Annals of Botany XX: 1–17, 2023

https://doi.org/10.1093/aob/mcad003, available online at www.academic.oup.com/aob

Variations in phenological, physiological, plant architectural and yield-related

traits, their associations with grain yield and genetic basis
Yibo Li1,2, Fulu Tao1,2,3,*, Yuanfeng Hao4, Jingyang Tong4, Yonggui Xiao4, Zhonghu He4,* and

Downloaded from https://academic.oup.com/aob/advance-article/doi/10.1093/aob/mcad003/6992592 by guest on 05 March 2023

Matthew Reynolds5,

Key Laboratory of Land Surface Pattern and Simulation, Institute of Geographic Sciences and Natural Resources Research,
1

CAS, Beijing 100101, China, 2University of Chinese Academy of Sciences, Beijing 100049, China, 3Natural Resources Institute
Finland (Luke), Helsinki, Finland, 4Institute of Crop Sciences, Chinese Academy of Agricultural Sciences, Beijing 100081,
China, and 5International Maize and Wheat Improvement Center (CIMMYT), Texcoco, Mexico
*
For correspondence. E-mail taofl@igsnrr.ac.cn or hezhonghu02@caas.cn

Received: 28 October 2022 Returned for revision: 12 December 2022 Editorial decision: 4 January 2023 Accepted: 9 January 2023

• Background and Aims Physiological and morphological traits play essential roles in wheat (Triticum aestivum)
growth and development. In particular, photosynthesis is a limitation to yield. Increasing photosynthesis in wheat
has been identified as an important strategy to increase yield. However, the genotypic variations and the genomic
regions governing morphological, architectural and photosynthesis traits remain unexplored.
• Methods Here, we conducted a large-scale investigation of the phenological, physiological, plant architec-
tural and yield-related traits, involving 32 traits for 166 wheat lines during 2018–2020 in four environments, and
performed a genome-wide association study with wheat 90K and 660K single nucleotide polymorphism (SNP)
arrays.
• Key Results These traits exhibited considerable genotypic variations in the wheat diversity panel. Higher yield
was associated with higher net photosynthetic rate (r = 0.41, P < 0.01), thousand-grain weight (r = 0.36, P < 0.01)
and truncated and lanceolate shape, but shorter plant height (r = −0.63, P < 0.01), flag leaf angle (r = −0.49,
P < 0.01) and spike number per square metre (r = −0.22, P < 0.01). Genome-wide association mapping discovered
1236 significant stable loci detected in the four environments among the 32 traits using SNP markers. Trait values
have a cumulative effect as the number of the favourable alleles increases, and significant progress has been
made in determining phenotypic values and favourable alleles over the years. Eleven elite cultivars and 14 traits
associated with grain yield per plot (GY) were identified as potential parental lines and as target traits to develop
high-yielding cultivars.
• Conclusions This study provides new insights into the phenotypic and genetic elucidation of physiological and
morphological traits in wheat and their associations with GY, paving the way for discovering their underlying gene
control and for developing enhanced ideotypes in wheat breeding.

Key words: Genetic variation, ideotypes, photosynthetic traits, Triticum aestivum, yield potential.

INTRODUCTION 1960s, improved grain yield has mostly been due to an in-
crease in harvest index, now close to the theoretical maximum.
Wheat (Triticum aestivum) accounts for about 20 % of hu- Another important strategy to increase crop yield further and
mans’ daily protein and calorie intake globally (FAO, 2019). solve the food crisis is improving the photosynthetic efficiency
The world’s population is expected to reach 9.6 billion by 2050. of crops. This strategy represents the core of a possible second
However, the global growth rate of wheat productivity is only green revolution (Xu & Shen, 2002; Driever et al., 2014).
1.1 % per annum (Dixon et al., 2009), and has even stagnated Improvement in any of the canopy photosynthesis contributors
in some regions (Ray et al., 2019). An annual gain of only about reflects a potential increase in yield and biomass production
2 % in grain yield may meet the projected global requirement (Takai et al., 2013). As the primary determinant of plant prod-
for wheat (Li et al., 2019a). Therefore, increasing annual yield uctivity, photosynthesis still has the potential to improve radi-
gain is critical for food security (Lopes et al., 2012; Abbai et ation use efficiency (RUE) (Long et al., 2006; Zhu et al., 2010;
al., 2020; Xiao et al., 2022). Today, the optimization of agro- Parry et al., 2011; Raines, 2011; Li et al., 2022a). Studies have
nomic management and sustainable intensification is increas- shown a positive relationship between photosynthesis, bio-
ingly accompanied by genomic and phenomics technologies to mass and yield (Fischer et al., 1998; Kruger & Volin, 2006).
further improve yield productivity (Godfray et al., 2010; Fu et Theoretically, photosynthesis could be improved by increasing
al., 2020; Pang et al., 2020; Welcker et al., 2022). the photosynthetic rate per unit leaf area and optimizing light
A quantitative understanding of the mechanisms influencing interception and utilization by modifying architecture and
yield gains is important for major food crops (Rizzo et al., photosynthetic duration (Driever et al., 2014). By unravelling
2022). Since the green revolution between 1950 and the late

© The Author(s) 2023. Published by Oxford University Press on behalf of the Annals of Botany Company.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/
by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
2 Li et al. — Variations in wheat traits and their genetic basis

the genetics of complicated characteristics and having a better introduction of novel adaptive alleles into genetic germplasms
understanding of the molecular processes of genes that sup- may improve grain production of old or newly produced wheat
port desirable features, the novel genomic areas or candidate cultivars to further balance global supply and demand (Rahimi
genes discovered could be utilized to enhance crops (Cui et al., et al., 2019). MAS, relying on identifying agronomic trait loci
2011; Li et al., 2022b). Many genes and hotspot genomic re- and the characterization of their genetic architecture, has been
gions influencing target traits for crop improvement have been applied for this purpose. Moreover, genome-wide association
identified thanks to recent improvements in DNA sequencing studies (GWAS) have become popular as a method to identify
technologies (Azadi et al., 2015). By matching crop genotypes marker–trait correlations (Wang et al., 2014; Li et al., 2018a).
to target environments, adopting agroecological genetics and High-resolution GWAS mapping has been applied to crops

Downloaded from https://academic.oup.com/aob/advance-article/doi/10.1093/aob/mcad003/6992592 by guest on 05 March 2023

genomics perspectives may maximize communal yield (Abbai to investigate the complex genetic architecture of polygenic
et al., 2020). Studies of photosynthesis using molecular modifi- traits (Cavanagh et al., 2013; Fu et al., 2020). Complex poly-
cation techniques, such as improving ribulose-1,5 bisphosphate genic traits in wheat, such as yield and bread-making quality,
carboxylase/oxygenase activity, faster regeneration of ribulose- have been improved through QTLs/genes identified in associ-
1,5-bisphosphate, and introducing carbon-concentrating mech- ation mapping studies (Crossa et al., 2007; Arif et al., 2012).
anisms, have proposed increases in the photosynthetic rate per However, few studies have comprehensively investigated the
unit leaf area (Parry et al., 2011; Bailey et al., 2019). Thanks to natural variation in phenological and physiological processes,
high-throughput phenotyping, we can now map the genetic re- plant architecture, yield-related traits, as well as their genetic
gions (and genes) that control variation in physiological traits, basis.
such as photosynthesis, that would otherwise be impossible to Therefore, the objectives of this study were to (1) investigate
score in large panels under field conditions (Flood et al., 2011). the phenotypic and genetic relationships among physiological
Natural variations in the rate of photosynthesis have been re- processes and yield, (2) identify stable loci for grain yield
ported in previous studies (Hubbart et al., 2007; Kanemura et per plot (GY) and physiological process traits using GWAS
al., 2007; Ohsumi et al., 2007; Gu et al., 2012), indicating that based on high-density single nucleotide polymorphism (SNP)
photosynthesis improvement is certainly recognized as an im- markers, (3) detect the available loci of traits in breeding for
portant target and is expected to be possible, even though gen- high yield and photosynthesis, and (4) characterize the 166 elite
etic analysis for photosynthesis lags behing that for sink size wheat cultivars and identify potential parent sources to develop
(Takai et al., 2013). Therefore, identifying genomic regions as- new, high-yielding cultivars.
sociated with high RUE and photosynthetic capacity will help
in breeding high-yielding cultivars (Driever et al., 2014).
RUE optimization can be achieved by improving light con- MATERIALS AND METHODS
version efficiency into harvestable grains and by increasing the
interception of photosynthetically active radiation (PAR) (Beer Plant materials and field trials
et al., 2010; Li et al., 2021). Flag leaves, as the ‘functional
leaves’ in wheat production, are the primary photosynthetic This study’s association panel comprised 166 representative
organ contributing 45–58 % of photosynthetic performance elite wheat (Triticum aestivum L.) cultivars chosen from more
during the grain-filling stage (Sharma et al., 2003; Khaliq & than 400 cultivars (Supplementary Data Table S1), including
Irshad, 2008). Flag leaf angle (FLANG) determines the amount 141 from the Yellow and Huai River Valley Zone (YHV), the
of incident light received for photosynthesis by leaves. Breeders major wheat-producing region of China (15.3 million hec-
have utilized this trait to optimize plant architecture (Morinaka tares, accounting for ~65 % of national wheat production),
et al., 2006). Studies have also shown a positive correlation in and 22 from five other countries (Liu et al., 2017; Zhu et al.,
cereals of flag leaf width (FLW), flag leaf length (FLL) and flag 2019; Fu et al., 2020). These cultivars represent the genetic
leaf area (FLA) with thousand-grain weight (TGW) (Wang et diversity of the YHV. These cultivars were divided into five
al., 2011b). Thus, a comprehensive understanding of flag leaf groups as follows: nine released during 1947–1979 (P1), 13
physiological and morphological traits will provide new in- in the 1980s (P2), 36 in the 1990s (P3), 59 in the 2000s (P4)
sights related to RUE and yield. Identifying quantitative trait and 13 in the 2010s (P5). The association panel was grown in
loci (QTLs) controlling flag leaf traits is an essential way to four environments and years, including Zhoukou (33°37ʹN,
enhance marker-assisted selection (MAS) for wheat yield im- 114°38ʹE; ZK19) and Xinxiang (35°18ʹN, 113°51ʹE; XX19)
provement through QTL pyramiding (Yang et al., 2016). during 2018–2019, and Luohe (33°36ʹN, 113°58ʹE; LH20)
A crop ideotype consists of several morphological and and Xinxiang (35°18ʹN, 113°51ʹE; XX20) during 2019–2020.
physiological traits that contribute to enhanced yield or better Field trials were conducted in a randomized block design with
performance than in the current crop cultivars. Recognizing two replications at all locations under well-watered conditions.
crop ideotypic traits has become a priority for breeding high- Irrigation was applied at least three times between four critical
yielding cultivars and for designing ideotypes (Tao et al., 2017a; stages (sowing at GS21, tillering at GS27, booting at GS31 and
Bodner et al., 2018). Some agronomic traits, such as crop archi- flowering at GS73) using diffuse irrigation. The area of each
tecture, phenological date, spike and grain-related morpho- plot was 4.2 m2 (3 m in length, 1.4 m in width). Sowing density
logical characteristics, also play crucial roles in yield formation was 270 seeds m−2 with a row spacing of 20 cm. Agricultural
(Jaiswal et al., 2016; Sun et al., 2017). Thus, understanding management, including fertilizer application, irrigation, weed
the natural variation among traits may promote new variation management, insect pest control throughout the growing season
for desirable traits, in which case additional genetic modifi- and harvesting for all environments, were kept the same in all
cations may be targeted to further improve crop yields. The four environments.
 Li et al. — Variations in wheat traits and their genetic basis 3

Phenotypic trait evaluation Ten spikes were randomly harvested in each plot at anthesis,
10 d after flowering and 20 d after flowering to calculate the
At the flowering stage, the main tillers of five representative
grain-filling rate (GFR). The stay-green trait, an indicator of
plants from each plot were used for phenotypic evaluation of
maintaining green character, was determined as the percentage
the flag leaf. The distance from the tip to the base was regarded
of decline in SPAD value at the grain-filling stage compared
as flag leaf length (FLL, cm), and the widest section of the flag
with the flowering stage. Days to anthesis and maturity from
leaf was flag leaf width (FLW, cm). Flag leaf area (FLA, cm2)
sowing were noted when more than 50 % of the plants of each
was calculated using the equation 0.77 × FLL × FLW (Li et
plot displayed anthesis and maturity at Zadoks GS65 and GS92
al., 2021). Flag leaf angle (FLANG, °) was measured as the
stages, respectively (Zadoks, 1974). Moreover, the thermal

Downloaded from https://academic.oup.com/aob/advance-article/doi/10.1093/aob/mcad003/6992592 by guest on 05 March 2023

angle between the flag leaf midrib and the stem below the spike,
time from sowing to flowering stage (TTF) and from sowing to
with a smaller leaf angle indicating more upright leaves. Flag
maturity stage (TTM) were calculated. The shape of the wheat
leaf biomass (FLB, g) was measured after drying for 72 h at
spike was acuminate, fusiform, linear, truncated and lanceolate
75 °C in an oven to a constant weight. Flag specific leaf area
(Bonnett, 1966; Backhaus et al., 2022), assigned values of 1,
(FSLA, cm2 g−1) was defined as leaf area ratio to leaf biomass.
2, 3, 4 and 5, respectively. The larger the spike shape, the more
Chlorophyll content SPAD meter readings (Minolta Camera
it tends to be truncated and lanceolate in shape. The smaller
Co., Osaka, Japan) were used to assess the relative chloro-
the spike shape, the more likely it is to be fusiform and acu-
phyll contents at the flowering stage. SPAD readings from five
minate in shape. A detailed image and description are provided
plants per plot were averaged at each flag leaf’s top, middle and
in Supplementary Data Fig. S2.
bottom.
The leaves, spikes and stems were clipped and weighed to de-
Leaf gas exchanges and chlorophyll fluorescence parameters
termine the fresh weight at the anthesis stage. They were dried
of flag leaves, including light-saturated net photosynthetic rate
in an oven for 72 h at 75 °C to a constant weight. Leaf water con-
(Pn, μmol CO2 m−2 s−1), stomatal conductance (Gs, mol H2O
tent (LWC), spike water content (SPWC) and stem water con-
m−2 s−1), transpiration rate (Tr, mmol H2O m−2 s−1), intercellular
tent (STWC) were calculated as the ratio of the corresponding
CO2 concentration (Ci, µmol CO2 m−2 s−1) and the maximum
water content of dry biomass. Leaf dry weights (LDWS), spike
quantum yield of photosystem II (PSII) photochemistry (Fvʹ/
dry weights (SPDWS), stem dry weights (STDWS) and total
Fmʹ), were measured during clear sky mornings (0900–1100 h)
dry weights (TDW) were measured after drying. After harvest,
using an infrared gas analyser (LI-6400XT; Li-Cor, Lincoln,
the kernel number per spike (KN, kernels per spike) and the
NE, USA) with a fluorescence leaf chamber (LI-6400-40;
spike number per square meter (spikes m–2) were measured.
Li-Cor). The saturating photosynthetic photon flux density was
After air-drying, thousand-grain weight (TGW) and grain yield
set to 1200 µmol m−2 s−1. Water use efficiency (WUE) was cal-
per plot (GY) were calculated.
culated as the ratio of Pn to Tr, and intrinsic water use efficiency
(iWUE) as the ratio of Pn to Gs.
The gas exchange parameters, such as photosynthetic rate Phenotypic data analysis
(Pn) and other related traits (Gs), are sensitive to microclimate
fluctuations, such as vapour pressure deficit (VPD) or leaf tem- Analysis of variance (ANOVA) was performed for the pheno-
perature (Cowan, 1977; Buckley et al., 2002; Supplementary typic data using SAS software (Version 9.2; SAS Institute Inc.,
Data Fig. S1). The modified methodology for microclimatic Cary, NC, USA). Broad-sense heritability was estimated using
differences utilizes a statistical covariant model using the en- the formula:
vironmental variable as a quantitative co-regressor (Gu et al.,
2012). The photosynthetic trait value of the ith genetic line was h2 = σG2 /(σG2 + σGE
(2) 2
/e + σε2 /(re))
expressed in a statistical covariant model using the environ- where e and r are the number of environments and replica-
mental characteristics as a quantitative co-regressor: tions per environment, respectively. Mean squares were used
µijk = µ + Gi + Ej + (GE)ij + Bk + bxijk + eijk to estimate the variance components for genotypes (σG2 ), geno-
(1) type × environment interaction (σGE 2
) and residual error (σε2
where u is the general mean. Gi is the genetic effect of the ith ), respectively. Best linear unbiased estimations (BLUE) for
genotype, Eij is the treatment effect, which represents the four phenotypic data across the four environments were extracted
environments, (GE)ij is the genetype × treatment interaction, Bk by implementing the ANOVA function in the QTL IciMapping
is the block effect, b is the effect of the environmental variable, software (Version 4.1). Correlation analyses and t-tests were
xijk is the environmental variable during measurement and eijk is performed using SAS software.
the residual effect. This method allows the observed photosyn- The random forest model (randomForest package in R) was
thetic traits to be statistically adjusted to the same value (such used to assess trait contributions to GY and TGW. The relative
as the average) of the climatic variable (Tleaf and VPD) that effect of the parameter estimates for each predictor was com-
changed during the field measurements. Here, we used VPD pared with the impact of all parameter estimates in the model to
as the most crucial environmental component in this study evaluate the relative importance of traits. Self-organizing maps
(Supplementary Data Fig. S1). (SOMs, kohonen package in R) were then used to identify the
Plant height (PH, cm) was measured at anthesis as the dis- traits suitable for GY and evaluate the favourable traits of 166
tance from the top of the canopy to the soil; height was meas- elite cultivars. SOMs represent a non-linear multivariate stat-
ured for five plants in each plot and averaged. Leaf area index istical method helpful in visualizing multidimensional data.
(LAI) was recorded using a Sunscan Canopy Analysis System This method allows the transformation of high-dimensional
(Delta T Devices Ltd, Cambridge, UK) with 64 photodiodes. data to low-dimensional data, preserving the input data’s main
4 Li et al. — Variations in wheat traits and their genetic basis

characteristics. By visual comparison of the maps, cultivars light-saturated net photosynthetic rate, stomatal conductance,
with similar distributions were detected to identify the trait cor- intercellular CO2 concentration, transpiration rate, water use
relations. Thus, this method was used to visualize and explore efficiency, intrinsic water use efficiency, maximum quantum
data properties and separate the data set into clusters of similar yield of PSII photochemistry, leaf area index, grain filling
characteristics. rate, stay green trait, spike shape, leaf water content, spike
water content, stem water content, leaf dry weights, spike dry
weights, stem dry weights, total dry weights, spike number
DNA extraction and physical map construction per square metre, kernel number per spike, thousand-grain
weight and grain yield, but decreasing plant height, flag

Downloaded from https://academic.oup.com/aob/advance-article/doi/10.1093/aob/mcad003/6992592 by guest on 05 March 2023

DNA isolation and genotyping were performed as reported
elsewhere (Li et al., 2019a; Fu et al., 2020). A modified leaf angle, thermal time from sowing to flowering stage and
cetyltrimethyl ammonium bromide (CTAB) protocol was used thermal time from sowing to maturity stage, were defined as
to extract genomic DNA from young leaves. The DNA sam- favourable alleles. QTL allelic frequencies and effects were
ples were genotyped with the Illumina 90K wheat SNP array analysed based on the representative markers of each locus.
(containing 81 587 SNPs) and the Affymetrix 660K wheat SNP The average BLUE values for two genotypes at each locus
array (containing 630 517 SNPs) by CapitalBio Technology were used to compare the subgroup effects by a t-test. The
Co., Ltd (http://www.capitalbiotech.com/). Heterozygous association panel was grouped based on the favourable alleles
genotypes were considered as missing data. SNPs with minor for each trait. Further, the correlation between the average
allele frequency (MAF) <5 % and missing data >20 % were phenotypic values of each group (average of BLUE values of
excluded from further analysis. Linkage disequilibrium (LD) accessions within each group) and the number of favourable
among SNPs was analysed using the full matrix and sliding alleles was analysed with the ‘rcorr’ function of the ‘Hmisc’
window options implemented in Tassel software (Version 5.0) package in R (Version 3.6.2).
by 12 323 SNPs evenly distributed on 21 wheat chromosomes.
Population structure was analysed using 2000 evenly distrib-
RESULTS
uted polymorphic SNPs with the Structure software (Version
2.3.4). A neighbour-joining tree was constructed, and prin-
cipal compenents analysis (PCA) was performed with Tassel to Phenotypic variation, heritability and contribution to yield of
verify the population stratification (Liu et al., 2017). wheat traits
The 32 traits among 166 wheat accessions across the four en-
vironments showed substantial variations (Table 1), following
Genome-wide association study a normal distribution (Supplementary Data Fig. S3). The co-
The GWAS was performed with the mean values of two rep- efficient of variation (CV) of phenotypic data ranged from
licates in each environment, the BLUE values across four en- 3.74 to 55.10 %, averaging 23.17 % across four environments.
vironments for each trait, and the SNP markers from the wheat Large phenotypic trait diversity among the accessions is ideal
90K and 660K SNP arrays. The analysis was conducted using for conducting GWAS. ANOVA revealed significant effects
GAPIT (Genome Association and Prediction Integrated Tool; of genotype, environment and genotype × environment inter-
Version 3). Kinship was set as the random effects and prin- actions on all the traits. Some traits presented high broad-sense
cipal components (PCs) as fixed effects to control the popu- heritabilities (h2), such as TDWS (0.91), PH (0.94), GFR (0.84),
lation structure and familial relatedness. Manhattan plots and TTM (0.84), FLW (0.85), FLANG (0.84) and TGW (0.91),
quantile–quantile plots (Q-Q plots) were generated using the indicating that most of these traits were mainly controlled by
R package ‘CMplot’. Bonferroni–Holm correction for mul- genetic factors.
tiple testing (α = 0.05) was too conserved for the present Correlation analysis for the BLUE value of each trait in-
study’s traits. Therefore, markers with an adjusted −log10(P- dicated that TGW was significantly and positively correlated
value) ≥ 3.0 were considered to show significant marker–trait with net photosynthetic rate (Pn), stomatal conductance (Gs),
associations (MTAs) to maximize the chances of identifying all spike shape (SS), leaf water content (LWC), spike water con-
possible QTLs. This threshold was also used in some wheat as- tent (SPWC), leaf dry weight (LDWS), flag leaf width (FLW)
sociation studies on complex traits (Gao et al., 2015; Li et al., and SPAD, but negatively correlated with thermal time from
2018b; Fu et al., 2020). The most significant SNP markers were sowing to flowering stage (TTF), flag specific leaf area (FSLA)
chosen for each common locus as the representative markers. A and spike number (SN) (Supplementary Data Fig. S4). GY was
locus detected in at least two environments was declared stable. significantly and positively correlated with net photosynthetic
Loci or QTLs reported in previous studies were considered the rate (Pn), stomatal conductance (Gs), transpiration rate (Tr),
same as those in the present study if the tightly linked or signifi- spike shape (SS), spike water content (SPWC), flag leaf width
cantly associated markers were within less than one LD. (FLW), SPAD and TGW, but negatively correlated with plant
height (PH), flag leaf length (FLL), flag leaf angle (FLANG)
and spike number (SN).
Analyses of allele frequencies and effects of alleles
QTL allelic frequencies were analysed based on the SNP Contributions of wheat traits to GY and TGW
at each QTL peak. The alleles increasing flag leaf length,
flag leaf width, flag leaf area, flag leaf biomass, flag leaf The relative importance of traits in determining GY and
specific leaf area, chlorophyll content SPAD meter reading, TGW was applied with a random forest model with 500
 Li et al. — Variations in wheat traits and their genetic basis 5

Table 1. Natural variation, broad-sense heritability and ANOVA for 32 traits across 166 wheat cultivars grown under four environments.

Trait Descriptive analysis Analysis of variance

Mean SD CV h 2
Genotype (G) Environment (E) G × E interaction

LWC 0.615 0.123 19.921 0.513 0.009*** 4.831*** 0.013***

STWC 0.664 0.062 9.387 0.588 0.004*** 0.830*** 0.002***
SPWC 0.540 0.124 22.937 0.380 0.008*** 4.167*** 0.008***

Downloaded from https://academic.oup.com/aob/advance-article/doi/10.1093/aob/mcad003/6992592 by guest on 05 March 2023

LDWS 0.408 0.145 35.591 0.875 0.327*** 4.528*** 0.046***
STDWS 1.234 0.506 40.969 0.918 3.510*** 33.430*** 0.311***
SPDWS 0.668 0.290 43.418 0.850 0.617*** 30.245*** 0.107***
TDWS 2.300 0.870 37.824 0.914 8.828*** 114.376*** 0.827***
Pn 21.378 2.984 13.957 0.618 29.334*** 157.363*** 16.617***
Gs 0.406 0.123 30.196 0.567 0.0383*** 397.838*** 0.027***
Tr 4.377 1.167 26.655 0.483 1560.667*** 103027.400*** 1522.654***
Ci 277.696 28.952 10.426 0.561 2.948*** 196.401*** 2.048***
WUE 5.200 1.243 23.896 0.553 3.330*** 229.081*** 2.330***
iWUE 57.059 14.303 25.067 0.542 499.572*** 17204.040*** 372.040***
Fvʹ/Fmʹ 0.602 0.050 8.256 0.458 0.007*** 0.542*** 0.006***
PH 93.041 12.822 13.781 0.938 1096.407*** 3067.014*** 72.145***
LAI 4.527 0.581 12.831 0.634 1.191*** 7.332*** 0.590***
GFR 0.003 0.001 27.083 0.838 0.00*** 0.00*** 0.00***
SGT 0.401 0.221 55.104 0.655 0.153*** 4.325*** 0.077***
TTF 1748.705 67.611 3.866 0.825 15886.880*** 606800.900*** 3104.403***
TTM 2395.350 89.535 3.738 0.836 28217.510*** 1723054.000*** 5217.478***
SS 3.467 1.180 34.031 0.620 4.607*** 5.442*** 2.679***
FLL 19.985 4.534 22.689 0.797 47.135*** 5541.081*** 11.355***
FLW 1.700 0.239 14.047 0.852 0.256*** 6.063*** 0.043***
FLA 22.234 7.415 33.349 0.711 92.409*** 14503.800*** 34.508***
FLB 0.143 0.034 24.141 0.694 0.004*** 0.014*** 0.002***
FSLA 175.442 67.386 38.409 0.444 4404.124*** 1116339.75*** 5131.530***
FLANG 2.015 0.691 34.311 0.837 46.460*** 642.406*** 12.498***
SPAD 54.829 3.268 5.960 0.765 2.429*** 9.524*** 0.458***
SN 680.791 169.901 24.956 0.692 166.990*** 1915.948*** 15.897***
KN 40.327 5.516 13.678 0.505 71224.880*** 6338188.000*** 28664.100***
TGW 41.344 5.396 13.051 0.908 55.014*** 4846.434*** 48.658***
GY 3.396 0.612 18.035 0.868 1.375*** 78.803*** 0.191***

Mean, mean value; SD, standard deviation; CV, coefficient of variation; h2, broad-sense heritability. ***P < 0.001. LWC, leaf water content; SPWC, spike
water content; STWC, stem water content; LDWS, leaf dry weights; SPDWS, spike dry weights; STDWS, stem dry weights; TDWS, total dry weights; Pn, light-
saturated net photosynthetic rate; Gs, stomatal conductance; Tr, transpiration rate; Ci, intercellular CO2 concentration; WUE, water use efficiency; iWUE, intrinsic
water use efficiency; Fvʹ/Fmʹ, maximum quantum yield of PSII photochemistry; PH, plant height; LAI, leaf area index; GFR, grain filling rate; SGT, stay green
trait; TTF, thermal time from sowing to flowering stage; TTM, thermal time from sowing to maturity stage; SS, spike shape; FLL, flag leaf length; FLW, flag leaf
width; FLA, flag leaf area; FLB, flag leaf biomass; FSLA, flag leaf specific leaf area; FLANG, flag leaf angle; SPAD, chlorophyll content SPAD meter reading;
SN, spike number per square metre; KN, kernel number per spike; TGW, thousand-grain weight; GY, grain yield.

classification trees. Model accuracy was 60.5 and 38.7 % for flag leaf width were the most significant predictors of thousand-
GY and TGW, respectively (Fig. 1). Fourteen morphological grain weight, accounting for 36.4 % of the total variation. Plant
and physiological traits conferring GY and TGW were selected height, spike shape, photosynthetic rate, SPAD, flag leaf width
based on the random forest results. Spike shape, plant height, and leaf water content contributed significantly to GY and
flag leaf angle and thousand-grain weight were the most signifi- TGW. Thus, the random forest model suggested that promoting
cant predictors of grain yield, accounting for 45.9 % of the total photosynthesis traits, spike shape, SPAD and leaf water content
variation. Spike number, spike shape, leaf water content and and reducing plant height could improve GY and TGW.
6 Li et al. — Variations in wheat traits and their genetic basis

A B
SS SN
16.7 % P 10.5 %
Ci 13 H SG
T
10
SS
% .5 % .2
3.8 % 5.0 %

FL .6 %
4.0 C

LW 0 %
5.5 TM
AN
LW
%

8.

C
T
G
TGW
%
LAI

7.1 %

FLW
7.7 %
%
KN

Downloaded from https://academic.oup.com/aob/advance-article/doi/10.1093/aob/mcad003/6992592 by guest on 05 March 2023

4.1

5.7
GY TGW
(60.5%)

7.0 %
(38.7%)

7.6 %
SPAD
Pn
%
C

%
SPW
2

PH
5.8
5 .

6.5 L

7.5 LA
FS
FL
% %
%

%
5 E
U5. 5
6.
iW 6.4 Pn 6.
5.6
%
Tr
% 6.5
% ST 9 %
5.9 % 6.7 % WC
FLW TFF
SPAD LDWC

Fig. 1. Random forest analysis predictors of grain yield (A) and TGW (B) in 209 wheat cultivars. The relative importance of each predictor is ranked in order and
presented for GY (A) and TGW (B). Arrow colours indicate the direction of correlation (blue, positive; red, negative) for the continuous variables. The values rep-
resent the contribution of the trait grain yield or TGW.LWC, leaf water content; SPWC, spike water content; STWC, stem water content; LDWS, leaf dry weights;
Pn, light-saturated net photosynthetic rate; Tr, transpiration rate; Ci, intercellular CO2 concentration; iWUE, intrinsic water use efficiency; PH, plant height; LAI,
leaf area index; SGT, stay green trait; TTF, thermal time from sowing to flowering stage; TTM, thermal time from sowing to maturity stage; SS, spike shape;
FLL, flag leaf length; FLW, flag leaf width; FSLA, flag leaf specific leaf area; FLANG, flag leaf angle; SPAD, chlorophyll content SPAD meter reading; SN, spike
number per square meter; TGW, thousand-grain weight; GY, grain yield.

Genomic variation, LD and population structure water use efficiency (iWUE), maximum quantum yield of PSII
photochemistry (Fvʹ/Fmʹ), plant height (PH), leaf area index
A total of 373 106 high-quality SNPs from the two SNP
(LAI), grain filling rate (GFR), stay green trait (SGT), thermal
arrays were used for GWAS. Approximately 39.8, 49.3 and
time from sowing to flowering stage (TTF), thermal time from
10.8 % of the markers were in sub-genomes A, B and D, respect-
sowing to maturity stage (TTM), spike shape (SS), leaf water
ively. The number of SNPs per chromosome ranged from 2374
content (LWC), spike water content (SPWC), stem water con-
on 4D to 46 708 on 3B. These markers covered a total phys-
tent (STWC), leaf dry weights (LDWS), spike dry weights
ical distance of 14 061.15 Mb, with a genome-wide average of
(SPDWS), stem dry weights (STDWS), total dry weights
26 SNPs per Mb. The principal component, neighbour-joining
(TDWS), spike number per square metre (SN), kernel number
phylogenetic tree and kinship analyses are presented in Fig.
per spike (KN), thousand-grain weight (TGW) and grain yield
2A–C, namely subgroup I (62 cultivars), subgroup II (54 cul-
(GY), respectively. Manhattan plots for 32 traits (Fig. 3; Fig.
tivars) and subgroup II (50 cultivars). LD decay varied among
S5) using MLM in the four environments and BLUE values
the sub-genomes and across the chromosomes. LD decay in the
showed the location of SNPs and the associated SNPs, and Q-
B sub-genome dropped quickly. The average genome-wide ex-
Q plots for the traits are shown in Fig. S6. The loci of 32 traits
tent of LD was 8 Mb, with an average of 6, 4 and 11 for the A,
detected in at least two out of the five environments (including
B and D sub-genomes, respectively.
BLUE) are summarized in Table S2. To conclude, the GWAS
results are reliable and efficient in detecting the loci for GY and
Genome-wide association studies related traits.
We performed a GWAS for photosynthetic, morphological
and agronomic traits using the mixed linear model (MLM)
Pleiotropic loci
method. These significant and stable SNPs were located on
21 chromosomes and explained 4.93–17.77 % of the pheno- We further investigated the pleiotropic loci for these traits
typic variance among environments (Fig. 3; Supplementary (Fig. 4). A total of 47 pleiotropic loci were associated with
Data Fig. S5). A total of 1236 stable loci were detected for three or more traits and GY/TGW on chromosomes 1A (7 loci),
the 32 traits. In SNP-GWAS, 59, 42, 44, 54, 38, 42, 32, 19, 1B (2 loci), 1D (2 loci), 2A (3 loci), 2B (4 loci), 3A (2 loci), 3B
13, 8, 11, 9, 6, 7, 115, 34, 58, 24, 53, 34, 39, 28, 23, 6, 68, (4 loci), 3D, 4A, 4B (4 loci), 4D, 5A (3 loci), 5B (4 loci), 6B (3
67, 73, 70, 35, 25, 55 and 45 loci were detected for flag leaf loci), 6D, 7A (2 loci), 7B and 7D (2 loci) based on the common
length (FLL), flag length width (FLW), flag leaf area (FLA), loci detected by GWAS (Supplementary Data Table S3). The
flag leaf biomass (FLB), flag leaf specific area (FSLA), flag interval 703.20–708.77 Mb on chromosome 5A was associated
leaf angle (FLANG), SPAD, net photosynthetic rate (Pn), sto- with plant height (PH), flag leaf angle (FLANG), GY, TGW,
matal conductance (Gs), intercellular CO2 concentration (Ci), stem dry weights (STDWS), flag leaf biomass (FLB), leaf dry
transpiration rate (Tr), water use efficiency (WUE), intrinsic weights (LDWS), stem dry weights (STDWS) and total dry
 Li et al. — Variations in wheat traits and their genetic basis 7
Color key
and histogram
A

2000
1500
Count
1000
200

500
100

400
PC3
0

0
300 –3 –2 –1 0 1 2 3
–300 –200 –100

200
Value
100

PC2

Downloaded from https://academic.oup.com/aob/advance-article/doi/10.1093/aob/mcad003/6992592 by guest on 05 March 2023

0
–100
–200
–300
–400 –300 –200 –100 0 100 200 300
PC1

B C 0 Mb 92 Mb 184 Mb 276 Mb 368 Mb 460 Mb 552 Mb 644 Mb 736 Mb 828 Mb

1.0
1A
0.9 1B
Whole genome 1D
0.8 A genome
Pairwise LD (I 2)

2A
0.7 B genome 2B
2D
0.6 D genome 3A
3B
0.5 3D
4A 0
0.4 4B 1
4D 41
0.3 5A 81
5B 121
0.2 5D 161
6A 201
0.1 6B 241
281
0 6D
7A 321
0 20 40 60 80 100 120 140 160 180 200 7B 361
7D >361
Physical distance (Mb)

Fig. 2. (A) Population structure of 166 wheat accessions revealed by principal component (top left), neighbour-joining tree (bottom left) and kinship (right) ana-
lyses. (B) Linkage disequilibrium (LD) decay across the whole genome and A, B and D sub-genomes and (C) distribution of SNPs with minor allele frequency
>0.05 and missing data <80 %.

weights (TDWS), sharing the same region with GY. Twenty- loci for the thermal time from sowing to flowering stage (TTF)
three pleiotropic loci were associated with GY, among which 12 or flag leaf length (FLL), nine for stem dry weights (STDWS),
were related to plant height (PH), seven to flag leaf length (FLL) and four for spike dry weights (SPDWS) or SPAD are crucial in
and six to grain filling rate (GFR). Seven spike shape (SS) loci on determining GY or TGW.
chromosomes 1A (GENE_1785_626), 1B (AX_109097017), 2B
(AX_109602295), 3A (AX_108765521), 4B (AX_111005064) Distributions and the cumulative effect of favourable alleles
and 5A (AX_108899874, RAC875_c18335_443) were located
in pleiotropic loci. Four net photosynthetic rate (Pn) loci on Favourable allele frequencies of the identified QTLs associ-
chromosomes 1A (AX_109095224), 2A (AX_111046029), 3A ated with 32 traits ranged from 0.05 to 0.94 (average 0.48). The
(AX_108765521) and 3D (AX_111656541) were also located in frequencies of the favourable allele of QTLs for net photosyn-
pleiotropic loci. Our observations indicated that 11 pleiotropic thetic rate (Pn), stomatal conductance (Gs), intercellular CO2
8 Li et al. — Variations in wheat traits and their genetic basis

PH.1 PH.2 PH.3 PH.4 PH.BLUE FLW.1 FLW.2 FLW.3 FLW.4 FLW.BLUE
8 8

6 6
–log10 (P)

–log10 (P)
4 4

2 2

0 0
Chr 1D 2B 3A 3D 4B 5A 5D 6B 7A 7D

Downloaded from https://academic.oup.com/aob/advance-article/doi/10.1093/aob/mcad003/6992592 by guest on 05 March 2023

Chr 1D 2B 3A 3D 4B 5A 5D 6B 7A 7D

SS.1 SS.2 SS.3 SS.4 SS.BLUE LWC.1 LWC.2 LWC.3 LWC.4 LWC.BLUE
8 8

6 6
–log10 (P)

–log10 (P)
4 4

2 2

0 0
Chr 1D 2B 3A 3D 4B 5A 5D 6B 7A 7D Chr 1D 2B 3A 3D 4B 5A 5D 6B 7A 7D

Pn.1 Pn.2 Pn.3 Pn.4 Pn.BLUE TGW.1 TGW.2 TGW.3 TGW.4 TGW.BLUE
8 8

6 6
–log10 (P)

–log10 (P)

4 4

2 2

0 0
Chr 1D 2B 3A 3D 4B 5A 5D 6B 7A 7D Chr 1D 2B 3A 3D 4B 5A 5D 6B 7A 7D

SPAD.1 SPAD.2 SPAD.3 SPAD.4 SPAD.BLUE GY.1 GY.2 GY.3 GY.4 TGY.BLUE
8 8

6 6
–log10 (P)

–log10 (P)

4 4

2 2

0 0
Chr 1D 2B 3A 3D 4B 5A 5D 6B 7A 7D Chr 1D 2B 3A 3D 4B 5A 5D 6B 7A 7D

Fig. 3. Manhattan plots for a genome-wide association study of the traits for GY and TGW in 166 wheat accessions under multiple environments.PH, plant height;
SS, spike shape; Pn, light-saturated net photosynthetic rate; SPAD, chlorophyll content SPAD meter reading; FLW, flag leaf width; LWC, leaf water content; TGW,
thousand-grain weight; GY, grain yield.

concentration (Ci), water use efficiency (WUE), flag length number per square metre (SN), kernel number per spike (KN)
width (FLW), flag leaf angle (FLANG), plant height (PH), leaf and GW (0.18–0.49). These results indicate that most of the
area index (LAI), thermal time from sowing to flowering stage accessions possessed alleles that increased net photosynthetic
(TTF), spike shape (SS), leaf water content (LWC), spike water rate (Pn), stomatal conductance (Gs), intercellular CO2 con-
content (SPWC), stem water content (STWC), TGW and GY centration (Ci), water use efficiency (WUE), flag length width
were higher (average from 0.53 to 0.92) than those for transpir- (FLW), leaf area index (LAI), spike shape (SS), leaf water con-
ation rate (Tr), intrinsic water use efficiency (iWUE), maximum tent (LWC), spike water content (SPWC), stem water content
quantum yield of PSII photochemistry (Fvʹ/Fmʹ), flag leaf (STWC), TGW and GY, but reduced flag leaf angle (FLANG),
length (FLL), flag leaf area (FLA), flag leaf biomass (FLB), flag plant height (PH) and thermal time from sowing to flowering
leaf specific area (FSLA), SPAD, grain filling rate (GFR), stay stage (TTF) (Supplementary Data Fig. S7).
green trait (SGT), thermal time from sowing to maturity stage The number of increasing-effect alleles in each accession
(TTM), leaf dry weights (LDWS), spike dry weights (SPDWS), was simulated further to investigate the effects of combined al-
stem dry weights (STDWS), total dry weights (TDWS), spike leles on those traits. Linear regression analysis was performed
 Li et al. — Variations in wheat traits and their genetic basis 9

A B

STDWS
iWUE

FLW

S
DW
Tr

SP

WS
TTM
Ci

TTF

TD
TR
G
s

Fv TGW

B
’/F TDWS

FL
m’ STDWC
WS STDWS
TTM LD SS
SPWC
SPDWS
SPAD
SGT SN SN
SGT
Pn

Traits
PH
Pn FSLA

Downloaded from https://academic.oup.com/aob/advance-article/doi/10.1093/aob/mcad003/6992592 by guest on 05 March 2023

LWC
LDWS
LAI
GY GFR KN
iWUE
GY
GS
G LWC GFF
FLAN Fv'/Fm'
FSLA
F KN FLW
TT FLL
FLB
FL FLANG
A
SS

FLA
Ci
SP
L
FL

AD

Chr 1A 1B 1D 2A 2B 2D 3A 3B 3D 4A 4B 4D 5A 5B 5D 6A 6B 6D 7A 7B 7D
LA
W

STWC
TG

I
C

PH
SPW

Fig. 4. The pleiotropic loci for traits of both traits controlled by the same SNP (A) and the position of pleiotropic loci (B).LWC, leaf water content; SPWC, spike
water content; STWC, stem water content; LDWS, leaf dry weights; SPDWS, spike dry weights; STDWS, stem dry weights; TDWS, total dry weights; Pn, light-
saturated net photosynthetic rate; Gs, stomatal conductance; Tr, transpiration rate; Ci, intercellular CO2 concentration; WUE, water use efficiency; iWUE, intrinsic
water use efficiency; Fvʹ/Fmʹ, maximum quantum yield of PSII photochemistry; PH, plant height; LAI, leaf area index; GFR, grain filling rate; SGT, stay green
trait; TTF, thermal time from sowing to flowering stage; TTM, thermal time from sowing to maturity stage; SS, spike shape; FLL, flag leaf length; FLW, flag leaf
width; FLA, flag leaf area; FLB, flag leaf biomass; FSLA, flag leaf specific leaf area; FLANG, flag leaf angle; SPAD, chlorophyll content SPAD meter reading;
SN, spike number per square metre; KN, kernel number per spike; TGW, thousand-grain weight; GY, grain yield.

using the BLUE values to further investigate the relation- not net photosynthetic rate, leaf area index (LAI), stay green
ships between trait values and the number of trait-increasing trait (SGT), thermal time from sowing to maturity stage (TTM),
alleles. Favourable alleles for 32 traits at each locus showed leaf water content (LWC) or biomass-related traits (Fig. 6;
significant and positive effects on the phenotypic trait values Fig. S10).
(Fig. 5; Supplementary Data Fig. S8). Significant correlations
(P < 0.01) were observed between trait values and the number
of trait-increasing alleles. For many traits, the coefficients of Ideotypes for wheat breeding
determination (R2) between the trait’s values and the number Data for the 32 wheat traits were inputted into the SOM system
of favourable alleles in each accession were >0.62. This result to classify the traits for wheat ideotype breeding based on the 166
suggests that QTLs with additive effects controlled many traits. wheat lines. The number of nodes was set as 35, the number of
Stem water content (STWC) and thermal time from sowing rows as eight and the number of columns as five. The SOM results
to flowering stage (TTF) had an R2 < 0.54, indicating that the are presented in Supplementary Data Table S4 and Fig. 7. Figure
environment affected the expression of QTLs more for these 7A shows five different clusters with similar composition of trait
traits. values. Figure 7B shows 32 component maps representing the
component values by visually identifying the correlation among
traits for the 35 nodes. Clusters 1, 2, 3, 4 and 5 consisted of 13,
Genetic progress
60, 20, 15 and 58 cultivars, respectively. Cultivars in Cluster 1
The genetic progress of 32 traits was investigated to ex- were characterized by high GY, TGW, net photosynthetic rate
plore the role of yield-associated loci in improving GY (Fig. (Pn), stomatal conductance (Gs), transpiration rate (Tr), intercel-
6; Supplementary Data Fig. S9). The cultivars released after lular CO2 concentration (Ci), maximum quantum yield of PSII
2010 (13) demonstrated an increase in Pn (11.78 %), Gs photochemistry (Fvʹ/Fmʹ), grain filling rate (GFR), stay green
(36.62 %), Tr (29.89 %), Ci (7.65 %), Fvʹ/Fmʹ (4.83 %), LAI trait (SGT), thermal time from sowing to maturity stage (TTM),
(6.51 %), SGT (10.43 %), LWC (11.87 %), SPWC (10.05 %), flag length width (FLW), SPAD, spike shape (SS), leaf water
LDWS (26.21 %), SPDWS (27.71 %), 5.69 % (TDWS), FLW content (LWC), spike water content (SPWC), stem water content
20.28 %), FLB (2.89 %), SPAD (5.97 %), KN (4.92 %) and (STWC), leaf dry weights (LDWS), spike dry weights (SPDWS),
TGW (15.50 %) relative to the cultivars released before 1980 stem dry weights (STDWS), total dry weights (TDWS), flag
(nine), accompanied by a 47.1 % increase in GY. However, leaf biomass (FLB) and kernel number per spike (KN) values;
the cultivars released after 2010 (13) showed a decrease in low water use efficiency (WUE), intrinsic water use efficiency
WUE (13.48 %), iWUE (20.14 %), PH (29.65 %), GFR (iWUE), plant height (PH), leaf area index (LAI), thermal time
(14.21 %), STDWS (6.80 %), FLL (13.06 %), FLA (11.47 %), from sowing to flowering stage (TTF), flag leaf specific area
Pn (11.78 %), FSLA (15.50 %), FLANG (43.04 %) and SN (FSLA) and spike number per square metre (SN) values; and
(15.99 %) relative to the cultivars released before 1980 (nine). median flag leaf length (FLL), flag leaf angle (FLANG) and flag
In addition, the year of release (five groups) had a significant leaf area (FLA) values. These traits were relatively important in
effect on the number of favourable alleles for phenotypic traits, determining grain yield and thousand weight. The selected top-
such as plant height (PH), spike shape (SS), flag length width performing cultivars (Sunong 6, Zheng 9023, Zhoumai 30, Zhou
(FLW), spike number per square metre (SN), TGW and GY, but 8425B, Zimai 12, Lumai 23, Linmai 2, Lankao 906, Linmai
10 Li et al. — Variations in wheat traits and their genetic basis

180 2.5
A B
160
2
140

FLW (cm)
1.5
PH (cm)

120

100 1
80

Downloaded from https://academic.oup.com/aob/advance-article/doi/10.1093/aob/mcad003/6992592 by guest on 05 March 2023

y = –0.53x + 137.35 0.5 y = 0.018x + 1.29
60 R 2 = 0.59 R 2 = 0.62
40 0
0 8 16 24 32 40 48 56 64 72 80 88 96 3 11 19 27 35
6 80
C D
5
70
4
60

LWC (%)
3
SS

50
2

1 Y = 0.065x + 1.94 40 y = 0.67x + 51.48

R 2 = 0.78 R 2 = 0.85
0 30
0 8 16 24 32 40 –1 6 13 20 27
30 60
E F
25 50
Pn (µmol m–2 s–1)

20 40
TGW (g)

15 30

10 20

5 y = 0.31x + 17.64 10 y = 0.40x + 28.62

R 2 = 0.62 R 2 = 0.73
0 0
–1 4 9 14 19 0 5 10 15 20 25 30 35 40 45 50 55
65 5
G H
60 4
GY (kg plot–1)

55 3
SPAD

50 2

45 y = 0.24x + 50.70 1 y = 0.04x + 2.20

R 2 = 0.69 R 2 = 0.63
40 0
0 8 16 24 32 3 11 19 27 35
Number of increasing-effect alleles Number of increasing-effect alleles

Fig. 5. Effect and distribution of favourable alleles of trait-associated markers contributing significantly to GY and TGW.PH, plant height; SS, spike shape; Pn,
light-saturated net photosynthetic rate; SPAD, chlorophyll content SPAD meter reading; FLW, flag leaf width; LWC, leaf water content; TGW, thousand-grain
weight; GY, grain yield.

4, Jining 16, Wanmai 33 and Lankao 2) were from the Yellow to different environments. Our findings suggest the use of cul-
and Huai Valleys Winter Wheat Zone where the trials were con- tivars in Cluster 1 as potential parents to develop high-yielding
ducted, suggesting that these wheat lines have some adaptability cultivars with desirable traits.
 Li et al. — Variations in wheat traits and their genetic basis 11

A B I J
a b a a a a b a a a a
200 2.5 ab a 50

FLW (number)
a

PH (number)
b 100 40

FLW (cm)
150 c
PH (cm)

b b bc 2.0 30
c 50
100 20
1.5
0 10
50
1.0 0
C D b a a a a
K L
b b a 80 50 a a a

LWC (number)
6 b

SS (number)
c 40 c 30

LWC (%)
d 70 30 20

Downloaded from https://academic.oup.com/aob/advance-article/doi/10.1093/aob/mcad003/6992592 by guest on 05 March 2023

4
SS

60 20 10
2 50 10
0 0 Class
40 –10 1947–1979
E F M N
b b b b 1980–1989
ab b ab a
Pn (µmol m–2 s–1)

32 c a a a a

TGW (number)
60 1990–1999

Pn (number)
28 20 60 b
TGW (g)

a 50 2000–2009
24 40
40 10 2010–2016
20 20
16 30
0
20 0
H O P
G bc bc ab a 5 b a a a
GY (kg per plot)

b 40

SPAD (number)
c c a

GY (number)
e d c
60 4 30 40
SPAD

3 20
50 20
2 10
0 0
40 1

Fig. 6. Genetic progress of the traits contributing significantly to GY and TGW. Violin plots A–H, phenotypic changes in PH, FLW, SS, LWC, Pn, TGW, SPAD
and GY, respectively; I–P, changes in number of increasing-effect alleles for PH, FLW, SS, LWC, Pn, TGW, SPAD and GY, respectively. PH, plant height; SS,
spike shape; Pn, light-saturated net photosynthetic rate; SPAD, chlorophyll content SPAD meter reading; FLW, flag leaf width; LWC, leaf water content; TGW,
thousand-grain weight; GY, grain yield. Different lowercase letters indicate significantly different at 0.05 level.

A B
Pn Gs Tr Ci WUE iWUE Fv'/Fm' PH
1 1 1 1.5 1 1 1 2.5
0.5 0.5 0.5 1 0.5 0.5 2
0.5 0.5 1.5
0 0 0
Codes plot –0.5
0 0 0
–0.5 0 1
–0.5 –0.5 –0.5 –0.5 0.5
–1 –1 –1 –1 –0.5 0
–1.5 –1 –1.5 –1 –1.5 –0.5
Cluster 4
LAI GFR SGT TTF TTM SS LWC SPWC
1 1.5 1 1.5 1 1 1 1
Cluster 5 0.5 0.5
0.5 1 1 0.5 0.5
0 0.5 0 0
0.5 0.5 0 0
–0.5 0 –0.5 –0.5
–1 0 0 –0.5 –1 –1 –0.5
–1.5 –0.5 –0.5 –0.5 –1 –1.5 –1.5 –1

STWC LDWS SPDWS STDWS TDWS FLL FLW FLA

1 2 2.5 3 1.5 1.5 1.5
0.5 2 2 2 1 1 1
Cluster 3 0 1 1.5 0.5 0.5 0.5
1 1 1
–0.5 0 0.5 0 0 0
–1 0 0 0 –0.5 –0.5 –0.5
Cluster 2 –1.5 –1 –0.5 –1 –1 –1

FLB FSLA FLANG SPAD KN SN TGW GY

Cluster 1 2 1.5 1 0.5 1.5 1 0.5 1
1.5 1 0.5 1 0 0.5
1 0.5 0 0.5 0 0
0.5 0 0 –0.5 –0.5
0 –0.5 –0.5 0 –1 –1
–0.5 –1 –0.5 –0.5 –1
–1.5
–1 –1.5 –1 –1 –1 –2 –1.5 –2

Fig. 7. Clustering of 166 wheat cultivars using a self-organizing map (SOM) algorithm (A) and the matrices of components (B).LWC, leaf water content; SPWC,
spike water content; STWC, stem water content; LDWS, leaf dry weights; SPDWS, spike dry weights; STDWS, stem dry weights; TDWS, total dry weights; Pn,
light-saturated net photosynthetic rate; Gs, stomatal conductance; Tr, transpiration rate; Ci, intercellular CO2 concentration; WUE, water use efficiency; iWUE,
intrinsic water use efficiency; Fvʹ/Fmʹ, maximum quantum yield of PSII photochemistry; PH, plant height; LAI, leaf area index; GFR, grain filling rate; SGT, stay
green trait; TTF, thermal time from sowing to flowering stage; TTM, thermal time from sowing to maturity stage; SS, spike shape; FLL, flag leaf length; FLW,
flag leaf width; FLA, flag leaf area; FLB, flag leaf biomass; FSLA, flag leaf specific leaf area; FLANG, flag leaf angle; SPAD, Chlorophyll content SPAD meter
reading; SN, spike number per square meter; KN, kernel number per spike; TGW, thousand-grain weight; GY, grain yield.

DISCUSSION Complex morphological and physiological traits at the tissue or

whole-plant level led to high RUE; therefore, a comprehensive
Natural variation of all traits evaluation based on multiple traits can help analyse yield for-
mation and screen germplasm for breeding (Ding et al., 2018;
MAS has several advantages over conventional breeding, con- Li et al., 2019a). Moreover, there is still much to learn from the
strained mainly by time and resources; however, the desirable natural variation in photosynthetic capacity and performance,
traits originate mostly by chance without fully understanding which demands further exploration of the underlying physio-
the underlying physiological mechanisms (Fischer et al., 1998). logical and genetic mechanisms among species and cultivars
12 Li et al. — Variations in wheat traits and their genetic basis

(Flood et al., 2011). A previous study indicated underutilized Previous studies that assessed the variations in wheat traits
photosynthetic capacity among existing wheat cultivars using measured limited germplasm samples (Xue et al., 2002; Sadras
64 elite wheat cultivars (Driever et al., 2014). Although intro- et al., 2012) and used different experimental approaches
gressions to improve photosynthesis have been performed since (Chytyk et al., 2011). Only a few researchers have reported
the 1960s, the process remains quite limited (Ojima, 1974). QTLs associated with photosynthetic characteristics (Teng et
The limited availability of critical targets and well-defined mo- al., 2004; Adachi et al., 2011), probably because the measure-
lecular markers associated with traits have constrained breeding ment of gas exchange parameters is laborious. In addition, the
programmes (Flood et al., 2011). Our study is the largest inves- environment, especially the microclimate during growth and
tigation of the phenological, physiological, plant architectural measurement, inevitably fluctuates under natural conditions

Downloaded from https://academic.oup.com/aob/advance-article/doi/10.1093/aob/mcad003/6992592 by guest on 05 March 2023

and yield-related traits of wheat to date. We identified 14 crit- and affects the phenotypic values (Flood et al., 2011; Gu et al.,
ical traits with the random forest model as predictors of grain 2012). Therefore, in this study, we used a statistical approach
yield variations, and therefore as potential traits for breeding via VPD to correct for microclimate variations. Based on this,
consideration. we identified a significant positive correlation between photo-
Intraspecific crop diversity could be a valuable biological re- synthesis rate and stomatal conductance, indicating that photo-
source for better understanding and maintaining crop resilience synthesis rate and stomatal conductance were significantly
to extreme weather (Heider et al., 2021). We aimed to uncover correlated with GY and TGW. In a panel of 215 elite rice culti-
the natural variation among 166 wheat cultivars to identify vars, crop biomass accumulation was positively correlated with
physiological parameters and morphological traits signifi- photosynthetic traits (Qu et al., 2017). However, no correlation
cantly correlated with biomass accumulation and yield, laying was found between photosynthetic capacity and GY in 64 elite
a foundational framework for gene discovery. The loci for plant wheat cultivars (Driever et al., 2014).
height were associated with many traits in this study (Fig. 4). Biomass accumulation, an integrative measure of carbon fix-
The QTLs for height are coincident with those for grain yield ation by above-ground structures, can indicate canopy photo-
(Foulkes et al., 2011). Reduced plant height during the Green synthesis (Qu et al., 2017). A previous study indicated that
Revolution may have been accompanied by screening for traits improved water use efficiency might be related to improved
favouring yield accumulation. Our data confirmed substantial radiation interception efficiency, photosynthetic efficiency and
variations among the different traits under four environmental a longer green-canopy duration (stay-green traits), all of which
conditions, indicating genetic diversity that can be potentially were found to be improved in modern cultivars (Voss-Fels et
exploited for wheat yield improvement. Besides, the high her- al., 2019). Our results indicated Pn was positively correlated
itability (h2) of many traits observed in this study suggests their with LDWS and SPDWS, and TGW was positively correlated
potential in wheat breeding, which requires identification of the with LDW, SPDWS, STDWS and TDWS. Meanwhile, chloro-
associated candidate QTLs for MAS. Therefore, conducting a phyll fluorescence parameters directly or indirectly reflect plant
GWAS for high RUE using an SNP array will enrich the geno- photosynthesis and are important for investigating the func-
type–phenotype map and identify valuable gene resources tional mechanism of PSII (Baker, 2008; Liang et al., 2009; Yin
for molecular breeding to enhance productivity in the face of et al., 2010). There is a positive correlation between relative
increasing food demand (Li et al., 2020). chlorophyll fluorescence in the leaf and GY (Von Korff et al.,
2008). However, the current study revealed a positive correl-
ation between the chlorophyll fluorescence parameter Fvʹ/Fmʹ
Role of photosynthesis in yield gains and LWC, despite the experiments being conducted under well-
Wheat yield potential has previously increased mainly due watered conditions. The changes in photosynthetic characteris-
to improvements in harvest index rather than increased bio- tics associated with breeding progress are valuable benchmarks
mass. Although further large increases in HI are unlikely, for further improvement (Sadras et al., 2012). Furthermore,
there is an opportunity to increase productive biomass and har- exploiting existing natural variation in photosynthetic capacity
vestable grain (Reynolds et al., 2011; Li et al., 2022a). Even and biomass accumulation laid a foundation for increasing
small increases in net photosynthesis rate could result in large yield. Our results indicated GY increased linearly from 1957
increases in biomass and, thus, yield. Therefore, increasing to 2016, but had no significant correlation with Pn or biomass,
photosynthetic capacity and efficiency are bottlenecks to indicating photosynthesis rate could be further exploited.
improving crop biomass production and yield potential (Zhu
et al., 2010; Qu et al., 2017). Improvements in features related Wheat ideotypes
to canopy architecture, such as semidwarf architecture, more
erect leaves and larger LAI, have played a significant role in A feasible method to improve yield potential is the devel-
traditional crop breeding; however, no studies have addressed opment of new cultivars (Sakuma & Schnurbusch, 2020).
the improvement in leaf photosynthetic capacity per se (Peng However, the genetic basis of grain yield is largely unclear, and
et al., 1999; Hedden, 2003). Our random forest models dem- the application of MAS for grain yield in current breeding re-
onstrated that net photosynthetic rate and SPAD contributed mains limited. The breeding of crop ideotypes was first pro-
greatly to GY and TGW (Fig. 1). The SPAD measurement posed by Donald (1968). Understanding the basic physiology of
can be utilized as an appropriate tool for photosynthetic gen- the traits associated with yield potential can guide hybridization
etic analysis (Takai et al., 2010). Therefore, complex physio- and selection strategies for these relatively heritable constituent
logical traits related to photosynthetic efficiency must be parameters (Reynolds et al., 2011). Yield gains are presumably
incorporated as additional criteria to accelerate the genetic a function of improved coordination among physiological agro-
gains related to yield. nomic traits (He & Bonjean, 2010; Valluru et al., 2017; Abbai
 Li et al. — Variations in wheat traits and their genetic basis 13

et al., 2020). The results indicated that these traits contributed Genetic correlation between grain yield and yield-related traits
to yield formation to some extent. Studies have suggested that
Combined phenomics and genomic methods are necessary
the best plant architecture consists of more erect leaves in the
to evaluate the progress of breeding strategies (Rizzo et al.,
upper canopy, and more horizontal leaves in the middle and
2022; Welcker et al., 2022). Studies have reported GY-related
lower canopy (Parry et al., 2011; Ort et al., 2015). Consistent
QTLs on all 21 wheat chromosomes (Kumar et al., 2007; Wang
with previous studies, our findings demonstrated a positive
et al., 2011a; Azadi et al., 2015; Sukumaran et al., 2015).
correlation between vertical flag leaf angle and GY. However,
Our findings identified 45 stable loci for GY. The 1A locus
SOM results showed that the ideal plant architectural flag leaf
(AX_110507437) for GY in our present study is at a similar pos-
angle was not the lowest when multiple traits were considered,

Downloaded from https://academic.oup.com/aob/advance-article/doi/10.1093/aob/mcad003/6992592 by guest on 05 March 2023

ition to gwm357 on the consensus linkage map (Maccaferri et
indicating specific feedback among characteristics (Reynolds
al., 2015), indicating that these two loci are probably the same.
et al., 2015; Li et al., 2020). It is noteworthy that Donald’s
Six loci (AX_110387060, AX_110418502, AX_94546135,
ideotypes and our study focused on above-ground parts, while
AX_111210290, AX_111492146 and AX_95257733) for GY
roots are important and main organs that absorb water and min-
(Li et al., 2018a) were consistent with those identified in this
erals and detect soil stress signals (Paez-Garcia et al., 2015).
study on chromosomes 1A, 1A, 2A, 2B, 3A and 3D, respect-
Their traits determine absorption capacity and stress response
ively. Maccaferri et al (2015) also detected loci related to KN on
(Fang et al., 2017; Li et al., 2019b). Therefore, root traits and
chromosomes 1A, 1B, 2A, 2B, 2D, 3A, 4B, 5A, 5B, 6B, 7A and
the trade-offs strategies between above- and below-ground
7B; the loci on 1A, 1B, 2A, 2D, 3A and 5A are likely to be the
traits need to be further studied in the future.
same as AX_111579941, IWB52449, IWB58750, IWB57054,
Another way to increase wheat yield productivity is to opti-
AX_108992368 and IWB8258 in our study since they are located
mize the source–sink ratio (Bustos et al., 2013). In the current
in the same positions. The 1A QTL has a position similar to an-
study, the ideal type confirmed a high correlation between the
other KN locus identified in our study (AX_108992368) (Azadi
sink (KN, TGW and GY) and the source (Pn, SPAD and Fvʹ/
et al., 2015; Gao et al., 2015). The locus AX_111579941 on 1A
Fmʹ). Increasing grain filling duration for more light intercep-
is about one LD from a QTL (Wang et al., 2011a). In addition,
tion and utilization will also improve yield potential (Brooks
loci related to TGW such as AX_111102999, AX_109862024,
et al., 2001; Semenov & Stratonovitch, 2013). For wheat, flag
GENE_1785_118, AX_111548406, AX_110958315,
leaves account for 45–58 % of the photosynthetic perform-
AX_108769612, AX_94761192 and AX_108929087 on 2B, 2B,
ance during the grain-filling stage (Khaliq et al., 2008). Earlier
3B, 4B, 5A, 5B, 6B and 6B were found at similar positions (Li
flowering should be improved in parallel in areas with high tem-
et al., 2018a). The locus AX_110958315 on 5A is about one LD
peratures around 40% of the temperature zones experience high
from a TGW-related QTL (Gao et al., 2015). By contrast, the
temperatures of more than 30 °C during grain filling that affects
locus AX_111548406 on 4B for TGW was identified at a similar
grain filling rate and final yield (Rane et al., 2007; Reynolds et
position (Liu et al., 2014). As for flag leaf traits, a pleiotropic
al., 2011). Previous studies have suggested that early flowering
locus for FLL and FLA is at the same position on chromosome
could shift the grain-filling period to relatively cooler conditions
2B as AX_111634394 (Wu et al., 2016). Six identical loci to our
to avoid heat stress, leading to increased reproductive growth
study were on chromosomes 1A, 1D, 2A, 2B, 3A, 4A, 4D, 5B,
duration and wheat yield (Tao & Zhang, 2013; Tao et al., 2017b).
7A and 7B for FLL (Li et al., 2018a). An FLL-related QTL on
Our current findings indicated that TGW and GY were signifi-
chromosome 2B overlapped with our FLL locus AX_111634394
cantly and positively correlated with spike shape, but negatively
(Wu et al., 2016). Identified Qflw-3A (FLW), Qflw-7D (FLW)
correlated with spike number. The lanceolate panicle was re-
and Qfla-3A (FLA) associated with flag leaf morphology are
lated to high grain yield. A trade-off exists in grain weight and
at positions similar to AX_94541532, AX_110227474 and
grain number (Sadras, 2007; Sadras and Slafer, 2012; Xiao et
AX_109838665 in our study (Yan et al., 2020), respectively.
al., 2022). We found a positive correlation between plant height
The loci AX_94541532 and AX_111476049 are the same as
and grain number (Supplementary Data Fig. S3). A previous re-
FLW-QTLs on chromosome 6B (Wu et al., 2016). In addition,
port from a similar environment has suggested that significant
some Rht loci can cause a reduction in plant height, which has
improvement in yield was mainly because grain weight contrib-
an impact on grain yield (Xiao et al., 2022). Rht-D1b is widely
utes to yield potential and stability (Zheng et al., 2011).
present in the Yellow and Huai Valleys Winter Wheat Zone (Gao
Plant water status is another important factor that plays a crit-
et al., 2017). The locus AX_89703298 related to PH on chromo-
ical role in yield formation. Maintaining a certain relative water
some 4D of this study is at the same position as Rht-D1 (Xue et
content could improve GY and stability (Matin et al., 1989;
al., 2002), indicating the influence of Rht-D1b on PH in the cur-
Teulat et al., 2003). Physiological and morphological traits, such
rent study (Gao et al., 2015; Sun et al., 2017; Li et al., 2018a).
as flag leaf photosynthesis, transpiration and yield, are closely
Loci AX_94638736 and AX_110476771 on chromosome 2A
related to plant water status (González et al., 2008). We tried to
are at a position similar to QUIL.caas-2AS.1 and QPH.caas-
keep the same field management under well-watered conditions
2AL (co-localized with QUIL.caas-2AL), respectively (Li et al.,
in our experiment. LWC and SPWC showed a significant rela-
2018a). QTLs for PH were identified on 3A, close to our locus
tionship with TGW and GY. To conclude, the best-yielding cul-
AX_108900466 related to PH (Cui et al., 2011; Lopes et al.,
tivars revealed higher water content, early flowering with longer
2012). The 5A locus BobWhite_c47401_491 is about one LD
reproductive growth duration, higher flag leaf width, photosyn-
from a QTL (Li et al., 2016), and near Rht12 (Reynolds et al.,
thetic capacity, high grain filling rate and stay-green features
2015). Vrn-B1 is at a position similar to locus AX_94715126
but lower flag leaf length, flag leaf angle, plant height and spike
(Fu et al., 2005). However, there is no reported correlation be-
numbers. Our findings provide information about ideotypic
tween PH and vernalization. The remaining loci are likely to be
traits, supporting breeding of climate-resilient crop cultivars.
14 Li et al. — Variations in wheat traits and their genetic basis

potentially novel MTAs responsible for PH. These identified description of different spike shapes. Fig. S3: Phenotypic vari-
significant SNPs indicated the reliability of the GWAS. ations for the 32 wheat traits measured under multiple envir-
Crop yield is a quantitative trait controlled by many other onments. Fig. S4: Correlation of 32 wheat traits by best linear
plant traits, mainly polygenic in nature (Wu et al., 2012; Li et al., unbiased estimations for each trait across four environments.
2020). The potential yield increase associated with these traits Fig. S5: Manhattan plots for a genome-wide association study
remains relatively untapped. In this study, a significant correl- of the traits in 166 wheat accessions under multiple environ-
ation was observed between the number of favourable alleles ments. Fig. S6: Quantile–quantile plots of GWAS for 32 traits.
and yield-related traits, suggesting that pyramiding the favour- Fig. S7: Favourable allele frequencies of the identified QTLs.
able alleles effectively could improve those traits and support Fig. S8: Effect and distribution of favourable alleles of trait-

Downloaded from https://academic.oup.com/aob/advance-article/doi/10.1093/aob/mcad003/6992592 by guest on 05 March 2023

breeding of ideotypes (Fig. 5). The GWAS of our study was also associated markers. Fig. S9: Phenotypic changes in the trait
reliable and efficient in detecting loci for GY and related traits. values for 166 wheat cultivars released over the past 70 years.
Theoretical analysis indicated that a 50 % increase in wheat yield Fig. S10: Changes in numbers of increasing-effect alleles for
potential is possible through genetic improvement of RUE. A 32 trait values over the past 70 years. Table S1: Information of
crop’s structural and reproductive aspects must be ameliorated in the 166 wheat accessions used in GWAS. Table S2: Loci for 32
parallel to achieve the desired impact (Ort et al., 2015). We found traits identified by GWAS. Table S3: Distribution of pleiotropic
that the loci related to GY and TGW were pleiotropically correl- loci associated with three or more grain yield-related traits on
ated with PH, photosynthetic capacity, flag leaf traits, GFR and wheat chromosomes. Table S4: 32 wheat trait values within
SS. Through co-localization, early flowering is likely to benefit each cluster.
TGW at lower temperatures. In addition, the year of release had a
significant effect on some easily observed agronomic traits (PH,
FLW, FLANG, TGW and GY), but no effect on complex traits ACKNOWLEDGEMENTS
(Pn, Gs, GFR, SGT and biomass traits), suggesting that there
should be still opportunities for further improvement of those We thank Shuaipeng Fei, Shibo Li, Cunli Li, Yongjun Lv and
traits. Thus, our study identifies cultivars with favourable alleles Yongtao Zhao for their help during this study. We thank TopEdit
as potential parents and molecular markers closely linked to the (www.topeditsci.com) for linguistic assistance during the prep-
above QTLs for assisted selection to combine more favourable aration of the manuscript. The authors declare no competing
alleles and develop new wheat cultivars with ideal traits. interests.

CONCLUSIONS AUTHOR CONTRIBUTIONS

Our GWAS is shown to be a powerful approach for genetic dis- F.T. and Z.H. designed the study. Y.L. carried out the pheno-
section of phenological, physiological, plant architectural and typic and growth analysis and performed the gas-exchange
yield-related traits in wheat based on a high-resolution physical measurements. Y.H. and J.T performed the molecular experi-
map. For the first objective, grain yield was positively correl- ments. Y.L. and F.T. drafted the original manuscript. Z.H. and
ated with the net photosynthetic rate, spike shape and thousand- R.M. provided suggestions for writing the manuscript. All au-
grain weight but negatively correlated with plant height, flag thors discussed the results and revised the manuscript.
leaf angle and spike number per square metre. For the second
objective, 1236 significant stable loci associated with these
were traits identified by SNP-GWAS, and which provide in- FUNDING
valuable sources. Linear regression indicated apparent cumu-
This study was supported by the National Key Research and
lative effects of favourable alleles for increasing corresponding
Development Program of China (2020YFE0202300) and the
traits. For the third objective, complex traits such as Pn could be
National Natural Science Foundation of China (Project Nos.
further used to break through the current wheat yield plateau.
31761143006, 41571493).
The identified QTLs had considerable additive effects on trait
performance (Fig. 5) with predictable phenotypes; thus, these
QTLs have great potential to be exploited in breeding new high-
yielding cultivars. For the fourth objective, 11 Chinese culti- LITERATURE CITED
vars could be used as parents for new high-yielding cultivars Abbai R, Singh V, Snowdon R, Kumar A, Schnurbusch T. 2020. Seeking
with desirable traits. Our study highlighted that variations in crops with balanced parts for the ideal whole. Trends in Plant Science 25:
different traits, their associations with grain yield and their gen- 1189–1193.
Adachi S, Tsuru Y, Nito N, et al. 2011. Identification and characteriza-
etic basis are important for yield improvement through MAS. tion of genomic regions on chromosomes 4 and 8 that control the rate
of photosynthesis in rice leaves. Journal of Experimental Botany 62:
1927–1938.
SUPPLEMENTARY DATA Arif MAR, Nagel N, Newmann K, Kobiljski B, Lohwasser U, Borner A.
2012. Genetic studies of seed longevity in hexaploid wheat using segrega-
Supplementary data are available online at https://academic. tion and association mapping approaches. Euphytica 1869: 1–13.
oup.com/aob and consist of the following. Fig. S1: Correlation Azadi A, Mardi M, Hervan EM, et al. 2015. QTL mapping of yield and
between net photosynthesis, intercellular CO2 concentration, yield components under normal and salt-stress conditions in bread wheat
(Triticum aestivum L.). Plant Molecular Biology Reporter 33: 102–120.
stomatal conductance, transpiration rate and vapour pres- Backhaus AE, Lister A, Tomkins M, et al. 2022. High expression of the
sure deficit VPD or Tleaf. Fig. S2: Image information and MADS-box gene VRT2 increases the number of rudimentary basal
 Li et al. — Variations in wheat traits and their genetic basis 15

spikelets in wheat. Plant Physiology 189: 1536–1552. doi:10.1093/ Gao F, Wen W, Liu J, et al. 2015. Genome-wide linkage mapping of QTL for
plphys/kiac156. yield components, plant height and yield-related physiological traits in
Bailey SJ, Parker JE, Ainsworth EA, Oldroyd GED, Schroeder JI. 2019. the Chinese wheat cross Zhou 8425B/Chinese spring. Frontiers in Plant
Genetic strategies for improving crop yields. Nature 575: 109–118. Science 6: 1099.
Baker NR. 2008. Chlorophyll fluorescence: a probe of photosynthesis in Gao F, Ma D, Yin G, et al. 2017. Genetic progress in grain yield and
vivo. Annual Review of Plant Biology 59: 89–113. doi:10.1146/annurev. physiological traits in Chinese wheat cultivars of southern yellow
arplant.59.032607.092759. and Huai Valley since 1950. Crop Science 57: 760–773. doi:10.2135/
Beer C, Reichstein M, Tomelleri E, et al. 2010. Terrestrial gross carbon di- cropsci2016.05.0362.
oxide uptake: Global distribution and covariation with climate. Science Godfray HCJ, Beddington JR, Crute IR, et al. 2010. Food Security: the chal-
329: 834–838. doi:10.1126/science.1184984. lenge of feeding 9 billion people. Science 327: 812–818.

Downloaded from https://academic.oup.com/aob/advance-article/doi/10.1093/aob/mcad003/6992592 by guest on 05 March 2023

Bodner G, Kronberga A, Lepse L, et al. 2018. Trait identification of faba González AI, Martín IL, Ayerbe L. 2008. Yield and osmotic adjustment
bean ideotypes for northern European environments. European Journal of capacity of barley under terminal water-stress conditions. Journal of
Agronomy 96: 1–12. doi:10.1016/j.eja.2018.02.008. Agronomy and Crop Science 194: 81–91. doi:10.1111/j.1439-037x.2007.
Bonnett OT. 1966. Inflorescences of maize, wheat, rye, barley, and oats: their 00289.x.
initiation and development. Champaign: Bulletin (University of Illinois Gu J, Yin X, Struik PC, Stomph TJ, Wang H. 2012. Using chromosome
(Urbana-Champaign campus) Agricultural Experiment Station). introgression lines to map quantitative trait loci for photosynthesis param-
Brooks RJ, Semenov MA, Jamieson PD. 2001. Simplifying Sirius: sensitivity ana- eters in rice (Oryza sativa L.) leaves under drought and well-watered field
lysis and development of a meta-model for wheat yield prediction. European conditions. Journal of Experimental Botany 63: 455–469.
Journal of Agronomy 14: 43–60. doi:10.1016/s1161-0301(00)00089-7. He ZH, Bonjean APA. 2010. Cereals in China. Mexico DF: CIMMYT, 51–68.
Buckley TN, Mott KA. 2002. Stomatal water relations and the control of hy- Hedden P. 2003. Constructing dwarf rice. Nature Biotechnology 21: 873–874.
draulic supply and demand. Berlin: Springer. doi:10.1038/nbt0803-873.
Bustos DV, Hasan AK, Reynold MP, Calderini DF. 2013. Combining high Heider B, Struelens Q, Faye M, Flores C, Dangles O. 2021. Intraspecific
grain number and weight through a DH-population to improve grain yield diversity as a reservoir for heat-stress tolerance in sweet potato. Nature
potential of wheat in high-yielding environments. Field Crops Research Climate Change 11: 1–6.
145: 106–115. Hubbart S, Peng S, Horton P, Chen Y, Murchie E. 2007. Trends in leaf
Cavanagh CR, Chao S, Wang S, et al. 2013. Genome-wide comparative di- photosynthesis in historical rice varieties developed in the Philippines
versity uncovers multiple targets of selection for improvement in hexa- since 1966. Journal of Experimental Botany 58: 3429–3438.
ploid wheat landraces and cultivars. Proceedings of the National Academy Jaiswal V, Gahlaut V, Meher PK, Mir RR, Gupta PK. 2016. Genome wide
of Sciences of the United States of America 110: 8057–8062. single locus single trait, multi-locus and multi-trait association mapping
Chytyk CJ, Hucl PJ, Gray GR. 2011. Leaf photosynthetic properties and bio- for some important agronomic traits in common wheat (Triticum aestivum
mass accumulation of selected western Canadian spring wheat cultivars. L.). PLoS One 11: e0159343.
Canadian Journal of Plant Science 91: 305–314. doi:10.4141/cjps09163. Kanemura T, Homma K, Ohsumi A, Tatsuhiko S, Horie T. 2007. Evaluation
Cowan IR. 1977. Stomatal behavior and environment. Advances in Botanical of genotypic variation in leaf photosynthetic rate and its associated factors
Research 4: 117–228. using rice diversity research set of germplasm. Photosynthesis Research
Crossa J, Burgueno J, Dreisigacker S, et al. 2007. Association analysis of 94: 23–30.
historical bread wheat germplasm using additive genetic covariance of re- Khaliq I, Irshad A, Ahsan AM. 2008. Awns and flag leaf contribution to-
latives and population structure. Genetics 177: 1889–1913. wards grain yield in spring wheat (Triticum aestivum L.). Cereal Research
Cui F, Li J, Ding A, et al. 2011. Conditional QTL mapping for plant height Communications 36: 65–76.
with respect to the length of the spike and internode in two mapping Kruger EL, Volin JC. 2006. Reexamining the empirical relation between plant
populations of wheat. Theoretical and Applied Genetics 122: 1517–1536. growth and leaf photosynthesis. Functional Plant Biology 33: 421–429.
doi:10.1007/s00122-011-1551-6. doi:10.1071/fp05310.
Ding T, Yang Z, Wei X, Yuan F, Yin S, Wang B. 2018. Evaluation of salt- Kumar N, Kulwal PL, Balyan HS, Gupta PK. 2007. QTL mapping for yield
tolerant germplasm and screening of the salt-tolerance traits of sweet sor- and yield contributing traits in two mapping populations of bread wheat.
ghum in the germination stage. Functional Plant Biology 45: 1073–1081. Molecular Breeding 19: 163–177.
doi:10.1071/fp18009. Li C, Bai G, Carver BF, Chao S, Wang Z. 2016. Mapping quantitative trait
Dixon J, Braun HJ, Kosina P, Crouch J. 2009. Wheat facts and futures 2009. loci for plant adaptation and morphology traits in wheat using single nu-
Mexico: CIMMYT. cleotide polymorphisms. Euphytica 208: 299–312.
Donald CM. 1968. The breeding of crop ideotypes. Euphytica 17: 385–403. Li F, Wen W, He Z, et al. 2018a. Genome-wide linkage mapping of yield re-
doi:10.1007/bf00056241. lated traits in three Chinese bread wheat populations using high-density
Driever SM, Lawson T, Andralojc PJ, Raines CA, Parry MAJ. 2014. SNP markers. Theoretical and Applied Genetics 131: 1903–1924.
Natural variation in photosynthetic capacity, growth, and yield in 64 field- doi:10.1007/s00122-018-3122-6.
grown wheat genotypes. Journal of Experimental Botany 65: 4959–4973. Li J, Liu J, Wen W, et al. 2018b. Genome-wide association mapping of
doi:10.1093/jxb/eru253. vitamins b1 and b2 in common wheat. The Crop Journal 6: 263–270.
Fang Y, Du Y, Wang J, et al. 2017. Moderate drought stress affected root doi:10.1016/j.cj.2017.08.002.
growth and grain yield in old, modern and newly released cultivars of Li F, Wen W, Liu J, Zhang Y, Xia X. 2019a. Genetic architecture of grain
winter wheat. Frontiers in Plant Science 8: 672. yield in bread wheat based on genome-wide association studies. BMC
FAO. 2019. FAOSTAT 2019: FAO Statistical Databases (accessed 23 May, Plant Biology 19: 168.
2019). http://faostat.fao.org/. Li L, Peng Z, Mao X, et al. 2019b. Genome-wide association study reveals
Fischer RA, Rees D, Sayre KD, Lu ZM, Condon AG, Saavedra AL. 1998. genomic regions controlling root and shoot traits at late growth stages in
Wheat yield progress associated with higher stomatal conductance and wheat. Annals of Botany 124: 993–1006. doi:10.1093/aob/mcz041.
photosynthetic rate, and cooler canopies. Crop Science 38: 1467–1475. Li L, Peng Z, Mao X, Wang J, Jing R. 2020. Genetic insights into natural
doi:10.2135/cropsci1998.0011183x003800060011x. variation underlying salt tolerance in wheat. Journal of Experimental
Flood PJ, Harbinson J, Aarts MGM. 2011. Natural genetic variation in plant Botany 72: 1135–1150.
photosynthesis. Trends in Plant Science 16: 327–335. Li Y, Hou R, Tao F. 2021. Wheat morpho-physiological traits and radiation use
Foulkes MJ, Slafer GA, Davies WJ, et al. 2011. Raising yield potential of efficiency under interactive effects of warming and tillage management.
wheat. III. Optimizing partitioning to grain while maintaining lodging re- Plant, Cell & Environment 44: 2386–2401.
sistance. Journal of Experimental Botany 62: 469–486. Li Y, Tao F, Hao Y, et al. 2022a. Wheat traits and the associated loci confer-
Fu D, Szucs P, Yan L, et al. 2005. Large deletions within the first intron in ring radiation use efficiency. Plant Journal 112: 565–582. doi:10.1111/
VRN-1 are associated with spring growth habit in barley and wheat. tpj.15954.
Molecular Genetics and Genomics 273: 54–65. Li A, Hao C, Wang Z, et al. 2022b. Wheat breeding history reveals syn-
Fu LP, Wu JC, Yang SR, et al. 2020. Genome-wide association analysis of ergistic selection of pleiotropic genomic sites for plant architec-
stem water-soluble carbohydrate content in bread wheat. Theoretical and ture and grain yield. Molecular Plant 15: 504–519. doi:10.1016/j.
Applied Genetics 133: 2897–2914. doi:10.1007/s00122-020-03640-x. molp.2022.01.004.
16 Li et al. — Variations in wheat traits and their genetic basis

Liang Y, Zhang K, Zhao L, et al. 2009. Identification of chromosome regions Reynolds MP, Pellegrineschi A, Skovmand B. 2015. Sink-limitation to yield
conferring dry matter accumulation and photosynthesis in wheat (Triticum and biomass: A summary of some investigations in spring wheat. Annals
aestivum L.). Euphytica 171: 145–156. doi:10.1007/s10681-009-0024-3. of Applied Biology 146: 39–49.
Liu G, Jia L, Lu L, et al. 2014. Mapping QTLs of yield-related traits using Rizzo G, Monzon JP, Tenorio FA, Howard R, Cassman KG, Grassini P.
RIL population derived from common wheat and Tibetan semi-wild 2022. Climate and agronomy, not genetics, underpin recent maize yield
wheat. Theoretical and Applied Genetics 127: 2415–2432. doi:10.1007/ gains in favorable environments. Proceedings of the National Academy of
s00122-014-2387-7. Sciences of the United States of America 119: e2113629119.
Liu J, He Z, Rasheed A, et al. 2017. Genome-wide association mapping of Sadras VO. 2007. Evolutionary aspects of the trade-off between seed size
black point reaction in common wheat (Triticum aestivum L.). BMC Plant and number in crops. Field Crops Research 100: 125–138. doi:10.1016/j.
Biology 17: 220. fcr.2006.07.004.

Downloaded from https://academic.oup.com/aob/advance-article/doi/10.1093/aob/mcad003/6992592 by guest on 05 March 2023

Long SP, Zhu XG, Naidu SL, Ort DR. 2006. Can improvement in Sadras VO, Slafer GA. 2012. Environmental modulation of yield components
photosynthesis increase crop yields? Plant, Cell & Environment 29: in cereals: heritabilities reveal a hierarchy of phenotypic plasticities. Field
315–330. Crops Research 127: 215–224. doi:10.1016/j.fcr.2011.11.014.
Lopes MS, Reynolds MP, Manes Y, Singh RP, Crossa J, Braun HJ. 2012. Sadras VO, Lawson C, Montoro A. 2012. Photosynthetic traits in Australian
Genetic yield gains and changes in associated traits of CIMMYT spring wheat varieties released between 1958 and 2007. Field Crops Research
bread wheat in a ‘historic’ set representing 30 years of breeding. Crop 134: 19–29. doi:10.1016/j.fcr.2012.04.012.
Science 52: 1123–1131. doi:10.2135/cropsci2011.09.0467. Sakuma S, Schnurbusch T. 2020. Of floral fortune: tinkering with the grain
Maccaferri M, Zhang J, Bulli P, et al. 2015. A genome-wide association yield potential of cereal crops. New Phytologist 225: 1873–1882.
study of resistance to stripe rust (Puccinia striiformis f. sp. tritici) in a Semenov MA, Stratonovitch P. 2013. Designing high-yielding wheat
worldwide collection of hexaploidy spring wheat (Triticum aestivum L.). ideotypes for a changing climate. Food and Energy Security 2: 185–196.
Genetics 5: 449–465. doi:10.1002/fes3.34.
Matin MA, Brown JH, Ferguson H. 1989. Leaf water potential, relative water Sharma SN, Sain RS, Sharma PK. 2003. The genetic control of flag leaf
content, and diffusive resistance as screening techniques for drought re- length in normal and late sown durum wheat. Journal of Agricultural
sistance in barley. Agronomy Journal 81: 100–105. doi:10.2134/agronj19 Science 141: 323–331.
89.00021962008100010018x. Sukumaran S, Dreisigacker S, Lopes M, Chavez P, Reynolds MP. 2015.
Morinaka Y, Sakamoto T, Inukai Y, et al. 2006. Morphological alter- Genome wide association study for grain yield and related traits in an
ation caused by brassinosteroid insensitivity increases the biomass and elite spring wheat population grown in temperate irrigated environments.
grain production of rice. Plant Physiology 141: 924–931. doi:10.1104/ Theoretical and Applied Genetics 128: 353–363.
pp.106.077081. Sun C, Zhang F, Yan X, et al. 2017. Genome-wide association study for 13
Ohsumi A, Hamasaki A, Nakagawa H, Yoshida H, Shiraiwa T, Horie T. agronomic traits reveals distribution of superior alleles in bread wheat
2007. A model explaining genotypic and ontogenetic variation of leaf from the Yellow and Huai Valley of China. Plant Biotechnology Journal
photosynthetic rate in rice (Oryza sativa) based on leaf nitrogen content 15: 953–969. doi:10.1111/pbi.12690.
and stomatal conductance. Annals of Botany 99: 265–273. doi:10.1093/ Takai T, Kondo M, Yano M, Yamamoto T. 2010. A quantitative trait locus
aob/mcl253. for chlorophyll content and its association with leaf photosynthesis in rice.
Ojima M. 1974. Improvement of photosynthetic capacity in soybean variety. Rice 3: 172–180. doi:10.1007/s12284-010-9047-6.
JARQ-Japan Agricultural Research Quarterly 8: 6–12. Takai T, Ohsumi A, Arai Y, et al. 2013. QTL analysis of leaf photo-
Ort DR, Merchant SS, Alric J, et al. 2015. Redesigning photosynthesis to synthesis in rice. JARQ-Japan Agricultural Research Quarterly 47:
sustainably meet global food and bioenergy demand. Proceedings of 227–235.
the National Academy of Sciences of the United States of America 112: Tao F, Zhang Z. 2013. Climate change, wheat productivity and water use in
8529–8536. the North China plain: A new super-ensemble-based probabilistic projec-
Paez-Garcia A, Motes CM, Scheible WR, Chen R, Blancaflor EB, Monteros tion. Agricultural and Forest Meteorology 170: 146–165. doi:10.1016/j.
MJ. 2015. Root traits and phenotyping strategies for plant improvement. agrformet.2011.10.003.
Plants 15: 334–355. Tao F, Rötter RP, Palosuo T, et al. 2017a. Designing future barley ideotypes
Pang YL, Liu CX, Wang DF, et al. 2020. High-resolution genome-wide asso- using a crop model ensemble. European Journal of Agronomy 82: 144–
ciation study identifies genomic regions and candidate genes for important 162. doi:10.1016/j.eja.2016.10.012.
agronomic traits in wheat. Molecular Plant 13: 1311–1327. doi:10.1016/j. Tao F, Xiao D, Zhang S, Zhang Z, Rötter RP. 2017b. Wheat yield benefited
molp.2020.07.008. from increases in minimum temperature in the Huang-Huai-Hai plain of
Parry MAJ, Reynolds M, Salvucci ME, et al. 2011. Raising yield potential China in the past three decades. Agricultural and Forest Meteorology 239:
of wheat. II. Increasing photosynthetic capacity and efficiency. Journal of 1–14. doi:10.1016/j.agrformet.2017.02.033.
Experimental Botany 62: 453–467. Teng S, Qian Q, Zeng D, et al. 2004. QTL analysis of leaf photosynthetic rate
Peng J, Richards DE, Hartley NM, et al. 1999. ‘Green revolution’ genes and related physiological traits in rice (Oryza sativa L.). Euphytica 135:
encode mutant gibberellin response modulators. Nature 400: 256–261. 1–7. doi:10.1023/b:euph.0000009487.89270.e9.
doi:10.1038/22307. Teulat B, Zoumarou-Wallis N, Rotter B, Salem MB, Bahri H, This D. 2003.
Qu M, Zheng G, Hamdani S, et al. 2017. Leaf photosynthetic parameters QTL for relative water content in field-grown barley and their stability
related to biomass accumulation in a global rice diversity survey. Plant across Mediterranean environments. Theoretical and Applied Genetics
Physiology 175: 248–258. doi:10.1104/pp.17.00332. 108: 181–188.
Rahimi Y, Bihamta MR, Taleei A, Alipour H, Ingvarsson PK. 2019. Valluru R, Reynolds MP, Davies WJ, Sukumaran S. 2017. Phenotypic and
Genome-wide association study of agronomic traits in bread wheat reveals genome-wide association analysis of spike ethylene in diverse wheat
novel putative alleles for future breeding programs. BMC Plant Biology genotypes under heat stress. New Phytologist 214: 271–283.
19: 541. Von Korff M, Grando S, Del Greco A, This D, Baum M, Ceccarelli S. 2008.
Raines CA. 2011. Increasing photosynthetic carbon assimilation in C3 plants Quantitative trait loci associated with adaptation to Mediterranean dry-
to improve crop yield: current and future strategies. Plant Physiology 155: land conditions in barley. Theoretical and Applied Genetics 117: 653–669.
36–42. doi:10.1007/s00122-008-0787-2.
Rane J, Pannu RK, Sohu VS, et al. 2007. Performance of yield and sta- Voss-Fels KP, Stahl A, Wittkop B, et al. 2019. Breeding improves wheat
bility of advanced wheat genotypes under heat stressed environments productivity under contrasting agrochemical input levels. Nature Plants 5:
of Indo-Gangetic Plains. Crop Science 47: 1561–1573. doi:10.2135/ 706–714. doi:10.1038/s41477-019-0445-5.
cropsci2006.07.0479. Wang J, Liu W, Wang H, et al. 2011a. QTL mapping of yield-related traits in
Ray DK, West PC, Clark M, Gerber JS, Prishchepov AV, Chatterjee S. the wheat germplasm 3228. Euphytica 177: 277–292.
2019. Climate change has likely already affected global food production. Wang P, Zhou G, Yu H, Yu S. 2011b. Fine mapping a major QTL for flag leaf
PLoS One 14: e0217148. doi:10.1371/journal.pone.0217148. size and yield-related traits in rice. Theoretical and Applied Genetics 123:
Reynolds M, Foulkes MJ, Slafer GA, et al. 2011. Raising yield potential of 1319–1330. doi:10.1007/s00122-011-1669-6.
wheat. I. Overview of a consortium approach and breeding strategies. Wang S, Wong D, Forrest K, et al. 2014. Characterization of poly-
Journal of Experimental Botany 62: 439–452. ploid wheat genomic diversity using a high-density 90000 single
 Li et al. — Variations in wheat traits and their genetic basis 17

nucleotide polymorphism array. Plant Biotechnology Journal 12: 787– wheat by genome-wide association study and linkage mapping. Molecular
796. doi:10.1111/pbi.12183. Breeding 40: 58.
Welcker C, Spencer N, Turc O, et al. 2022. Physiological adaptive traits are Yang D, Liu Y, Cheng H, et al. 2016. Genetic dissection of flag leaf morph-
a potential allele reservoir for maize genetic progress under challenging ology in wheat (Triticum aestivum L.) under diverse water regimes. BMC
conditions. Nature Communications 13: 3225. Genetics 17: 94.
Wu XS, Chang XP, Jing RL. 2012. Genetic insight into yield-associated traits Yin Z, Meng F, Song H, He X, Xu X, Yu D. 2010. Mapping quantitative
of wheat grown in multiple rain-fed environments. PLoS One 7: e31249. trait loci associated with chlorophyll a fluorescence parameters in soy-
doi:10.1371/journal.pone.0031249. bean (Glycine max (L.) Merr.). Planta 231: 875–885. doi:10.1007/
Wu Q, Chen Y, Fu L, et al. 2016. QTL mapping of flag leaf traits in common s00425-009-1094-0.
wheat using an integrated high-density SSR and SNP genetic linkage map. Zadoks JC, Chang TT, Konzak CF. 1974. A decimal code for growth stages

Downloaded from https://academic.oup.com/aob/advance-article/doi/10.1093/aob/mcad003/6992592 by guest on 05 March 2023

Euphytica 208: 337–351. of cereals. Weed Research 14: 415–421. doi:10.1111/j.1365-3180.1974.
Xiao J, Liu B, Yao Y, et al. 2022. Wheat genomic study for genetic improve- tb01084.x.
ment of traits in China. Science China Life Sciences 65: 1718–1775. Zheng TC, Zhang XK, Yin GH, et al. 2011. Genetic gains in grain yield,
Xu DQ, Shen YK. 2002. Photosynthetic efficiency and crop yield. In: net photosynthesis and stomatal conductance achieved in Henan province
Pessarakli M, ed. Handbook of plant and crop physiology, 2nd edn. New of China between 1981 and 2008. Field Crops Research 122: 225–233.
York: Marcel Dekker Inc, 821–834. doi:10.1016/j.fcr.2011.03.015.
Xue QW, Soundararajan M, Weiss A, Arkebauer TJ, Baenziger PS. Zhu XG, Long SP, Ort DR. 2010. Improving photosynthetic efficiency for
2002. Genotypic variation of gas exchange parameters and carbon iso- greater yield. Annual Review of Plant Biology 61: 235–261. doi:10.1146/
tope discrimination in winter wheat. Journal of Plant Physiology 159: annurev-arplant-042809-112206.
891–898. Zhu Z, Hao Y, Mergoum M, et al. 2019. Breeding wheat for resistance to fu-
Yan X, Zhao L, Ren Y, Zhang N, Dong Z, Chen F. 2020. Identification of sarium head blight in the global north: China, USA, and Canada. The Crop
genetic loci and a candidate gene related to flag leaf traits in common Journal 7: 730–738. doi:10.1016/j.cj.2019.06.003.