Validation-Of-Qgs10-A-Quantitative-Trait-Locus-For-Grain-Size-On-The-Long-Arm-Of-Chromosome-10-In-Rice-Oryza-Sativa-L.pdf

Journal of Integrative Agriculture 2017, 16(1): 16–26
Available online at www.sciencedirect.com

ScienceDirect

RESEARCH ARTICLE

Validation of qGS10, a quantitative trait locus for grain size on the
long arm of chromosome 10 in rice (Oryza sativa L.)
WANG Zhen*, CHEN Jun-yu*, ZHU Yu-jun, FAN Ye-yang, ZHUANG Jie-yun
State Key Laboratory of Rice Biology and Chinese National Center for Rice Improvement, China National Rice Research Institute,
Hangzhou 310006, P.R.China

Abstract
Grain size is a major determinant of grain weight and a trait having important impact on grain quality in rice. The objective
of this study is to detect QTLs for grain size in rice and identify important QTLs that have not been well characterized before.
The QTL mapping was first performed using three recombinant inbred line populations derived from indica rice crosses
Teqing/IRBB lines, Zhenshan 97/Milyang 46, Xieqingzao/Milyang 46. Fourteen QTLs for grain length and 10 QTLs for grain
width were detected, including seven shared by two populations and 17 found in one population. Three of the seven common QTLs were found to coincide in position with those that have been cloned and the four others remained to be clarified.
One of them, qGS10 located in the interval RM6100–RM228 on the long arm of chromosome 10, was validated using F2:3
populations and near isogenic lines derived from residual heterozygotes for the interval RM6100–RM228. The QTL was
found to have a considerable effect on grain size and grain weight, and a small effect on grain number. This region was
also previously detected for quality traits in rice in a number of studies, providing a good candidate for functional analysis
and breeding utilization.
Keywords: grain size, quantitative trait locus, residual heterozygote, rice (Oryza sativa L.)

in rice is determined by three components, i.e., number of

1. Introduction
Rice (Oryza sativa L.) is one of the most important cereal
crops, feeding half of the world’s population. Grain yield

panicles per plant, number of grains per panicle and grain
weight. Grain size is a major determinant of grain weight,
and a trait having important impact on the market value of
rice grain. Long and slender grains are preferred in the
major segment of the international market, whereas short
and round grains are favored in northern China, Korea and
Japan (Calingacion et al. 2014). In addition, slender grains

Received 11 January, 2016 Accepted 25 April, 2016
WANG Zhen, Tel: +86-571-63370197, Fax: +86-571-63370364,
E-mail: ; Correspondence ZHUANG Jie-yun,
Tel: +86-571-63370369, Fax: +86-571-63370364,
E-mail:
*
These authors contributed equally to this study.

are more likely to have lower grain chalkiness thus a better

© 2017, CAAS. Published by Elsevier Ltd. This is an open
access article under the CC BY-NC-ND license (http://
creativecommons.org/licenses/by-nc-nd/4.0/)
doi: 10.1016/S2095-3119(16)61410-7

major negative regulator controlling grain length and weight

appearance quality (Wang et al. 2005).
Over the last two decades, a large number of quantitative
trait loci (QTLs) for grain size and grain weight in rice were
detected and some of them were cloned since 2006. GS3, a

is the first QTL cloned for grain size (Fan et al. 2006). Eight
more QTLs were cloned up to date, including GL3.1/qGL3


WANG Zhen et al. Journal of Integrative Agriculture 2017, 16(1): 16–26

(Qi et al. 2012; Zhang et al. 2012), TGW6 (Ishimaru et al.
2013), GW6a (Song et al. 2015), and GW7/GL7 (Wang
S K et al. 2015; Wang Y X et al. 2015) determining grain
length and weight, and GW2 (Song et al. 2007), qSW5/
GW5 (Shomura et al. 2008; Weng et al. 2008), GS5 (Li Y
et al. 2011), and GW8 (Wang et al. 2012b) responsible for
grain width and weight.
It has been commonly applied that QTLs exhibiting major
and consistent effects in primary mapping populations were
first targeted for cloning. As a result, QTLs that have been
cloned for yield traits in rice, either those for grain size and
grain weight described above, or others associated with
grain number (Ikeda et al. 2013; Yan et al. 2013), all showed
large effects for the trait under study. Because few QTLs of
this kind is available, it is not uncommon that different groups
separately make great efforts on the same QTL (Shomura
et al. 2008; Weng et al. 2008; Qi et al. 2012; Zhang et al.
2012; Ikeda et al. 2013; Wang S K et al. 2015; Wang Y X
et al. 2015). Diversifying rice crosses in constructing populations for primary QTL mapping may facilitate the detection
of new QTLs and alleviate the shortage of candidate QTLs
for cloning.
In the present study, QTL mapping for grain size in rice
was performed using three primary populations, followed by
the validation of one QTL region. Fourteen QTLs for grain

length and 10 QTLs for grain width were detected in three
recombinant inbred line (RIL) populations derived from the
indica rice crosses Zhenshan 97/Milyang 46 (ZM), Xieqingzao/Milyang 46 (XM) and Teqing/IRBB lines (TI). One QTL
shared by different populations and located in a region that
was away from those that have been cloned was selected
for validation. Two lines of the TI population were crossed
to develop an F2:3 population and three sets of near isogenic
lines (NILs). The target QTL, qGS10 located in the interval
RM6100–RM228 on the long arm of chromosome 10, was
validated to have a considerable effect on grain size and
grain weight.

2. Materials and methods
2.1. Plant materials
The three RIL populations used in this study have been
reported by Mei et al. (2013). Both the female and male
parents of the TI population are indica rice restorer lines,
of which the male parent included six IRBB lines (IRBB50,
IRBB51, IRBB52, IRBB54, IRBB55, and IRBB59) that
are NILs in the genetic background of IR24 (Huang et al.
1997a). The numbers of RILs included in the TI population
are 122 for Teqing/IRBB52, 77 for Teqing/IRBB59, two for
Teqing/IRBB50, and one each for Teqing/IRBB51, Teqing/

17

IRBB54 and Teqing/IRBB55. The female parents of the
ZM and XM populations, Zhenshan 97 and Xieqingzao, are
maintainer lines of the commercial three-line indica rice hybrid Shanyou 10 and Xieyou 46, respectively, and the common male parent Milyang 46 is the restorer line of the two
hybrids (Mei et al. 2013). In the rice zone of middle-lower

reaches of Yangtze River in China, Zhenshan 97 and
Xieqingzao are used as early-season rice, and Milyang 46,
Teqing and IR24 are grown as middle-season rice.
Development of secondary populations for the validation
of qGS10 were described below and illustrated in Fig. 1.
Two lines of the TI population, having distinct phenotypes in
grain size and different genotypes in the interval RM6100–
RM228 on the long arm of rice chromosome 10, were
selected and crossed. 120 F2 plants were produced and
assayed using the four markers in the qGS10 region, i.e.,
RM6100, RM3773, RM3123, and RM228. Plants that were
heterozygous in all the four marker loci were identified as
residual heterozygotes (RHs) for qGS10. They were then
subjected to genotyping with 122 polymorphic SSR markers
located in other regions. One plant was selected, remained
to be heterozygous in the target region and identified to be
heterozygous and homozygous at 19 and 103 marker loci
in the background, respectively. This plant was selfed to
produce a F2-type population and then a F2:3-type population.
Teqing/IRBB lines
Two lines differing in the interval
RM6100–RM228 on chromosome 10
Cross
F1
Selfing
F2

Marker assay

A residual heterozygote (RH) for

the interval RM6100–RM228
Selfing
RH-F2
Marker assay
RH-F2:3

Marker assay
Three plants that were
heterozygous in RM6100–RM228
Selfing
New RH-F2
Marker assay
Non-recombinant homozygotes
Selfing
Three sets of NILs

Fig. 1 Construction of a residual heterozygote (RH)-derived
F2:3 population and three sets of near isogenic lines (NILs).


18

WANG Zhen et al. Journal of Integrative Agriculture 2017, 16(1): 16–26

The F2:3 population consisting of 307 individuals was used
for QTL analysis on grain size and yield traits. In the mean
time, three F3 plants that were heterozygous at the four
marker loci in the target region and at four or five marker loci
in the background were selected. New RH-F2 populations
were developed and assayed with the segregating markers. Plants showing no recombination in the target region

were identified in each population. Selfing seeds of these
plants resulted in the development of three sets of NILs, of
which each consisted of two homozygous genotypic groups
differing in the target region.

2.2. Trait measurement
All the rice populations were planted in the middle rice
growing season in the paddy field of the China National Rice
Research Institute located in Hangzhou, Zhejiang, China.
The three RIL populations were tested for two years, including 2008 and 2009 for TI, 2009 and 2010 for ZM, and 2003
and 2009 for XM. The 307 F3 families and the three NIL sets
were tested for one year in 2013 and 2015, respectively. A
randomized complete block design with two replications was
used. In each replication, one line was grown in a single row
of 12 plants, except that six-row plots with 12 plants per row
was employed for XM in 2003. The planting density was
16.7 cm between plants and 26.7 cm between rows. Field
management followed the normal agricultural practice. At
maturity, five middle plants of each row/plot were harvested
in bulk for trait measurement. The three RIL populations
were only measured for grain length and width, and the F3
families and three NIL sets were measured for seven traits
including grain length (GL), grain width (GW), 1 000-grain
weight (TGW), number of panicle per plant (NP), number
of grains per panicle (NGP), number of spikelet per panicle
(NSP), and grain yield per plant (GY). The grain length and
width were estimated by the Rice Product Quality Inspection
and Supervision Testing Center of the Ministry of Agriculture
of China according to the National Standard GB/T 178911999 (1999) for the three RIL populations, and measured
using an automatic instrument (Model SC-G, Wanshen Ltd.,

Hangzhou, China) for the RH-derived F2:3 population and
the three NIL sets.

2.3. DNA marker analysis
Total DNA was extracted following the conventional method
(Lu and Zheng 1992) for plants assayed with 126 markers
and using the mini-preparation protocol (Zheng et al. 1995)
for other plants. PCR amplification was performed according
to Chen et al. (1997). The products of the SSR markers were
visualized on 6% non-denaturing polyacrylamide gels using
silver staining. All the SSR markers were selected from the

Gramene database ( />
2.4. Data analysis
In each trial, phenotypic values of the two replications were
averaged for each line and used for data analysis. Basic
descriptive statistics, including mean trait value, standard
deviation, coefficient of variation, the minimum and maximum trait values, skewness, and kurtosis, were computed
for each population in each trial.
Linkage maps for the three RIL populations used in this
study were constructed previously (Mei et al. 2013), in which
the genetic distance in centiMorgan (cM) was derived using
Kosambi function. The TI, ZM and XM maps are 1 197.7,
1 814.7 and 2 080.4 cM in length, consisting of 127, 256
and 240 DNA markers, respectively (Appendix A). All the
12 chromosomes are well-covered in the ZM and XM maps,
whereas the major segment of chromosomes 1 and 4 are
un-covered in the TI map due to parental monomorphism.
As compared to the ZM, the map distance is generally expanded in XM and compressed in TI. For the RH-derived
F2 population, Mapmaker/Exp 3.0 (Lander et al. 1987) was

used for map construction, with the genetic distances in cM
also derived using the Kosambi function.
For the three RIL populations in which the phenotypic
data were available for two years, QTL analysis was performed using QTL Network 2.0 (Yang et al. 2008). Critical
F values for genome-wise type I error were calculated with
1 000 permutation test and used for claiming a significant
event. Significant level of P<0.05 was used for candidate
interval selection, putative QTL detection and QTL effect
estimation. The phenotypic variance explained (R2) by a single QTL or genotype-by-environmental (GE) interaction, as
well as the overall R2 jointly explained by all the QTLs or GE
interactions detected for a given trait in a given population,
were calculated by Markov Chain Monte Carlo algorithm. In
the genome scan, testing window of 10 cM, filtration window
of 10 cM and walk speed of 1 cM were chosen. QTLs were
designated following the rules proposed by McCouch and
CGSNL (2008).
For the RH-F2:3 population in which phenotyping was
conducted for the F3 families in one year, QTL analysis was
performed with composite interval mapping (CIM) in Windows QTL Cartographer 2.5 (Wang et al. 2012a). Critical
LOD values for genome-wise Type I error of P<0.05 were
determined with 1 000 permutation test.
For the NIL populations, two-way ANOVA were performed
to test phenotypic differences between the two homozygous genotypic groups in each NIL set. The analysis was
performed using SAS procedure general linear model
(GLM) (SAS Institute 1999) as previously described (Dai
et al. 2008). Given the detection of a significant difference


19

WANG Zhen et al. Journal of Integrative Agriculture 2017, 16(1): 16–26

(P<0.05), the same data were used to estimate the genetic
effect of the QTL, including additive effect and the proportion
of phenotypic variance explained.

QTLs GS3 (Fan et al. 2006) and qSW5/GW5 (Shomura et al.
2008; Weng et al. 2008), respectively. R2 values of other
QTLs were much smaller, ranging from 0.31 to 4.89% for
GL and 0.93 to 4.37% for GW.
In the ZM population, eight QTLs for GL and four QTLs
for GW were detected. For GL, qGL3.1 and qGL3.3 had
the highest two R2 of 14.40 and 11.18%, followed by the
R2 of 8.11% contributed by qGL6 located in the interval
RZ398–RM204 on the short arm of chromosome 6. The
five other QTLs for GL had R2 ranging from 0.64 to 5.39%.
For GW, qGW10.2 located in the interval RG561–RM228
on the long arm of chromosome 10 had the highest contribution of 8.04%, and the other three QTLs had R2 ranging
from 2.00 to 2.95%.
In the XM population, four QTLs for GL and two QTLs for
GW were detected. The two QTLs showing major effects in
the TI population, qGL3.2 and qGW5, had the largest R2 of
19.57 and 9.87% for GL and GW, respectively. The three
other QTLs detected for GL, qGL3.1, qGL3.3 and qGL6,
had R2 ranging from 1.32 to 3.32%. The remaining QTL
that was detected for GW, qGW3.3, contributed 1.00% to
the phenotypic variance.
The seven QTLs shared by different populations were
screened to identify candidate regions for validation. Three
QTLs which coincided in position with those that have been

cloned were excluded. They were major QTLs qGL3.2 and
qGW5 described above, and qGW3.2 which matched the
cloned QTL qGL3/GL3.1 (Qi et al. 2012; Zhang et al. 2012)
and shared by the TI and ZM populations. Three other QTLs
were detected to have relatively large and small effects in
the ZM and XM populations, respectively. The qGL3.1 and
qGL3.3 having opposite allelic directions were located in
either side of the qGL3.2 region, providing candidates for
separating multiple QTLs for the same trait. The qGL6 was

3. Results
3.1. QTLs for grain length and width detected in three
RIL populations
The three RIL populations were each tested for two years
in the middle rice season in Hangzhou, China. Descriptive
statistics of grain length (GL) and grain width (GW) in the six
trials are presented in Table 1. The two traits were always
continuously distributed with low skewness and kurtosis,
showing typical pattern of quantitative variation. In the ZM
population, the coefficients of variation in two years were
0.057 and 0.056 for GL, and 0.039 and 0.041 for GW,
which were much lower than the values of 0.075–0.079 for
GL and 0.064–0.073 for GW in the two other populations,
respectively. It is noted that the parental differences were
also smaller for ZM than the two other populations.
QTLs and GE interactions were determined with QTL
Network 2.0 (Yang et al. 2008), in which year was taken
as the environmental factor. A total of 18 QTLs for GL and
13 QTLs for GW were detected in the three populations,
of which none showed significant GE interaction. Among

these QTLs, four for GL and three for GW were shared by
two populations and the others were detected in a single
population, thus the numbers of QTLs for GL and GW were
reduced to 14 and 10, respectively (Table 2).
In the TI population, six QTLs for GL and seven QTLs
for GW were detected. Notably, 56.71 and 59.51% of the
phenotypic variance of GL and GW were explained by
qGL3.2 and qGW5 which coincided in position with cloned

Table 1 Phenotypic performance of grain length (GL) and grain width (GW) in the three recombinant inbred line populations
Trait
GL (mm)

Population1)

Year

Mean

SD

CV

Range

Skewness

Kurtosis

TI

2008
2009
2003
2009
2009
2010
2008
2009
2003
2009
2009
2010

6.38
5.91
6.14
6.27
5.76
5.81
2.55
2.47
2.33
2.51
2.67
2.74

0.505
0.464
0.476

0.472
0.327
0.326
0.187
0.173
0.148
0.162
0.105
0.112

0.079
0.079
0.078
0.075
0.057
0.056
0.073
0.070
0.064
0.065
0.039
0.041

5.5–7.6
5.2–7.0
5.1–7.7
5.1–7.5
5.0–7.0
5.0–6.9
2.2–3.0

2.1–2.9
2.0–2.8
2.1–3.0
2.2–3.0
2.2–3.1

0.21
0.25
0.08
–0.05
0.64
0.50
0.43
0.40
0.36
0.33
–0.15
–0.50

–1.15
–1.17
–0.35
–0.57
1.62
0.94
–0.70
–0.65
0.12
–0.46
0.85

1.50

XM
ZM
GW (mm)

TI
XM
ZM

1)

Parents
Female
Male2)
5.9
6.7/7.0
5.3
6.4/6.7
6.2
5.7
6.3
5.6
5.7
5.7
5.7
5.5
2.8
2.5/2.3
2.7

2.2/2.2
2.1
2.5
2.4
2.6
2.6
2.6
2.8
2.6

TI, 204 lines of Teqing/IRBB lines, including 122 of Teqing/IRBB52, 77 of Teqing/IRBB59, two of Teqing/IRBB50, and one each of
Teqing/IRBB51, Teqing/IRBB54 and Teqing/IRBB55; XM, 209 lines of Xieqingzao/Milyang 46; ZM, 230 lines of Zhenshan 97/Milyang 46.
The same as below.
2)
For TI, trait values of the male parents IRBB52 and IRBB59 are listed before and after “/”, respectively.


The RH-derived F2 population
used for the first validation of
qGW10.2 was segregated at
marker loci RM6100, RM3773,
RM3123, and RM228 on the
long arm of chromosome 10.
In other regions, this population was segregated at 19
loci and homozygous at 103
loci (Fig. 2). It is noted that
the regions covering the major grain-size QTLs GS3 and
qSW5/GW5 which were segregated in the TI population had
become homozygous. In the
population consisting of 307

F2:3 lines, the coefficients of
variation for GL and GW were
estimated as 0.017 and 0.018,
respectively, much lower than
the values of 0.070–0.079 in
the original population TI. This
was in agreement with the removal of segregation at major
QTLs GS3 and qSW5/GW5 in
the new population.
Results of QTL analysis using the F2:3 population are presented in Table 3. In the target region RM6100–RM228,

<0.0001
<0.0001
<0.0001

<0.0001

RM3773–RM3123

0.0001
<0.0001

RM5672–RM3859
RM242–RM107

RM437–RM18038
RM190–RM587
RM5647–RM25

<0.0001

RM18038–RM18189

<0.0001
0.0001
<0.0001

<0.0001

RM15139–RM15303

RM13576–RM263
RM232–RM15139
RM16–RM15644

<0.0001
<0.0001

RM12178–RM12210
RM71–RM327

TI population2)
P

–0.027

–0.158
–0.035
–0.025

0.030
–0.025
–0.027

0.044
–0.067

0.115

0.430

0.063
0.069

A

2.62

59.51
3.55
0.96

0.93
4.37
1.20

0. 31
4.89

3.95

56.71

3.49
0.49

R2 (%)

RM5348–RM1859
RG561–RM228

RZ403–RM168

RM171–RM1108
RM246–RG101

RM274–RZ225
RZ398–RM204

RZ328–RZ575
RM401–RM3643

RM218–RM232

<0.0001
<0.0001

<0.0001

<0.0001

<0.0001

<0.0001
<0.0001

<0.0001
0.0002

<0.0001

0.024
–0.038

–0.026

–0.071
–0.025

0.098
–0.092

0.113
0.055

–0.097

ZM population
Interval
P
A

RM1195–RM5359 <0.0001 –0.051
RM294A–RM294B <0.0001 –0.058

2.00
8.04

2.95

5.01
2.35

5.09
8.11

14.40
0.64

11.18

R2 (%)
4.73
5.39

XM population
P
A

R2 (%)

RZ575–R1927

CDO82–RG182

RM204–RM197

<0.0001 –0.034
<0.0001 0.082

<0.0001 –0.086

1.00
9.87

1.32

RZ517–RZ399
0.0001 –0.080 3.32
RZ696–RG445A <0.0001 –0.275 19.57
RZ519–RZ328
<0.0001 0.132 2.06

Interval

qSW5/GW5

GS3
qGL3/GL3.1

qSW5/GW5

GS3

Cloned
QTL3)

2)

QTLs are designated as proposed by McCouch and CGSNL (2008). The same as below.
A, additive effect of replacing a maternal allele by a paternal allele. Positive value, male>female; negative value, maleQTL. The same as below.
3)
Cloned QTLs located in the given region.

1)

GW

qGL1.1
qGL1.2
qGL1.3
qGL2
qGL3.1
qGL3.2
qGL3.3
qGL4
qGL5.1
qGL5.2
qGL6
qGL7
qGL9
qGL10

qGW1
qGW2
qGW3.1
qGW3.2
qGW3.3
qGW5
qGW6
qGW8
qGW10.1
qGW10.2

GL

Interval

3.2. QTLs for grain size and
yield traits detected in a
RH-derived F2:3 population

QTL1)

located in the region harboring
florigen genes Hd3 and RFT1
(Tsuji et al. 2011), providing
candidates for determining the
pleiotropism of Hd3 and RFT1.
The remaining common QTL,
qGW10.2, was located in
a region where no QTL for
grain size has been cloned,

contributing 2.62% to the phenotypic variance in TI under
the segregation of major QTL
qGW5 and having the largest
R2 of 8.04% in ZM. This QTL
was taken for validation using
populations with more homogenous background.

Trait

Table 2 QTLs detected for grain length and grain width in three RIL populations

20
WANG Zhen et al. Journal of Integrative Agriculture 2017, 16(1): 16–26


21

WANG Zhen et al. Journal of Integrative Agriculture 2017, 16(1): 16–26

Teqing homozygous region

Heterozygous region

QTLs detected in the TI population:
Mb

1

3
RM110

RM236
RM3732

4
RM35
RM6466
RM23

8
12

RM71
Centromere

Centromere
RM24

16

24

RM327NIL2&3
RM262NIL2&3
RM13495
RM13576
RM263

28

RM6

20

GS3

RM16252

RM6301
RM14629

RM335NIL1 GW5
RM16335NIL1qSW5

RM218
RM232

Centromere

RM153
RM611
RM13
RM592
RM437
RM18038
RM18189
RM249
Centromere

6

RM146
RM164NIL1&2

RM15303
RM16

RM15717
RM15935
RM6759
RM16048
RM570NIL3

RM303
RM3474
RM6992
RM349
RM3333

RM469
RM589
RM190
RM587
RM584
RM6119
RM276
RM549
RM3330
Centromere
RM7193
RM3827

RM18927
RM3321
RM274
RM334

RM15644

RM207

RM11869

36

5

4
RM14302

RM15139
Centromere

RM240

32

Monomorphic region

Grain width

Grain length

2

0

IR24 homozygous region

RM20361
RM20591
RM340
RM20731

RM12178NIL2
RM12210

40
44
7

9

8
RM1243
RM5672
RM3859
Centromere
RM214
RM11
RM10

RM70
RM18

RM5647
RM25
RM310
RM547
RM22755
Centromere
RM23001
RM210
RM23325
SR11
RM264

10
RM23662
RM5688
Centromere
RM8206
RM219
RM524
RM1896
RM566
RM434
RM242
RM107
RM1026

11

RM24992
RM216
Centromere
RM3152
RM1859
RM1375
RM6704
RM6100
RM3773
RM3123
RM228

12
RM3863
RM2459
RM167

Centromere
RM287NIL1

RM20
RM27610
RM3246
YL155
Centromere
RM511

pTA248
RM254
RM224

RM28313

RM1233
RM5926

RM12NIL2&3

RM28597

Fig. 2 Genotypic composition of the RH-derived F2 population. In the three sets of near isogenic lines (NILs) derived from progenies
of this population, a few markers in the background region remained to be segregated. Superscripts NIL1, NIL2 and NIL3 are
attached to the background markers segregated in the given NIL population.

qGW10.2, qGL10 and qTGW10 associated with the three
traits for grain size were detected, explaining 23.88, 14.35
and 11.58% of the phenotypic variance for GW, GL and
TGW, respectively. The enhancing alleles of the three QTLs
were all derived from Teqing, the same as what was found
for qGW10.2 in the TI population. While the additive effects
of 0.025 and 0.027 mm estimated for qGW10.2 in the two
populations were almost identical, the R2 was increased from
2.62 to 23.88%. For ease of description, the three putative
QTLs for grain size detected in this region were integrated
as qGS10. Regarding the other four traits analyzed, qGS10
was found to affect NGP only, with the enhancing allele

derived from IR24.
Additional QTLs were detected in seven other regions,
of which none had significant effects in the TI population.

Three of the regions were found to affect two or more traits.
The RM14302–RM6301 interval at the top of chromosome 3
showed significant effects on the three traits for grain size,
with R2 ranging from 4.04 to 4.93%. The RM146–RM164
interval on the long arm of chromosome 5 was associated
with the other four traits, with R2 ranging from 4.77 to 11.75%.
RM216 on the short arm of chromosome 10 was shown to
influence grain width and spikelet number, explaining 4.13
and 3.93% of the phenotypic variance, respectively. Four


22

WANG Zhen et al. Journal of Integrative Agriculture 2017, 16(1): 16–26

Table 3 QTLs for grain size and yield traits detected in a RH-derived F2:3 population
Chr.
3

4
5

10

11
12

Interval
RM14302–RM6301
RM14302–RM6301

RM14302–RM6301
RM16252–RM335
RM3321
RM146–RM164
RM146–RM164
RM146–RM164
RM146–RM164
RM216
RM216
RM3123–RM228
RM3773–RM3123
RM3773–RM3123
RM6100–RM3773
RM287–pTA248
RM28597

QTL
qGW3
qGL3
qTGW3
qNSP4
qTGW5
qNP5
qNGP5
qNSP5
qGY5
qGW10.1
qNSP10
qGW10.2
qGL10

qTGW10
qNGP10
qNGP11
qGY12

LOD
4.62
3.23
5.25
2.97
3.97
3.57
9.96
8.81
3.42
3.77
3.02
19.07
9.46
8.29
3.44
2.76
3.13

A1)
–0.014
–0.032
–0.26
–0.85
–0.20

–0.23
5.84
4.96
0.83
0.011
2.43
–0.025
–0.068
–0.34
3.27
3.61
0.31

D2)
–0.005
–0.016
–0.01
–4.22
–0.04
0.27
–0.00
–1.80
0.13
0.001
–2.24
0.006
0.007
–0.04
–1.65
1.20

1.10

D/[A]3)
–0.38
–0.49
–0.05
–4.98
–0.18
1.19
–0.00
–0.36
0.15
0.06
–0.92
0.25
0.11
–0.13
–0.50
0.33
3.58

R2 (%)
6.93
4.04
6.42
4.33
4.95
5.22
11.59
11.75

4.77
4.13
3.93
23.88
14.35
11.58
4.12
4.53
4.26

1)

Additive effect of replacing a Teqing allele by an IR24 allele. The same as below.
Dominance effect.
3)
Degree of dominance.
2)

other regions which covered RM16252–RM335, RM3321,
RM287–pTA248, and RM28597 on chromosomes 4, 5, 11
and 12, respectively, were each detected for a single trait
with R2 ranging as 4.26–4.95%.

3.3. QTLs for grain size and yield traits detected in
the three NIL sets
Three sets of NILs, each consisting of two homozygous
genotypes differing in the target interval RM6100–RM228,
were used to further confirm the effects of qGS10 on grain
size. They were each descended from a F3 plant of the F2:3
population described above. Four or five markers in the

background region remained to be segregated in each NIL
set (Fig. 2), of which none was found to be associated with
the grain-size traits in the F2:3 population.
The seven traits analyzed in the previous study were
measured using the three NIL populations. All the traits
were continuously distributed, but one-gene segregating
mode was observed for the three traits for grain size. Distributions of the two genotypes were largely discrete for GW
in NIL1 and NIL2, for GL in NIL2 and NIL3, and for TGW in
all the three populations, in which the Teqing homozygous
lines were clustered towards to the area of higher values
and the IR24 homozygous lines to the lower values (Fig. 3).
These results indicate that allelic differences in the qGS10
region could be the major source for grain size variation in
the three NIL populations.
Influence of the genotypic difference in the qGS10 region
on the seven traits was tested with two-way ANOVA and the
results are shown in Table 4. Highly significant (P<0.01)
effects on the three traits for grain size were detected in all

the three populations except for GW in NIL3. It was also
shown that the allele derived from Teqing always increased
the trait values of GW, GL and TGW, which was in agreement
with the results generated from the two previous studies.
For GW, the significant additive effects detected in NIL1
and NIL2 were 0.031 and 0.029 mm, explaining 53.53 and
57.25% of the phenotypic variance, respectively. For the two
traits having significant variations in all the three populations,
the additive effects and R2 ranged as 0.027–0.071 mm and
16.94–65.90% for GL, and 0.42–0.73 g and 26.68–63.35%
for TGW, respectively.

Significant effects of the qGS10 region on the other four
traits were only detected for NGP (P=0.0289) and NSP
(P=0.0365) in NIL3, with the enhancing allele derived from
IR24. Higher trait values on NGP and NSP of the IR24
over the Teqing allele also appeared in NIL1 and NIL2,
although no statistical significance was reached. These
results indicate that the qGS10 region has an influence
on grain number, but the effect is so small that a statistical
significance was not always achievable. It is thus concluded
that the qGS10 region had considerable effects on grain size
and minor effects on grain number with the Teqing allele increasing grain size and decreasing grain number. Trade-off
between the two traits resulted in residual enhancing effects
of the Teqing allele on grain yield, as it was shown in Table 4
that the Teqing allele tended to have a higher grain yield in
all the three populations.

4. Discussion
Due to insufficient molecular marker and high genotyping
cost in the early time of molecular mapping, segregating


WANG Zhen et al. Journal of Integrative Agriculture 2017, 16(1): 16–26

populations were commonly constructed from crosses
between two distinct rice lines, either belong to different
species (Cai et al. 2002; Li et al. 2004), subspecies (Huang
et al. 1997b; Redona and Mackill 1998) or ecological types
(Tan et al. 2000; Rabiei et al. 2004). Since many QTLs were
simultaneously segregated in a population, the influence
of individual QTLs was diluted and only those having large

effects could be consistently detected. Nowadays, marker
polymorphism and genotyping efficiency is not longer a
problem (Cobb et al. 2013), laying a solid foundation for
the detection of QTLs underlying natural variation between
genetically close-related rice cultivars.
Among the three RIL populations used in the present
study, XM and ZM were derived from crosses between
early-season indica rice Zhenshan 97 and Xieqingzao and
middle-season indica rice Milyang 46, respectively, whereas

IR24 homozygote

No. of lines

NIL1
16
14
12
10
8
6
4
2
0

2.02 2.04 2.06 2.08 2.10 2.12

No. of lines

NIL2

10
8
6
4
2
0

1.98 2.00 2.02 2.04 2.06 2.08

the female and male parents for TI were both middle-season
indica rice. A more homogeneous genetic background in
TI has resulted in the detection of more QTLs and higher
R2 values (Table 2). A QTL region detected for grain size
in the TI and ZM populations, qGW10.2/qGL10 located in a
region away from those that have been cloned, was selected
for validation. In the absence of major-QTL segregation,
the target region showed consistent effects on grain size
with single-gene segregation pattern. These results were
in support of the common understanding that removal of the
masking effect of major QTLs could facilitate the detection
of QTLs having relatively small effects (Uga et al. 2007;
Yan et al. 2014).
The QTL for grain size newly validated in this study,
qGS10, exerted pleiotropic effects on grain number albeit with much lower R2. Opposite allelic directions were

Teqing homozygote

12

12

10

10

8

8

6

6

4

4

2

2

0

6.63 6.66 6.69 6.72 6.75 6.78 6.81

0

8

8

6

6

4

4

2

2

0

6.55 6.60 6.65 6.70 6.75 6.80 6.85 6.90

0

12

12

10

10

10

8

8

8

6

6

6

4

4

4

2

2

2

No. of lines

NIL3
12

0

2.04 2.06 2.08 2.10 2.12 2.14
Grain width (mm)

0

23

6.66 6.71 6.76 6.81 6.86 6.91
Grain length (mm)

Fig. 3 Distribution of the three grain-size traits in the three NIL populations.

0

21.3 21.9 22.5 23.1 23.7 24.3

20.6 21.1 21.6 22.1 22.6 23.1 23.6

22.7 23.3 23.9 24.5 25.1 25.7
1 000-grain weight (g)


24

WANG Zhen et al. Journal of Integrative Agriculture 2017, 16(1): 16–26

Table 4 Effects of the qGS10 region detected in three NIL populations
Population
NIL1

No. of lines1)
NIL
NILIR24
23
24
Teqing

NIL2

15

12

NIL3

25

14

1)
2)

Trait2)
GW
GL
TGW
NP
NGP
NSP
GY

GW
GL
TGW
NP
NGP
NSP
GY
GW
GL
TGW
NP
NGP
NSP
GY

Mean±SD
NILTeqing
2.068±0.024
6.696±0.062
22.74±0.52
8.71±0.70
146.93±8.02
155.23±8.13
29.19±2.78
2.052±0.021
6.719±0.064
22.75±0.55
8.09±0.74
153.08±11.99
161.11±12.37

28.41±2.45
2.068±0.027
6.813±0.050
23.64±0.66
9.75±0.96
141.16±7.93
148.78±8.28
31.04±2.73

NILIR24
2.006±0.027
6.643±0.042
21.45±0.44
8.81±0.80
152.46±10.91
160.38±11.13
28.92±3.10
1.994±0.018
6.597±0.103
21.28±0.42
8.17±0.64
158.43±9.11
166.38±9.81
27.87±2.19
2.057±0.015
6.672±0.028
22.79±0.32
9.17±1.04
147.61±9.42
155.14±9.62

29.40±3.16

P

A

R² (%)

<0.0001
0.0011
<0.0001
0.6512
0.0555
0.0753
0.7583
<0.0001
0.0008
<0.0001
0.7892
0.2156
0.2349
0.5523
0.1588
<0.0001
<0.0001
0.0891
0.0289
0.0365
0.0979

–0.031
–0.027
–0.640

53.53
16.94
58.33

–0.029
–0.061
–0.730

57.25
31.24
63.35

–0.071
–0.420

65.90
26.68

3.220
3.180

6.93
5.88

NILTeqing and NILIR24 are near isogenic lines with Teqing and IR24 homozygous genotypes in the qGS10 region, respectively.
GW, grain width (mm); GL, grain length (mm); TGW, 1 000-grain weight (g); NP, number of panicle per plant; NGP, number of grain

per panicle; NSP, number of spikelet per panicle; GY, grain yield per plant (g).

observed, with the Teqing allele enhancing grain size
but reducing grain number. The qGS10 region was also
previously reported to affect grain chalkiness and endosperm transparency in the TI and ZM populations, with the
Teqing and Zhenshan 97 alleles associated with inferior
appearance quality, i.e., high grain chalkiness and low
endosperm transparency (Mei et al. 2013). In addition,
QTLs for grain size were detected in five other studies
(Huang et al. 1997b; Cai et al. 2002; Li et al. 2004; Li S Q
et al. 2011; Nelson et al. 2011). These results suggest
that the qGS10 region is a good target for studying the
genetic control of grain quality, examining the trade-off
between grain size and grain number, and analyzing the
genetic drag between grain size and appearance quality.
Moreover, qGS10 showed significant effects even between
cultivars of similar ecological adaption, such as Teqing and
IR24 used in the present study, thus it could have a broad
application in rice breeding.

in increasing the efficiency of detecting QTLs for complex
traits. One QTL, qGS10 located in the RM6100–RM228
interval on the long arm of chromosome 10, was newly
validated in this study. It was found to have a considerable
effect on grain size and a small effect on grain number.
This region was also previously detected for quality traits
in rice in a number of studies, providing a good candidate
for functional analysis and breeding utilization.

5. Conclusion

References

Using three RIL populations of indica rice, 14 QTLs for GL
and 10 QTLs for GW were detected. Seven of them were
shared by different populations, including three QTLs coinciding in position with those that have been clone and four
QTLs worthy of further investigation. Results also indicate
that a more homogeneous genetic background could result

Acknowledgements
This work was supported by the National Natural Science
Foundation of China (31521064), the Chinese 863 Program
(2014AA10A603), and a project of the China National Rice
Research Institute (2014RG003-1).
Appendix associated with this paper can be available on
/>
Cai H W, Morishima H. 2002. QTL clusters reflect character
associations in wild and cultivated rice. Theoretical and
Applied Genetics, 104, 1217–1228.
Calingacion M, Laborte A, Nelson A, Resurreccion A,
Concepcion J C, Daygon V D, Mumm R, Reinke R, Dipti S,
Bassinello P Z, Manful J, Sophany S, Lara K C, Bao J, Xie


WANG Zhen et al. Journal of Integrative Agriculture 2017, 16(1): 16–26

L, Loaiza K, El-hissewy A, Gayin J, Sharma N, Rajeswari S,
et al. 2014. Diversity of global rice markets and the science
required for consumer-targeted rice breeding. PLOS ONE,
9, e85106.

Chen X, Temnykh S, Xu Y, Cho Y G, McCouch S R. 1997.
Development of a microsatellite framework map providing
genome-wide coverage in rice (Oryza sativa L.). Theoretical
and Applied Genetics, 95, 553–567.
Cobb J, DeClerck G, Greenberg A, Clark R, McCouch S. 2013.
Next-generation phenotyping: Requirements and strategies
for enhancing our understanding of genotype-phenotype
relationships and its relevance to crop improvement.
Theoretical and Applied Genetics, 126, 867–887.
Dai W M, Zhang K Q, Wu J R, Wang L, Duan B W, Zheng K L,
Cai R, Zhuang J Y. 2008. Validating a segment on the short
arm of chromosome 6 responsible for genetic variation in
the hull silicon content and yield traits of rice. Euphytica,
160, 317–324.
Fan C, Xing Y, Mao H, Xu C, Li X, Zhang Q. 2006. GS3, a
major QTL for grain length and weight and minor QTL
for grain width and thickness in rice, encodes a putative
transmembrane protein. Theoretical and Applied Genetics,
112, 1164–1171.
GB/T17891-1999. 1999. High Quality Rice Grain. Bureau of
Quality and Technical Supervision of China. (in Chinese)
Huang N, Angeles E R, Domingo J, Singh S, Zhang G,
Kumaravadivel N, Bennett J, Khush G S. 1997a. Pyramiding
of bacterial blight resistance genes in rice: Marker-assisted
selection using RFLP and PCR. Theoretical and Applied
Genetics, 95, 313–320.
Huang N, Parco A, Mew T, Magpantay G, McCouch S,
Guiderdoni E, Xu J, Subudhi P, Angeles E R, Khush G S.
1997b. RFLP mapping of isozymes, RAPD and QTLs for
grain shape, brown plant hopper resistance in a doubled

haploid rice population. Molecular Breeding, 3, 105–113.
Ikeda M, Miura K, Aya K, Kitano H, Matsuoka M. 2013. Genes
offering the potential for designing yield-related traits in rice.
Current Opinion in Plant Biology, 16, 213–220.
Ishimaru K, Hirotsu N, Madoka Y, Murakami N, Hara N, Onodera
H, Kashiwagi T, Ujiie K, Shimizu B I, Onishi A, Miyagawa
H, Katoh E. 2013. Loss of function of the IAA-glucose
hydrolase gene TGW6 enhances rice grain weight and
increases yield. Nature Genetics, 45, 707–711.
Lander E, Green P, Abrahamson J, Barlow A, Daley M,
Lincoln S, Newburg L. 1987. MAPMAKER: An interactive
computer package for constructing primary genetic maps
of experimental and natural populations. Genomics, 1,
174–181.
Li J, Xiao J, Grandillo S, Jiang L, Wan Y, Deng Q, Yuan L,
McCouch S R. 2004. QTL detection for rice grain quality traits
using an interspecific backcross population derived from
cultivated Asian (O. sativa L.) and African (O. glaberrima
S.) rice. Genome, 47, 697–704.
Li S Q, Cui G K, Guan C R, Wang J, Liang G H. 2011. QTL
detection for rice grain shape using chromosome single
segment substitution lines. Rice Science, 18, 273–278.

25

Li Y, Fan C, Xing Y Z, Jiang Y, Luo L, Sun L, Shao D, Xu C, Li
X, Xiao J, He Y, Zhang Q. 2011. Natural variation in GS5
plays an important role in regulating grain size and yield in
rice. Nature Genetics, 43, 1266–1270.
Lu Y J, Zheng K L. 1992. A simple method for isolation of rice

DNA. Chinese Journal of Rice Science, 1, 47–48.
McCouch S R, CGSNL (Committee on Gene Symbolization,
Nomenclature and Linkage, Rice Genetics Cooperative).
2008. Gene nomenclature system for rice. Rice, 1, 72–84.
Mei D Y, Zhu Y J, Yu Y H, Fan Y Y, Huang D R, Zhuang J
Y. 2013. Quantitative trait loci for grain chalkiness and
endosperm transparency detected in three recombinant
inbred line populations of indica rice. Journal of Integrative
Agriculture, 12, 1–11.
Nelson J C, McClung A M, Fjellstrom R G, Moldenhauer K A
K, Boza E, Jodari F, Oard J H, Linscombe S, Scheffler B
E, Yeater K M. 2011. Mapping QTL main and interaction
influences on milling quality in elite US rice germplasm.
Theoretical and Applied Genetics, 122, 291–309.
Qi P, Lin Y S, Song X J, Shen J B, Huang W, Shan J X, Zhu M Z,
Jiang L, Gao J P, Lin H X. 2012. The novel quantitative trait
locus GL3.1 controls rice grain size and yield by regulating
Cyclin-T1; 3. Cell Research, 22, 1666–1680.
Rabiei B, Valizadeh M, Ghareyazie B, Moghaddam M, Ali A J.
2004. Identification of QTLs for rice grain size and shape
of Iranian cultivars using SSR markers. Euphytica, 137,
325–332.
Redona E D, Mackill D J. 1998. Quantitative trait locus analysis
for rice panicle and grain characteristics. Theoretical and
Applied Genetics, 96, 957–963.
SAS Institute. 1999. SAS/STAT User’s Guide. SAS Institute,
Cary.
Shomura A, Izawa T, Ebana K, Ebitani T, Kanegae H, Konishi
S, Yano M. 2008. Deletion in a gene associated with grain
size increased yields during rice domestication. Nature

Genetics, 40, 1023–1028.
Song X J, Huang W, Shi M, Zhu M Z, Lin H X. 2007. A QTL
for rice grain width and weight encodes a previously
unknown RING type E3 ubiquitin ligase. Nature Genetics,
39, 623–630.
Song X J, Kuroha T, Ayano M, Furuta T, Nagai K, Komeda
N, Segami S, Miura K, Ogawa D, Kamura T, Suzuki T,
Higashiyama T, Yamasaki M, Mori H, Inukai Y, Wu J,
Kitano H, Sakakibara H, Jacobsen S E, Ashikari M. 2015.
Rare allele of a previously unidentified histone H4 acetyl
transferase enhances grain weight, yield, and plant biomass
in rice. Proceedings of the National Academy of Sciences
of the United States of America, 121, 76–81.
Tan Y F, Xing Y Z, Li J X, Yu S B, Xu C G, Zhang Q. 2000.
Genetic bases of appearance quality of rice grains in
Shanyou 63, an elite rice hybrid. Theoretical and Applied
Genetics, 101, 823–829.
Tsuji H, Taoka K I, Shimamoto K. 2011. Regulation of flowering
in rice: Two florigen genes, a complex gene network, and
natural variation. Current Opinion in Plant Biology, 14,
45–52.


26

WANG Zhen et al. Journal of Integrative Agriculture 2017, 16(1): 16–26

Uga Y, Nonoue Y, Liang Z, Lin H, Yamamoto S, Yamanouchi U,
Yano M. 2007. Accumulation of additive effects generates
a strong photoperiod sensitivity in the extremely lateheading rice cultivar “Nona Bokra”. Theoretical and Applied

Genetics, 114, 1457–1466.
Wang D, Zhang X F, Zhu Z W, Chen N, Min J, Yao Q, Yan J
L, Liao X Y. 2005. Correlation analysis of rice grain quality
characteristics. Acta Agronomica Sinica, 31, 1086–1091.
Wang S, Basten C J, Zeng Z B. 2012a. Windows QTL
Cartographer 2.5. Department of Statistics. North Carolina
State University, Raleigh, NC, USA.
Wang S, Wu K, Yuan Q, Liu X, Liu Z, Lin X, Zeng R, Zhu H, Dong
G, Qian Q, Zhang G, Fu X. 2012b. Control of grain size,
shape and quality by OsSPL16 in rice. Nature Genetics,
44, 950–954.
Wang S K, Li S, Liu Q, Wu K, Zhang J Q ,Wang S A , Wang Y,
Chen X B, Zhang Y, Gao C X, Wang F, Huang H X, Fu X
D. 2015. The OsSPL16-GW7 regulatory module determines
grain shape and simultaneously improves rice yield and
grain quality. Nature Genetics, 47, 949–954.
Wang Y X, Xiong G S, Hu J, Jiang L, Yu H, Xu J, Fang Y X,
Zeng L J, Xu E B, Xu J, Ye W J, Meng X B, Liu R F, Chen
H Q, Jing Y H, Wang Y H, Zhu X D, Li J Y, Qian Q. 2015.
Copy number variation at the GL7 locus contributes to grain
size diversity in rice. Nature Genetics, 47, 944–948.
Weng J, Gu S, Wan X, Gao H, Guo T, Su N, Lei C, Zhang X,
Cheng Z, Guo X, Wang J, Jiang L, Zhai H, Wan J. 2008.
Isolation and initial characterization of GW5, a major QTL

associated with rice grain width and weight. Cell Research,
18, 1199–1209.
Yan B, Tondi Yacouba N, Chen J, Wang Y, Gao G, Zhang
Q, Liu X, He Y. 2014. Analysis of minor quantitative trait
loci for eating and cooking quality traits in rice using a

recombinant inbred line population derived from two indica
cultivars with similar amylose content. Molecular Breeding,
34, 2151–2163.
Yan W H, Liu H Y, Zhou X C, Li Q P, Zhang J, Lu L, Liu T M,
Liu H J, Zhang C J, Zhang Z Y, Shen G J, Yao W, Chen
H X, Yu S B, Xie W B, Xing Y Z. 2013. Natural variation in
Ghd7.1 plays an important role in grain yield and adaptation
in rice. Cell Research, 23, 969–971.
Yang J, Hu C, Hu H, Yu R, Xia Z, Ye X, Zhu J. 2008. QTL
Network: Mapping and visualizing genetic architecture of
complex traits in experimental populations. Bioinformatics,
24, 721–723.
Zhang X, Wang J, Huang J, Lan H, Wang G, Yin C, Wu Y, Tang
H, Qian Q, Li J, Zhang H. 2012. Rare allele of OsPPKL1
associated with grain length causes extra-large grain and a
significant yield increase in rice. Proceedings of the National
Academy of Sciences of the United States of America, 109,
21534–21539.
Zheng K L, Huang N, Bennett J, Khush G S. 1995. PCRbased Marker-Assisted Selection in Rice Breeding:
IRRI Discussion Paper Series No.12. International Rice
Research Institute, Los Banos, Phillipines.
(Managing editor WANG Ning)