) and by
studies demonstrating high transferability of markers from
apple to pear (Celton et al., 2009; Gasic et al., 2009).
Therefore, significant associations identified by genome-wide
AM in apples (genome has been sequenced) could be readily
tested and validated in pear (genome has not yet been
sequenced) through candidate-gene AM.
One plausible method for identification and genotyping
of SNPs in candidate genes could be through GBS of an
AM population (Fig. 2). However, this method will require
access to high genome coverage for reliable detection of
polymorphisms for plant species where a reference genome
is not available. For plant species lacking a whole genome
sequence, a two-step strategy may be preferable. This will
entail identifying SNPs in candidate genes using high-
throughput sequencing of a small set of a population, and
subsequently genotyping-identified SNPs in a large associa-
tion population using SNP genotyping assays; e.g., Illumina
GoldenGate.
Potential of CNV-based association mapping
In addition to DNA polymorphism, there is growing
evidence for presence of structural variations among
genomes of different individuals within a species, resulting
in phenotypic variations (Swanson-Wagner et al., 2010).
Variations in genome structure consist of chromosomal
rearrangements (inversions and translocations), segmental
and gene duplications, and CNVs in the form of deletions,
duplications, and insertions. These structural variations
may account for phenotypic variation not captured by
DNA marker-based genetic studies. It has been demon-
strated in several studies that variations in genome structure
contribute extensively to genetic variability, influencing
phenotypic variation in plants, livestock, and humans
(Springer et al., 2009; Liu et al., 2010; Stranger et al.,
2011). Chromosomal rearrangements and CNVs in DNA
contribute greatly to total structural variations within
a genome. CNVs are segments of the genome that are >1
kb in length, have >90% sequence identity, and vary in
copy number when compared to a reference genome. CNVs
affect more nucleotides per genome than sequence varia-
tions found in SNPs (Zhang et al., 2009).
4056 | Khan
12 of 16 andand
| Khan Korban
Korban
Studies in domesticated livestock and in humans suggest
that CNVs contribute to important production and disease
traits, either through dosage effects or via non-additive
effects of allele combinations (Springer et al., 2009;
Henshall et al., 2010). In addition to dosage effects, CNVs
could influence gene function by positional effects and by
unmasking mutant alleles of genes when the functional copy
is deleted (Stankiewicz and Lupski, 2010). In contrast to
currently known sequence variations, including SNPs and
SSRs, levels of structural variations, especially of CNVs
that likely contribute to extensive phenotypic diversity and
plasticity of crops, have not been given much attention in
plants. According to Springer et al. (2009), CNVs could
provide yet another valuable resource beyond traditional
microsatellite and SNP variations for pursuing genomic
studies. Hence, investigating the role, structure, and signif-
icance of CNVs in higher plants would lead to new critical
and fundamental information and knowledge on plant
phenotypic outcomes. As genome-wide SNP data are
generated, they can be readily reanalysed to identify CNVs,
and subsequently used for AM studies as well. In recent
studies in cattle and humans, statistical methodologies used
for SNP association studies have also been used to identify
CNVs associated with traits of interest. As CNVs often
include entire genes and their regulatory regions, they are
likely to play major roles in control of expression of
phenotypic traits (Kato et al., 2008). As SNPs are used in
AM studies, CNVs could also prove useful in studying
phenotypic traits. Therefore, it is prudent to capitalize on
CNVs both as markers for mapping efforts as well as
potential functional variants, i.e. analogous to functional
nucleotide polymorphisms.
Often, there are two concerns in utilizing CNVs in AM
studies: (1) to what extent are CNVs inherited rather than
arise as new mutations in each generation, and (2) the
extent to which inherited CNVs arise from common
polymorphisms compared to rare variants. In humans,
studies have revealed that CNVs follow Mendelian in-
heritance, and ;80% of copy number differences between
any two individuals appear to arise from common copy
number polymorphisms (CNPs) with segregating allelic
frequencies of >5%, and > 90% of these CNPs are reported
to arise from those segregating at an allele frequency of
above 1%. Another concern over rare variants is their
difficulty in identifying them, although with high-through-
put sequencing, this is no longer a major concern.
When choosing perennial forest and fruit trees for CNV
studies, presence of high levels of heterozygosity reduces
the power of detecting CNVs and is a valid concern. This
is similar to studies on humans, wherein heterozygosity is
also high, and it could be argued that the availability of
higher quality reference genome(s) would have had
contributed to successful read-depth-based CNV identifi-
cation. As reference genome sequences for perennial trees
are not as well refined as in human or other model species,
reliable identification of CNVs requires higher genome
sequence coverage. Therefore, enhancing the quality of
the reference genome sequence and increasing the rese-
quencing coverage depth could help to avoid problems
related to the failure of the read-mapping algorithm to align
reads to the reference sequence. However, an alternative
approach would be to perform low-coverage sequencing
on large numbers of accessions from an association
panel and then identify the most diverse accessions, which
could then be subjected to deeper genome coverage
sequencing.
Genome-wide AM studies of both common and rare
CNVs, as well as common SNPs can now be pursued
(McCarroll et al., 2008). Thus, CNVs can be reliably
identified using the GBS method as described by Elshire
et al. (2011) with high coverage depth (Fig. 2) and
subsequently used either as functional variations and/or as
neutral markers in association mapping studies.
Conclusions
Recently, increasing numbers of genomic resources have
become available for forest trees and fruit crops. Whole
genome sequences of strawberry (Shulaev et al., 2011),
cacao (Argout et al., 2011), apple (Velasco et al., 2010),
peach (Sosinski et al., 2010), papaya (Ming et al., 2008),
grape (Velasco et al., 2007), and poplar (Tuskan et al.,
2006) have recently become available. Moreover, complete
genome sequences of two citrus species (Gmitter, 2010),
sweet orange and Clementine mandarin, will soon become
available, along with genome sequences of raspberry (Price
et al., 2011), coffee (Wincker et al., 2011), Chinese chestnut
(Barakat et al., 2010), and blueberry (Brown et al., 2010),
among others. Although many of these are merely draft
sequences, high-throughput sequencing integrated with
barcoded multiplexing of samples can drastically reduce
sequencing and genotyping costs. All of these recent
developments will aid not only in expanding genetic and
genomic resources, but also in refining genome sequences of
these fruit species. Ultimately, these genomes can be used as
references to identify SNPs and CNVs, and will significantly
limit flaws of SNP- and CNV-based genome-wide and
candidate gene-based AM studies, especially in economi-
cally important perennial species.
Presence of considerable levels of synteny among many
rosaceous species suggest that even for those plant species
with fewer genomic resources, candidate-gene AM coupled
with QTL mapping studies and comparative mapping
would be feasible and highly valuable. Knowledge acquired
in one species can then be extended to others. For example,
establishing apple as a model by increasing and refining its
genome sequence and identifying regions associated with
important traits could even be used for other related species
of Rosaceae and the sub-tribe Pyreae, particularly in pear,
by conducting comparative genomics. This can lead to
a better understanding of genome structure and polymor-
phism, and will also provide large resources of molecular
markers for germplasm evaluation, breeding, and pursuing
positional cloning efforts.
 Association
Association
Acknowledgements
This work was partially funded by the USDA-NIFA-SCRI
(grant AG 2009-51181-06023) and the University of Illinois
Office of Research (projects 65-325 and 875-922. We thank
Dr. Kenneth Olsen and his research group (Washington
University) for critical reading of the manuscript and their
helpful suggestions.
References
Abdurakhmonov IY, Abdukarimov A. 2008. Application of
association mapping to understanding the genetic diversity of plant
germplasm resources. International Journal of Plant Genomics 2008,
574927.
Agrama HA, Eizenga GC, Yan W. 2007. Association mapping of
yield and its components in rice cultivars. Molecular Breeding 19,
341356.
Alkan C, Kidd JM, Marques-Bonet T, et al. 2009. Personalized
copy number and segmental duplication maps using next-generation
sequencing. Nature Genetics 41, 10611067.
Altshuler D, Pollara VJ, Cowles CR, Van Etten WJ, Baldwin J,
Linton L, Lander ES. 2000. An SNP map of the human genome
generated by reduced representation shotgun sequencing. Nature
407, 513516.
Aranzana MJ, Abbassi EK, Howad W, Arus P. 2010. Genetic
variation, population structure and linkage disequilibrium in peach
commercial varieties. BMC Genetics 11, 69.
Aranzana MJ, Kim S, Zhao KY, et al. 2005. Genome-wide
association mapping in Arabidopsis identifies previously known
flowering time and pathogen resistance genes. PLoS Genetics 1,
531539.
Arai-Kichise Y, Shiwa Y, Nagasaki H, Ebana K, Yoshikawa H,
Yano M, Wakasa K. 2011. Discovery of genome-wide DNA
polymorphisms in a landrace cultivar of japonica rice by whole-
genome sequencing. Plant and Cell Physiology 52, 274282.
Argout X, Salse J, Aury J-M, et al. 2011. The genome of
Theobroma cacao. Nature Genetics. 43, 101108.
Baird NA, Etter PD, Atwood TS, et al. 2008. Rapid SNP discovery
and genetic mapping using sequenced RAD markers. PLoS One 3,
e3376.
Barakat A, Addo-Quaye C, Ficklin S, Saski C, Staton M,
Hebard F, Miller W, Schuster S, Carlson JE. 2010. Sequencing the
genome of Chinese chestnut (Castanea mollissima) for ecosystem
restoration. Abstracts of the American Society of Plant Biologists
Annual Meetings, 31 July to 4 August 2010, Montreal, Canada.
Abstract no. P06009.
Barnaud A, Laucou V, This P, Lacombe T, Doligez A. 2010.
Linkage disequilibrium in wild French grapevine, Vitis vinifera L.
subsp.silvestris. Heredity 104, 431437.
Bernardo R. 2008. Molecular markers and selection for complex traits
in plants: learning from the last 20 years. Crop Science 48,
16491664.
Bradbury PJ, Zhang Z, Kroon DE, Casstevens TM,
Ramdoss Y, Buckler ES. 2007. TASSEL: software for association
mapping
mapping in forest
in forest treestrees
andand
fruit fruit
crops | 13| of
crops 4057
16
mapping of complex traits in diverse samples. Bioinformatics 23,
26332635.
Breseghello F, Sorrells ME. 2006a. Association analysis as
a strategy for improvement of quantitative traits in plants. Crop
Science 46, 13231330.
Breseghello F, Sorrells MS. 2006b. Association mapping of kernel
size and milling quality in wheat (Triticum aestivum L.) cultivars.
Genetics 172, 11651177.
Brown A, Colman S, Lommel S, Rowland L, Diener S,
Windham E, Burke M. 2010. Construction of a blueberry (Vaccinium
corymbosium) draft genomic sequence using multiple platforms.
Abstracts of the Plant and Animal Genome Conference XVIII, 913
January 2010, San Diego, California. Abstract no. P-105.
Brown GR, Bassoni DL, Gill GP, Fontana JR, Wheeler NC,
Megraw RA, Davis MF, Sewell MM, Tuskan GA, Neale DB. 2003.
Identification of quantitative trait loci influencing wood property traits in
loblolly pine III QTL verification and candidate gene mapping. Genetics
164, 15371546.
Buckler ES, Holland JB, Bradbury PJ, et al. 2009. The genetic
architecture of maize flowering time. Science 325, 714718.
Celton J-M, Chagne D, Tustin S, Terakami S, Nishitani C,
Yamamoto T, Gardiner S. 2009. Update on comparative genome
mapping between Malus and Pyrus. BMC Research Notes 2, 182188.
Cevik V, Ryder CD, Popovich A, Manning K, King GJ,
Seymour G. 2009. A FRUITFULL-like gene is associated with genetic
variation for fruit flesh firmness in apple (Malus domestica Borkh.). Tree
Genetics and Genomes 6, 271279.
Chagne D, Crowhurst RN, Troggio M, et al. 2012. Genome-wide
SNP detection, validation, and development of an 8K SNP array for
apple. PLoS One 7, e31745.
Cockram J, White J, Zuluaga DL, et al. 2010. Genome-wide
association mapping to candidate polymorphism resolution in the
unsequenced barley genome. Proceedings of the National Academy of
Sciences, USA 107, 2161121616.
Dillon SK, Nolan M, Li W, Bell C, Wu HX, Southertonet SG. 2010.
Allelic variation in cell wall candidate genes affecting solid wood
properties in natural populations and land races of Pinus radiate.
Genetics 185, 14771487.
Dirlewanger E, Cosson P, Tavaud M, Aranzana MJ, Poizat C,
Zanetto A, Arus P, Laigret F. 2002. Development of microsatellite
markers in peach [Prunus persica (L.) Batsch] and their use in genetic
diversity analysis in peach and sweet cherry (Prunus avium L.).
Theoretical and Applied Genetics 105, 127138.
Eckert AJ, Bower AD, Wegrzyn JL, Pande B, Jermstad KD,
Krutovsky KV, Clair JBS, Neale DB. 2009. Association genetics of
coastal Douglas-fir (Pseudotsuga menziesii var. menziesii, Pinaceae). I.
Cold-hardiness related traits. Genetics 182, 12891302.
Eckert AJ, van Heerwaarden J, Wegrzyn JL, Nelson CD, Ross-
Ibarra J, Gonzalez-Martnez SC, Neale DB. 2010. Patterns of
population structure and environmental associations to aridity across the
range of loblolly pine (Pinus taeda L., Pinaceae). Genetics 185, 969982.
Ehrenreich IM, Hanzawa Y, Chou L, Roe JL, Kover PX,
Purugganan MD. 2009. Candidate gene association mapping of
Arabidopsis flowering time. Genetics 183, 325335.
4058 | Khan
14 of 16 andand
| Khan Korban
Korban
Elshire RJ, Glaubitz JC, Sun Q, et al. 2011. A robust, simple
genotyping-by-sequencing (GBS) approach for high diversity species.
PLoS One 6, e19379.
Emanuelli F, Battilana J, Costantini L, Cunff LL, Boursiquot J-M,
This P, Grando MS. 2010. A candidate gene association study on
muscat flavor in grapevine (Vitis vinifera L.). BMC Plant Biology 10,
241.
Ersoz ES, Wright MH, Gonzalez-Martinez SC, Langley CH,
Neale DB. 2010. Evolution of disease response genes in loblolly pine:
insights from candidate genes. PLoS One 5, e14234.
Ersoz ES, Yu J, Buckler ES. 2007. Applications of linkage
disequilibrium and association mapping in crop plants. In: RK
Varshney, R Tuberosa, eds, Genomics-assisted crop improvement:
vol. 1 Genomics approaches and platforms. The Netherlands:
Springer, pp 97119.
Fournier-Level A, Lacombe T, Le Cunff L, Boursiquot JM,
This P. 2010. Evolution of the VvMYbA gene family, the major
determinant of berry colour in cultivated grapevine (Vitis vinifera L.).
Heredity 104, 351362.
Gasic K, Han Y, Kertbundit S, Shulaev V, Iezzoni A, Stover E,
Bell R, Wisniewski M, Korban S. 2009. Characteristics and
transferability of new apple EST-derived SSRs to other Rosaceae
species. Molecular Breeding 23, 397411.
Gmitter FG. 2010. The haploid mandarin and diploid sweet orange
genome sequences. Abstracts of the Plant and Animal Genome
Conference XVIII, 913 January 2010, San Diego, California. Abstract
no. w-146.
GonzalezMartnez SC, Huber DA, Ersoz E, Davis JM, Neale DB.
2008. Association genetics in Pinus taeda L. II. Carbon isotope
discrimination. Heredity 101, 1926.
GonzalezMartnez SC, Wheeler NC, Ersoz E, Nelson CD,
Neale DB. 2007. Association genetics in Pinus taeda L. I. Wood
property traits. Genetics 175, 399409.
Gupta PK, Rustgi S, Kulwal PL. 2005. Linkage disequilibrium and
association studies in higher plants: present status and future
prospects. Plant Molecular Biology 57, 461485.
Hall D, Luquez V, Garcia MV, St Onge KR, Jansson S,
Ingvarsson PK. 2007. Adaptive population differentiation in
phenology across a latitudinal gradient in European aspen (Populus
tremula L.): a comparison of neutral markers, candidate genes and
phenotypic traits. Evolution 61, 28492860.
Henshall JM, Whan VA, Norris BJ. 2010. Reconstructing CNV
genotypes using segregation analysis: combining pedigree information
with CNV assay. Genetics, Selection, Evolution 42, 34.
Heuertz M, De Paoli E, Kallman T, Larsson H, Jurman I,
Morgante M, Lascoux M, Gyllenstrand N. 2006. Multilocus
patterns of nucleotide diversity, linkage disequilibrium and
demographic history of Norway spruce [Picea abies (L.) Karst].
Genetics 174, 20952105.
Houel C, Bounon R, Chab J, et al. 2010. Patterns of sequence
polymorphism in the fleshless berry locus in cultivated and wild Vitis
vinifera accessions. BMC Plant Biology 10, 284.
Huang X, Feng Q, Qian Q, et al. 2009. High-throughput genotyping
by whole-genome resequencing. Genome Research 19, 10681076.
Huang X, Wei X, Sang T, et al. 2010. Genome-wide association
studies of 14 agronomic traits in rice landraces. Nature Genetics 42,
961967.
Hyten DL, Cannon SB, Song Q, Weeks N, Fickus EW,
Shoemaker RC, Specht JE, Farmer AD, May GD, Cregan PB.
2010. High-throughput SNP discovery through deep resequencing of
a reduced representation library to anchor and orient scaffolds in the
soybean whole genome sequence. BMC Genomics 11, 38.
Ingvarsson PK. 2005. Nucleotide polymorphism and linkage
disequilbrium within and among natural populations of European
aspen (Populus tremula L., Salicaceae). Genetics 169, 945953.
Ingvarsson PK, Garcia MV, Hall D, Luquez V, Jansson S. 2006.
Clinal variation in phyB2, a candidate gene for day-length-induced
growth cessation and bud set, across a latitudinal gradient in
European aspen (Populus tremula). Genetics 172, 18451853.
Ingvarsson PK, Garcia MV, Luquez V, Hall D, Jansson S. 2008.
Nucleotide polymorphism and phenotypic associations within and
around the phytochrome B2 locus in European aspen (Populus
tremula, Salicaceae). Genetics 178, 22172226.
Kato M, Nakamura Y, Tsunoda T. 2008. MOCSphaser: a haplotype
inference tool from a mixture of copy number variation and single
nucleotide polymorphism data. Bioinformatics 24, 16451646.
Khan MA, Han Y, Korban SS. 2012. A high-throughput apple SNP
genotyping assay using the GoldenGate platform. Gene 494,
196201.
Kim S, Plagnol V, Hu TT, Toomajian C, Clark RM, Ossowski S,
Ecker JR, Weigel D, Nordborg M. 2007. Recombination and
linkage disequilibrium in Arabidopsis thaliana. Nature Genetics 39,
11511155.
Komulainen P, Brown GR, Mikkonen M, Karhu A, Garcia-
Gil MR, OMalley D, Lee B, Neale DB, Savolainen O. 2003.
Comparing EST-based genetic maps between Pinus sylvestris and. P.
taeda. Theoretical and Applied Genetics 107, 667678.
Kover PX, Valdar W, Trakalo J, Scarcelli N, Ehrenreich IM,
Purugganan MD, Durrant C, Mott R. 2009. A multiparent advanced
generation inter-cross to fine-map quantitative traits in Arabidopsis
thaliana. PLoS Genetics 5, e1000551.
Krutovsky KV, Neale DB. 2005. Nucleotide diversity and linkage
disequilibrium in cold-hardiness- and wood quality-related candidate
genes in Douglas fir. Genetics 171, 20292041.
Liu GE, Hou Y, Zhu B, et al. 2010. Analysis of copy number variations
among diverse cattle breeds. Genome Research 20, 693703.
Maccaferri M, Sanguineti MC, Demontis A, et al. 2011.
Association mapping in durum wheat grown across a broad range of
water regimes. Journal of Experimental Botany 62, 409438.
McCarroll SA, Kuruvilla FG, Korn JM, et al. 2008. Integrated
detection and population-genetic analysis of SNPs and copy number
variation. Nature Genetics 40, 11661174.
McMullen MD, Kresovich S, Villeda HS, et al. 2009. Genetic
properties of the maize nested association mapping population.
Science 325, 737740.
Ming R, Hou S, Feng Y, et al. 2008. The draft genome of the
transgenic tropical fruit tree papaya (Carica papaya L.). Nature 452,
991996.
 Association
Association
Murray SC, Rooney WL, Hamblin MT, Mitchell SE, Kresovich S.
2009. Sweet sorghum genetic diversity and association mapping for
brix and height. The Plant Genome 2, 4862.
Myles S, Boyko AR, Owens CL, et al. 2011. Genetic structure and
domestication history of the grape. Proceedings of the National
Academy of Sciences, USA 108, 34573458.
Myles S, Peiffer J, Brown PJ, Ersoz ES, Zhang Z, Costich DE,
Buckler ES. 2009. Association mapping: critical considerations shift
from genotyping to experimental design. The Plant Cell 21, 21942202.
Neale DB. 2007. Genomics to tree breeding and forest health.
Current Opinion in Genetics & Development 17, 539544.
Neale DB, Ingvarsson PK. 2008. Population, quantitative and
comparative genomics of adaptation in forest trees. Current Opinion in
Plant Biology 11, 149155.
Neale DB, Kremer A. 2011. Forest tree genomics: growing
resources and applications. Nature Reviews Genetics 12, 111122.
Neale DB, Savolainen O. 2004. Association genetics of complex
traits in conifers. Trends in Plant Science 9, 325330.
Olsen KM, Halldorsdottir SS, Stinchcombe JR, et al. 2004.
Linkage disequilibrium mapping of Arabidopsis CRY2 flowering time
alleles. Genetics 167, 13611369.
Paterson AH, Lander ES, Hewitt JD, Peterson S, Lincoln SE,
Tanksley SD. 1988. Resolution of quantitative traits into Mendelian
factors by using a complete linkage map of restriction fragment length
polymorphisms. Nature 335, 721726.
Pavy N, Paule S, Parsons L, et al. 2005. Generation, annotation,
analysis and database integration of 16,500 white spruce EST
clusters. BMC Genomics 6, 144.
Pavy N, Pegas B, Beauseigle S, et al. 2008. Enhancing genetic
mapping of complex genomes through the design of highly-
multiplexed SNP arrays: application to the large and unsequenced
genomes of white spruce and black spruce. BMC Genomics 9, 21.
Petit RJ, Hampe A. 2006. Some evolutionary consequences of being
a tree. Annual Review of Ecology, Evolution, and Systematics 37,
187214.
Pflieger S, Lefebvre V, Causse M. 2001. The candidate gene
approach in plant genetics: a review. Molecular Breeding 7, 275291.
Price JC, Ward JA, Clement MJ, Weber CA, Lewers KS,
Hagen W, Haynes B, Swanson J-D, Udall JA. 2011. Whole
genome sequencing of the highly heterozygous diploid red raspberry,
Rubus idaeus cv Heritage. Abstracts of the Plant and Animal Genome
XIX Conference. 1519 January 2012, San Diego, California. Abstract
no. w-251.
Pritchard JK, Stephens M, Donnelly P. 2000. Inference of population
structure using multilocus genotype data. Genetics 155, 945959.
Quesada T, Gopal V, Cumbie WP, Eckert AJ, Wegrzyn JL,
Neale DB, Goldfarb B, Huber DA, Casella G, Davis JM. 2010.
Association mapping of quantitative disease resistance in a natural
population of loblolly pine (Pinus taeda L.). Genetics 186, 677686.
Ravel C, Praud S, Murigneux A, Linossier L, Dardevet M,
Balfourier F, Dufour P, Brunel D, Charmet G. 2006. Identification
of Glu-B11 as a candidate gene for the quantity of high-molecular-
weight glutenin in bread wheat (Triticum aestivum L.) by means of an
association study. Theoretical and Applied Genetics 112, 738743.
mapping
mapping in forest
in forest treestrees
andand
fruit fruit
crops | 15| of
crops 4059
16
Rikkerink EHA, Oraguzie NC, Gardiner SE. 2007. Prospects of
association mapping in perennial horticultural crops. In: NC Oraguzie,
EHA Rikkerink, SE Gardiner, HN Silva, eds, Association mapping in
plants. New York: Springer, pp. 249269.
Rungis D, Hamberger B, Berube Y, Wilkin J, Bohlmann J,
Ritland K. 2005. Efficient genetic mapping of single nucleotide
polymorphisms based upon DNA mismatch digestion. Molecular
Breeding 16, 261270.
Savolainen O, Pyhajarvi T. 2007. Genomic diversity in forest trees.
Current Opinion in Plant Biology 10, 162167.
Sax K. 1923. The association of size differences with seed-coat
pattern and pigmentation in Phaseolus vulgaris. Genetics 8, 552560.
Shulaev V, Korban SS, Sosinski B, et al. 2008. Multiple models for
Rosaceae genomics. Plant Physiology 147, 9851003.
Shulaev V, Sargent DJ, Crowhurst RN, et al. 2011. The genome of
woodland strawberry (Fragaria vesca). Nature Genetics 43, 109116.
Sorkheh K, Malysheva-Otto LV, Wirthensohn MG, Tarkesh-
Esfahani S, Martnez-Gomez P. 2008. Linkage disequilibrium,
genetic association mapping and gene localization in crop plants.
Genetics and Molecular Biology 31, 805814.
Sosinski B, Verde I, Morgante M, Rokhsar D. 2010. The
international peach genome initiative. A first draft of the peach genome
sequence and its use for genetic diversity analysis in peach. Abstracts
of the 5th International Rosaceae Genomics Conference. November
2010, Stellenbosch, South Africa. O46.
Springer NM, Ying K, Fu Y, et al. 2009. Maize inbreds exhibit high
levels of copy number variation (CNV) and presence/absence variation
(PAV) in genome content. PLoS Genetics 5, e1000734.
Stankiewicz P, Lupski JR. 2010. Structural variation in the human
genome and its role in disease. Annual Reviews of Medicine 6,
437455.
Stracke S, Haseneyer G, Veyrieras JB, Geiger HH, Sauer S,
Graner A, Piepho HP. 2009. Association mapping reveals gene
action and interactions in the determination of flowering time in barley.
Theoretical and Applied Genetics 118, 259273.
Stranger BE, Stahl EA, Raj T. 2011. Progress and promise of
genome-wide association studies for human complex trait genetics.
Genetics 187, 367383.
Su Z, Li X, Hao Z, et al. 2011. Association analysis of the nced and
rab28 genes with phenotypic traits under water stress in maize. Plant
Molecular Biology Reporter 29, 714722.
Swanson-Wagner RA, Eichten SR, Kumari S, Tiffin P, Stein JC,
Ware D, Springer NM. 2010. Pervasive gene content variation and
copy number variation in maize and its undomesticated. progenitor.
Genome Research 20, 16891699.
Thornsberry JM, Goodman MM, Doebley J, Kresovich S,
Nielsen D, Buckler ES. 2001. Dwarf8 polymorphisms associate with
variation in flowering time. Nature Genetics 28, 286289.
Thumma BR, Matheson BA, Zhang D, Meeske C, Meder R,
Downes GM, Southerton SG. 2009. Identification of a cis-acting
regulatory polymorphism in a eucalypt cobra-like gene affecting
cellulose content. Genetics 183, 11531164.
Thumma BR, Nolan MF, Evans R, Moran GF. 2005.
Polymorphisms in cinnamoyl CoA reductase (CCR) are associated with
4060 | Khan
16 of 16 andand
| Khan Korban
Korban
variation in microfibril angle in Eucalyptus spp. Genetics 171,
12571265.
Tommasini L, Schnurbusch T, Fossati D, Mascher F, Keller B.
2007. Association mapping of Stagonospora nodorum blotch
resistance in modern European winter wheat varieties. Theoretical and
Applied Genetics 115, 697708.
Tuskan GA, DiFazio S, Jansson S, et al. 2006. The genome of
black cottonwood, Populus trichocarpa (Torr. & Gray). Science 313,
15961604.
Velasco R, Zharkikh A, Affourtit J, et al. 2010. The genome of the
domesticated apple (Malus 3 domestica Borkh.). Nature Genetics 42,
833839.
Velasco R, Zharkikh A, Troggio M, et al. 2007. A high quality draft
consensus sequence of the genome of a heterozygous grapevine
variety. PLoS One 2, e1326.
Wincker P, Albert VA, Andrade AA, et al. 2011. Sequencing the
coffee genome. Abstracts of the Plant and Animal Genome
Conference XIX, 1519 January 2011, San Diego, California. Abstract
no. w-152.
Wu X, Ren C, Joshi T, Vuong T, Xu D, Nguyen H. 2010. SNP
discovery by high-throughput sequencing in soybean. BMC Genomics
11, 469.
Zhang Z, Ersoz E, Lai CQ, et al. 2010. Mixed linear model approach
adapted for genome-wide association studies. Nature Genetics 42,
355360.
Zhang F, Gu W, Hurles ME, Lupski JR. 2009. Copy number
variation in human health, disease, and evolution. Annual Review of
Genomics Human Genetics 10, 451481.
Zhu C, Gore M, Buckler ES, Yu J. 2008. Status and prospects of
association mapping in plants. The Plant Genome 1, 520.