Bd

PMa

b

c

a

ARRA

KQQSBG

1

tamoSlabp

Qt

((G

J

0h

Field Crops Research 134 (2012) 185–192

Contents lists available at SciVerse ScienceDirect

Field Crops Research

jou rn al h om epage: www.elsev ier .com/ locate / fc r

ulk segregant analysis: “An effective approach for mapping consistent-effectrought grain yield QTLs in rice”

rashant Vikrama, B.P. Mallikarjuna Swamya, Shalabh Dixita, Helaluddin Ahmeda,1,.T. Sta Cruza, Alok K. Singhb, Guoyou Yec, Arvind Kumara,∗

Plant Breeding Genetics & Biotechnology Division, International Rice Research Institute, DAPO Box 7777, Metro Manila, PhilippinesDepartment of Genetics and Plant Breeding, V.B.S. Purvanchal University, Jaunpur, IndiaCrop Research Informatics Laboratory, International Rice Research Institute, DAPO Box 7777, Metro Manila, Philippines

r t i c l e i n f o

rticle history:eceived 26 February 2012eceived in revised form 30 May 2012ccepted 30 May 2012

eywords:TLsuantitative trait locielective genotyping

a b s t r a c t

Mapping QTLs for grain yield under drought in rice involves phenotyping and genotyping of large map-ping populations. The huge cost incurred in genotyping could be considered as a bottleneck in theprocess. Whole population genotyping (WPG), selective genotyping (SG), and bulk segregant analysis(BSA) approaches were employed for the identification of major grain-yield QTLs under drought in rice inthe past few years. The efficiency of different QTL mapping approaches in identifying major-effect grain-yield QTLs under drought in rice was compared using phenotypic and genotypic data of two recombinantinbred line populations, Basmati334/Swarna and N22/MTU1010. All three genotyping approaches wereefficient in identifying consistent-effect QTLs with an additive effect of 10% or more. Comparative anal-

ulk segregant analysisrain yield under drought

ysis revealed that SG and BSA required 63.5% and 92.1% fewer data points, respectively, than WPG inthe N22/MTU1010 F3:4 mapping population. The BSA approach successfully detected consistent-effectdrought grain-yield QTLs qDTY1.1 and qDTY8.1 detected by WPG and SG. Unlike SG, BSA did not lead to anupward estimation of the additive effect and phenotypic variance. The results clearly demonstrate thatBSA is the most efficient and effort saving genotyping approach for identifying major grain yield QTLs

under drought.

. Introduction

Advances in molecular marker technology have enabled fast-rack improvement of crop plants in recent years. Marker-basedpproaches, including marker-assisted backcrossing and QTL pyra-iding, have been applied in cereals for improving the tolerance

f biotic and abiotic stresses (Collard and Mackill, 2008; Ye andmith, 2010). Drought is an important abiotic stress causing hugeosses in rice yields. Progress in breeding rice for drought toler-

nce could be made more efficient by applying marker-assistedreeding (Bernier et al., 2007). To unveil the genetic basis of a com-lex trait such as grain yield under drought, it is a prerequisite to

Abbreviations: BSA, bulk segregant analysis; MAS, marker-assisted selection;TLs, quantitative trait loci; SG, selective genotyping; WPG, whole population geno-

yping.∗ Corresponding author. Tel.: +63 2 580 5600x2586; fax: +63 2 580 5699.

E-mail addresses: p.vikram@irri.org (P. Vikram), m.swamy@irri.orgB.P.M. Swamy), s.dixit@cgiar.org (S. Dixit), helaluddinahmed@hotmail.comH. Ahmed), m.stacruz@cgiar.org (M.T.S. Cruz), alok6619@gmail.com (A.K. Singh),.Ye@cgiar.org (G. Ye), akumar@cgiar.org (A. Kumar).1 Present address: Plant Breeding Division, Bangladesh Rice Research Institute,

oydebpur, Gazipur 1701, Bangladesh.

378-4290/$ – see front matter © 2012 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.fcr.2012.05.012

© 2012 Elsevier B.V. All rights reserved.

genotype and phenotype mapping populations consisting of a largenumber of individuals. The costs associated with large-scale phe-notyping and genotyping are a bottleneck to conducting a study onseveral populations at any particular time in identifying QTLs withconsistent effects across multiple environments and genetic back-grounds. With little alternative available to phenotype in differentseasons/environments, the costs involved with genotyping couldbe effectively reduced by successfully applying different genotyp-ing approaches.

In a model cereal crop such as rice (Oryza sativa) with a genomesize of 389 Mb, it is a common practice to use 200–400 lines (recom-binant inbred lines, backcross inbred lines, doubled haploids) as amapping population. A dense map covering all 12 chromosomeswith an average genetic distance of 10–15 cM has been developedto identify the precise QTLs. In most of the QTL mapping experi-ments, the whole genome is scanned (Gomez et al., 2010; Lanceraset al., 2004; Xu et al., 2005) to ultimately find only a few sig-nificant markers showing association with the trait under study.To reduce the extra burden and costs associated with genotyp-

ing of large mapping populations, alternative strategies have beenproposed. A trait-based genotypic analysis called “selective geno-typing” (SG) was suggested by Lebowitz et al. (1987), in whichprogenies are categorized into distinct classes based on the trait

1 s Rese

vctTysw2acTaoleuBhtaKa1w2

cbBauapteesdSt(

SdtaSfvbdae

2

I3ddl2pB

86 P. Vikram et al. / Field Crop

alues and marker allele frequencies are compared between thelasses. In this approach, a subset of the population is used for geno-yping instead of the whole population (Lander and Botstein, 1989).his approach was used for mapping major-effect drought grain-ield QTL qDTY12.1 using only 169 (38.7%) lines from a populationize of 436 (Bernier et al., 2007); later, this QTL was reconfirmedith only 4.5% of the lines from the whole population (Navabi et al.,

009). Another time and effort saving approach is bulk segregantnalysis (BSA) (Michelmoore et al., 1991). In BSA, DNA of progeniesorresponding to the phenotypic extremes is extracted and pooled.herefore, only two pools of extreme lines along with the parentsre genotyped for the identification of markers linked with the traitf interest. BSA was first applied to the identification of markersinked with disease resistance. Initially, it was applied to the dis-ases in which resistance was mostly governed by major genes,sually qualitative in nature (Michelmoore et al., 1991). Recently,SA has been applied for quantitative traits also such as QTLs foreat tolerance in rice (Zhang et al., 2009), salt tolerance in Egyp-ian cotton (El-Kadi et al., 2006), and drought tolerance in wheatnd maize (Altinkut and Gozukirmizi, 2003; Quarrie et al., 1999;anagaraj et al., 2010; Venuprasad et al., 2009; Vikram et al., 2011),s well as QTLs for grain yield under drought in maize (Quarrie et al.,999). BSA has been used for the identification of QTLs associatedith high grain yield under drought in rice (Venuprasad et al., 2009,

011; Vikram et al., 2011).Whole population genotyping involves markers from all the 12

hromosomes so that they represent whole genome and geneticackground is taken in to account while QTL analysis. Contrarily,SA and SG (selective genotyping) approaches may not considerll recombination options during QTL identification due to thenavailability of additional marker data on the same chromosomend other chromosomes. However, these approaches have beenroven to be powerful enough to identify major QTLs that are wor-hy for MAS (Bernier et al., 2007; Venuprasad et al., 2009; Vikramt al., 2011). BSA and SG might fail to detect QTLs with smallerffects because only the extreme or tail lines are used for analy-is. Also, interactions between different loci are least likely to beetected through these approaches. QTL analysis through BSA orG approaches does not takes in to account the background so thathey might tilt upward the magnitude of phenotypic variance, LODlogarithm of odds) score, and additive effects.

The study was undertaken to perform a comparative analysis ofG and BSA in identifying major-effect QTLs for grain yield underrought and compare the fluctuations in phenotypic variance, addi-ive effect, LOD, and level of significance values obtained in BSAnd SG compared with WPG (whole population genotyping). WPG,G, and BSA approaches were compared in a population derivedrom a traditional variety (Basmati334) and a popular lowland riceariety (Swarna). Comparative analysis of these approaches mighte biased if only one population is taken into account. To vali-ate the results and perform a comprehensive analysis, phenotypicnd genotypic data of a rice population applied to QTL mapping inarlier studies were used (Vikram et al., 2011).

. Materials and methods

The study was conducted at the International Rice Researchnstitute (IRRI), Los Banos, Laguna, Philippines. An F3:4 Basmati34/Swarna population was phenotyped for grain yield underrought and genotyped through WPG, SG, and BSA. Basmati334 is arought-tolerant local landrace whereas Swarna is a popular low-

and rice cultivar in India (Sivaranjani et al., 2010; Verulkar et al.,010). The phenotypic and genotypic data of the N22/MTU1010opulation were also used for comparison of WPG, SG, andSA (Vikram et al., 2011). Basmati334/Swarna population was

arch 134 (2012) 185–192

screened under drought stress in wet season 2008 (WS2008) andN22/MTU1010 in dry season of 2009 (DS2009). WS2008 experi-ments were shown on June 17, 2008 whereas, DS2010 experimentson December 9, 2009.

2.1. Phenotyping of Basmati334/Swarna population

The F3:4 Basmati334/Swarna RIL population was phenotypedfor grain yield under drought stress as well as irrigated/non-stressconditions (Supplementary Table 1). The numbers of lines usedfor drought screening were 204 during WS2008 and 367 duringDS2010. The 204 lines used in WS2008 were a subset of 367 linesthat were phenotyped in DS2010. Screening for grain yield underdrought in the lowland rice ecosystem was carried out at IRRI usinga standard phenotyping protocol (Kumar et al., 2008; Venuprasadet al., 2007). Data for grain yield (g m−2 converted to kg ha−1), daysto 50% flowering (DTF), plant height (PH), biomass (BIO), and har-vest index (HI) were recorded.

2.2. Genotyping of Basmati334/Swarna population

Leaf samples were collected from a whole F4 plot of the non-stress experiment and bulked. DNA was extracted by the modifiedCTAB (cetyl tri-methyl ammonium bromide) method Murray andThompson, 1980). DNA was quantified on 0.8% agarose gel to adjustthe concentration to 25 ng �L−1. Quantified DNA was subjected toPCR amplification with a 15-�L reaction mixture involving 50 ngDNA, 1× PCR buffer, 100 �M dNTPs, 250 �M primers, and 1 unitTaq polymerase enzyme. Eight percent non-denaturing PAGE wasused for the resolution of PCR products (Sambrook et al., 1989). Aparental polymorphism survey between Basmati334 and Swarnawas carried out with 880 simple sequence repeat (SSR) markers(Temnykh et al., 2001; McCouch et al., 2002; IRGSP, 2005). Geno-typing of the Basmati334/Swarna population of 367 lines was doneby 71 polymorphic SSR markers spread throughout the 12 chromo-somes of the rice genome. BSA was also done for QTL identificationwith only 10% of the lines from the population (5% with high yieldand 5% with low yield under drought stress). Grain yield data fromthe stress trial of DS2010 were used for selecting lines to consti-tute BSA. DNA of all the selected lines was quantified and bulkedin equal quantity to prepare high- and low-yield bulks. DNA ofthese two bulks was screened with 203 polymorphic SSR mark-ers along with parents – Basmati334 and Swarna (Fig. 1). A similarprocedure for BSA in the N22/MTU1010 population was carried out(Vikram et al., 2011). Polymorphic markers from all the 12 chromo-somes were selected at equal distance so that they represent wholegenome. SG was carried out with the same number of markers usedin WPG. Both RILpulations-Basmati334/Swarna and N22/MTU1010presented in the study were analyzed through whole genome andBSA both.

2.3. Statistical analysis

Statistical analysis was performed through SASv9.1.3 (SASInstitute Inc., 2004). The REML algorithm of PROC MIXED of SASv9.1.3 was used for the determination of mean values of entriesusing a model in which lines were treated as a fixed effect andreplications and blocks within replications as random. Inbred linemean-based broad-sense heritability (H) was computed as:

Vg

H =

Vg + Ve/r

where Vg is genetic variance, Ve is error variance, and r is the num-ber of replications.

P. Vikram et al. / Field Crops Research 134 (2012) 185–192 187

r iden

2

wm1o((mIsgdppwfiiamwiw

2

wal2lciRa

y

Fig. 1. Bulked segregant analysis strategy used fo

.4. Linkage map construction and QTL analysis

A genetic linkage map of the Basmati334/Swarna populationas constructed through Mapdisto software using the Kosambiap function and an LOD score of 3.0 as the threshold (Kosambi,

944; Lorieux, 2007). All the non-linked markers were placedn their respective chromosomes based on the physical positionIRGSP, 2005). A similar procedure was followed by earlier workersAmrawathi et al., 2008). QTL network software v2.1 was used for

ixed model-based composite interval mapping (Yang et al., 2008).n this procedure, candidate marker intervals were determined andelected intervals were used as a co-factor in a one-dimensionalenome scan. The significance level for the determination of candi-ate intervals, putative QTL detection, and QTL effects was set at arobability level of 0.05. The threshold was determined using 1000ermutations. For the genome scan, a window size of 10 cM andalk speed of 1 cM were followed. QTL analysis results were con-rmed with QGene software to validate the magnitudes of the QTLs

dentified through three different strategies. Phenotypic variancend LOD of the QTLs were estimated through composite intervalapping using QGene software (Nelson, 1997). A LOD value of 2.5as considered as threshold to determine significance of QTLs. Sim-

lar threshold value was used because all QTLs used in this analysisere found significant in analysis through QTL network software.

.5. Analysis for SG and BSA

QTL analysis with a lesser number of lines in both populationsas carried out to work out the efficiency of SG and BSA. For

nalysis through SG, 36.5% of the lines (10% highest yielding, 10%owest yielding, and 16.5% random lines) were used (Bernier et al.,007). The linkage map constructed using WPG was also used to

ocate the QTLs identified by SG analysis. QTL analysis for BSA wasarried out with three markers – RM210, RM223, and RM339 –n the Basmati334/Swarna population and with eight markers –

M315, RM431, RM212, RM3825, RM11943, RM12023, RM12091,nd RM12146 – in the N22/MTU1010 population.

To estimate the size of population for QTL analysis for grainield under drought, QTL analysis with 90%, 80%, 70%, 60%, and

tifying QTLs for grain yield under drought in rice.

50% population size was performed in both populations. For thispurpose, a whole population was sorted based on grain yield underdrought. Equal numbers of individuals were omitted from high-yielding lines, low-yielding lines, and lines with medium yield fromthe stress trials to make sure that the lines selected were randomand distribution remained normal.

3. Results

The results of QTL identification for grain yield, days to 50% flow-ering, plant height, and harvest index under drought by WPG inthe Basmati334/Swarna and N22/MTU1010 populations are pre-sented in Supplementary Tables 2 and 3. QTLs for grain yield underdrought identified through WPG, SG, and BSA in both populationsare presented in Table 1.

3.1. QTLs identified for grain yield under drought with WPG

A consistent-effect QTL (qDTY8.1) for grain yield under droughtlocated between the marker intervals RM339 and RM210 on chro-mosome 8 was detected in the Basmati 334/Swarna population.This QTL explained phenotypic variance (R2) of 15.6% and 7.1%,with an additive effect (additive effect as percent of trial mean)of 16.8% and 22.6%, during WS2008 (wet season 2008) and DS2010(dry season 2010), respectively (Table 1). Two QTLs, qDTY1.1 andqDTY10.1, were found significant for grain yield under drought inthe N22/MTU1010 population. qDTY1.1 explained phenotypic vari-ance of 11.9% and 4.5%, with an additive effect of 18.1% and 10.0%during DS2009 and DS2010, respectively. qDTY10.1 explained phe-notypic variance of 5.2% and 3.2% and an additive effect of 13.2%and 12.6% during DS2009 and DS2010, respectively (Table 1).

3.2. QTLs identified for grain yield under drought with SG

qDTY8.1 in Basmati334/Swarna and qDTY1.1 in N22/MTU1010

were detected by the SG approach. The QTL qDTY8.1 explained phe-notypic variance of 21.9% and 16.5%, with an additive effect of25.4% and 39.8% during WS2008 and DS2010, respectively (Table 1).The QTL qDTY1.1 explained phenotypic variance of 17.7% and 12.9%

188 P. Vikram et al. / Field Crops Research 134 (2012) 185–192

Table 1QTLs for grain yield identified through SG and BSA and effect of WPG, SG, and BSA approach on F-value, additive effect, LOD, and R2.

Population Approach Season QTL Chr. Peak intervala F-value AE (%)c LOD R2

Basmati 334/Swarna

WPG WS2008 qDTY8.1 8 RM339–RM210 14.1 −16.8b 7.1 15.6DS2010 15.4 −22.6 5.1 7.1

SG WS2008 qDTY8.1 8 RM339–RM210 10.7 −25.4 4.3 21.9DS2010 15.8 −39.8 5.6 16.5

BSA WS2008 qDTY8.1 8 RM223–RM210 9.5 −12.2 2.2 4.9DS2010 21.1 −20.8 4.7 5.8

N22/MTU1010

WPG DS2009 qDTY1.1 1 RM11943–RM12091 36.7 18.1 9.3 11.9DS2010 18.5 10.0 3.6 4.5

WPG DS2009 qDTY10.1 10 RM216–RM304 19.9 −13.2 3.9 5.2DS2010 15.7 −12.6 2.6 3.2

SG DS2009 qDTY1.1 1 RM11943–RM431 22.8 28.8 5.6 17.7DS2010 15.3 10.3 3.9 12.9

SG DS2009 qDTY1.1a 1 RM12091–RM12146 – – – –DS2010 16.1 10.8 2.8 9.4

BSA DS2009 qDTY1.1 1 RM11943–RM431 34.7 18.3 9.3 11.9DS2010 13.9 10.1 3.6 4.5

phenona.

effect

dtD

3

wgvqarwNopwD

3a

eNwWiqaaS(tewliwmta

a Peak interval, peak of the QTL interval was between these marker intervals. R2,b Negative sign indicates that allele contribution is from susceptible parent Swarc AE% (additive effect %), additive effect as the percentage of trial mean [additive

uring DS2009 and DS2010, respectively. Additive effects con-ributed by qDTY1.1 were 28.8% and 10.3% during DS2009 andS2010, respectively.

.3. QTLs identified for grain yield under drought with BSA

qDTY8.1 in Basmati334/Swarna and qDTY1.1 in N22/MTU1010ere detected by BSA. RM210 was identified to be associated with

rain yield under drought in the Basmati334/Swarna populationia BSA. Three markers were run on the whole population to detectDTY8.1. It explained phenotypic variance of 4.9% and 5.8% anddditive effects of 12.2% and 20.8% during WS2008 and DS2010,espectively (Table 1). In the case of qDTY1.1, RM315 and RM431ere detected as linked markers for grain yield under drought in the22/MTU1010 population. Eight polymorphic markers were runn the whole population to determine QTL boundaries (Fig. 1). Thehenotypic variance explained by qDTY1.1 in DS2009 and DS2010as 11.9% and 4.5%, respectively. Additive effects in DS2009 andS2010 were 18.3% and 10.1%, respectively (Table 1).

.4. Magnitude and efficiency of QTLs in different genotypingpproaches

All three approaches successfully detected two consistent-ffect QTLs, qDTY8.1 and qDTY1.1, in Basmati334/Swarna and22/MTU1010 populations, respectively (Table 1). qDTY10.1, whichas found in the N22/MTU1010 population, was detected only inPG. The F-value of qDTY8.1 was lesser in WS2008 and greater

n DS2010 in SG in comparison with WPG. The additive effect ofDTY8.1 was highest in SG, followed by WPG and BSA in WS2008nd DS2010. The LOD value was highest in WPG and SG in WS2008nd DS2010, respectively. The phenotypic variance was highest inG, followed by WPG and BSA in WS2008 and DS2010, respectivelyTable 1). Comparative analysis of the parameters of qDTY1.1 revealshat the F-value is lower in SG than in WPG (Table 1). The additiveffect in SG was highest, followed by BSA and WPG in DS2009 asell as DS2010 (Table 1). It is notable that the additive effect was

east in WPG in both years. The LOD value of qDTY1.1 was the samen WPG and BSA in DS2009 and DS2010. But LOD estimated by SG

as lower during DS2009 and higher in DS2010 than LOD esti-ated with WPG and BSA. The phenotypic variance of qDTY1.1 was

he same in WPG and BSA, whereas it was greater in SG in DS2009nd DS2010 (Table 1).

typic variance explained by the QTL.

/trial mean × 100].

3.5. QTL effect analysis with different population sizes

QTL analysis for grain yield under drought was carried out withdifferent population sizes. qDTY1.1 identified with phenotypic data(under drought stress) of DS2009 with an original population sizeof 362 was also successfully detected in a lesser population size of218 lines. The same QTL was detected with phenotypic data (underdrought stress) of DS2010 in the original population but failed to bedetected in a reduced population size (Table 2). A similar trend wasobserved for qDTY8.1 in the Basmati 334/Swarna population duringDS2008. Additive effect decreased with a decrease in populationsize. The trend in DS2010 for this QTL was similar for a decreasein population size up to 30%, but thereafter it became constantup to a 50% decrease in population size and then decreased again(Table 2). The additive effect of qDTY10.1 did not show any patternwith respect to population size.

4. Discussion

Marker-assisted selection using consistent-effect drought grainyield QTLs is could be an alternative approach for improving ricegrain yield under drought stress situations (Swamy et al., 2011;Swamy and Kumar, 2011; Vikram et al., 2012). QTLs have beenidentified in the past few years for different drought-related traitsthrough phenotyping and genotyping of large mapping populations(Babu et al., 2003; Gomez et al., 2010). Phenotyping of map-ping populations under managed drought stress condition thattoo at target environments is a difficult task and involves largeexpense and efforts and is considered a significant bottleneck.Over last many years, very limited progress has been made onthe development of high-throughput, cost-effective phenotypingmethodologies. In the process, genotyping of populations with alarge number of markers for each of the 12 rice chromosomes wasfollowed. Polymorphic markers were selected at equal distanceson all the chromosomes and used for a whole-genome scan. Traitselection is another important concern in drought molecular breed-ing. QTLs for a number of physiological and morphological traitshave been identified but they have limited implications in breed-ing drought-tolerant rice varieties. Large-effect QTLs for grain yieldunder drought have been recently identified (Bernier et al., 2007;Venuprasad et al., 2009) and their introgression is in process at

IRRI (IRRI, unpublished). Genotyping of different populations andtheir screening in multiple environments could be made effectivethrough an effort and time saving genotyping strategy that can beapplied to multiple populations simultaneously. Therefore, three

P. Vikram et al. / Field Crops Research 134 (2012) 185–192 189

Table 2Comparative analysis of QTL effects in different population sizes with a decrease in population size.

Population QTL Population size Year F-value AE (%)a P-value AE% in 100% population size

Basmati334/Swarna qDTY8.1

183 (∼90%) WS2008 26.7 −15.8 <0.001 −16.8331 (∼90%) DS2010 24.1 −23.2 <0.001 −22.6162 (∼80%) WS2008 28.6 −15.6 <0.001 −16.8295 (∼80%) DS2010 18.0 −17.2 <0.001 −22.6141 (∼70%) WS2008 27.7 −14.9 <0.001 −16.8259 (∼70%) DS2010 17.1 −16.9 <0.001 −22.6120 (∼60%) WS2008 17.1 −11.3 <0.001 −16.8223 (∼60%) DS2010 11.5 −17.1 <0.001 −22.6

99 (∼50%) WS2008 13.4 −11.0 <0.001 −16.8187 (∼50%) DS2010 11.4 −17.1 <0.001 −22.6

78 (∼40%) WS2008 8.2 −9.5 0.004 −16.8151 (∼40%) DS2010 8.8 −15.7 <0.001 −22.6

N22/MTU1010

qDTY1.1

326 (∼90%) DS2009 27.9 14.1 <0.001 18.1DS2010 12.9 8.7 <0.001 10.0

290 (∼80%) DS2009 17.2 11.0 <0.001 18.1DS2010 – – – 10.0

254 (∼70%) DS2009 14.3 9.9 <0.001 18.1DS2010 – – – 10.0

218 (∼60%) DS2009 9.8 7.7 0.002 18.1DS2010 – – – 10.0

182 (∼50%) DS2009 – – – 18.1DS2010 – – – 10.0

qDTY10.1

326 (∼90%) DS2009 16.6 −10.7 <0.001 −13.2DS2010 – – – −12.6

290 (∼80%) DS2009 17.2 −9.7 <0.001 −13.2DS2010 – – – −12.6

254 (∼70%) DS2009 8.5 −9.9 <0.001 −13.2DS2010 8.4 −2.5 0.003 −12.6

218 (∼60%) DS2009 – – – −13.2DS2010 10.2 −10.0 0.002 −12.6

182 (∼50%) DS2009 9.3 −8.4 0.002 −13.2S201

d as “A

ged

4

WfQ(msppcQmrmBpt3sud(uds(h

D

a AE% (additive effect %), additive effect as the percentage of trial mean calculate

enotyping approaches, WPG, SG, and BSA, were compared for theirfficiency in identifying consistent-effect QTLs for grain yield underrought.

.1. QTL identification using WPG, SG, and BSA

qDTY8.1 was detected in the Basmati334/Swarna population byPG. The increased yield at the qDTY8.1 locus was contributed

rom the susceptible parent Swarna. Contributions of the positiveTL allele from a drought-susceptible parent were also reported

Bernier et al., 2007; Kumar et al., 2007). The positive QTL allele forajor-effect drought grain yield QTL qDTY12.1 was contributed by

usceptible parent Way Rarem (Bernier et al., 2007). A susceptiblearent (CT9993) was found to increase the grain yield of tolerantarent IR62266 (Kumar et al., 2007). qDTY1.1 was also detected suc-essfully through WPG in the N22/MTU1010 population (Table 1).TL mapping in large populations through WPG involves high cost,uch time, and much labor (Kanagaraj et al., 2010). SG has been

eported as an alternative strategy in the identification of QTLs withajor effects (Darvasi and Soller, 1992; Lander and Botstein, 1989;

ernier et al., 2007). The simulation of QTL mapping with differentopulation sizes demonstrated that QTLs explaining 15% or more ofhe phenotypic variance for population size could be detected using0 out of 500 lines with a power of 0.8 (Navabi et al., 2009). In ourtudy, the consistent-effect QTLs qDTY1.1 and qDTY8.1 explainingp to 11.9% and 15.6% of the phenotypic variance were successfullyetected through SG of 36.5% of the lines (Table 1). Shashidhar et al.2005) first applied BSA for the identification of QTLs for grain yieldnder drought in rice. Large-effect QTL qDTY3.1 for grain yield under

rought was also detected by BSA (Venuprasad et al., 2009). In ourtudy, qDTY8.1 and qDTY1.1 were successfully detected through BSATable 1; Fig. 1). BSA has been reviewed as a suitable strategy forighly heritable traits that follow Mendelian inheritance (Tuberosa

0 – – – −12.6

dditive effect/trial mean × 100′′ .

et al., 2002). However, it has also been applied in the identificationof QTLs for traits having low heritability (Grattapaglia et al., 1996).Even for traits such as grain yield under drought with moderate tohigh heritability, BSA can be used to identify the major-effect QTLs(Venuprasad et al., 2009).

4.2. Comparative analysis of WPG, SG, and BSA

Comparative analysis of the parameters of qDTY8.1 and qDTY1.1clearly indicates that the estimated additive effect was higher in SGthan in WPG (Table 1). Phenotypic variance of qDTY8.1 and qDTY1.1was also highest in SG among all three approaches (Table 1). Over-estimation of the phenotypic variance in the SG approach was alsoreported (Vales et al., 2005). qDTY8.1 and qDTY1.1 were identified athigher F-values with WPG than with SG except in DS2010 in theBasmati 334/Swarna population (Table 1). Such overestimation ofthe additive effect in SG is due to the use of a selected number ofgenotypes (Collard et al., 2005). For consistent-effect QTLs for grainyield under drought, WPG was comparable with the BSA approach.Phenotypic variance of qDTY8.1 was less in BSA than in WPG. On theother hand, this value was the same for qDTY1.1 (Table 1). For bothqDTY8.1 and qDTY1.1, the additive effect under BSA was similar tothat under WPG but lower than that under SG (Table 1).

Another QTL, qDTY10.1 in the N22/MTU1010 population, was notsignificant for grain yield under drought in SG and BSA but wassignificant in WPG (Table 1). It is interesting to note that, at the samelocus, a QTL (qDTF10.1) for days to 50% flowering under droughtstress as well as non-stress conditions were detected (Vikram et al.,2011). It is noteworthy, that the selection of lines for SG was based

on only grain yield under drought and not on days to 50% flowering.It is also reported that selection for SG should be based on only onetrait (Darvasi and Soller, 1992). In BSA, its contribution to the traitdifference between the two extreme pools (made for grain yield

1 s Research 134 (2012) 185–192

ufumSg(acd

stmtttsQiSt2

4c

twiirtcda1aea3WuTmdeyotl2inu

4

ddBgt(

Fig. 2. Graph showing variation in additive effects against population sizes. Y-axis

90 P. Vikram et al. / Field Crop

nder drought) was too small to be detected. Further, to confirm theact that a QTL effect is due to days to 50% flowering or grain yieldnder drought, covariate analysis was carried out to predict theean grain yield after keeping days to 50% flowering as a covariate.

ingle marker analysis was carried out with the predicted meanrain yield under drought and it was found to be non-significantdata not presented). This analysis confirms that qDTY10.1 is not

true QTL for grain yield under drought and the additive effectontributed by this QTL is due to earliness rather than tolerance ofrought.

Precise application of SG and BSA depends on the proper clas-ification of individuals into distinct groups that roughly reflecthe genotypes of target QTLs (QTLs to be identified). This will be

ore effective if phenotype is a good indicator of the QTL geno-ype. Genetically, the effect of the target QTL must be so large thathe collective effects of other QTLs and their interactions with thearget QTL do not result in significant distortion of the relation-hip between the QTL genotype and the phenotypic performance.TL studies suggest that a major QTL is more likely to be present

n the populations derived from contrasting parents, implying thatG and BSA will be more successful when applied to such popula-ions (Bernier et al., 2007; Venuprasad et al., 2009; Semagn et al.,010).

.3. Effect of population size in the identification ofonsistent-effect drought QTLs

The size of the mapping population to be used for the detec-ion of drought yield QTLs is an important factor for concludinghether SG and BSA are advantageous over WPG. Increasing the

nitial population size and reducing the selection proportion willncrease the difference between the two extreme groups and as aesult improve detection power. The most extreme genotypes arehose that contain desirable QTL alleles and occur in low frequen-ies in the mapping population. Population sizes used for mappingrought yield QTLs (qDTY1.1, qDTY8.1, and qDTY10.1) were plottedgainst their additive effects (Fig. 2). qDTY1.1, which showed an8.1% additive effect with population size of 362, was detected withs low as a 218 population size. This QTL showed a 10% additiveffect with population size of 362 with phenotypic data of DS2010nd could not be detected with a population size of less than26 (Table 2). qDTY8.1 showed 15.8% and 23.2% additive effect inS2008 and DS2010, respectively, and was significantly detected

sing 40% of the population sizes in respective years (Table 2).hese results reveal the importance of a large population size inapping drought yield QTLs. Several studies conducted in past for

rought related traits with population sizes less than 250 (Babut al., 2003; Lanceras et al., 2004) could not identify a drought grainield QTL with the effect, large enough to deploy in MAS. On thether hand, most of the studies which identified large and consis-ent effect QTLs involved populations with more than 300 inbredines (Bernier et al., 2007; Venuprasad et al., 2009; Vikram et al.,011). Based on the results of present study and previous reports

t could be concluded that a population size of 300–350 recombi-ant inbred lines is good enough to map major QTLs for grain yieldnder drought.

.4. BSA: an effective approach for grain yield under drought

BSA is a powerful, effort-saving, and cost-effective approach foretecting consistent-effect QTLs, even for traits showing continualistribution such as grain yield under drought. QTL mapping using

SA led to the identification of several consistent-effect QTLs forrain yield under drought at IRRI (Table 3). There are a few sampleso be genotyped, which in turn reduces the number of data pointsSupplementary Table 4). Linked markers identified via BSA can be

corresponds to the additive effects (AE%, additive effects as percentage of trial mean)and X-axis population size. (A) qDTY1.1; (B) qDTY8.1 and (C) qDTY10.1 .

confirmed by genotyping of tail lines that are also used for makingDNA pools (Kanagaraj et al., 2010). The BSA approach comparedwith WPG provides opportunities to genotype multiple populationsas revealed by the number of data points generated in BSA and WPGin our study with the N22/MTU1010 population (SupplementaryTable 4). For large-effect QTLs, it might be more important to testtheir effects in multiple backgrounds. Several QTLs for grain yieldunder drought were identified in three N22-derived populationssimultaneously because of the ease of application and effectiveness(Vikram et al., 2011).

The precision and QTL detection power of BSA can be increasedby repeating it with bulks having a large number of samples in thesecond stage. Alternatively, four DNA bulks could be used instead ofjust two for each phenotypic extreme consisting of the first and sec-ond 5% most tolerant and susceptible lines to reduce false-positivebands (Xu et al., 2000). In BSA, only a few markers were used for QTLanalysis so that there was little option to control the backgroundeffects of other segregating QTLs using a cofactor. Depending uponthe magnitude and direction of the effect of background QTLs, theeffect of the target QTL might be over- or under-estimated com-pared to conventional WPG. Therefore, BSA has a limitation indetermining the epistatic interactions for the QTLs. It is advisableto include a few markers per chromosome so that the backgroundeffect can be partially removed.

BSA could be applied with single nucleotide polymorphism(SNP) markers also. Recently, the combined use of high-density

SNPs and BSA has been applied in Arabidopsis for the identificationof QTLs associated with sulfur and selenium ionomics (Beckeret al., 2011). A large number of SNP markers can be applied onparents and bulks to pin point few regions associated with the trait

P. Vikram et al. / Field Crops Research 134 (2012) 185–192 191

Table 3Major QTLs for grain yield under drought in rice identified and validated in different populations at IRRI via BSA.

Population QTL PV/GV (%) Reference

N22/Swarna qDTY1.1 14.3 Vikram et al. (2011)N22/IR64 qDTY1.1 19.3 Vikram et al. (2011)N22/MTU1010 qDTY1.1 13.4 Vikram et al. (2011)Dhagaddeshi/Swarna qDTY1.1 32.0 Ghimire et al. (2012)Dhagaddeshi/IR64 qDTY1.1 9.3 Ghimire et al. (2012)

Apo/SwarnaqDTY2.1

a 16.0 Venuprasad et al. (2009)qDTY3.1

a 31.0 Venuprasad et al. (2009)qDTY6.1

a 66.0 Venuprasad et al. (2011)Apo/IR72 qDTY6.1

a 63.0 Venuprasad et al. (2011)Vandana/IR72 qDTY6.1

a 40.0 Venuprasad et al. (2011)Basmati334/Swarna qDTY8.1 15.6 IRRI (unpublished)IR77298-5-6-18/Sabitri qDTY3.2 10.5 IRRI (unpublished)

P

otldsm(

5

pRtoaeegi8tw

A

itQoc

A

t

R

A

A

B

V, phenotypic variance; GV, genotypic variance.a QTLs explained on the basis of genotypic variances.

f interest. This increases the probability of identifying markershat are quite closer to the target QTL with minimum unwantedinkages. Several SNP genotyping platforms have recently beeneveloped in rice (Affymetrix, Illumina, perlegen) and can beuccessfully used for QTL identification studies. In addition to QTLapping, BSA has been applied to transcriptomics in recent years

Pandit et al., 2010; Kadam et al., 2012).

. Conclusions

Identifying genomic regions associated with drought requireshenotyping and genotyping of large mapping populations.esources including time, effort and cost involved in genotypinghe large mapping populations is a bottleneck in the identificationf drought grain yield QTLs. The three mapping strategies, WPG, SG,nd BSA, were compared with respect to the QTL effects. SG over-stimated QTL effects compared with WPG. BSA was found to be anffective approach in the identification of consistent-effect droughtrain yield QTLs. A QTL for grain yield under drought was identifiedn a Basmati334/Swarna population through BSA on chromosome

(qDTY8.1). BSA can be applied in multiple rice populations simul-aneously to identify consistent-effect drought grain yield QTLsorthy for MAS.

cknowledgments

The authors thank the Generation Challenge Program for provid-ng financial support for this study. The authors also acknowledgehe technical support received from Marilyn Del Valle, Lenieuiatchon, Jocelyn Guevarra, and Paul Maturan in the completionf this study. Experiments conducted in our study comply with theurrent laws of the Philippines.

ppendix A. Supplementary data

Supplementary data associated with this article can be found, inhe online version, at http://dx.doi.org/10.1016/j.fcr.2012.05.012.

eferences

ltinkut, A., Gozukirmizi, N., 2003. Search for microsatellites associated with waterstress tolerance in wheat through bulked segregant analysis. Mol. Biotechnol.23, 97–106.

mrawathi, Y., Singh, R., Singh, A.K., Singh, V.P., Mohapatra, T., Sharma, T.R., Singh,N.K., 2008. Mapping of quantitative trait loci for basmati quality traits in rice(Oryza sativa L.). Mol. Breeding 21, 49–65.

abu, R.C., Nguyen, B.D., Chamarerk, V., Shanmugasundaram, P., Chezhian, P.,Jeyaprakash, P., Ganesh, S.K., Palchamy, A., Sadasivam, S., Sarkarung, S., Wade,L.J., Nguyen, H.T., 2003. Genetic analysis of drought resistance in rice by molec-ular markers: association between secondary traits and field performance. CropSci. 43, 1457–1469.

Becker, A., Chao, D.Y., Zhang, X., Salt, D.E., Baxter, I., 2011. Bulksegregant analysis using SNP microarrays. PLoS One 6, e15993,http://dx.doi.org/10.1371/journal.pone.0015993.

Bernier, J., Kumar, A., Venuprasad, R., Spaner, D., Atlin, G.N., 2007. A large-effect QTLfor grain yield under reproductive-stage drought stress in upland rice. Crop Sci.47, 507–516.

Collard, B.C.Y., Jahufer, M.Z.Z., Brouwer, J.B., Pang, E.C.K., 2005. An introduction tomarkers, quantitative trait loci (QTL) mapping and marker assisted selection forcrop improvement: the basic concepts. Euphytica 142, 169–196.

Collard, B.C.Y., Mackill, D.J., 2008. Marker-assisted selection: an approach for pre-cision plant breeding in the twenty first century. Philos. Trans. Roy. Soc. B 363,557–572.

Darvasi, A., Soller, M., 1992. Selective genotyping for determination of linkagebetween a marker locus and a quantitative trait locus. Theor. Appl. Genet. 85,353–359.

El-Kadi, D.A., Afiah, S.A., Aly, M.A., Badran, A.E., 2006. Bulked segregant analysisto develop molecular markers for salt tolerance in Egyptian cotton. Arab. J.Biotechnol. 9, 129–142.

Ghimire, K.H., Quiatchon, L.A., Vikram, P., Swamy, B.P.M., Dixit, S., Ahmed, H.U., Her-nandez, J.E., Borromeo, T.H., Kumar, A., 2012. Identification and mapping of aQTL (qDTY1.1) with a consistent effect on grain yield under drought. Feild CropsRes. 131, 88–96.

Gomez, S.M., Boopathi, N.M., Kumar, S.S., Ramasubramanian, T., Chengsong, Z.,Jeyaprakash, P., Senthil, A., Babu, R.C., 2010. Molecular mapping and locationof QTLs for drought-resistance traits in indica rice (Oryza sativa L.) lines adaptedto target environments. Acta Physiol. Plant 32, 355–364.

Grattapaglia, D., Bertolucci, F.L.G., Sederoff, R., 1996. Genetic mapping of quantitativetrait loci controlling growth and wood quality traits in Eucalyptus grandis usinga material half-sib family and RAPD markers. Genetics, 1205–1214.

IRGSP: International Rice Genome Sequencing Project, 2005. The map-basedsequence of the rice genome. Nature 436, 793–800.

Kadam, S., Singh, K., Shukla, S., Goel, S., Vikram, P., Pawar, V., Gaikwad, K.,Chopra, R.K., Singh, N.K., 2012. Genomic associations for drought toler-ance on the short arm of wheat chromosome 4B. Func. Integr. Genomic.,http://dx.doi.org/10.1007/s10142-012-0276-1.

Kanagaraj, P., Prince, K.S.J., Sheeba, J.A., Biji, K.R., Paul, S.B., Senthil, A., Babu, R.C.,2010. Microsatellite markers linked to drought resistance in rice (Oryza sativaL.). Curr. Sci. 98, 836–839.

Kosambi, D.D., 1944. The estimation of a map distance from recombination values.Ann. Eugenics 12, 172–175.

Kumar, R., Venuprasad, R., Atlin, G.N., 2007. Genetic analysis of rainfed lowland ricedrought tolerance under naturally-occurring stress in eastern India: heritabilityand QTL effects. Field Crops Res. 103, 42–52.

Kumar, A., Bernier, J., Verulkar, S., Lafitte, H.R., Atlin, G.N., 2008. Breeding for droughttolerance: direct selection for yield, response to selection and use of drought-tolerant donors in upland and lowland-adapted populations. Field Crops Res.107, 221–231.

Lanceras, J.C., Pantuwan, G., Jongdee, B., Toojinda, T., 2004. Quantitative trait lociassociated with drought tolerance at reproductive stage in rice. Plant Physiol.135, 384–399.

Lander, E.S., Botstein, D., 1989. Mapping Mendelian factors underlying quantitativetraits using RFLP linkage maps. Genetics 121, 185–199.

Lebowitz, R.J., Soller, M., Beckmann, J.S., 1987. Trait-based analyses for the detectionof linkage between marker loci and quantitative trait loci in crosses betweeninbred lines. Theor. Appl. Genet. 73, 556–562.

Lorieux, M., 2007. MapDisto, a free user-friendly program for computing geneticmaps. In: Computer Demonstration Given at the Plant and Animal Genome XVConference, Jan 13–17, San Diego, CA, p. 958, http://mapdisto.free.fr.

McCouch, S.R., Teytelman, L., Xu, Y., Lobos, K.B., Clare, K., Walton, M., Fu, B., Maghi-

rang, R., Li, Z., Xing, Y., Zhang, Q., Kono, I., Yano, M., Fjellstrom, R., DeClerck,G., Schneider, D., Cartinhour, S., Ware, D., Stein, L., 2002. Development andmapping of 2240 new SSR markers for rice (Oryza sativa L.). DNA Res. 9,199–207.

1 s Rese

M

M

N

N

P

Q

S

SS

S

S

S

S

T

T

92 P. Vikram et al. / Field Crop

ichelmoore, R.W., Paran, I., Kesseli, R.V., 1991. Identification of markers linked todisease resistance genes by bulked segregant analysis: a rapid method to detectmarkers in specific genomic regions by using segregating populations. Proc. Natl.Acad. Sci. USA 88, 9828–9832.

urray, M.G., Thompson, W.F., 1980. Rapid isolation of high molecular weight plantDNA. Nucleic Acids Res. 8, 4321–4326.

avabi, A., Mather, D.E., Bernier, J., Spaner, D.M., Atlin, G.N., 2009. QTL detectionwith bidirectional and unidirectional selective genotyping: marker-based andtrait-based analyses. Theor. Appl. Genet. 118, 347–358.

elson, J.C., 1997. QGENE: software for marker-based genomic analysis and breed-ing. Mol. Breeding 3, 239–245.

andit, A., Rai, V., Bal, S., Sinha, S., Kumar, V., Chauhan, M., Gautam, R.K., Singh, R.,Sharma, P.C., Singh, A.K., Gaikwad, K., Sharma, T.R., Mohapatra, T., Singh, N.K.,2010. Combining QTL mapping and transcriptome profiling of bulked RILs foridentification of functional polymorphism for salt tolerance genes in rice (Oryzasativa L.). Mol. Genet. Genomics 284, 121–136.

uarrie, S., Lazic-jancic, V., Kovacevic, D., Steed, A., Pekic, S., 1999. Bulk segregantanalysis with molecular markers and its use for improving drought resistancein maize. J. Exp. Bot. 50, 1299–1306.

ambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning: a Laboratory Manual,2nd ed. Cold Spring Harbor, New York.

AS Institute Inc., 2004. SAS Online Doc 9.1.3. SAS Institute Inc., Cary, NC.emagn, K., Bjørnstad, A., Xu, Y., 2010. The genetic dissection of quantitative traits

in crops. Electron. J. Biotechnol. 13, 1–45.hashidhar, H.E., Vinod, M.S., Sudhir, N., Sharma, G.V., Krishnamurthy, K., 2005.

Markers linked to grain yield using bulked segregant analysis approach in rice(Oryza sativa L.). Rice Genet. Newsl. 22, 69–71.

ivaranjani, A.K.P., Pandey, M.K., Sudharshan, I., Kumar, G.R., SheshuMadhav, M.,Sundaram, R.M., Varaprasad, G.S., Shobharani, N., 2010. Assessment of geneticdiversity among basmati and non-basmati aromatic rices of India using SSRmarkers. Curr. Sci. 99, 221–226.

wamy, B.P.M., Vikram, P., Dixit, S., Ahmed, H.U., Kumar, A., 2011. Meta-analysis ofgrain yield QTL identified during agricultural drought in grasses showed con-sensus. BMC Genomics 12, 319.

wamy, B.P.M., Kumar, A., 2011. Sustainable rice yield in water short drought proneenvironments: conventional and molecular approaches. In: Teang Shui, Lee (Ed.),Irrigation Systems and Practices in Challenging Environments. InTech, Croatia,pp. 149–168.

emnykh, S., Declerck, G., Lukashova, A., Lipovich, L., Cartinhour, S., McCouch, S.,

2001. Computational and experimental analysis of microsatellites in rice (Oryzasativa L.): frequency, length variation, transposon associations, and geneticmarker potential. Genome Res. 11, 1441–1452.

uberosa, R., Salvi, S., Sanguineti, M.C., Landi, P., Maccaferri, M., Conti, S.,2002. Mapping QTLs regulating morpho-physiological traits and yield: case

arch 134 (2012) 185–192

studies, shortcomings and perspectives in drought stressed maize. Ann. Bot. 89,941–963.

Vales, M.I., Schön, C.C., Capettini, F., Chen, X.M., Corey, A.E., Mather, D.E., Mundt,C.C., Richardson, K.L., Sandoval-Islas, J.S., Utz, H.F., Hayes, P.M., 2005. Effect ofpopulation size on estimation of QTL: a test using resistance to barley stripe rust.Theor. Appl. Genet. 111, 1260–1270.

Venuprasad, R., Bool, M.E., Quiatchon, L., Sta Cruz, M.T., Amante, M., Atlin, G.N.,2011. A large-effect QTL for rice grain yield under upland drought stress onchromosome 1. Mol. Breeding, http://dx.doi.org/10.1007/s11032-011-9642-2.

Venuprasad, R., Dalid, C.O., Del Valle, M., Zhao, D., Espiritu, M., Sta Cruz, M.T., Amante,M., Kumar, A., Atlin, G.N., 2009. Indentification and characterization of large-effect quantitative trait loci for grain yield under lowland drought stress in riceusing bulk-segregant analysis. Theor. Appl. Genet. 120, 177–190.

Venuprasad, R., Lafitte, H.R., Atlin, G.N., 2007. Response to direct selection for grainyield under drought stress in rice. Crop Sci. 47, 285–293.

Verulkar, S.B., Mandal, N.P., Dwivedi, J.L., Singh, B.N., Sinha, P.K., Mahato, R.N., Swain,P., Dongre, P., Payasi, D., Singh, O.N., Bose, L.K., Robin, S., Babu, R.C., Senthil,S., Jain, A., Shashidhar, H.E., Hittalmani, S., Vera Cruz, C., Paris, T., Hijmans, R.,Raman, A., Haefele, S., Serraj, R., Atlin, G., Kumar, A., 2010. Breeding resilientand productive rice genotypes adapted to drought-prone rainfed ecosystems ofIndia. Field Crops Res. 117, 197–208.

Vikram, P., Swamy, B.P.M., Dixit, S., Sta Cruz, M.T., Ahmed, H.U., Singh, A.K., Kumar,A., 2011. qDTY1.1 , a major QTL for rice grain yield under reproductive-stagedrought stress with a consistent effect in multiple elite genetic backgrounds.BMC Genetics 12, 89.

Vikram, P., Kumar, A., Singh, A.K., Singh, N.K., 2012. Rice: genomics-assisted breedingfor drought tolerance. In: Tuteja, N., Gill, S.S., Tiburico, A.F., Tuteja, R. (Eds.),Improving Crop Resistance to Abiotic Stress. Wiley-VCH Verlag GmbH & Co.KGaA, Germany, pp. 715–731.

Xu, J.L., Lafitte, H.R., Gao, Y.M., Fu, B.Y., Torres, R., Li, Z.K., 2005. QTLs for droughtescape and tolerance identified in a set of random introgression lines of rice.Theor. Appl. Genet. 111, 1642–1650.

Xu, K., Xu, X., Ronald, P.C., Mackill, D.J., 2000. A high-resolution linkage map of thevicinity of the rice submergence tolerance locus Sub1. Theor. Appl. Genet. 263,681–689.

Yang, J., Hu, C.C., Hu, H., Yu, R., Xia, Z., Ye, X., Zhu, J., 2008. QTL network: mapping andvisualizing genetic architecture of complex traits in experimental populations.Bioinformatics 24, 721–723.

Ye, G., Smith, K.F., 2010. Marker-assisted gene pyramiding for cultivar

development. Plant Breeding Reviews, vol. 3. John Wiley and Sons,pp. 219–256.

Zhang, G.L., Chen, L.Y., Xiao, G.Y., Xiao, Y.H., Chen, X.B., Zhang, S.T., 2009. Bulkedsegregant analysis to detect QTL related to heat tolerance in rice (Oryza sativaL.) using SSR markers. Agric. Sci. China 8, 482–487.