Published May 11, 2017
original research
Quantitative Trait Loci Mapping for Spike
Characteristics in Hexaploid Wheat
Yaopeng Zhou, Benjamin Conway, Daniela Miller, David Marshall, Aaron Cooper,
Paul Murphy, Shiaoman Chao, Gina Brown-Guedira, and José Costa*
Abstract
Core Ideas
Wheat (Triticum aestivum L.) spike characteristics determine the
number of grains produced on each spike and constitute key
components of grain yield. Understanding of the genetic basis of
spike characteristics in wheat, however, is limited. In this study,
genotyping-by-sequencing (GBS) and the iSelect 9K assay were
used on a doubled-haploid (DH) soft red winter wheat population
that showed a wide range of phenotypic variation for spike traits.
A genetic map spanning 2934.1 cM with an average interval
length of 3.4 cM was constructed. Quantitative trait loci (QTL)
analysis involving additive effects, epistasis (QQ) and QTL 
environment (QE), and epistasis  environment (QQE) interactions
detected a total of 109 QTL, 13 QE, and 20 QQ interactions in
five environments. Spike characteristics were mainly determined by
additive effects and were fine-tuned by QQ, QE, and QQE. Major QTL QSl.cz-1A/QFsn.cz-1A explained up to 30.9% of the phenotypic variation for spike length (SL) and fertile spikelet number,
QGsp.cz-2B.1 explained up to 15.6% of the phenotypic variation
of grain number per spikelet, and QSc.cz-5A.3 explained up to
80.2% of the phenotypic variation for spike compactness. Additionally, QTL for correlated spike characteristics formed QTL clusters on chromosomes 1A, 5A, 2B, 3B, 5B, 1D, and 5D. This study
expands the understanding of the genetic basis of spike characteristics in hexaploid wheat. A number of stable QTL detected in this
study have potential to be used in marker-assisted selection. Additionally, the genetic map generated in this study could be used
to study other traits of economic importance.
Published in Plant Genome
Volume 10. doi: 10.3835/plantgenome2016.10.0101
© Crop Science Society of America
5585 Guilford Rd., Madison, WI 53711 USA
This is an open access article distributed under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
the plant genome

july
2017

vol.
10,
no. 2 •
•
•
A high-density wheat genetic map was generated
using GBS and the 9K iSelect Array.
Spike characteristics were mainly determined by
additive effects.
Major QTL of spike characteristics were identified on
chromosomes 1A, 2B, and 5A.
W
heat is a major food crop across the globe.
Improving its yield potential is crucial to meeting the food demand from an increasing world population. Grain yield of wheat is largely determined by yield
components out of which the three most important are
spikes per unit area, grains per spike, and grain weight
(Dilbirligi et al., 2006; Mengistu et al., 2012). Previous
Y. Zhou, D. Miller, and A. Cooper, Dep. of Plant Sci. and Landscape Architecture, Univ. of Maryland, 2102 Plant Sci. Bldg.,
College Park, MD 20742; J. Costa, USDA–ARS, 5601 Sunnyside
Avenue GWCC 4-2232, Beltsville, MD 20705; B. Conway, Dep.
of Soil and Crop Sci., Colorado State Univ., Plant Sci. Bldg., Fort
Collins, CO 80523; D. Marshall, USDA–ARS, Plant Sci. Research
Unit, Raleigh, NC 27695; P. Murphy, Dep. of Crop Sci., NC State
Univ., Greenhouse Unit 3, Raleigh, NC 27695; G. Brown-Guedira,
USDA–ARS, Plant Sci. Research Unit, Raleigh, NC 27695; S. Chao,
USDA–ARS, Cereal Crops Research Unit, Fargo, ND 58102. Received 6 Oct. 2016. Accepted 17 Feb. 2017. *Corresponding
author ([email protected]). Assigned to Associate Editor
Katrien Devos.
Abbreviations: DH, doubled-haploid; FSN, fertile spikelet number
per spike; GBS, genotyping-by-sequencing; GSP, grains per spikelet;
ICIM, inclusive composite interval mapping; KASP, Kompetitive Allele
Specific PCR; LOD, logarithm of odds; MCIM, mixed-model-based
composite interval mapping; PCR, polymerase chain reaction; QE,
QTL  environment; QQ, epistasis; QQE, epistasis  environment;
QTL, quantitative trait loci; RIL, recombinant inbred line; SC, spike
compactness; SL, spike length; SSN, sterile spikelet number per
spike; TSN, total spikelet number per spike.
1 of 15
studies have shown that grain yield variation is mostly
associated with increased grain number where grain
number, expressed as grains per square meter, is the
product of spikes per square meter and grains per spike.
There appears, however, to be less opportunity for genetic
yield improvement by selecting heavier grains (Fischer,
2011; Frederick and Bauer, 1999). Increases in grains per
spike and spikes per square meter have contributed the
most to wheat yield improvement in the past decades
(Ma et al., 2007). Spike characteristics, including SL, total
spikelet number per spike (TSN), fertile spikelet number
per spike (FSN), sterile spikelet number per spike (SSN),
spike compactness (SC), and grains per spikelet (GSP)
determine the number of grains per spike and thus, to a
certain extent, determine the yield potential.
Spike characteristics are quantitative traits under
QTL control and subject to environmental influence
(Cui et al., 2012; Ma et al., 2007). Genetic dissection of
spike characteristics could facilitate improving grain
yield potential of wheat. Several domestication genes,
such as Q, compactum (C), and sphaerococcum (S1) are
related to wheat spike morphology and have been identified on chromosomes 5A, 2D, and 3D, respectively (Faris
et al., 2003; Faris and Gill, 2002; Johnson et al., 2007a;
Rao, 1977). The Q gene confers a free-threshing spike
and pleiotropically influences many other domestication
related traits including plant height, glume keeledness,
rachis toughness, spike type, and spike emergence time
(Faris et al., 2003; Simons et al., 2006; Sormacheva et al.,
2014). The C gene is located on the long arm of chromosome 2D near the centromere and affects spike compactness, grain size, grain shape, and grain number per spike
(Johnson et al., 2007a). The S1 gene confers rigid short
culms, straight flag leaves, dense spikes, hemispherical
glumes, and small spherical grains (Rao, 1977). In addition to these loci, previous studies have identified other
genomic regions associated with spike-related traits on
all 21 wheat chromosomes (Borner et al., 2002; Cui et
al., 2012; Deng et al., 2011; Kumar et al., 2007; Ma et al.,
2007; Marza et al., 2006; Wang et al., 2011). For example,
Cui et al. (2012) detected 190 QTL across all wheat chromosomes for seven spike-related traits in two recombinant inbred line populations. 18 of the detected QTL
had major effects and were significant across multiple
environments. Ma et al. (2007) investigated the additive, dominant and epistatic effects of QTL for SL, FSN,
SSN, TSN, and SC in a recombinant inbred line (RIL)
population and also from an immortalized F2 population
derived from the same parents and found 18 genomic
regions on chromosomes 1A, 1B, 2D, 3B, 4A, 5A, 5B, and
7A associated with spike characteristics. Additionally,
Kumar et al. (2007) identified QTL for SL on chromosomes 1A, 1B, 1D, 2B, 2D, 4A, 5A, and 5D and QTL for
TSN on 2D, 4A, 4D, 5A, and 6A. These results demonstrated that multiple loci with unequal effects can affect
spike traits and that epistasis and dominance effects can
also influence genetic architecture of spike characteristics. Furthermore, mapping wheat spike characteristics
2 of 15
QTL as Mendelian factors in wheat was also reported.
Deng et al. (2011) investigated wheat spike traits in a
F2 population, derived from the cross between an elite
cultivar Laizhou953 and an introgression line 05210
(in Laizhou953 background). This population showed
a clear 3:1 segregation ratio for spike number per plant,
SL, and grain number per spike. The underlying QTL
was mapped to chromosome 4B, spanned 6.2 cM, and
explained 30.1 to 67.6% of the phenotypic variation in
two environments.
High-density genetic maps are critical for QTL mapping. Driven by the decreasing cost of next-generation
sequencing technology, high-throughput genotyping
methods are starting to be employed to dissect quantitative traits in wheat with improved map resolution (Li et
al., 2013, 2015a,b; Lin et al., 2015; Spindel et al., 2013). For
example, using GBS in a 155 RIL population, Lin et al.
(2015) delimited a major preharvest sprouting resistance
QTL to a 2.9-cM interval on 4AL. Li et al. (2015b) constructed a consensus map using three RIL populations
genotyped by GBS and validated 15 published QTL and
three well-characterized rust resistance genes (Sr58/Lr46/
Yr29, Sr2/Yr30/Lr27, and Sr57/Lr34/Yr18). Additionally, Li
et al. (2013) positioned the major Hessian fly resistance
H34 in a 4.5-cM region on 6B in a RIL population genotyped with the iSelect 9K SNP Array.
In this study, a DH population of 124 lines derived
from two soft red winter wheat genotypes that showed a
wide range of phenotypic variation for spike characteristics was used for mapping. The objectives of this research
were to construct a genetic map with markers aligned to
the wheat chromosomes survey sequence to identify QTL
responsible for the observed phenotypic variation, gain
a better understating of the genetic architecture of spike
characteristics, and develop DNA markers closely linked
to the QTL to be used for marker-assisted selection.
Materials and Methods
Plant Materials
A DH population of 124 lines was established from
the cross of the soft red winter wheat germplasm line
MD01W233-06-1 = PI 658682 (MD233) (Costa et al.,
2010) and soft red winter wheat cultivar Southern
States 8641 (SS8641) (Johnson et al., 2007b). MD233 is
a breeding line released by the Maryland Agricultural
Experiment Station in 2009 with enhanced Fusarium
head blight resistance. It was produced by crossing the
soft red winter wheat cultivar McCormick [VA92-5139 (IN71761A4-31-5-48//VA71-54-147/‘McNair 1813’)/
AL870365 (‘Coker 747*2/‘Amigo’)] (PI632691) (Griffey
et al., 2005) with ‘Choptank’ (‘Coker 9803’/‘Freedom’)
(PI 639724) (Costa et al., 2006). MD233 carries the RhtD1b dwarfing gene, the Ppd-D1b photoperiod sensitive
allele, as well as the 1RS:1AL translocation. SS8641 is
a photoperiod insensitive cultivar (Ppd-D1a) released
by the University of Georgia Experiment Station in
2007, with high yield and multiple-disease resistance
the plant genome

july
2017

vol.
10,
no.
2
(Johnson et al., 2007b). It is a medium-maturing, whitechaffed, medium-tall line derived from the cross ‘GA
881130/2*GA 881582’. GA 881130 is the source of Hessian fly resistance gene H13 and the rust resistance genes
Lr37/Yr17/Sr38 present on the translocation segment
between the short arms of T. ventricosum 2NS and the
bread wheat chromosome 2AS (Helguera et al., 2003).
Field Experiments and Phenotyping
The DH mapping population and parents were evaluated in field plots with two replications in a randomized
complete block design at five location–year environments
in Maryland (Clarksville, MD, 2013 [C13]; Clarksville,
MD, 2014 [C14]; Queenstown, MD, 2013 [Q13]; Queenstown, MD, 2014 [Q14]) and North Carolina (Kinston,
NC, 2014 [K14]). Plots at Maryland locations consisted
of seven rows 3.2 m long and 15.2 cm apart planted at 18
seeds per 0.305-m row. Plots at North Carolina had seven
rows 19.1 cm apart with 24 seeds per 0.305-m row. Growing season rainfall and temperature data were obtained
from respective research farms for Clarksville, MD, and
Queenstown, MD, and from the National Oceanic and
Atmospheric Administration measurements for Kinston,
NC (http://www.noaa.gov) (Supplemental File S1). Soil
fertility management followed recommended management practices for each location. All environments were
sprayed with the metconazole fungicide (0.11 mL m−2)
(Caramba, BASF) at anthesis to reduce potential infection by Fusarium graminearum.
Ten plants in the middle rows from each plot were
randomly selected for evaluation of spike traits. Traits
examined included SL (cm) measured from the base of
the rachis to the top of the uppermost spikelet, FSN, and
SSN. A spikelet is sterile if none of its florets set grains.
Otherwise, it is fertile. Total spikelet number per spike
was equal to FSN plus SSN. Spike compactness was
derived by dividing TSN by SL, and GSP was derived by
dividing grain number per spike by FSN.
Phenotypic Data Analysis
Phenotypic data analysis was performed using SAS version 9.3 (SAS Institute, 2011). The phenotypic value for
SL, FSN, SSN, TSN, SC, and GSP for 10 plants from each
DH line in each replication was averaged before analyses. Analysis of variance for SL, FSN, SSN, TSN, SC, and
GSP was performed separately for each environment
and for the five environments combined by the PROC
GLM procedure in SAS. The linear model for ANOVA
for single environment analysis was Yij = µ + gi + rj + Eij,
where µ is the overall mean, Yij is the phenotypic value
of the ith DH line in jth replication, gi is the fixed effect
of the ith DH line, rj is the fixed effects of jth replication, and Eij is the random effects of error associated
with Yij, and for combined analysis Yijk = µ + gi + rjk +
ek + Eijk, where µ is the overall mean, Yijk is the phenotypic value of the ith DH line in jth replication of kth
environment, gi is the fixed effect of the ith DH line, rjk
is the fixed effects of jth replication of kth environment,
zhou et al.: qtls for spike characteristics in wheat ek is the fixed effect of the kth environment, and Eijk is
the random effect of error associated with Yijk. Pearson’s
correlation coefficients were calculated by the PROC
CORR procedure in SAS to detect the association among
spike traits. Broad-sense heritability (H2), defined as
, where
is the
is the genotype  envivariance of genotypic effect,
ronment variance, and e and r are the number of environments and replicates, respectively, for each trait was
calculated on a family-mean basis by the PROC MIXED
procedure in SAS, as described by Holland et al. (2003).
Simple-Sequence Repeat Genotyping
and 9K iSelect Array
Approximately 25 mg of leaf tissue of the parents and
124 DH lines were collected from 2- to 3-wk-old seedlings for genomic DNA extraction, which was performed
according to the protocol of Pallotta et al. (2003). For
all SSR markers, the polymerase chain reaction (PCR)
master mix consisted of 2 μL of 20 ng μL−1 genomic DNA
template, 0.40 μL of a 10 μM mixture of forward and
reverse primers, 0.18 μL (0.9 U) of Taq polymerase, 1.20
μL of 10 buffer (10 mM Tris-HCL, 50 mM KCl, and 1.5
mM MgCl2, pH 8.3), 0.96 μL of a 100-μM mixture of
deoxyribonucleotide triphosphates (dNTPs), and 7.26 μL
of water, bringing the total reaction volume to 12 μL. A
touchdown profile was used that consisted of an initial
denaturation at 95C followed by 15 cycles of 95C for 45
s, 65C for 45 s decreasing by 1C each cycle, and 72C
for 60 s, followed by 25 cycles of 50C annealing temperature. The forward primers were 5-modified to include
one of the following fluorescent dyes: 6-FAM, VIC, NED,
or PET. Amplifications were performed using an Eppendorf Mastercycler (Eppendorf AG). Sizing of PCR products was performed by capillary electrophoresis using an
ABI3130xl Genetic Analyzer (Applied BioSystems, Foster
City, CA). Analysis of PCR fragments was performed
using GeneMarker 1.60 software (SoftGenetics, LLC).
The SNP genotyping was performed on the 9K
iSelect SNP genotyping array containing 9000 wheat
SNP markers developed by Illumina Inc. This assay was
designed under the protocols of the International Wheat
SNP Consortium (Cavanagh et al., 2013). In addition, 24
Kompetitive Allele Specific PCR (KASP) markers derived
from SSR markers were also used in genotyping.
Genotyping-by-Sequencing
Genotyping-by-sequencing libraries were prepared
according to Poland et al. (2012) using the PstI-MspI twoenzyme combination for genome complexity reduction.
A set of 96 barcodes having the PstI overhang was used
to barcode samples. The population was sequenced in
combination with samples from another project in two
lanes on the Illumina HiSeq2500 each having 95 samples
and one blank control.
The SNP identification and calling was performed
using the TASSEL 4 GBS pipeline (Glaubitz et al., 2014).
Reads were aligned using the Burrows–Wheeler aligner
3 of 15
version 0.6.2 against a pseudo reference genome developed
from the T. aestivum Chinese Spring chromosome survey
sequence (International Wheat Genome Sequencing Consortium, 2014). The pseudo reference consisted of 41 molecules, one for each chromosome arm except for chromosome 3B, which was a single molecule. Each pseudo molecule consisted of concatenated sequences for chromosome
specific contigs with a string of 64 Ns inserted between
contigs. A text file having the start and end position of
each contig within the pseudo molecule was created and
used to identify SNP containing contigs and to determine
location of SNPs on contigs. The SNPs were named in the
following format: ChromArm_contig number_SNP position within the contig. The SNP in tags aligning to contigs
having map positions from the POPSEQ map (International Wheat Genome Sequencing Consortium; http://
wheat-urgi.versailles.inra.fr/Seq-Repository/PublicationIWGSC) were used to compare order of the MD233 
SS8641 linkage map with the POPSEQ map.
maximum was used as the estimate of the QTL location.
Quantitative trait loci were named following the Catalogue of Gene Symbols for Wheat (McIntosh et al., 2013).
Quantitative trait loci with .cz and .czm in their names
were detected by ICIM and MCIM, respectively.
Linkage Map Construction
Phenotypic Analysis
Markers with >20% missing rate and those that were
monomorphic and distorted (differing significantly from
the expected 1:1 segregation ratio) were eliminated from
the analyses. The remaining polymorphic markers were
clustered into linkage groups using the MAP function
in software IciMapping version 4.0 with a logarithm of
odds (LOD) score of 10 (Li et al., 2008). Linkage groups
from the same chromosome were merged referencing
the marker chromosomal location on the 9K SNP consensus map (Cavanagh et al., 2013), the SSR consensus
map (Somers et al., 2004), and the POPSEQ map. Markers from the same chromosome were reordered and
the genetic distance recalculated with RECORD and
COUNT algorithm in IciMapping version 4.0. Recombination frequencies were converted to centimorgans (cM)
using the Kosambi mapping function.
Quantitative Trait Loci Detection
Two methods were used to map QTL for spike characteristics. First, inclusive composite interval mapping (ICIM)
was performed separately for each environment. Quantitative trait loci with additive effects were detected by the
ICIM-ADD module of IciMapping version 4.0 (Li et al.,
2008). The walking speed for all traits under study was
1 cM. Reference LOD values were determined by 1000
permutations (Doerge, 2002). Type I error to determine
the LOD from the permutation test was 0.05 and the LOD
threshold to declare the presence of a significant QTL
was 3.0. Second, mixed-model-based composite interval
mapping (MCIM) or a full-QTL model was used to estimate additive, QE, QQ, and QQE interaction effects with
QTLNetwork version 2.1 (Wang et al., 1999; Yang et al.,
2007). Additive, QE, QQ, and QQE effects were estimated
by the Monte Carlo Markov chain method with a scanning speed of 1-cM step with the experiment-wise type I
error for putative QTL detection of 0.05. In both methods,
the position at which the LOD score curve reached its
4 of 15
Comparative Mapping
Flanking markers of QTL detected in multiple environments were used to obtain their associated contig
sequences in the wheat survey sequence database (http://
wheat-urgi.versailles.inra.fr/). Contig sequences were
blasted against genomic sequence of Oryza sativa L.
subsp. japonica Kato and Brachypodium distachyon (L.)
Beauv. using BLASTn of Ensemble Plants (http://plants.
ensembl.org/index.html). An E-value of 1  10−10 were
used to retain the best match. A list of orthologous genes
are listed in the Supplemental File S10.
Results
Phenotypic distribution for spike traits is shown in Fig. 1.
Mean values of evaluated traits at each environment are
shown in Supplemental File S2. SS8641 had longer spikes,
also more fertile and total spikelets per spike as well
as more grains per spikelet; whereas MD233 had more
sterile spikelets per spike. The compactness was similar
between the parents. In all environments, the DH population showed significant variation and transgressive
segregation was observed with data distributed beyond
the parental values (Supplemental File S2). ANOVA
results showed that significant differences existed among
DH lines and among environments at p < 0.001 level
for six spike traits (Supplemental File S3). Estimates of
heritability (on a family-mean basis) of the traits were
high, ranging from 0.88 to 0.95. The TSN had the highest
heritability of 0.95, whereas GSP had the lowest (Supplemental File S3). Significant Pearson correlation coefficients were observed among the spike traits in different
environments (Table 1). Spike length showed significant
positive correlation with fertile and total spikelet numbers and negative correlation with compactness across
all five environments. There was a high positive correlation between the number of fertile and total spikelets (r
= 0.87–0.91). In almost all of the environments, SC was
positively correlated with number of total and sterile
spikelets. A significant positive correlation between SC
and FSN (r = 0.27) only occurred in the K14 environment. Grains per spikelet was negatively correlated with
SSN and had no significant relationships with either SL
or FSN, except in Q13. Significant negative correlations
were also observed between GSP and TSN in C13 and
C14 as well as GSP and SC in C13, C14, and Q13.
Genotyping
Approximately 207 million reads were obtained for the
population from the Illumina HiSeq 2500. The number
of reads assigned to each DH line ranged from 998,050 to
the plant genome

july
2017

vol.
10,
no.
2
Fig. 1. Phenotype distribution for spike length (cm), sterile spikelets per spike, fertile spikelets per spike, total spikelets per spike, spike
compactness, and grains per spikelet at (A) Clarksville, MD, 2013; (B) Clarksville, MD, 2014; (C) Queenstown, MD 2013; (D) Queenstown, MD 2014; and (E) Kinston, NC, 2014.
2,888,396 with an average of 1,682,947. Six samples of the
population parents were included to give a higher depth
of coverage and the average number of reads for parents
was 5,565,484. For the population and parents, 94% of
reads mapped to the pseudo reference genome. The GBS
SNP genotypes from the TASSEL 4 pipeline were filtered
for markers homozygous and polymorphic between the
parents and segregating in the mapping population.
A total of 3160 SNPs with <20% missing data that did
not significantly deviate from the 1:1 segregation ratio
expected for a DH population were retained for linkage
mapping. Of the 8,632 functional SNPs from the iSelect
Assay, 1,791 were polymorphic between the parents and
had <20% missing data. These SNP markers, in conjunction with 24 SSR, four KASP markers, and the morphological red coleoptile trait, were used for construction of
the linkage map.
Using 4,980 polymorphic markers, a linkage map was
constructed assigning a total of 4,973 markers to 21 wheat
chromosomes (Table 2). After placing cosegregating markers into bins, the linkage map contained 860 unique loci
(Table 2; Fig. 2). The genetic linkage map spanned 2934.1
cM with an average interval length of 3.4 cM, which was
the average distance between two adjacent unique loci.
The A, B, and D genomes covered total distances of 1023.3,
965.9, and 944.9 cM, respectively. Distribution of mapped
markers and unique loci varied across the three genomes.
The D genome had the fewest markers and loci, with only
7.6% of mapped markers and 15.6% of unique loci, respectively, compared with 45 and 43.7% of mapped markers and unique loci, respectively, for the A genome and
47.1 and 40.7% for the B genome. Similarly, the A and B
genomes had higher map resolution (smaller average interval length) than the D genome. Addition of GBS markers
greatly improved resolution of the linkage map. Of the
860 mapped unique loci, 449 (52.2%) were located by GBS
markers only and 115 (13.4%) by iSelect markers only
(Fig. 3; Supplemental File S4). In particular, the number of
unique loci located on D-genome chromosomes increased
by 60% with the inclusion of GBS markers.
During map construction pairwise recombination
fractions between loci and their associated LOD scores
were calculated. Closely linked loci that showed small
recombination fractions and large LOD scores were placed
in the same linkage group. The assignment of marker to
linkage groups was examined by the plot of estimated
recombination fractions vs. LOD scores (Supplemental File
zhou et al.: qtls for spike characteristics in wheat 5 of 15
Linkage Map Construction and
Quality Assessment
Table 1. Pearson correlation coefficients among spike
length (SL; cm), fertile spikelet number (FSN), sterile
spikelet number (SSN), total spikelet number (TSN),
spike compactness (SC), and grains per spikelet (GSP)
at Clarksville, MD, 2013 (C13) and 2014 (C14), Queenstown, MD, 2013 (Q13) and 2014 (Q14), and Kinston,
NC, 2014 (K14).
Environments
C13
C14
Q13
Q14
K14
Traits
SSN
SL
SSN
FSN
TSN
SC
SL
SSN
FSN
TSN
SC
SL
SSN
FSN
TSN
SC
SL
SSN
FSN
TSN
SC
SL
SSN
FSN
TSN
SC
−0.26**
–
–
–
–
0.00
–
–
–
–
−0.18*
–
–
–
–
0.03
–
–
–
–
−0.05
–
–
–
–
FSN
0.68***
−0.26**
–
–
–
0.66***
−0.09
–
–
–
0.71***
−0.14
–
–
–
0.71***
−0.19*
–
–
–
0.62***
−0.06
–
–
–
TSN
0.58***
0.17
0.91***
–
–
0.61***
0.37***
0.89***
–
–
0.58***
0.37***
0.87***
–
–
0.71***
0.27**
0.90***
–
–
0.55***
0.38***
0.90***
–
–
SC
GSP
−0.55***
0.47***
0.15
0.36***
–
−0.60***
0.36***
0.10
0.25**
0.14
−0.39***
−0.05
−0.22*
−0.38***
0.07
−0.62***
0.05
−0.23*
−0.31***
0.25**
−0.62***
0.23**
−0.09
−0.36***
0.06
−0.39***
0.15
−0.03
−0.12
0.13
−0.40***
0.16
−0.03
−0.16
−0.56***
0.56***
0.07
0.35***
–
−0.60***
0.27**
0.01
0.13
–
−0.50***
0.45***
0.27**
0.44***
–
* Significantly different from zero at the 0.05 probability level.
** Significantly different from zero at the 0.01 probability level.
*** Significantly different from zero at the 0.001 probability level
S5). Loci with smaller recombination fractions and large
LOD scores produced a hot region along the diagonal indicating the proper chromosome assignment of loci. Additionally, very good overall agreement was observed when
our linkage map was compared with the marker order of
the wheat SNP consensus map of Cavanagh et al. (2013)
and wheat POPSEQ map (http://wheat-urgi.versailles.
inra.fr/) (Fig. 4, 5) except for the 4A, 5A, and 5B chromosomes that showed reverse orientation compared with the
consensus map and negative Spearman’s rank correlation
coefficients. For the majority of linkage groups, the Spearman’s rank correlation coefficients between our genetic
map and the SNP consensus map and POPSEQ map were
>0.90. Exceptions included linkage groups corresponding to the D-genome chromosomes and chromosome
4B, where the coefficients could not be calculated or may
be biased because of the very small number of common
markers (Supplemental File S7).
6 of 15
The 1RS.1AL translocated chromosome segregating
in this population affected observed recombination in
the 1A linkage group. All of the recombination on the 1A
linkage group was associated with markers located on
the long arm of the chromosome. A total of 213 markers associated with sequences from the short arm and
the proximal portion of the long arm were cosegregating in our population, indicating lack of recombination
between the 1RS arm from rye (Secale cereale L.) with
wheat 1AS. Less than 6% (31) of the 521 polymorphic
markers on this chromosome represented unique loci
compared with 13 to 26% of markers representing unique
loci on other A-genome chromosomes (Table 2).
Quantitative Trait Loci with Additive Effects
Combined QTL analysis using the MCIM method
detected a total of 45 regions with significant additive
effects on the six spike traits in this study (Table 3). The
number of QTL for each trait ranged from five for SC
and GSP to 10 for TSN. The QTL analysis performed for
each individual environment using the ICIM method
detected a total of 109 putative additive QTL for the
six spike traits (Supplemental File S9). These QTL were
located on 15 chromosomes and mostly formed QTL
clusters (Supplemental File S8). The majority of QTL
detected in individual environments were located within
20 cM of those identified across environments by MCIM
(Table 3). Quantitative trait loci detected in at least four
environments and explaining 10% or more of the variation in an environment were considered to have major
effects. Major QTL were identified on eight chromosomes and mostly affected multiple spike traits: 1A (SL,
FSN, TSN), 5A (SL, SC), 2B (SSN, SC), 3B (SL), 5B (SC),
2D (SSN, TSN), and 5D (TSN, SC). In all cases, the major
QTLs identified also had the largest additive effects for
the respective traits (Table 3). The additive effects from
MCIM represented the average effect of each QTL across
environments and were smaller than those observed in
some individual environments.
Quantitative Trait Loci  Environment
and Epistatic Interactions: QE, QQ, QQE
In this study, we used a MCIM method to estimate the
QE, QQ, and QQE interactions. Thirteen QE interactions
involving eight marker intervals were detected for SSN,
FSN, and TSN (Table 4). Six intervals were also significant for additive effects. The other two involved nonsignificant QTL (LOD < 3) for additive effects. Results from
location–year environments Q14, K14, and Q13 each had
six, five, and two QE interactions, respectively. No QE
interaction was detected in C13 and C14. The contribution of QE interactions ranged from 0.6 to 2.2%. Twenty
pairs of QQ interactions were detected for all six spike
traits in the DH population and three QQE interactions
were also identified (Table 4; Supplemental File S10).
Twelve intervals involved in epistasis were significant
for additive effects. The heritability estimates of epistasis
the plant genome

july
2017

vol.
10,
no.
2
Table 2. Summary of MD233  SS8641 linkage map containing iSelect and genotyping-by-sequencing (GBS)
single-nucleotide polymorphism, simple-sequence repeat (SSR), Kompetitive Allele Specific polymerase chain reaction (KASP) markers.
Chromosome
or genome
1A
2A
3A
4A
5A
6A
7A
A genome
1B
2B
3B
4B
5B
6B
7B
B genome
1D
2D
3D
4D
5D
6D
7D
D genome
Total
No. of
iSelect markers
No. of
GBS markers
286
74
142
69
87
123
102
883
80
206
141
13
155
117
84
796
23
21
12
0
15
23
16
110
1789
232
224
189
202
129
119
260
1355
174
308
345
107
274
168
161
1537
32
93
16
7
20
58
37
263
3155
No. of
SSR or KASP markers
3
1
2
1
2
0
3
12
3
2
2
1
1
1
0
10
0
2
1
1
1
0
1
7
29
Total
markers
521
299
333
272
218
242
365
2250
257
516
488
121
430
286
245
2343
55
116
29
8
36
81
55
380
4973
No. of
unique loci
31
58
63
43
58
33
90
376
47
65
66
28
64
36
44
350
18
34
12
7
20
22
21
134
860
Genetic length
Average interval
length
——————— cM ———————
67.7
2.2
123.9
2.1
216.8
3.4
158.3
3.7
190
3.3
95.1
2.9
171.5
1.9
1023.3
2.7
144.1
3.1
139.8
2.2
134
2
134.5
4.8
123.2
1.9
112.4
3.1
177.8
4
965.9
2.8
85.6
4.8
125.9
3.7
72.7
6.1
76.8
11
179.5
9
144.6
6.6
259.8
12.4
944.9
7.1
2934.1
3.4
Fig. 2. Genetic linkage map of the MD233  SS8641 cross. Each bar represents one group of cosegregating markers.
zhou et al.: qtls for spike characteristics in wheat 7 of 15
Fig. 3. Unique loci located by iSelect and genotyping-bysequencing (GBS) single-nucleotide polymorphism, simplesequence repeat, Kompetitive Allele Specific polymerase
chain reaction (KASP) markers in the genetic map of MD233
 SS8641 cross. A unique locus represents a unique position
mapped by cosegregated markers.
ranged from 0.3 to 4.9%, while those for the QQE interactions ranged from 0.9 to 1.4%.
Discussion
The genetic architecture of quantitative traits is complex.
Quantitative traits are controlled by large-effect QTL
and many loci with small effects. The effects of the same
QTL allele, however, usually vary in their magnitude and
direction with different genetic backgrounds and environments. This is known as genotype  genotype (QQ)
interactions (epistasis) and genotype  environment interactions (Mackay et al., 2009). In this study, we evaluated
quantitative traits expressed in a soft red winter wheat DH
population in five environments. The QTL analyses were
performed to identify QTL with additive genetic effects on
six spike traits and to investigate their interaction effects.
To date, few QTL mapping studies on wheat spike characteristics have been conducted that integrate additive, QQ,
QE interaction, and QQE interaction effects.
Quantitative Trait Loci for Spike Characteristics
Using combined and separate QTL analyses, we detected
major QTL for SL, FSN, TSN, and GSP that colocalized
to a 5-cM region on chromosome 1A (Table 3; Supplemental File S9). This region contains a common SSR
marker, Xbarc28, which was also associated with QTL
for spike length (phenotypic variation explained [PVE]
= 10.8%) by another study (Marza et al., 2006), implying
8 of 15
it is possibly the same underlying QTL. However, in our
study, one of the parents (MD233) carries the 1RS:1AL
translocation, whereas Marza et al. (2006) used material carrying the 1BL.1RS translocation. The well-known
chromosomal segment transferred from rye into cultivated wheat confers superior traits such as improved
grain yield and adaptive traits, high grain number per
spike, grain weight, as well as better disease and insect
resistance (Kim et al., 2004; Qi et al., 2016; Rabinovich,
1998). However, the agronomic benefits of the 1RS segment were not consistent across genetic backgrounds
(Villareal et al., 1998). In our population, the translocation contributed to smaller spikes as evidenced by
MD233 alleles being associated with reduced length, total
number of spikelets, as well as fewer grains per spikelet.
Therefore, the genes influencing spike traits found in this
region could be, partially if not fully, associated with the
1RS:1AL translocation. Further separation of the underlying QTL for SL, FSN, TSN, and GSP that clustered
in this region is hindered by the lack of recombination
between the 1RS arm from rye and 1AS of wheat.
Additional major QTL for SL were identified on the
long arm of chromosome 5A. The vernalization response
gene Vrn-A1 and the major wheat domestication gene Q
are also located on chromosome arm 5AL (Kato et al.,
1998). The wheat VRN1 genes, together with photoperiod
response genes (PPD1 genes) and earliness per se genes
determine flowering time of wheat and hence, in part,
confer wide adaptation to diverse growing environments
around the world (Snape et al., 2001). The Q gene is a
well-known domestication locus conferring the freethreshing character and is responsible for many other
domestication-related traits such as rachis fragility, glume
shape and tenacity, spike length, plant height, and spike
emergence time (Faris et al., 2003; Simons et al., 2006;
Sormacheva et al., 2014). To determine if the SL QTL on
chromosome 5AL were in the same regions where Vrn-A1
and Q are located, genomic sequences of Vrn-A1 (Yan et
al., 2003) and Q (Simons et al., 2006) were used to BLAST
against the wheat draft genome (version IWGSC 1.0) and
Popseq. Two identical sequences that were annotated as
Traes_5AL_13E2DEC48 and Traes_5AL_06DC5CDA7
were identified for Vrn-A1 and Q, respectively. However,
the contigs on which the annotated genes reside were
not assembled into the draft genome and we were not
able to locate their physical position on 5AL. BLAST and
sequence alignment showed that the flanking markers of
QTL identified in this region did not match any annotated genes nor overlap with Vrn-A1 and Q. Thus, we are
not able to determine if one or both of these two genes
contributed to SL variation in this study or if a new locus
other than Vrn-A1 and Q was involved.
Additionally, a consistent QTL for SL (QSl.czm-3B)
was identified on chromosome 3B in C13, C14, Q13, and
K14 environments. Quantitative trait loci for FSN, TSN,
and GSP were also identified in the same region in some
environments (Supplemental File S9). Li et al. (2007)
detected QTL for grain yield and grain number per
the plant genome

july
2017

vol.
10,
no.
2
Fig. 4. Marker order relationship between genetic map from the MD233  SS8641 cross and the 9K iSelect consensus map for all
21chromosomes of hexaploid wheat. Unique loci are represented by spots.
Fig. 5. Marker order relationship between genetic map from the MD233  SS8641 cross and the POPSEQ map for all 21chromosomes
of hexaploid wheat. Unique loci are represented by spots.
zhou et al.: qtls for spike characteristics in wheat 9 of 15
Table 3. Quantitative trait loci (QTL) detected for spike length (SL; cm), sterile spikelet number per spike (SSN), fertile
spikelet number (FSN), total spikelet number per spike (TSN), spikelet compactness (SC), and grain number per
spikelet (GSP) in the MD233  SS8641 population using a mixed-model based composite interval mapping method.
QTL name
Peak
position
Left marker
Right marker
Additive
effect
P-value
Environments†
QSl.czm-1A¶
QSl.czm-3A
QSl.czm-5A
QSl.czm-6A
QSl.czm-3B
QSl.czm-2D.1
QSl.czm-2D.2
QSl.czm-5D.1
QSl.czm-5D.2
QSsn.czm-1A
QSsn.czm-2A
QSsn.czm-3A
QSsn.czm-5A
QSsn.czm-2B
QSsn.czm-5B
QSsn.czm-2D
QSsn.czm-3D
QFsn.czm-1A
QFsn.czm-2A
QFsn.czm-3A
QFsn.czm-3B
QFsn.czm-1D
QFsn.czm-2D.1
QFsn.czm-2D.2
QFsn.czm-5D
QTsn.czm-1A
QTsn.czm-2A
QTsn.czm-3B.1
QTsn.czm-3B.2
QTsn.czm-5B
QTsn.czm-1D
QTsn.czm-2D.1
QTsn.czm-2D.2
QTsn.czm-2D.3
QTsn.czm-5D
QSc.czm-3A
QSc.czm-5A.1
QSc.czm-5A.2
QSc.czm-2B
QSc.czm-5D
QGsp.czm-2A
QGsp.czm-3B
QGsp.czm-5B
QGsp.czm-2D
QGsp.czm-3D
cM
0
2.8
78.9
88.2
45.3
28.5
118.5
44.1
94.5
8.7
7.9
169.8
92.5
69
114
67.7
36.6
0.9
44.8
173.9
38.3
52.6
58
99.7
28.1
0.9
49.8
41.8
81.3
81.9
52.6
34.4
63.7
119.4
19.7
141.1
59.1
134.4
65.1
96.6
82.7
72.5
60.1
118.5
16.6
Xwmc496
3AS_3398062_644
5DL_4584184_14443
6AL_5833398_2845
3B_10706474_9184
2DS_5386510_5609
2DL_9813904_943
5DS_2782527_6993
IWA4274
Xbarc28
2AS_5254682_4662
3AL_4449191_960
5AL_2758417_1165
2BL_8018724_14746
5BL_10922184_4499
XPpd-D1
3DS_2577014_1698
1AL_3962319_3787
2AS_5260825_4791
3AL_4456364_278
3B_10610499_1037
1DS_1890379_6849
2DS_5375380_1169
IWA1072
5DS_2782527_6993
1AL_3962319_3787
2AL_6357025_932
3B_10730729_1056
3B_10723152_5820
5BL_10908775_610
1DS_1890379_6849
2DS_5351730_1417
XPpd-D1
IWA3571
IWA8360
3AL_4406528_3920
5AL_2793993_2759
5AL_2808837_1616
2BS_5242129_6102
5DL_4604235_7923
2AL_6355802_8344
3B_10769285_4136
IWA4103
2DL_9813904_943
3DS_2571866_1401
1AL_3962319_3787
IWA6977
5AL_460494_596
6AL_5828239_2711
3B_10416442_2670
2DS_5351730_1417
IWA3571
Xgdm136
5DL_4604235_7923
1AL_3980322_212
2AS_5309087_9086
3AL_4456364_278
1AS_612486_1691
2BL_7993505_9892
5BL_10872018_3053
2DS_5382880_5243
IWA4559
Xbarc28
2AS_5307952_3491
3AL_4249917_1899
3B_10730729_1056
1DL_2227188_3400
XPpd-D1
2DL_9727927_2650
Xgdm136
Xbarc28
IWA3842
3B_10749124_15377
3B_10397959_8732
5BL_10879329_11196
1DL_2227188_3400
2DS_5375380_1169
2DS_5382880_5243
IWA8325
5DS_2730791_8400
3AL_471273_270
Xbarc100
5AL_2806290_2571
Xgwm319
5DL_1208228_231
2AL_6350115_2943
3B_10776028_6083
5BL_10915544_8072
IWA3571
3DS_2603133_61
0.15
0.1
0.19
0.08
0.12
−0.1
−0.04
0.1
0.08
−0.05
0.05
−0.1
−0.11
0.15
0.14
−0.18
0.09
0.4
−0.14
0.21
0.23
0.18
−0.16
−0.15
0.26
0.35
−0.14
0.26
−0.14
0.11
0.16
−0.11
−0.34
−0.19
0.31
0.04
−0.06
0.02
0.06
−0.04
−0.02
0.03
−0.07
0.05
−0.04
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
0.003
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
0.01
<0.001
<0.001
<0.001
<0.001
C13, C14, Q13, K14
C13, K14
ALL
C13, Q13
C13, C14, Q13, K14
C13, K14
C14
C14
Q14, K14
C13
K14
Q14, C13
K14
ALL
C14, Q14
C13, Q13, Q13, K14
C13, K14
C13, C14, Q13, K14
C14
–
Q13, Q14, K14
C13, K14
C13, Q13
C14, K14
C14, Q14
C13, C14, Q13, K14
C14
Q13
–
–
C13, K14
K14
ALL
C14, Q13, K14
C14 Q13, Q14, K14
C13, C14, Q13
ALL
–
ALL
ALL
–
C14
ALL
C13
–
LOD‡ range
PVE§ range
5.7–11.6
9.2–23.6
3.7–5.0
6.5–7.3
4.0–8.2
6.7–13.1
3.7–5.0
6.5–7.3
3.0–6.6
6.0–12.1
2.7–2.9
3.7–5.5
3
5.6
6.1
13.5
4.0–4.7
5.8–12.0
2.6
7
3.8
6
3.9–8.1
9.8–13.8
5.8
8.4
3.2–8.5
5.1–18.3
4.8–6.0
6.8–10.0
3.6–16.3 7.5–30
2.7
3.7–6.8
9.4–20.6 16.7–30.9
4.2
4.1
–
–
2.8–4.7
4.3–8.9
3.6–10.4
6.3–12.4
3.3–5.3
6.3–10.4
3.6–6.7
3.7–11.6
2.7–7.4
7.7–12.6
5.9–13.2 8.4–20.0
6.2
8.8
3.5
6.4
–
–
–
–
3.1–5.7
5.7–7.4
3.8
5.2
4.2–9.4
5.9–20.9
3.8–4.9
5.1–7.4
2.6–11.6 5.0–18.4
3.6–5.0
7.4–8.8
3.8–27.7 9.3–80.2
–
–
5.8–9.4 13.2–21.6
2.8–7.4
5.6–16.8
–
–
4.3
7.7
3.4–7.7
7.9–14.5
7.5
15.5
–
–
† Environments having a QTL peak within 20 cM according to inclusive composite interval mapping results.
‡ LOD, logarithm of odds.
§ PVE, phenotypic variation explained.
¶ Major QTL noted in bold were significant in at least four environments and explained at least 10% of the variation in one environment.
10 of 15
the plant genome

july
2017

vol.
10,
no.
2
Table 4. Quantitative trait loci  environment interactions influencing spike length (LS; cm), fertile spikelet number
(FSN), sterile spikelet number (SSN), total spikelet number (TSN), spike compactness (SC), and grains per spikelet
(GSP) at Clarksville, MD, 2013 (C13) and 2014 (C14); Queenstown, MD, 2013 (Q13) and 2014 (Q14); and Kinston,
NC, 2014 (K14).
Trait
SSN
SSN
SSN
FSN
FSN
FSN
TSN
TSN
Chromosome
Right marker
Left marker
Position
5A
2B
2D§
1A§
2A§
1D§
1A§
2D§
5AL_2758417_1165
2BL_8018724_14746
XPpd-D1
1AL_3962319_3787
2AS_5260825_4791
1DS_1890379_6849
1AL_3962319_3787
2DS_5351730_1417
5AL_2755044_1686
2BL_7993505_9892
2DS_5382880_5243
Xbarc28
2AS_5307952_3491
1DL_2227188_3400
Xbarc28
2DS_5375380_1169
92.5
69
67.7
0.9
44.8
52.6
0.9
34.4
Q13†
Q14†
K14†
0.05*
−0.05*
−0.22***
−0.12***
0.14**
−0.10*
−0.15***
0.19**
0.11*
0.13*
−0.16**
0.10*
−0.12*
H 2(ae)‡
%
0.9
0.7
2.0
2.2
0.6
0.9
1.2
1.8
* Significant at the 0.05 probability level
**Significant at the 0.01 probability level
***Significant at the 0.001 probability level
† The additive  environment interaction effect at each environment.
‡ H 2(ae) is heritability estimate of the additive  environment interaction effect across environments.
§ Interval with significant additive effect.
spike on 3B in two environments using a population of
recombinant inbred lines derived from two winter wheat
cultivars. In addition, Wang et al. (2009) also found this
region to be significant for grain filling rate and yieldrelated traits over multiple environments.
Three QTL on 2D (Supplemental File S9) were of
special interest because these were the only three loci
where MD233 alleles were associated with a longer spike.
In a QTL mapping study for spike-related traits, Ma et
al. (2007) detected two QTL on chromosome 2D flanked
by marker Xgwm261 for SL and SC in the cross of winter
genotypes Nanda 2419 and Wangshuibai, where the QTL
linked to Xgwm261 explained 8.8 to 23.2% of the phenotypic variation. In our study, Xgwm261 was 2.3 and 13.3
cM away from QSl.cz-2D.1/QFsn.cz-2D.1 and QSl.cz-2D.2,
respectively. In addition, Xgwm261 was reported to flank
colocalized QTL and a QTL cluster for yield-related traits
including plant height, harvest index, days to maturity,
1000-grain weight, and grain weight per spike (Mason et
al., 2013). Taken together, these results suggest that the
region near the marker Xgwm261 on chromosome 2D
may harbor genes affecting grain yield. Another QTL,
QSl.cz-2D.3, mapped to the long arm of chromosome
2D and shared the same interval with major QTL QFsn.
cz-2D.3 and QTsn.cz-2D.4 in agreement with the QTL for
FSN and TSN report by Ma et al. (2007). The same position and genetic effects suggested the possibility of similar underlying QTL.
A few studies have documented QTL and genes for
SSN, FSN, and TSN (Cui et al., 2012; Ma et al., 2007).
Some previously reported QTL were confirmed in the
present study. A minor QTL, QSsn.cz-6D, was consistent with the QTL detected by Cui et al. (2012) who
also located a cluster of QTL for spike characteristics
on chromosome 2B corresponding with the major QTL
clusters identified in the present study. Quantitative trait
loci in this cluster were repeatedly detected in almost
all environments evaluated. At these loci, SS8641 contributed positive additive effects for SSN, TSN, and SC,
whereas MD233 was associated with positive GSP, suggesting that the SS8641 allele of this cluster may lower
spikelet fertility and increase TSN and SC by increasing
the number of sterile spikelets. In addition, a QTL cluster for FSN and TSN was identified on chromosome 5D
flanked by 5DL_74079_2181-Xgdm136 and was located
in the same region as previously reported QTL detected
by Li et al. (2007) and Cui et al. (2012). Cuthbert et al.
(2008) reported a QTL cluster for grain numbers per
spike, grain yield, 1000-grain weight, grain filling time,
and days to heading on chromosome 2D, which may correspond to the region of major QTL QSsn.cz-2D.3 identified in this study (Supplemental File S9).
At the distal end of chromosome 2D, we detected
three additional QTL, QFsn.cz-2D.4, QTsn.cz-2D.5, and
QGsp.cz-2D, which were closely linked. These QTL were
not located in the region containing the compactum (C)
locus, a spike-compacting gene on the long arm of chromosome 2D (Johnson et al., 2007a). The SS8641 alleles in
this region decreased FSN and TSN but increased GSP.
The association of this region with spike traits has not
been reported elsewhere. Furthermore, we found that the
major QTL QFsn.cz-2D.2 shared the interval with QTsn.
cz-2D.2 and QSsn.cz-2D.1 and overlapped with QSn.
cz-2D.2 and QTsn.cz-2D.3 at the locus Ppd-D1. The effects
of these QTL were possibly caused by the locus Ppd-D1,
which is a member of the Ppd1 genes known to confer
photoperiod sensitivity and influence agronomic traits
such as plant height, days to heading, and 1000-grain
weight (Guo et al., 2010). Recently, the Ppd-D1 locus was
shown to control photoperiod-dependent floral induction
zhou et al.: qtls for spike characteristics in wheat 11 of 15
and that it has a major inhibitory effect on paired spikelet
formation by regulating the expression of the FLOWERING LOCUS T (FT) (Boden et al., 2015).
The QTL cluster on chromosome 5A for SC included
the locus Xgwm304 that is neither close to the Q gene
nor the Vrn-A1 gene. However, it has also been related
to grain yield and 1000-grain weight by Cuthbert et al.
(2008) and SL and SC by Ma et al. (2007). In those two
studies, this region was identified as harboring major
QTL because of high PVE values similar to our results.
Thus, it is possible that this region may contribute to
grain yield by increasing spikelet numbers and grain
weight. Furthermore, a major QTL (QSc.cz-5A.3) residing
in a 0.9-cM region on 5AL stood out since it explained
80.2% of the phenotypic variation for SC in K14. It is
flanked by Xgwm304 and IWA4736. IWA4736 is a SNP
within the protein coding region of the annotated gene
Traes_5AL_4210A8A6E. This gene codes for asparagine
synthetase, which plays a crucial role in the synthesis
and transport of asparagine in plants (Gaufichon et al.,
2010). In addition to its role in N assimilation, asparagine synthetase has also been shown to be required for
the defense responses to microbial pathogens in tomato
(Solanum lycopersicum L.) (Olea et al., 2004) and pepper (Capsicum annuum L.) (Hwang et al., 2011). Another
SNP, IWA3445, was mapped to the same position as
IWA4736 but did not hit any annotated gene. Thus, we
postulate that Traes_5AL_4210A8A6E is a possible candidate gene for the QSc.cz-5A.3 QTL.
Sourdille et al. (2003) used a DH population derived
from the cross Courtot  Chinese Spring to study wheat
development traits and detected one QTL on the long
arm of chromosome 5D for SC. This QTL explained
13.6% of the phenotypic variation and was similar to the
genomic region IWA4274- 5DL_1208228_231, where two
major QTL for SL and SC were identified in the present study. Another QTL cluster comprising four major
QTL for GSP on chromosome 5B (Supplemental File S9)
coincided with the interval of the SL QTL QSl.ccsu-5B.2
identified by Kumar et al. (2007).
Chromosome 3A of wheat is known to contain QTL
for grain yield and other important agronomic traits.
Using a RIL population derived from the winter wheat
cultivar Cheyenne and its single chromosome substitution
line Cheyenne (WI3A) where chromosome 3A of Cheyenne was substituted for Wichita chromosome 3A, Mengistu et al. (2012) as well as Dilbirligi et al. (2006) detected
QTL for grain yield, plant height, spikes per square meter,
and grain number per spike and found that most of the
detected QTL on 3A were colocalized in two regions. We
detected five QTL on chromosome 3A for SC and SSN
among which QSsn.cz-3A.1 explained 13.8% of the phenotypic variation while the rest were minor QTL. Based
on the mapping positions of SSR markers used in the current and previous studies (Somers et al., 2004), these five
QTL and the QTL previously identified by Dilbirligi et al.
(2006) and Mengistu et al. (2012) are located in this region.
12 of 15
Genetic Complexity of Spike Characteristics
Most important agronomic traits are quantitative in
nature controlled by polygenes and influenced by the
environment. Understanding the genetic and environmental factors causing the phenotypic variation of
quantitative traits is essential for the genetic improvement of crops via knowledge-based breeding (Mackay,
2001; Würschum, 2012). In the present study, the effects
of major, minor, and epistatic QTL as well as their interactions with the environment and their relative contributions to spike characteristics were estimated (Fig.
6). Most QTL had additive effects and had the largest
genetic contribution to phenotypic variation. This agreed
with previous QTL studies involving QQ, QE, and QQE
interactions (Kuchel et al., 2007; Wu et al., 2012; Xing et
al., 2002; Zhang et al., 2014). Furthermore, QTL for spike
characteristics were not evenly distributed within and
across chromosomes and tended to cluster (Supplemental
File S8). We identified QTL clusters on chromosome 1A,
5A, 2B, 3B, 5B, 1D, 2D, and 5D where QTL for multiple
spike characteristics were colocalized or closely linked
within a 10-cM region. In most cases, each cluster contained at least one major QTL. The clustering of QTL also
partially explained the correlation among spike characteristics. In this study, SL was highly correlated with FSN
across environments (Table 1). This could be caused by
the colocalization of QSl.cz-1A and QFsn.cz-1A plus the
effects of closely linked QTL QSl.cz-3A.1, QSl.cz-3A.2,
and QFsn.cz-3A.1. Despite the slight difference in interpretation, characterizing the interaction at two or more
loci or epistasis is as important in quantitative genetics
as in classical genetics. We found that interactions (QE,
QQ, and QQE) served as modifiers for spike characteristics determination in this DH population. For example,
the interval 5AL_2775235_593–5AL_460494_596 on
chromosome 5A was not detected with significant
additive effects but contributed to SL through its interactions with 2DL_9813904_943-IWA3571, Xwmc496–
1AL_3962319_3787, and 5DS_2782527_6993-Xgdm136,
which were associated with significant additive effects
for FSN, GSP, and SL, respectively. Significant epistasis
was also detected between nonsignificant intervals, such
as 1BL_3802551_64689280–1BL_3886068_66500558
and 2BL_8084228_12492–2BL_8092402_18755, which
increased FSN and accounted for 4.9% of the phenotypic variation. Similar results were reported by Ma et
al. (2007), where the interaction of two nonsignificant
loci on chromosome 3D decreased TSN and FSN. These
results confirmed that loci without main effects may contribute to trait determination through epistasis (Li et al.,
2001). Additionally, we found that the SS8641 allele at the
interval 1AL_3962319_3787-Xbarc28 increased FSN and
TSN in K14 and these effects were enhanced by 21.5%
through the QE interaction. Although the effects and
contribution from QE, QQ, and QQE interactions were
relatively small compared with additive main effects,
they were important terms for fine-tuning the expression
the plant genome

july
2017

vol.
10,
no.
2
Fig. 6. Distribution of relative magnitude of additive (a), additive  environment (ae), epistasis (aa), and epistasis  environment interactions (aae) effects.
of spike traits. This is valuable information for pyramiding QTL in breeding programs.
Advantages of Genotyping-by-Sequencing
and Array-Based Single-Nucleotide
Polymorphism Genotyping
The advantages of GBS and array-based SNP genotyping stand out with the availability of the wheat draft
genome sequence. First, a genetic map with many
physically anchored SNPs was developed. By aligning
GBS and array-based SNPs with the draft genome, we
were able to locate the markers and QTL in the wheat
genome with improved accuracy and confidence as well
as assess the correctness of the marker order. Second,
a candidate gene for a QTL could be postulated when
its flanking SNPs are within a functional gene. For
example, GBS marker 5DS_2730791_8400, significant
for a TSN QTL (QTsn.cz-5D.1) in two environments (C13
and K14), was located within the protein coding gene
Traes_5DS_5C8D3789B. Its ortholog in Arabidopsis thaliana (L.) Heynh. is OBE2, which, together with OBE1, is
required for the maintenance and establishment of both
the shoot and root meristems, possibly by controlling the
expression of the meristem genes such as WUS, PLT1,
and PLT2 and of genes required for auxin responses
(http://www.uniprot.org/uniprot/Q9LUB7). The SNP
5DS_2730791_8400 generated a missense variant and
thus may affect a similar biological process in wheat.
Therefore, Traes_5DS_5C8D3789B could be a candidate
gene for the TSN QTL QTsn.cz-5D.1. Third, these SNPs
can be easily converted to KASP markers to be used by
breeders (Semagn et al., 2014).
Conclusions
This is one of the most comprehensive studies providing
new insights into the genetic basis of wheat spike characteristics. By taking advantage of the wheat genome survey
sequence, we were able to construct a highly reliable genetic
map for QTL analysis. This genetic map enabled us to
zhou et al.: qtls for spike characteristics in wheat locate 190 QTL for spike characteristics in a hexaploid winter wheat DH population. Consistent QTL, such as QGsp.
cz-2B.1 that explained 15.6% of the phenotypic variation
for GSP, QSl.cz-1A/QFsn.cz-1A that explained up to 30.9%
of the phenotypic variation for SL and FSN, as well as QSc.
cz-5A.3 that explained 80.2% of the phenotypic variation
for SC, have potential to be used in marker-assisted selection. The genetic determinant of spike characteristics was
mainly from additive QTL, but was also fine-tuned by QQ
and genetic  environment interactions. For example,
we found that the contribution of QFsn.cz-1A to FSN was
enhanced by 19.9% via QQ interaction with the interval
5DS_2782527_6993-Xgdm136. Additionally, QTL clusters
on chromosomes 1A, 5A, 2B, 3B, 5B, 1D, and 5D with synergistic or antagonistic genetic effects partially explained
the phenotypic correlation between spike characteristics.
These results provide valuable information for manipulating spike morphology for breeding purposes.
Supplemental Information Available
Supplemental information is available with the online
version of this manuscript.
Acknowledgments
This research was supported by the Department of Plant Science and
Landscape Architecture at the University of Maryland, College Park,
and the Maryland Grain Producers Utilization Board. Funding for GBS
was provided by the USDA National Institute of Food and Agriculture
(nifa.usda.gov) Triticeae Coordinated Agricultural Project (grant No.
2011-68002-30029) to GBG. The funders had no role in study design, data
collection and analysis, decision to publish, or preparation of the manuscript. We appreciate Dr. Jim Holland’s help for providing advice on SAS
code for heritability estimate. We would also like to thank Emily Snyder,
Hally Sablosky, Laura Brockdorff, Zeke Waisel, and Shiming Shen for
their excellent technical support.
References
Boden, S.A., C. Cavanagh, B.R. Cullis, K. Ramm, J. Greenwood, E.J.
Finnegan, B. Trevaskis, and S.M. Swain. 2015. Ppd-1 is a key regulator
of inflorescence architecture and paired spikelet development in wheat.
Nature Plants 1:14016. doi:10.1038/nplants.2014.16
Borner, A., E. Schumann, A. Furste, H. Coster, B. Leithold, S. Roder, and E.
Weber. 2002. Mapping of quantitative trait loci determining agronomic
13 of 15
important characters in hexaploid wheat (Triticum aestivum L.). Theor.
Appl. Genet. 105:921–936. doi:10.1007/s00122-002-0994-1
Cavanagh, C.R., S. Chao, S. Wang, B.E. Huang, S. Stephen, S. Kiani, K. Forrest, C. Saintenac, G.L. Brown-Guedira, A. Akhunova, D. See, G. Bai,
M. Pumphrey, L. Tomar, D. Wong, S. Kong, M. Reynolds, M.L. da Silva,
H. Bockelman, L. Talbert, J.A. Anderson, S. Dreisigacker, S. Baenziger,
A. Carter, V. Korzun, P.L. Morrell, J. Dubcovsky, M.K. Morell, M.E.
Sorrells, M.J. Hayden, and E. Akhunov. 2013. Genome-wide comparative diversity uncovers multiple targets of selection for improvement
in hexaploid wheat landraces and cultivars. Proc. Natl. Acad. Sci. USA
110:8057–8062. doi:10.1073/pnas.1217133110
Costa, J.M., H.E. Bockelman, G. Brown-Guedira, S.E. Cambron, X. Chen,
A. Cooper, C. Cowger, Y. Dong, A. Grybauskas, Y. Jin, J. Kolmer, J.P.
Murphy, C. Sneller, and E. Souza. 2010. Registration of the soft red
winter wheat germplasm MD01W233-06-1 resistant to fusarium head
blight. J. Plant Reg. 4:255. doi:10.3198/jpr2010.01.0034crg
Costa, J.M., C.A. Griffey, H.E. Bockelman, S.E. Cambron, X. Chen, A.
Cooper, C. Gaines, R.A. Graybosch, A. Grybauskas, R.J. Kratochvil,
D.L. Long, E. Shirley, and L. Whitcher. 2006. Registration of ‘Choptank’
wheat registration by CSSA. Crop Sci. 46:474–476. doi:10.2135/cropsci2005.04-0026
Cui, F., A. Ding, J. Li, C. Zhao, L. Wang, X. Wang, X. Qi, X. Li, G. Li, J. Gao,
and H. Wang. 2012. QTL detection of seven spike-related traits and
their genetic correlations in wheat using two related RIL populations.
Euphytica 186:177–192. doi:10.1007/s10681-011-0550-7
Cuthbert, J., D. Somers, A. Brûlé-Babel, P.D. Brown, and G. Crow. 2008.
Molecular mapping of quantitative trait loci for yield and yield components in spring wheat (Triticum aestivum L.). Theor. Appl. Genet.
117:595–608. doi:10.1007/s00122-008-0804-5
Deng, S., X. Wu, Y. Wu, R. Zhou, H. Wang, J. Jia, and S. Liu. 2011. Characterization and precise mapping of a QTL increasing spike number
with pleiotropic effects in wheat. Theor. Appl. Genet. 122:281–289.
doi:10.1007/s00122-010-1443-1
Dilbirligi, M., M. Erayman, B.T. Campbell, H.S. Randhawa, P.S. Baenziger,
I. Dweikat, and K.S. Gill. 2006. High-density mapping and comparative
analysis of agronomically important traits on wheat chromosome 3A.
Genomics 88:74–87. doi:10.1016/j.ygeno.2006.02.001
Doerge, R.W. 2002. Mapping and analysis of quantitative trait loci in experimental populations. Nat. Rev. Genet. 3:43–52. doi:10.1038/nrg703
Faris, J.D., J.P. Fellers, S.A. Brooks, and B.S. Gill. 2003. A bacterial artificial chromosome contig spanning the major domestication locus Q in
wheat and identification of a candidate gene. Genetics 164:311–321.
Faris, J.D., and B.S. Gill. 2002. Genomic targeting and high-resolution
mapping of the domestication gene Q in wheat. Genome 45:706–718.
doi:10.1139/g02-036
Fischer, R.A. 2011. Wheat physiology: A review of recent developments.
Crop Pasture Sci. 62:95–114. doi:10.1071/CP10344
Frederick, J.R., and P.J. Bauer. 1999. Physiological and numerical components
of wheat yield. In: E.H. Satorre and G.A. Slafer, editors, Wheat: Ecology
and physiology of yield determination. Haworth Press Inc., New York.
Gaufichon, L., M. Reisdorf-Cren, S.J. Rothstein, F. Chardon, and A. Suzuki.
2010. Biological functions of asparagine synthetase in plants. Plant Sci.
179:141–153. doi:10.1016/j.plantsci.2010.04.010
Glaubitz, J.C., T.M. Casstevens, F. Lu, J. Harriman, R.J. Elshire, Q. Sun,
and E.S. Buckler. 2014. TASSEL-GBS: A high capacity genotyping by
sequencing analysis pipeline. PLoS ONE 9:e90346. doi:10.1371/journal.
pone.0090346
Griffey, C.A., W.L. Rohrer, T.H. Pridgen, W.S. Brooks, J. Chen, J.A. Wilson,
D. Nabati, D.E. Brann, E.G. Rucker, H.D. Behl, M.E. Vaughn, W.L.
Sisson, T.R. Randall, R.A. Corbin, J.C. Kenner, D.W. Dunaway, R.M.
Pitman, H.E. Bockelman, C. Gaines, D.L. Long, D.V. McVey, S.E. Cambron, and L. Whitcher. 2005. Registration of ‘McCormick’ wheat registration by CSSA. Crop Sci. 45:417–419. doi:10.2135/cropsci2005.0417
14 of 15
Guo, Z., Y. Song, R. Zhou, Z. Ren, and J. Jia. 2010. Discovery, evaluation
and distribution of haplotypes of the wheat Ppd-D1 gene. New Phytol.
185:841–851. doi:10.1111/j.1469-8137.2009.03099.x
Helguera, M., I.A. Khan, J. Kolmer, D. Lijavetzky, L. Zhong-qi, and J.
Dubcovsky. 2003. PCR Assays for the Lr37-Yr17-Sr38 cluster of rust
resistance genes and their use to develop isogenic hard red spring wheat
lines. Crop Sci. 43:1839–1847.
Holland, J.B., W.E. Nyquist, and C.T. Cervantes-Martínez. 2003. Estimating and interpreting heritability for plant breeding: An update. Plant
Breed. Rev. 22:9–111.
Hwang, I.S., S.H. An, and B.K. Hwang. 2011. Pepper asparagine synthetase 1 (CaAS1) is required for plant nitrogen assimilation and defense
responses to microbial pathogens. Plant J. 67:749–762. doi:10.1111/
j.1365-313X.2011.04622.x
International Wheat Genome Sequencing Consortium. 2014. A chromosome-based draft sequence of the hexaploid bread wheat (Trititum aestivum) genome. Science 345:1251788.
Johnson, E.B., V.J. Nalam, R.S. Zemetra, and O. Riera-Lizarazu. 2007a.
Mapping the compactum locus in wheat (Triticum aestivum L.) and its
relationship to other spike morphology genes of the Triticeae. Euphytica
163:193–201. doi:10.1007/s10681-007-9628-7
Johnson, J.W., J.W. Buck, G.D. Buntin, and Z. Chen. 2007b. University
of Georgia Experiment Station and the USDA–ARS, Griffin. Annual
Wheat Newletter. 53:104–105.
Kato, K., H. Miura, M. Akiyama, M. Kuroshima, and S. Sawada.
1998. RFLP mapping of the three major genes, Vrn1, Q and B1, on
the long arm of chromosome 5A of wheat. Euphytica 101:91–95.
doi:10.1023/A:1018372231063
Kim, W., J.W. Johnson, P.S. Baenziger, A.J. Lukaszewski, and C.S. Gaines.
2004. Agronomic effect of wheat–rye translocation carrying rye chromatin (1R) from different sources. Crop Sci. 44:1254–1258. doi:10.2135/
cropsci2004.1254
Kuchel, H., K. Williams, P. Langridge, H.A. Eagles, and S.P. Jefferies. 2007.
Genetic dissection of grain yield in bread wheat. II. QTL-by-environment interaction. Theor. Appl. Genet. 115:1015–1027. doi:10.1007/
s00122-007-0628-8
Kumar, N., P.L. Kulwal, H.S. Balyan, and P.K. Gupta. 2007. QTL mapping
for yield and yield contributing traits in two mapping populations of
bread wheat. Mol. Breed. 19:163–177. doi:10.1007/s11032-006-9056-8
Li, C., M. Chen, S. Chao, J. Yu, and G. Bai. 2013. Identification of a novel
gene, H34, in wheat using recombinant inbred lines and single nucleotide polymorphism markers. Theor. Appl. Genet. 126:2065–2071.
doi:10.1007/s00122-013-2118-5
Li, G., Y. Wang, M.S. Chen, E. Edae, J. Poland, E. Akhunov, S. Chao, G. Bai,
B.F. Carver, and L. Yan. 2015a. Precisely mapping a major gene conferring resistance to Hessian fly in bread wheat using genotyping-bysequencing. BMC Genomics 16:108. doi:10.1186/s12864-015-1297-7
Li, H., J.M. Ribaut, Z. Li, and J. Wang. 2008. Inclusive composite interval
mapping (ICIM) for digenic epistasis of quantitative traits in biparental
populations. Theor. Appl. Genet. 116:243–260. doi:10.1007/s00122-0070663-5
Li, H., P. Vikram, R. Singh, A. Kilian, J. Carling, J. Song, J. Burgueno-Ferreira, S. Bhavani, J. Huerta-Espino, T. Payne, D. Sehgal, P. Wenzl, and S.
Singh. 2015b. A high density GBS map of bread wheat and its application for dissecting complex disease resistance traits. BMC Genomics
16:216. doi:10.1186/s12864-015-1424-5
Li, S., J. Jia, X. Wei, X. Zhang, L. Li, H. Chen, Y. Fan, H. Sun, X. Zhao, T.
Lei, Y. Xu, F. Jiang, H. Wang, and L. Li. 2007. A intervarietal genetic
map and QTL analysis for yield traits in wheat. Mol. Breed. 20:167–178.
doi:10.1007/s11032-007-9080-3
Li, Z.K., L.J. Luo, H.W. Mei, D.L. Wang, Q.Y. Shu, R. Tabien, D.B. Zhong,
C.S. Ying, J.W. Stansel, G.S. Khush, and A.H. Paterson. 2001. Overdominant epistatic loci are the primary genetic basis of inbreeding
depression and heterosis in rice. I. Biomass and grain yield. Genetics
158:1737–1753.
the plant genome

july
2017

vol.
10,
no.
2
Lin, M., S. Cai, S. Wang, S. Liu, G. Zhang, and G. Bai. 2015. Genotyping-bysequencing (GBS) identified SNP tightly linked to QTL for pre-harvest
sprouting resistance. Theor. Appl. Genet. 128:1385–1395. doi:10.1007/
s00122-015-2513-1
Ma, Z., D. Zhao, C. Zhang, Z. Zhang, S. Xue, F. Lin, Z. Kong, D. Tian, and
Q. Luo. 2007. Molecular genetic analysis of five spike-related traits
in wheat using RIL and immortalized F2 populations. Mol. Genet.
Genomics 277:31–42. doi:10.1007/s00438-006-0166-0
Mackay, T.F.C. 2001. The genetic architecture of quantitative traits. Annu.
Rev. Genet. 35:303–339. doi:10.1146/annurev.genet.35.102401.090633
Mackay, T.F., E.A. Stone, and J.F. Ayroles. 2009. The genetics of quantitative traits: Challenges and prospects. Nat. Rev. Genet. 10:565–577.
doi:10.1038/nrg2612
Marza, F., G.H. Bai, B.F. Carver, and W.C. Zhou. 2006. Quantitative trait
loci for yield and related traits in the wheat population Ning7840 ´
Clark. Theor. Appl. Genet. 112:688–698. doi:10.1007/s00122-005-0172-3
Mason, R.E., D.B. Hays, S. Mondal, A.M.H. Ibrahim, and B.R. Basnet. 2013.
QTL for yield, yield components and canopy temperature depression
in wheat under late sown field conditions. Euphytica 194:243–259.
doi:10.1007/s10681-013-0951-x
McIntosh, R.A., Y. Yamazaki, J. Dubcovsky, W. Rogers, C. Morris, R.
Appels, and X. Xia. 2013. Catalogue of gene symbols for wheat. In: Proc.
of 12th Int. Wheat Genetics Symp, Yokohama, Japan. 8–13 Sep. 2013 .
Mengistu, N., P.S. Baenziger, K.M. Eskridge, I. Dweikat, S.N. Wegulo, K.S.
Gill, and A. Mujeeb-Kazi. 2012. Validation of QTL for grain yieldrelated traits on wheat chromosome 3A using recombinant inbred chromosome lines. Crop Sci. 52:1622–1632. doi:10.2135/cropsci2011.12.0677
Olea, F., A. Pérez-García, F.R. Cantón, M.E. Rivera, R. Cañas, C. Ávila, F.M.
Cazorla, F.M. Cánovas, and A. de Vicente. 2004. Up-regulation and
localization of asparagine synthetase in tomato leaves infected by the
bacterial pathogen Pseudomonas syringae. Plant Cell Physiol. 45:770–
780. doi:10.1093/pcp/pch092
Pallotta, M.A., P. Warner, R.L. Fox, H. Kuchel, S.J. Jefferies, and P. Langridge. 2003. Marker assisted wheat breeding in the southern region of
Australia. In: Proc. of the Tenth Int. Wheat Genetics Symp., Paestum,
Italy. 1–6 Sept. 2003. p. 789-791
Poland, J.A., P.J. Brown, M.E. Sorrells, and J.L. Jannink. 2012. Development of high-density genetic maps for barley and wheat using a novel
two-enzyme genotyping-by-sequencing approach. PLoS ONE 7:e32253.
doi:10.1371/journal.pone.0032253
Qi, W., Y. Tang, W. Zhu, D. Li, C. Diao, L. Xu, J. Zeng, Y. Wang, X. Fan,
L. Sha, H. Zhang, Y. Zheng, Y. Zhou, and H. Kang. 2016. Molecular
cytogenetic characterization of a new wheat–rye 1BL•1RS translocation
line expressing superior stripe rust resistance and enhanced grain yield.
Planta 244:405–416. doi:10.1007/s00425-016-2517-3
Rabinovich, S.V. 1998. Importance of wheat–rye translocations for breeding modern cultivar of Triticum aestivum L. Euphytica 100:323–340.
doi:10.1023/A:1018361819215
Rao, M.V.P. 1977. Mapping of the sphaerococcum ‘s’ on chromosome 3D of
wheat. Cereal Res. Commun. 5:15–17.
SAS Institute. 2011. SAS system for Windows. v. 9.3. SAS Inst. Inc., Cary, NC.
Semagn, K., R. Babu, S. Hearne, and M. Olsen. 2014. Single nucleotide polymorphism genotyping using Kompetitive Allele Specific PCR (KASP):
Overview of the technology and its application in crop improvement.
Mol. Breed. 33:1–14. doi:10.1007/s11032-013-9917-x
Simons, K.J., J.P. Fellers, H.N. Trick, Z. Zhang, Y.S. Tai, B.S. Gill, and J.D.
Faris. 2006. Molecular characterization of the major wheat domestication gene Q. Genetics 172:547–555. doi:10.1534/genetics.105.044727
Snape, J., K. Butterworth, E. Whitechurch, and A.J. Worland. 2001. Waiting
for fine times: genetics of flowering time in wheat. In: Z. Bedö and L.
Láng, editors, Wheat in a global environment. Vol. 9. Springer, Netherlands. p. 67–74. doi:10.1007/978-94-017-3674-9_7
Somers, D., P. Isaac, and K. Edwards. 2004. A high-density microsatellite
consensus map for bread wheat (Triticum aestivum L.). Theor. Appl.
Genet. 109:1105–1114. doi:10.1007/s00122-004-1740-7
Sormacheva, I., K. Golovnina, V. Vavilova, K. Kosuge, N. Watanabe, A.
Blinov, and N.P. Goncharov. 2014. Q gene variability in wheat species with different spike morphology. Genet. Resour. Crop Evol. 62.
doi:10.1007/s10722-014-0195-1
Sourdille, P., T. Cadalen, H. Guyomarc’h, J.W. Snape, M.R. Perretant, G.
Charmet, C. Boeuf, S. Bernard, and M. Bernard. 2003. An update of the
Courtot ´ Chinese Spring intervarietal molecular marker linkage map
for the QTL detection of agronomic traits in wheat. Theor. Appl. Genet.
106:530–538. doi:10.1007/s00122-002-1044-8
Spindel, J., M. Wright, C. Chen, J. Cobb, J. Gage, S. Harrington, M. Lorieux,
N. Ahmadi, and S. McCouch. 2013. Bridging the genotyping gap: Using
genotyping by sequencing (GBS) to add high-density SNP markers and
new value to traditional bi-parental mapping and breeding populations.
Theor. Appl. Genet. 126:2699–2716. doi:10.1007/s00122-013-2166-x
Villareal, R.L., O. Bañuelos, A. Mujeeb-Kazi, and S. Rajaram. 1998.
Agronomic performance of chromosomes 1B and T1BL.1RS nearisolines in the spring bread wheat Seri M82. Euphytica 103:195–202.
doi:10.1023/A:1018392002909
Wang, D.L., J. Zhu, Z.K.L. Li, and A.H. Paterson. 1999. Mapping QTLs with
epistatic effects and QTL ´ environment interactions by mixed linear
model approaches. Theor. Appl. Genet. 99:1255–1264. doi:10.1007/
s001220051331
Wang, J., W. Liu, H. Wang, L. Li, J. Wu, X. Yang, X. Li, and A. Gao. 2011.
QTL mapping of yield-related traits in the wheat germplasm 3228.
Euphytica 177:277–292. doi:10.1007/s10681-010-0267-z
Wang, R.X., L. Hai, X.Y. Zhang, G.X. You, C.S. Yan, and S.H. Xiao. 2009.
QTL mapping for grain filling rate and yield-related traits in RILs of the
Chinese winter wheat population Heshangmai ´ Yu8679. Theor. Appl.
Genet. 118:313–325. doi:10.1007/s00122-008-0901-5
Wu, X., X. Chang, and R. Jing. 2012. Genetic insight into yield-associated
traits of wheat grown in multiple rain-fed environments. PLoS ONE
7:e31249. doi:10.1371/journal.pone.0031249
Würschum, T. 2012. Mapping QTL for agronomic traits in breeding populations. Theor. Appl. Genet. 125:201–210. doi:10.1007/s00122-012-1887-6
Xing, Z., F. Tan, P. Hua, L. Sun, G. Xu, and Q. Zhang. 2002. Characterization of the main effects, epistatic effects and their environmental interactions of QTLs on the genetic basis of yield traits in rice. Theor. Appl.
Genet. 105:248–257. doi:10.1007/s00122-002-0952-y
Yan, L., A. Loukoianov, G. Tranquilli, M. Helguera, T. Fahima, and J. Dubcovsky. 2003. Positional cloning of the wheat vernalization gene VRN1.
Proc. Natl. Acad. Sci. USA 100:6263–6268. doi:10.1073/pnas.0937399100
Yang, J., J. Zhu, and R.W. Williams. 2007. Mapping the genetic architecture
of complex traits in experimental populations. Bioinformatics 23:1527–
1536. doi:10.1093/bioinformatics/btm143
Zhang, J., J. Chen, C. Chu, W. Zhao, J. Wheeler, E. Souza, and R. Zemetra.
2014. Genetic dissection of QTL associated with grain yield in diverse
environments. Agronomy 4:556–578. doi:10.3390/agronomy4040556
zhou et al.: qtls for spike characteristics in wheat 15 of 15