Theor Appl Genet (2009) 119:281–292

DOI 10.1007/s00122-009-1037-y

ORIGINAL PAPER

A new interspeciWc, Gossypium hirsutum £ G. barbadense, RIL population: towards a uniWed consensus linkage map of tetraploid cotton

Jean-Marc Lacape · J. Jacobs · T. Arioli · R. Derijcker · N. Forestier-Chiron · D. Llewellyn · J. Jean · E. Thomas · C. Viot

Received: 9 January 2009 / Accepted: 5 April 2009 / Published online: 23 April 2009© Springer-Verlag 2009

Abstract We report the development of a new interspe-ciWc cotton recombinant inbred line (RIL) population of140 lines deriving from an interspeciWc cross betweenGossypium hirsutum (Gh) and G. barbadense (Gb), usingthe same two parents that have served for the constructionof a BC1 map and for the marker-assisted backcross selec-tion program underway at CIRAD. Two marker systems,microsatellites and AFLPs, were used. An important fea-ture of the RIL population was its marked segregation dis-tortion with a genome-wide bias to Gh alleles (parentalgenome ratio is 71/29). The RIL map displays an excellentcolinearity with the BC1 map, although it is severely con-tracted in terms of map size. Existence of 255 loci in com-mon (between 6 and 14 per chromosome) allowed theintegration of the two data sets. A consensus BC1–RIL mapbased upon 215 individuals (75 BC1 + 140 RIL) was built.

It consisted of 1,745 loci, spanned 3,637 cM, intermediatebetween the sizes of the two component maps, andconstituted a solid framework to cross align cotton mapsusing common markers. The new RIL population will befurther exploited for Wber property QTL mapping andeQTL mapping.

Introduction

Both of the cultivated AD allotetraploid (2n = 4x = 52)Gossypium species, G. hirsutum and G. barbadense, wereindependently domesticated for their epidermal seed tric-homes (Wendel et al. 1989) used as textile Wbers. Highlyproductive and agronomically versatile, G. hirsutumproduces Wbers of good quality (with present day cultivars)and comprises more than 90% of world production of cot-ton. Conversely, G. barbadense produces extra long staple(ELS) Wbers, which are much longer, stronger and Wner, butis much less productive and of narrower agronomic adapt-ability. Given their complementary agronomic and Wberquality characteristics, numerous attempts have been madeto combine their favorable properties by conventionalbreeding and selection from crosses between these twosexually compatible species. The limited success ofinterspeciWc G. hirsutum £ G. barbadense cultivars hasbeen caused in part by reduced fertility, diVerences inmaturity and hybrid breakdown often observed in the latergenerations.

Cotton molecular genetic research since the early 90shas demonstrated a workable level of molecular polymor-phism between G. hirsutum and G. barbadense genomes incontrast to the relatively low levels within each of thesespecies individually. This has encouraged the developmentof interspeciWc G. hirsutum £ G. barbadense framework

Communicated by C. Gebhardt.

Electronic supplementary material The online version of this article (doi:10.1007/s00122-009-1037-y) contains supplementary material, which is available to authorized users.

J.-M. Lacape (&) · N. Forestier-Chiron · J. Jean · E. Thomas · C. ViotCIRAD, UMR DAP, Avenue Agropolis, 34398 Montpellier Cedex 5, Francee-mail: marc.lacape@cirad.fr

J. Jacobs · R. DerijckerBayer BioScience N.V., Technologiepark 38, Ghent, Belgium

T. ArioliBayer CropScience, BioScience Research, Lubbock, TX, USA

D. LlewellynCSIRO, Plant Industry, Canberra, ACT 2601, Australia

123

282 Theor Appl Genet (2009) 119:281–292

mapping populations in several global cotton researchcenters, including in France at CIRAD from the crossGuazuncho 2 £ VH8 (BC1) (Nguyen et al. 2004), in theUSA from crosses race palmeri £ K101 (F2) andTM1 £ 3–79 (RIL) (Rong et al. 2004; Frelichowski et al.2006), and in China from crosses TM1 £ Hai7124 (BC1),CRI36 £ Hai7124 (F2) and Handan 208 £ Pima90 (BC1)(Guo et al. 2007; He et al. 2007; Yu et al. 2007). The mainfeatures of the resulting genetic maps are the following:(1) map size typically ranges from 3,500 to 5,700 cM. Therelatively large variation in map size is caused by diVer-ences in population structure and population size, as well asdiVerences in analysis tools, marker density and the qualityof the mapping data (2) PCR-based markers and particu-larly markers derived from genomic or EST-derived SSRare abundant in most of the maps (3) the various maps havea signiWcant and increasing number of markers in commonallowing comparative displays. Several websites oVerCMap as a tool for performing map comparisons,including the Cotton Marker Database (CMD) (http://www.cottonmarker.org/), CottonDB (http://cottondb.org/)and TropGeneDB http://tropgenedb.cirad.fr). Most of theseframework maps have also served for QTL mapping,mainly focusing on Wber quality QTL (Paterson et al. 2003;Mei et al. 2004; Chee et al. 2005; Lacape et al. 2005; Parket al. 2005; Frelichowski et al. 2006; Wang et al. 2006a; Heet al. 2008). Although the various maps show reasonableglobal consistency, an integrative approach for comparingQTL data from the various parental combinations and/orenvironments has met with limited success in revealingconsistent QTL locations. This is partly due to contradic-tions in locus orders between the maps at a higher resolu-tion (Lacape et al. 2005; Rong et al. 2007; Wang et al.2007).

As compared to F2 and backcross (BC) populations,recombinant inbred lines (RILs) present several advantagesfor gene and QTL mapping: the high level of homoeozyg-osity enables identical replication in progeny and eVec-tively renders RILs as perpetual populations. This allowsfor replicated studies by many investigators and in a multi-ple environments (Burr et al. 1988). The main advantage ofthe immortal status of a RIL population is that phenotypescan be estimated much more accurately. Drawbacks ofRILs are the long development timelines and, in somecases, reduced fertility in a subset of the lines. Untilrecently, the use of RILs has been limited in cotton. Todate, four RIL populations have served for genetic mappingin cotton, two derived from interspeciWc G. hirsutum £G. barbadense crosses (Frelichowski et al. 2006; He et al.2008) and two from intraspeciWc (G. hirsutum) crosses(Shen et al. 2007; Wang et al. 2007).

Here, we report the development of a new interspeciWc,G. hirsutum £ G. barbadense, RIL population, which

forms the basis of an ongoing international research projectaimed at a genetic and genomic dissection of cotton Wberquality (Génoplante project nr ANR-06-GPLA-018). Theparents of this new RIL population are highly contrastingfor Wber characteristics and were previously used for thedevelopment of a BC1 genetic framework map (Lacapeet al. 2003) and an advanced backcross marker-assistedselection program based on the BC1 materials is underwayat CIRAD (Lacape et al. 2005). As a Wrst milestone, thispaper reports on the development of a combined SSR-AFLP genetic map for the RIL population, and on themerging of this RIL genetic map with the earlier BC1 mapto create an integrated consensus genetic map permittingnumerous bridging points to other published cotton maps.

Materials and methods

Generation of the RIL population

The RIL population comprises 140 lines in an F6 to F9 (6F6, 19 F7, 89 F8 and 26 F9) stage of selWng through singleseed descent (SSD). The two parents, Guazuncho 2(G. hirsutum), and VH8-4602 (G. barbadense), were cho-sen from the germplasm bank of CIRAD for their excellentagronomic performance (Guazuncho 2) and superior Wberquality parameters (VH8) (Lacape et al. 2005). The sameG. hirsutum and G. barbadense parents also served todevelop the backcross BC1 mapping population (Lacapeet al. 2003) and further BC1, BC2, and BC2S1 generationsthat were used for QTL mapping of Wber quality parameters(Lacape et al. 2005) and leaf hairiness (Lacape and Nguyen2005).

DNA extraction and marker analysis

DNA was extracted from 140 RILs grown in the CIRADgreenhouse in Montpellier and from 133 RILs grown in theBayer greenhouse in Belgium. At CIRAD, DNA wasextracted from a pooled sample (25 mg) of the Wrstexpanded leaf of two to three young seedlings, groundusing a MM mixer mill (Qiagen). DNA was isolated usingthe MATAB (mixed alkylmethylammonium bromide)method as previously reported (Risterucci et al. 2000;Lacape et al. 2003), without further puriWcation steps. AtBayer, DNA was extracted from young leaf tissue from oneplant per line (100 mg), following the protocol of Dellaportaet al. (1983) with minor modiWcations.

Procedures for gel-based SSR detection (CIRAD) wereas detailed in Lacape et al. (2007). In total, 451 SSRs wereampliWed and electrophoresed in 197 diVerent PCR reac-tions, the majority of the SSRs being tested as 3£ PCRmultiplexes. The SSRs originate from various libraries

123

Theor Appl Genet (2009) 119:281–292 283

(predominantly BNL, CIR, NAU, MU, and JESPR) that areall described and their sequences posted on the website ofthe Cotton Marker Database, CMD (http://www.cottonmar-ker.org). Eighty percent of the SSRs were chosen from theBC1 map of the Guazuncho £ VH8 cross in order to obtainan even coverage of each of the 26 cotton chromosomes.Following this Wrst round of genotyping, additional SSRswere identiWed from an examination of other internationallinkage maps to increase the saturation of a few regions oflower coverage.

In total, 597 reliable SSR loci, deriving from 411 diVer-ent SSRs, were scored for their segregation over 140 RILs.Fragment sizes were estimated from their mobility relativeto a 30–330 bp size standard, and for those markers in com-mon with the BC1, the radiogram of the RILs was alignedwith the corresponding BC1 radiogram, and an identicallocus name was adopted for the 2 maps. The majority ofSSR loci were codominantly scored, though for a smallportion (<5%) of them, only a single parental allele couldbe traced and was thus dominantly scored.

AFLP analysis for 25 primer pairs (Bayer) was per-formed essentially as described in the protocol of Voset al. (1995), except that genomic DNA was Wrst ampli-Wed according to the whole genome ampliWcation (WGA)using the Illustra GenomePhi V2 DNA AmpliWcation Kit(GE Healthcare Bioscience, Piscataway, NJ). Pilot experi-ments showed that original genomic DNA and WGA-ampliWed DNA gave nearly identical AFLP band patterns.Bands that were not reproducible between genomic DNAand WGA-ampliWed DNA were excluded from furtheranalysis. In total, 952 AFLP markers were scored usingAFLP QuantarPro software (Keygene products N·V.), ofwhich 763 were used in genetic map construction. Basedon segregation patterns, around 15% of AFLP loci couldbe considered allelic and were coded as codominantmarkers. Part of the primer pairs, 8 out of 25, used in thisstudy were common with the ones used for the BC1 s, butdue to diVerences in revelation techniques and fragmentsize estimation (visual in the case of BC1 s and usingQuantarPro in the case of RILs), comparisons of RIL andBC proWles and of the positional correspondence amongAFLP loci were not made possible.

Linkage analysis and construction of the RIL map

In a preliminary screen of the complete set of 1,360 loci(597 SSRs and 763 AFLPs), the option “similarity of loci”of the software JoinMap® 4.0 (Stam and Van Ooijen 1995;Van Ooijen 2006) was used to identify a series of 238 pairsof similar AFLP loci that were discarded from the analysis.The Wnal SSR + AFLP data set comprises 1,122 loci, 597SSR and 525 non-redundant AFLPs that were eVectivelygrouped and ordered with JoinMap. Population type was

deWned as an “RI8” (Recombinant Inbred in generation F8)in JoinMap.

Locus localization derived from other maps was used toexamine RIL linkage groups as proposed by JoinMap, i.e.,to putatively assign them to chromosomes. The denomina-tion of the 26 chromosomes followed the most recentaccepted nomenclature of the 13 A (numbered c1–c13) and13 D (c14–c26) chromosomes of tetraploid AD cotton, bywhich the six formerly un-assigned linkage groups are nownumbered: A01 as chromosome 13, c13; A02 as c8, A03 asc11, D02 as c21, D03 as c24 and D08 as c19 (Wang et al.2006b). Despite the use of multiplex PCR and loading of 3SSRs, the comparison of the RILs and BC1 proWles wasstraightforward as the parents were the same. The compari-son with the other international maps relied on the exis-tence of several bridge loci per chromosome.

To maintain consistency of loci designation within RILgroups, we had to modify default calculation options in theJoinMap software. When necessary, a conservative high-LOD value (up to 15) was used to break apart large andmixed inconsistent groups into smaller, more tightly linkedand consistent ones. Smaller groups known to belong to agiven chromosome were then treated as separate data setsand resolved at LOD values between 1 and 4. Manualimprovements of the draft map were realized iteratively asnew loci were added to the map. When possible, these newloci were used as bridges with other maps and thus alterna-tive locus orders were sometimes tested accordingly. Wealso used the JoinMap statistical outputs (“genotype proba-bilities” and “mean square contributions”) as well asgraphical genotypes to come back to raw data and checkSSR radiograms and AFLP scans for possible coding errorsgenerating singletons. Last, in an advanced version of themap, graphical genotypes were also used to re-check radio-grams and gel scans to recode some dominant markers intocodominant markers based upon signal intensity (H codingswere generated from C or D) and complementary segrega-tion (a pair of AFLP bands, being co-segregating and co-localized on the map, would be associated as a singlecodominant locus).

Map distances were calculated using Kosambi’s map-ping function, a goodness-of-Wt jump threshold for removalof loci of three, pairwise recombination estimates smallerthan 0.50 and LOD scores larger than 0.5.

Map integration

After construction of the RIL linkage map, its integrationwith the BC1 map was accomplished with the programJoinMap v. 4.0 (Stam 1993; Van Ooijen 2006). Both BC1

(1,271 loci, as 5 of the 1,276 loci of the original map wereAFLP grouped in an un-assigned linkage group NL5) andRIL (800 loci) datasets were loaded into JoinMap and each

123

284 Theor Appl Genet (2009) 119:281–292

of the 26 chromosomes assembled separately (26 individualgroup nodes in each population). The option “Wxed orders”was occasionally used in a few cases when contradictionsbetween the BC1 and RIL orders occurred, to impose thebest locus order (veriWed across populations). Map integra-tion was realized for each chromosome using the JoinMapfunction “combine groups for map integration”. JoinMapused the data obtained from each population to estimaterecombination frequencies that were then used to determinethe arrangement of markers by minimizing recombinationevents (Stam 1993). The recombination frequency betweena given pair of markers was calculated after applying theappropriate weighting (LOD-based). Map distances werecalculated using function and values as mentioned earlier.

Graphical genotypes, used to check for singletons, weredisplayed using Excel formatting tools and GGT software(Van Berloo 2008). Comparative mapping and veriWcation

of consistency of marker order and distances between mapswere done using MapChart (Voorrips 2002).

All genetic maps were Wnally imported into the webdatabase of CIRAD, TropGeneDB, (http://tropgenedb.cirad.fr/en/cotton.html) for comparative visualization usingCMap tool.

Results

Construction of the RIL genetic map

The Wnal genetic map of the RIL population was computedfrom over 150,000 genotype data points, derived from1,122 markers in 140 RILs. It comprised 800 loci andspanned 2,044 cM (Table 1, Fig. 1). All chromosomesemerged as single linkage groups that were well saturated

Table 1 Descriptive parameters of the Guazuncho 2 £ VH8–RIL map (constructed using JoinMap 4.0) and the Guazuncho 2 £ VH8-consen-sus map obtained by integration of the BC1 and RIL data using JoinMap 4.0

Chromosome RIL map Consensus map Common BC1–RIL

cM No loci CO/100 cM cM No loci No loci

c1 97.9 23 2.5 128.3 50 9

c2 81.2 37 2.6 107.9 76 10

c3 76.6 33 2.7 116.5 71 11

c4 69.7 27 2.9 92.1 51 8

c5 108.9 43 3.7 138.3 81 14

c6 48.9 20 3.6 102.9 54 8

c7 74.5 21 2.7 113.0 65 7

c8 66.8 33 3.1 181.0 68 6

c9 73.3 42 4.6 131.2 80 10

c10 67.7 23 2.4 157.9 63 7

c11 66.3 30 4.2 245.2 91 9

c12 75.4 41 4.6 134.9 89 13

c13 72.9 35 3.0 136.8 86 8

Sum A 980.2 408 3.3 1786.0 925 120

c14 98.9 39 3.0 125.0 73 14

c15 71.7 25 2.9 150.7 64 8

c16 58.7 23 2.6 142.0 45 7

c17 45.8 15 3.4 82.2 42 6

c18 99.7 38 3.0 135.6 72 12

c19 75.9 25 3.9 199.8 76 12

c20 100.2 34 2.7 171.6 68 14

c21 167.2 47 2.4 190.0 76 14

c22 72.8 34 2.6 106.0 57 9

c23 60.9 30 3.3 128.5 65 9

c24 83.1 37 3.0 124.0 69 12

c25 53.4 23 2.9 155.5 60 10

c26 75.5 22 2.7 139.8 53 8

Sum D 1063.6 392 2.9 1850.6 820 135

Overall 2043.9 800 3.1 3636.6 1745 255

123

Theor Appl Genet (2009) 119:281–292 285

throughout their length, except for chromosome 21 whichemerged with two unbridgeable linkage groups. A gap dis-tance of 25 cM was extrapolated for c21 after comparisonswith other reference maps. The average distance between 2loci was 2.55 cM in the RIL map. The map comprised 24bridgeable gaps of at least 10 cM, of which 7 were largerthan 15 cM.

The RIL genetic map contained 281 AFLP loci, 517 SSRloci and two morphological marker loci. The two morpho-logical loci both mapped to the expected chromosomallocation: petal spot (R2) on c7 and petal color (Y1 or pet) onc13. The 517 SSR loci of the RIL map derived from 387primer pairs, originating from 5 SSR collections BNL(158), CIR (138), NAU (89), MU (76), and JESPR (36).Marker groupings and loci order on the RIL map were vali-dated by comparative alignments based on shared SSR lociwith the CIRAD BC1 map and other public maps.

The A and D subgenomes (Table 1) contributed verysimilarly to marker composition (408 and 392 loci, respec-tively) and map distance (980 and 1,064 cM, respectively).

Segregation distortion and residual heterozygosity

The RIL population displayed broad genetic distortiontowards the G. hirsutum (Guazuncho) parent. For SSRmarkers the allelic contribution of both parents could bedirectly calculated from the observed marker genotypes.For dominantly scored AFLP markers in ambiguous geno-

typic classes “C” and “D”, the allelic contribution of thetwo parents was calculated by extrapolation from SSR data.On average, any given RIL carried Guazuncho 2 alleles at70.8% of all loci and VH8 alleles at 29.2% of all loci (total800 loci). The contribution of Guazuncho 2 and VH8ranged from 95 and 5 to 32 and 68%, respectively. Only 15of the 140 RILs contained over 50% of G. barbadensealleles. Nearly all loci (99%) deviated signiWcantly from theexpected 1:1 parental contribution (�² test at P < 0.05,corresponding to 59/41% is met by only 6 of the 800 loci).The parental allele ratio for individual markers ranged from44/56 to 87/13% (G. hirsutum/G. barbadense) and thedistortions are evenly distributed across chromosomes,without apparent “hotspots”.

The RIL population displayed a higher-than-expectedlevel of residual heterozygosity: 4.9%, instead of theexpected 0.8% for an F8 population derived by SSD. Resid-ual heterozygosity in individual lines ranged from 0.1 to29.7%. Residual heterozygosity was similar for all chromo-somes, ranging from 4.5% (c22) to 6.6% (c6).

Genotypic content and population structure

The frequency and distribution of recombination crossovers(CO) was estimated from examination of the graphicalgenotypes for all RILs. In case of dominant marker codings(AFLP data), the most conservative interpretation of thecoding was used in order to minimize CO estimates. For

Fig. 1 Genetic map of the Guazuncho 2 £ VH8–RIL population. The26 chromosomes are arranged according to the 13 homoeologous pairsof A (c1–c13) and D (c14–c26) chromosomes, i.e. c1 (At genome) isfollowed by its homoeologous c15 (Dt genome), etc. The 670 markersinclude 281 AFLPs named by Eco/Mse primer pair code (see text forprimer nomenclature) followed by the fragment size (bp), 387 SSRs of

various sources (BNL, CIR, MU, NAU, JESPR, MGHES andSSR-containing RFLPs noted A, pAR, pGH, Gate, Coau and Unig) alldescribed at http://www.cottonmarker.org, and two morphologicalmarkers (R2 and pet). Underlined loci are duplicated loci. Mapdistances are given in centimorgans (cM)

BNL24400MUSS1614CIR0095

CIR19913

BNL169322NAU247425

NAU541132

CIR05547

E43M53-M234.954E35M54-M350.256MUSB0958MGHES10MUSS422

57

E32M49-M73.6E43M48-M157.158

E43M49-M-512.160BNL135063BNL309066NAU093470BNL388876CIR11482NAU241986

NAU208398

c1

MUSS1610BNL24403

BNL169312M15113MGHES3216

E43M52-M326.030E43M50-M292.732E43M53-M51.8MUSB081833

E31M48-M347.434JESPR20535NAU3690NAU325736

E32M51-M103.837MUSS42239MGHES1040E35M52-M284.942BNL135043BNL309048BNL0786E42M50-M164.254

NAU241962BNL334565E31M55-M109.068CIR10572

c15

CIR3810BNL143411Unig25A0214E31M48-M87.215BNL189718E32M54-M107.822JESPR17925MUSB101727CIR40132BNL334733E35M52-M276.334MUCS399E31M48-M290.0E43M48-M159.0E43M54-M368.2BNL3413

35

JESPR10136BNL3971E43M48-M249.337

E32M48-M111.9MUSB117838

E32M48-M109.4E43M54-M90.8E35M48-M211.1MUSB0958MUSB0399MUSB0749

39

E41M47-M477.140E31M48-M270.5NAU387541

NAU538457CIR37663NAU368464pAR85171NAU226575NAU124677NAU289881

c2

CIR2460

E32M48-M74.2E32M54-M195.212

E35M48-M99.217NAU373320NAU320922CIR38125E32M54-M286.927MUSB103431BNL189736E43M49-M69.640E43M54-M226.646JESPR16547E42M48-M327.148CIR04750CIR05851BNL1059E43M54-M430.052

E43M48-M131.953MUSB103554CIR28857E41M47-M326.058CMS1761E35M50-M314.762E35M52-M261.265BNL022668CIR23969CIR20272CIR26374NAU321476NAU359877E43M52-M182.379BNL3259NAU219081

CIR08487CIR22888CIR20292CIR097CIR21099

c14

BNL34080CIR0309BNL244315BNL244317M12521E36M47-M-532.8E31M48-M102.123

CIR058MUSB1101MUSB0641

24

BNL3140BNL1059NAU0884

26

E31M53-M113.1E43M48-M167.227

E32M54-M109.8E41M47-M452.731

E32M49-M160.932E32M51-M247.738E43M53-M431.939BNL022642BNL398945BNL346346E31M48-M79.349E41M47-M85.951NAU097252CIR26354BNL325961JESPR107NAU086270

CIR13371NAU216173CIR22877

c3

BNL34080

NAU380017BNL244321JESPR19525E43M52-M193.2JESPR10130

E32M48-M88.732BNL2471MUSB064133

E32M49-M185.235CIR25139NAU102840CIR18041NAU285942NAU289846

c17

CIR2180

NAU22354BNL25729CIR12213NAU229117BNL404921M7626CIR04829CIR29134NAU086935E32M49-M307.2E43M49-M84.9E43M52-M97.2E32M50-M195.4E35M52-M62.4E32M54-M282.7

36

E41M47-M125.2E31M48-M165.1E43M48-M411.0

37

BNL404740E35M50-M211.843CIR22245E35M52-M293.949E36M47-M84.657NAU223161CIR18762CIR22370

c4

CIR2180NAU2235BNL38734

CIR12216E43M54-M171.318NAU229123BNL404927E42M48-M145.131E41M47-M243.034BNL106138CIR224CIR048MUSS181

40

E35M50-M77.9E42M48-M219.741

E32M48-M80.5E43M53-M-541.543

JESPR23046E32M51-M229.9E43M48-M154.749

E32M54-M102.5E35M52-M159.7E43M50-M386.5E43M48-M105.4E43M54-M128.6

51

E31M53-M148.852MUSB0641MUSB109353

CIR29454JESPR06560E43M54-M176.262BNL409266BNL403067CIR25373

c22

BNL24400MUSS1614CIR0095

CIR19913

BNL169322NAU247425

NAU541132

CIR05547

E43M53-M234.954E35M54-M350.256MUSB0958MGHES10MUSS422

57

E32M49-M73.6E43M48-M157.158

E43M49-M-512.160BNL135063BNL309066NAU093470BNL388876CIR11482NAU241986

NAU208398

c1

MUSS1610BNL24403

BNL169312M15113MGHES3216

E43M52-M326.030E43M50-M292.732E43M53-M51.8MUSB081833

E31M48-M347.434JESPR20535NAU3690NAU325736

E32M51-M103.837MUSS42239MGHES1040E35M52-M284.942BNL135043BNL309048BNL0786E42M50-M164.254

NAU241962BNL334565E31M55-M109.068CIR10572

c15

CIR3810BNL143411Unig25A0214E31M48-M87.215BNL189718E32M54-M107.822JESPR17925MUSB101727CIR40132BNL334733E35M52-M276.334MUCS399E31M48-M290.0E43M48-M159.0E43M54-M368.2BNL3413

35

JESPR10136BNL3971E43M48-M249.337

E32M48-M111.9MUSB117838

E32M48-M109.4E43M54-M90.8E35M48-M211.1MUSB0958MUSB0399MUSB0749

39

E41M47-M477.140E31M48-M270.5NAU387541

NAU538457CIR37663NAU368464pAR85171NAU226575NAU124677NAU289881

c2

CIR2460

E32M48-M74.2E32M54-M195.212

E35M48-M99.217NAU373320NAU320922CIR38125E32M54-M286.927MUSB103431BNL189736E43M49-M69.640E43M54-M226.646JESPR16547E42M48-M327.148CIR04750CIR05851BNL1059E43M54-M430.052

E43M48-M131.953MUSB103554CIR28857E41M47-M326.058CMS1761E35M50-M314.762E35M52-M261.265BNL022668CIR23969CIR20272CIR26374NAU321476NAU359877E43M52-M182.379BNL3259NAU219081

CIR08487CIR22888CIR20292CIR097CIR21099

c14

BNL34080CIR0309BNL244315BNL244317M12521E36M47-M-532.8E31M48-M102.123

CIR058MUSB1101MUSB0641

24

BNL3140BNL1059NAU0884

26

E31M53-M113.1E43M48-M167.227

E32M54-M109.8E41M47-M452.731

E32M49-M160.932E32M51-M247.738E43M53-M431.939BNL022642BNL398945BNL346346E31M48-M79.349E41M47-M85.951NAU097252CIR26354BNL325961JESPR107NAU086270

CIR13371NAU216173CIR22877

c3

BNL34080

NAU380017BNL244321JESPR19525E43M52-M193.2JESPR10130

E32M48-M88.732BNL2471MUSB064133

E32M49-M185.235CIR25139NAU102840CIR18041NAU285942NAU289846

c17

CIR2180

NAU22354BNL25729CIR12213NAU229117BNL404921M7626CIR04829CIR29134NAU086935E32M49-M307.2E43M49-M84.9E43M52-M97.2E32M50-M195.4E35M52-M62.4E32M54-M282.7

36

E41M47-M125.2E31M48-M165.1E43M48-M411.0

37

BNL404740E35M50-M211.843CIR22245E35M52-M293.949E36M47-M84.657NAU223161CIR18762CIR22370

c4

CIR2180NAU2235BNL38734

CIR12216E43M54-M171.318NAU229123BNL404927E42M48-M145.131E41M47-M243.034BNL106138CIR224CIR048MUSS181

40

E35M50-M77.9E42M48-M219.741

E32M48-M80.5E43M53-M-541.543

JESPR23046E32M51-M229.9E43M48-M154.749

E32M54-M102.5E35M52-M159.7E43M50-M386.5E43M48-M105.4E43M54-M128.6

51

E31M53-M148.852MUSB0641MUSB109353

CIR29454JESPR06560E43M54-M176.262BNL409266BNL403067CIR25373

c22

123

286 Theor Appl Genet (2009) 119:281–292

example, if three successive loci were coded “ADA” or“HDH”, the “D” coding would be assumed to be an A andan H, respectively, and no CO would be inferred.

A total of 8,673 CO were counted for 800 loci across all140 RILs. On average, we counted 62 CO per RIL, repre-senting 2.4 CO per chromosome or 0.031 CO/cM (Table 1).The recombination frequency was fairly stable across mostchromosomes, ranging from 0.024 CO/cM on c10 and c21to 0.046 CO/cM on c9 and c12. Higher values like in c9 andc12, could be indicative of remaining errors in marker cod-ing or map order. On average, any given RIL presented atleast one recombination event on 21 of the 26 chromo-somes, meaning that on average Wve chromosomes per RILwere unrecombined, or completely of one single parentalorigin, with no commonality of the unrecombined chromo-somes in diVerent RILs. In the context of trait introgressionfrom G. barbadense, it is interesting to note that within thepopulation of RILs it was possible to Wnd most of the 26G. barbadense donor chromosomes in a non-recombinedform in at least one RIL. The frequencies of recombinationfound in the current cotton RIL population are similar tothose reported for maize and Arabidopsis RIL populations.In maize, an average of 57 CO per RIL (5.7 CO per chro-mosome, 0.032 CO/cM) was reported for 91 intermated

RILs genotyped with 857 (Fu et al. 2006). In Arabidopsis,1.13 and 1.86 CO per chromosome or 0.016 and 0.022 CO/cM were reported from two diVerent RIL populations (420and 423 lines) by Loudet et al. (2002) and Törjék et al.(2006) respectively.

Comparison of the RIL and BC1 maps from the same parents

Marker order in the current RIL map was very consistentwith the marker order of the previously published BC1 mapfrom the same parents (Nguyen et al. 2004). In contrast,average marker distance and total map size were markedlydiVerent between the two maps, with the RIL map beingmuch smaller. One reason for the size diVerences could bethe mapping algorithms used. The BC1 map (1,275 loci,5,735 cM) was constructed with MapMaker 3.0, while theRIL map was constructed using JoinMap 4.0. MapMakerand JoinMap use diVerent algorithms and, as a conse-quence, are known to generate diVerences in distances,typically shorter distances with JoinMap as compared toMapMaker (Bradeen et al. 2001). The distances of the BC1

map were recalculated using JoinMap 4.0 with the samemapping parameters (see “Material and methods”) as used

Fig. 1 continued

CIR224NAU11370

Gate3BG119E31M61-M61.4NAU122315

BNL407118BNL302922NAU202923NAU104225CIR06231CIR03436BNL244838E32M54-M245.040BNL187843E36M47-M87.949BNL054251JESPR24155JESPR20457E43M49-M260.0E43M48-M68.359

E31M48-M188.565NAU200174E35M54-M195.475E43M53-M308.6E32M50-M186.8E32M49-M517.0MUSB1035

76

CIR301CIR29477

BNL399579E31M53-M88.3E36M47-M158.580

E35M52-M107.581E35M48-M196.883JESPR06588E43M54-M316.489CIR18592BNL403093E31M53-M116.394NAU212198CIR25399E31M48-M87.8103CIR328109

c5

CIR2240

BNL34528

E42M48-M287.216CIR17618BNL169022MUCS58524NAU527331JESPR23033NAU104236CIR06238JESPR27342NAU0911BNL244843

MUSB103544MUSB0641CIR17945

E43M52-M373.846CIR22253BNL334756

NAU223165E31M55-M139.566CIR34468CMS2172E43M54-M140.273E43M49-M287.576

c19

CIR2030CIR0177BNL335912E43M49-M84.714E31M55-M220.216E32M50-M197.218BNL1440E35M52-M452.919

CIR40522E43M48-M335.323CIR07924E32M48-M236.4BNL106425

E31M61-M328.726E43M49-M250.527CIR29128A176330BNL269133BNL288436

CIR12849

c6

BNL33590BNL34364BNL106112CIR28018E43M48-M204.419BNL380625E32M49-M292.829JESPR22732BNL144034E32M51-M179.5E43M48-M445.4E32M49-M175.6MUSB0641MUSB1035E32M54-M89.3

38

CIR407E41M47-M329.239

JESPR27342BNL141745E42M50-M56.146E31M55-M58.147CIR33852NAU127753

c25

CIR0280

E43M52-M270.918E31M48-M461.919E43M52-M-553.6JESPR22820

E43M49-M302.0E35M50-M174.923

E32M48-M121.024CIR32325E43M48-M271.326NAU243230JESPR06532MUSB046333CIR16936NAU218639R243BNL159746CIR41250NAU104357E32M49-M245.059

CIR32075

c7

NAU09660

E31M53-M313.46Coau4E2113BNL112214E32M50-M215.416JESPR22821E32M51-M288.023NAU243229BNL298631CIR100NAU218636

MGHES75NAU0751E43M52-M387.4

38

E31M61-M86.9E35M52-M246.939

CIR10742E41M47-M77.854E32M51-M97.0NAU1020E35M50-M369.5

55

CIR17558E32M54-M260.059

c16

CIR244NAU09200

JESPR2324E42M48-M236.611NAU240713CIR37615E32M50-M295.217E41M47-M140.125E31M48-M172.627MUSS250NAU243928

E35M52-M152.031CIR11933BNL362736E36M47-M173.944E32M49-M130.845MUSB081846MUSB0958E35M54-M402.8E42M48-M206.2

47

E32M51-M345.6E32M54-M394.0CIR187

48

MUSB0139CIR29149

BNL038751BNL3792E43M50-M239.652

E35M50-M390.653E35M54-M148.954E31M61-M272.162NAU116466E43M49-M169.067

c8

BNL25970CIR026CIR3881

JESPR15711NAU240713BNL386019E36M47-M363.322JESPR07025CIR35430MUSS25035NAU243936E32M49-M109.438BNL296142MUSS50046E43M52-M438.649E41M47-M216.653CIR06154E43M52-M334.358E43M53-M57.859E43M52-M373.060BNL038761CIR289E43M54-M169.462

E42M47-M134.6E32M50-M238.363

MUSB0632BNL308464

MUSB0641E32M48-M378.165

E43M48-M276.3E31M61-M115.467

E42M48-M190.970BNL025273MUSB017578CIR06179BNL362782CIR34383

c24

E41M50-M401.50E31M55-M93.22BNL06863NAU315919E32M51-M102.220E32M50-M197.524E43M52-M72.8E32M48-M134.0E32M50-M171.3NAU0966MUSB0958E43M49-M222.9

25

E43M48-M187.326E31M53-M113.9E41M50-M166.527

CIR22728E32M50-M128.530MUSB104031BNL377934BNL303137MUSS54740NAU1360E31M53-M51.543

BNL284748E41M47-M129.1E31M48-M220.649

BNL358250NAU3365JESPR24852

M9253E43M48-M89.354E32M49-M159.656BNL1317BNL259057

MUSS08358MGHES4661E42M47-M179.3E32M49-M118.164

MUSS02266NAU086470BNL059772BNL405373

c9

NAU10350CIR2862BNL06863NAU32144BNL116112E41M47-M369.520MUSB097321E35M54-M128.8CIR060MUSB0632E43M54-M307.0E43M48-M183.9E35M52-M127.2

24

MUSB0641BNL269025

JESPR15127MUSB1040pAR12728

E32M51-M131.429NAU546532BNL351137BNL303139BNL259046JESPR114MUSS18949

MUSS15154BNL059757BNL4053CIR06159

BNL317361

c23

CIR224NAU11370

Gate3BG119E31M61-M61.4NAU122315

BNL407118BNL302922NAU202923NAU104225CIR06231CIR03436BNL244838E32M54-M245.040BNL187843E36M47-M87.949BNL054251JESPR24155JESPR20457E43M49-M260.0E43M48-M68.359

E31M48-M188.565NAU200174E35M54-M195.475E43M53-M308.6E32M50-M186.8E32M49-M517.0MUSB1035

76

CIR301CIR29477

BNL399579E31M53-M88.3E36M47-M158.580

E35M52-M107.581E35M48-M196.883JESPR06588E43M54-M316.489CIR18592BNL403093E31M53-M116.394NAU212198CIR25399E31M48-M87.8103CIR328109

c5

CIR2240

BNL34528

E42M48-M287.216CIR17618BNL169022MUCS58524NAU527331JESPR23033NAU104236CIR06238JESPR27342NAU0911BNL244843

MUSB103544MUSB0641CIR17945

E43M52-M373.846CIR22253BNL334756

NAU223165E31M55-M139.566CIR34468CMS2172E43M54-M140.273E43M49-M287.576

c19

CIR2030CIR0177BNL335912E43M49-M84.714E31M55-M220.216E32M50-M197.218BNL1440E35M52-M452.919

CIR40522E43M48-M335.323CIR07924E32M48-M236.4BNL106425

E31M61-M328.726E43M49-M250.527CIR29128A176330BNL269133BNL288436

CIR12849

c6

BNL33590BNL34364BNL106112CIR28018E43M48-M204.419BNL380625E32M49-M292.829JESPR22732BNL144034E32M51-M179.5E43M48-M445.4E32M49-M175.6MUSB0641MUSB1035E32M54-M89.3

38

CIR407E41M47-M329.239

JESPR27342BNL141745E42M50-M56.146E31M55-M58.147CIR33852NAU127753

c25

CIR0280

E43M52-M270.918E31M48-M461.919E43M52-M-553.6JESPR22820

E43M49-M302.0E35M50-M174.923

E32M48-M121.024CIR32325E43M48-M271.326NAU243230JESPR06532MUSB046333CIR16936NAU218639R243BNL159746CIR41250NAU104357E32M49-M245.059

CIR32075

c7

NAU09660

E31M53-M313.46Coau4E2113BNL112214E32M50-M215.416JESPR22821E32M51-M288.023NAU243229BNL298631CIR100NAU218636

MGHES75NAU0751E43M52-M387.4

38

E31M61-M86.9E35M52-M246.939

CIR10742E41M47-M77.854E32M51-M97.0NAU1020E35M50-M369.5

55

CIR17558E32M54-M260.059

c16

CIR244NAU09200

JESPR2324E42M48-M236.611NAU240713CIR37615E32M50-M295.217E41M47-M140.125E31M48-M172.627MUSS250NAU243928

E35M52-M152.031CIR11933BNL362736E36M47-M173.944E32M49-M130.845MUSB081846MUSB0958E35M54-M402.8E42M48-M206.2

47

E32M51-M345.6E32M54-M394.0CIR187

48

MUSB0139CIR29149

BNL038751BNL3792E43M50-M239.652

E35M50-M390.653E35M54-M148.954E31M61-M272.162NAU116466E43M49-M169.067

c8

BNL25970CIR026CIR3881

JESPR15711NAU240713BNL386019E36M47-M363.322JESPR07025CIR35430MUSS25035NAU243936E32M49-M109.438BNL296142MUSS50046E43M52-M438.649E41M47-M216.653CIR06154E43M52-M334.358E43M53-M57.859E43M52-M373.060BNL038761CIR289E43M54-M169.462

E42M47-M134.6E32M50-M238.363

MUSB0632BNL308464

MUSB0641E32M48-M378.165

E43M48-M276.3E31M61-M115.467

E42M48-M190.970BNL025273MUSB017578CIR06179BNL362782CIR34383

c24

E41M50-M401.50E31M55-M93.22BNL06863NAU315919E32M51-M102.220E32M50-M197.524E43M52-M72.8E32M48-M134.0E32M50-M171.3NAU0966MUSB0958E43M49-M222.9

25

E43M48-M187.326E31M53-M113.9E41M50-M166.527

CIR22728E32M50-M128.530MUSB104031BNL377934BNL303137MUSS54740NAU1360E31M53-M51.543

BNL284748E41M47-M129.1E31M48-M220.649

BNL358250NAU3365JESPR24852

M9253E43M48-M89.354E32M49-M159.656BNL1317BNL259057

MUSS08358MGHES4661E42M47-M179.3E32M49-M118.164

MUSS02266NAU086470BNL059772BNL405373

c9

NAU10350CIR2862BNL06863NAU32144BNL116112E41M47-M369.520MUSB097321E35M54-M128.8CIR060MUSB0632E43M54-M307.0E43M48-M183.9E35M52-M127.2

24

MUSB0641BNL269025

JESPR15127MUSB1040pAR12728

E32M51-M131.429NAU546532BNL351137BNL303139BNL259046JESPR114MUSS18949

MUSS15154BNL059757BNL4053CIR06159

BNL317361

c23

123

Theor Appl Genet (2009) 119:281–292 287

for the RIL map. This conWrmed that map distances gener-ated by JoinMap 4.0 were shorter, by 28% in our case, ascompared to the same data processed with MapMaker 3.0,resulting in a total adjusted size of the BC1 map of4,104 cM. The comparison of the RIL map with theadjusted BC1 map indicated that around 16% (670 cM) ofthe BC1 was not covered in the RIL map. This diVerencerelated to the existence in the BC1 map of more markersand of markers of diVerent marker systems (RFLP) as com-pared to the RIL map. The total length of syntenic segments(1,910 of 2,044 cM in the RIL map corresponding to 3,433of 4,104 cM in the BC1 map) indicated that the RIL maponly represented 56% of the length of the syntenic region ofthe BC1 map. Except for chromosomes 5, 13, 21, and 25which were not signiWcantly diVerent in length betweenBC1 and RIL map, the size contraction was fairly similaracross all chromosomes (not shown).

Integration of the BC1 and RIL maps

The nearly perfect conservation of marker order betweenthe RIL and BC1 maps facilitated their integration into aconsensus map based on the combined raw data, using the

command “combine groups for map integration” of theJoinMap software. The consensus map (Table 1) was builtfrom a total of 1,745 loci, including 255 SSR loci sharedbetween the two maps and 988 and 502 loci unique to theBC1 and RIL maps, respectively. The number of bridgingloci per chromosome (Table 1) was on average ten andranged from six (c8 and c17) to 14 (c5, c14, c20 and c21).Figure 2 displays a simpliWed view of the colinearity oflocus order between the two component maps (BC1 andRIL) and the consensus map for three chromosomes c1, c2,and c3 (complete set of 26 chromosomes shown in supple-mentary Fig. A1).

The 1,745 loci of the consensus map consisted of 781SSR, 715 AFLP, 190 RFLP, 2 morphological markers, and57 other markers. 28 BC1 and 43 RIL loci were excluded byJoinMap and not integrated into the consensus map (Fig.A2). On average, each chromosome contained 67 loci, andthe consensus map spanned a total distance of 3,637 cM,with an average distance of 2.1 cM between two loci. Thus,the map size and marker distances in the consensus map areintermediate between the distances covered in the BC1 andRIL maps. The 3,637 cM of the consensus are similar to the3,540 cM of a JoinMap-based SSR-rich interspeciWc

Fig. 1 continued

E35M54-M429.00E31M48-M190.7E35M48-M111.3E32M54-M376.8

6

M997E32M50-M100.5E41M47-M68.9E32M50-M360.5E43M50-M159.5

9

CIR17112BNL389516E36M47-M170.922E41M47-M374.625BNL116127E32M49-M254.829CIR05530CIR08237BNL296042NAU118244MUSS09647

CIR30557

BNL269162

NAU116968

c10

pAR4680CIR1871

BNL34829CIR09410MUSS46712

BNL114527NAU201728MUSB033830BNL257037CIR12144CIR30547BNL337950E32M51-M258.0E31M55-M264.953

JESPR24454BNL094655CIR17157E31M53-M131.058E41M50-M97.1E42M48-M140.559

TW1460BNL307163E31M61-M191.264E35M54-M69.0E31M55-M132.966

BNL016967JESPR23570MUSS09676JESPR19086BNL328090E43M50-M102.995BNL364697BNL255398E35M48-M173.8100

c20

CIR1960NAU21522CIR0036BNL10667BNL083615E43M48-M159.923M7125BNL359227M16328MUCS39931E43M54-M158.9E31M61-M177.332

MUSB064133E43M52-M175.3E32M49-M149.634

E43M52-M83.235E43M52-M368.9E41M47-M336.839

BNL280540E43M52-M-529.242E32M49-M112.5E41M50-M214.743

NAU102844MUSB1001MUSB100045

BNL325446BNL343160MUCS55761BNL314765CIR25466

c11

CIR0690

JESPR0657

CIR07712

BNL327917E35M52-M144.319

E31M53-M256.826E35M54-M85.627

BNL083633

E31M61-M311.347BNL155149CIR39851E31M48-M367.153E31M61-M345.654MUSB0641pAR78555

CIR06156E43M54-M125.457E35M52-M138.458E31M55-M456.861E32M50-M-525.2BNL280562

E31M53-M66.164NAU2110JESPR24470

CIR061NAU335472

NAU521773NAU2950E43M52-M336.974

NAU485578BNL1053JESPR25180

BNL168186BNL105395BNL1681102

BNL1034127BNL3171130E41M50-M453.8E32M51-M91.0134

BNL3147136NAU1366139CIR013143NAU1148144MUSB1197146

CIR254151

E43M48-M113.1160

BNL1705167

c21

JESPR300A16140

CIR2722CIR3628BNL353715CIR08120BNL167924M6325E31M61-M129.627E32M48-M401.128CMS37E43M49-M238.0BNL1673E32M51-M95.0

34

NAU4089BNL326136

CIR167BNL170739

E41M47-M190.146NAU127847BNL122749E32M54-M229.450E35M52-M339.3E43M49-M139.2E31M48-M242.2

54

E43M52-M167.8E35M54-M243.6MUSB1035

55

BNL2967CIR167CIR291

56

CIR13357BNL019361BNL359962JESPR27064MUSS02665BNL262167E43M50-M193.268E42M48-M309.269BNL326171MUSB111775

c12

NAU38960JESPR3001

E43M52-M274.48CIR2729

BNL353720E32M49-M463.122E43M52-M105.523

E41M47-M315.432BNL348234E43M52-M239.135BNL034138E35M50-M91.639MUCS06440E43M49-M358.846E32M49-M209.2E43M49-M319.1E35M48-M116.2

52

E31M48-M151.954MUSS30355BNL359957pGH33162

CIR17076

c26

BNL32810A13453NAU22857CIR0578MUSS20326E42M50-M259.528pet29JESPR15331MUSB0399CIR338E32M51-M347.1

32

BNL402933NAU102336E32M51-M185.2MUSB126337

MUSB0958E41M50-M100.5CIR079

38

BNL3623MUSB126339

MUSB0818E35M54-M406.840

E32M48-M145.541BNL2652CIR09642

CIR38045CIR34246E43M50-M362.347E41M50-M315.659BNL170762BNL325464pGH351BNL040965

MUSS18171CIR05473

c13

E36M47-M-552.00

BNL32814

CIR21611BNL2571JESPR20416

E42M48-M364.920NAU313021NAU386124NAU384326NAU344727CIR09934MUSS60336Gate1CD12NAU1215CIR277

37

Gate1AG05MUCS40541

MUSS20345NAU244346E31M55-M415.3BNL344555

BNL3445BNL328056

CIR04058BNL1079E43M52-M407.459

MUSB1263E35M48-M81.860

E32M50-M279.7E43M48-M215.562

E43M49-M144.564MUSB056466BNL265271BNL355882E43M49-M124.885MUSB098491BNL254497BNL0193100

c18

E35M54-M429.00E31M48-M190.7E35M48-M111.3E32M54-M376.8

6

M997E32M50-M100.5E41M47-M68.9E32M50-M360.5E43M50-M159.5

9

CIR17112BNL389516E36M47-M170.922E41M47-M374.625BNL116127E32M49-M254.829CIR05530CIR08237BNL296042NAU118244MUSS09647

CIR30557

BNL269162

NAU116968

c10

pAR4680CIR1871

BNL34829CIR09410MUSS46712

BNL114527NAU201728MUSB033830BNL257037CIR12144CIR30547BNL337950E32M51-M258.0E31M55-M264.953

JESPR24454BNL094655CIR17157E31M53-M131.058E41M50-M97.1E42M48-M140.559

TW1460BNL307163E31M61-M191.264E35M54-M69.0E31M55-M132.966

BNL016967JESPR23570MUSS09676JESPR19086BNL328090E43M50-M102.995BNL364697BNL255398E35M48-M173.8100

c20

CIR1960NAU21522CIR0036BNL10667BNL083615E43M48-M159.923M7125BNL359227M16328MUCS39931E43M54-M158.9E31M61-M177.332

MUSB064133E43M52-M175.3E32M49-M149.634

E43M52-M83.235E43M52-M368.9E41M47-M336.839

BNL280540E43M52-M-529.242E32M49-M112.5E41M50-M214.743

NAU102844MUSB1001MUSB100045

BNL325446BNL343160MUCS55761BNL314765CIR25466

c11

CIR0690

JESPR0657

CIR07712

BNL327917E35M52-M144.319

E31M53-M256.826E35M54-M85.627

BNL083633

E31M61-M311.347BNL155149CIR39851E31M48-M367.153E31M61-M345.654MUSB0641pAR78555

CIR06156E43M54-M125.457E35M52-M138.458E31M55-M456.861E32M50-M-525.2BNL280562

E31M53-M66.164NAU2110JESPR24470

CIR061NAU335472

NAU521773NAU2950E43M52-M336.974

NAU485578BNL1053JESPR25180

BNL168186BNL105395BNL1681102

BNL1034127BNL3171130E41M50-M453.8E32M51-M91.0134

BNL3147136NAU1366139CIR013143NAU1148144MUSB1197146

CIR254151

E43M48-M113.1160

BNL1705167

c21

JESPR300A16140

CIR2722CIR3628BNL353715CIR08120BNL167924M6325E31M61-M129.627E32M48-M401.128CMS37E43M49-M238.0BNL1673E32M51-M95.0

34

NAU4089BNL326136

CIR167BNL170739

E41M47-M190.146NAU127847BNL122749E32M54-M229.450E35M52-M339.3E43M49-M139.2E31M48-M242.2

54

E43M52-M167.8E35M54-M243.6MUSB1035

55

BNL2967CIR167CIR291

56

CIR13357BNL019361BNL359962JESPR27064MUSS02665BNL262167E43M50-M193.268E42M48-M309.269BNL326171MUSB111775

c12

NAU38960JESPR3001

E43M52-M274.48CIR2729

BNL353720E32M49-M463.122E43M52-M105.523

E41M47-M315.432BNL348234E43M52-M239.135BNL034138E35M50-M91.639MUCS06440E43M49-M358.846E32M49-M209.2E43M49-M319.1E35M48-M116.2

52

E31M48-M151.954MUSS30355BNL359957pGH33162

CIR17076

c26

BNL32810A13453NAU22857CIR0578MUSS20326E42M50-M259.528pet29JESPR15331MUSB0399CIR338E32M51-M347.1

32

BNL402933NAU102336E32M51-M185.2MUSB126337

MUSB0958E41M50-M100.5CIR079

38

BNL3623MUSB126339

MUSB0818E35M54-M406.840

E32M48-M145.541BNL2652CIR09642

CIR38045CIR34246E43M50-M362.347E41M50-M315.659BNL170762BNL325464pGH351BNL040965

MUSS18171CIR05473

c13

E36M47-M-552.00

BNL32814

CIR21611BNL2571JESPR20416

E42M48-M364.920NAU313021NAU386124NAU384326NAU344727CIR09934MUSS60336Gate1CD12NAU1215CIR277

37

Gate1AG05MUCS40541

MUSS20345NAU244346E31M55-M415.3BNL344555

BNL3445BNL328056

CIR04058BNL1079E43M52-M407.459

MUSB1263E35M48-M81.860

E32M50-M279.7E43M48-M215.562

E43M49-M144.564MUSB056466BNL265271BNL355882E43M49-M124.885MUSB098491BNL254497BNL0193100

c18

123

288 Theor Appl Genet (2009) 119:281–292

genetic map recently reported by Guo et al. (2007) from1,790 loci and 138 BC1 progeny.

Locus order is globally well conserved between the con-sensus and component maps, with a high overall consis-tency for the majority of the 26 chromosomes (Fig. 2, Fig.A1 and A2). However, a few notable diVerences in locusorder were observed:

c4. several loci in the bottom region of the BC1 map werenot integrated into the consensus map because of looselinkages (these include pAR574, CIR223, CIR142),

c5. an important shrinkage in the corresponding topregions of the BC1 and RIL maps was caused by uncer-tainties in the BC1 map (2 loci not integrated into theconsensus map, NAU1223 and NAU1042), and in theRIL map (5 SSR loci not integrated, NAU1137,Gate3BG11, NAU1223, BNL4071, BNL3029, andNAU1042),

c6. three loci of the bottom region of the BC1 map were notintegrated into the consensus map because of looselinkages (these include 2 SSRs, CIR128, CIR179, andthe RFLP, M16-147),

c7. the stretching in the top region of the BC1 map wascaused by the wrong position of the top distal locusCIR028 correctly relocated in the consensus map,

c8. 20 loci of the RIL map were not integrated into theconsensus map by JoinMap, including 16 AFLPs and 4SSRs (NAU0920, NAU2407, NAU0250, NAU2439),

c9. the stretching in the top region of the BC1 map wascaused by three AFLP loci and two SSRs (BNL1423and BNL1162); an overall re-arrangement in the con-sensus map was proposed in agreement with other pub-lished maps,

c21.four SSR loci of the RIL map were not integrated intothe consensus map by JoinMap (BNL1034, NAU1366,NAU1148, and MUSB197); length of consensus c21may be overestimated because of the existence of anextrapolated 25 cM gap in the RIL map,

c23.two loci (CIR198 and CIR019) positioned 42 cM apartin the top region of the BC1 map were repositioned inthe consensus map in agreement with other publishedmaps,

c25.the RIL map was lacking 1/3rd in the bottom regionbecause of a lack of scored markers; as a consequencelength of consensus c25 may be overestimated.

Variation of locus density along chromosomes wasassessed by counting loci in 5-cM intervals and visualizedin Fig. A3. The distribution of loci along chromosomes wasclearly not random. A single region of high marker density

Fig. 2 SimpliWed representation of the BC1 and RIL maps originatingfrom the Guazuncho 2 £ VH8 cross aligned to an integrated consensusmap (cons). RIL and consensus maps were produced using JoinMap.For consistency with published data, the BC1 map is unchanged fromearlier reports (Nguyen et al. 2004), i.e., obtained with MapMaker 3.0.

Three chromosomes (c1, c2 and c3) are shown; complete series of 26chromosomes available from supplementary Fig. A1. Presence of alocus is denoted by a horizontal line within the chromosome bar andonly loci in common between BC1 and RIL maps are connected

c1(BC1)

c1(cons)

c1(RIL)

c2(BC1)

c2(cons)

c2(RIL)

c3(BC1)

c3(RIL)

c3(cons)

123

Theor Appl Genet (2009) 119:281–292 289

Xanked by regions of lower density was observed on eachof the 26 chromosomes except for c9, c10, c16, and c19which displayed more uniform marker distribution.

Discussion

Segregation distortion

The cotton RIL population described here displays severe,genome-wide allelic distortion towards the G. hirsutum par-ent (71% G. hirsutum alleles and 29% G. barbadense alle-les). Biased transmission of parental alleles is a commonfeature of mosaic interspeciWc plant populations (Zamir andTadmor 1986). In cotton, a comparable biased transmissionwas observed in interspeciWc (G. hirsutum £ G. barba-dense) BC1 (Jiang et al. 2000) and RIL populations (Heet al. 2008 and Mauricio Ulloa, personal communication),with in all cases G. hirsutum alleles being favored. Themost probable explanation for the observed distortionrelates to an (unintended) selection bias towards theG. hirsutum parent during the SSD process. It is known thatinterspeciWc Gh £ Gb progenies suVer from fertility break-down, and this was also true for the current RIL population.Moreover, the G. barbadense parent used in this study hassome less favorable reproductive features, such as latenessin Xowering and reduced seed numbers per boll as com-pared to the G. hirsutum parent. We assume that ourattempts to produce progeny seed within a reasonable timeframe has resulted in inadvertent selection against individu-als bearing G. barbadense alleles that negatively inXuenceXowering time and fruiting properties. Distorted segrega-tion at a genome-wide scale has also been reported forinterspeciWc RIL populations in a number of other species,including chickpea (Cobos et al. 2006), tomato (De Vicenteand Tanksley 1993; Paran et al. 1995) and rice (Wang et al.1994; Lee et al. 2005; Jie et al. 2006). In all these cases, thepercent of skewed loci ranged from 54 to 80% and dis-played a scattered distribution throughout the genome. Thesegregation distortion, however, should not aVect the utilityof this population for mapping QTLs for signiWcant agro-nomic and Wber quality traits observed in the RILs.

Map distances in the RIL map

Despite the marker distortion it was possible to develop acoherent genetic map from the genotype data of 140 RILs.The RIL map provided good coverage of all 26 cotton chro-mosomes. However, after eliminating the eVect of the map-ping software, we observed that the map size and markerdistances of the RIL map were markedly reduced (56%)compared with the size and distances of syntenic regions inthe published BC1 map from the same cross. The size

reduction, in theory, should not relate to the diVerence inpopulation type (BC1 vs. RIL) and consequential diVerencein inferred numbers of meioses, because the mapping soft-ware compensates for these eVects. We have identiWed twofactors that may have contributed to the observed diVer-ences in map distances: (1) diVerent rates of missing dataand coding errors between the BC1 and RIL data sets and(2) the allelic imbalance in the RIL data. Hackett andBroadfoot (2003) have shown that coding errors increasemap size estimates, while missing data reduce map sizeestimates. AFLP codings are generally thought to be moreprone to coding errors (false positives, homoplasy) thanother marker systems. Since the BC1 map contained moreAFLP loci than the RIL map (465 vs. 281) and since AFLPmarkers for the RIL map were more stringently selected,we expect that more coding errors were included in the BC1

map, which may have resulted in overestimation of mapdistances in this map. Conversely, the presence of moremissing data in the RIL data set (6.0%) compared to theBC1 data set (2.2%), may have resulted in an underestima-tion of map distances in the RIL map. The RIL populationdisplayed strong distortion towards the G. hirsutum parent,while the BC1 population was more balanced. The biastowards one parent may have lowered the number of “eVec-tive” (detectable) recombinations in the RIL population,while this would not be the case in the BC1 population.DiVerences in map distances between various maps fromthe same species (even with identical population types),as well as a strong inXuence of the mapping softwarewere previously reported (Qi et al. 1996; He et al. 2001;Armstead et al. 2002; Knox and Ellis 2002). In cotton, thereported sizes of the seven most saturated interspeciWcgenetic maps range from 3,500 to 5,300 cM or even more(Nguyen et al. 2004; He et al. 2007; Guo et al. 2008; Zhanget al. 2008). Thus, the current RIL map and Guazuncho2 £ VH8 consensus map are positioned at the lower end ofcotton map sizes. We believe, together with Wang et al.(1994), that map distances can be expected to Xuctuateamong crosses, while marker assignation and order alongchromosomes are expected to be repeatable betweenstudies.

The Guazuncho 2 £ VH8 consensus map

Comparative analysis of genetic maps and map-associatedinformation (such as QTL data) is a complex and laboriousprocess, especially if data originating from diVerent labora-tories must be compared. In principle, two approaches maybe considered: (1) map comparison based on Wnal individ-ual map versions, using one selected map as a stable refer-ence (Truco et al. 2007), or (2) a true integration of the rawdata and re-computation of an integrated map (through asoftware package that uses the base data). Two recent

123

290 Theor Appl Genet (2009) 119:281–292

publications reported on the integration of cotton mapsusing the Wrst approach (Rong et al. 2007; Xu et al. 2008).In both cases, the race palmeri £ K101-F2 map of Ronget al. (2004) was considered as a basis for performing eithermeta-analysis of QTLs from diVerent experiments (Ronget al. 2007) or for the integration of genetic and physicalmaps of chromosomes 12 and 26 (Xu et al. 2008). In thispaper we describe map integration based on raw data. Weshow that, with exception of a few minor inconsistencies inlocus order, the consensus map faithfully represents thelocus order of the two component maps. We believe thatthis approach will help in providing a stronger basis forfuture comparative mapping and meta-QTL analysis.

Compared with the component maps, the Guazuncho £VH8 consensus map has improved genome coverage at twolevels: (1) higher marker density (one locus per 2 cM) and(2) reduced number of gaps (13 gaps exceed 15 cM ofwhich 5 span 20–25 cM). The distribution of loci alongindividual chromosomes is still unequal in the consensusmap. A single region of higher locus density is typicallyobserved on 20 of 26 chromosomes (Fig. A3). AFLP mark-ers are slightly over-represented in these locus-denseregions (56 vs. 41% overall). Clustering of EcoRI/MseIAFLP markers is expected to occur preferentially in centro-meric regions (Young et al. 1999). Thus, the marker-denseregions could indicate centromeres.

Alignment of the Guazuncho £ VH8 consensus mapwith other genetic maps provided additional evidence(more duplicated loci) to substantiate the 13 knownhomoeologous A/D relationships of tetraploid cotton(Fig. A2), bridged by 104 pairs of homoeologous dupli-cations (86 SSR and 18 RFLP), and also conWrmed the

known reciprocal translocations involving the two pairsof A chromosomes, c2 and c3 (12 duplicated loci), andc4 and c5 (6 loci).

The consensus map as a community resource

The Guazuncho £ VH8 consensus map reported here is ofparticular interest to the cotton research community due tothe fact that it contains the highest number of “common”markers when compared with other public cotton maps(Table 2 and CMap display from http://tropgenedb.cirad.fr/en/cotton.html).

Common markers were calculated for six genome-widegenetic maps of cotton, containing (after exclusion of para-logous duplications) 7,397 genetic loci. The following sixmaps were included:

– Guazuncho 2 £ VH8-consensus (GV-cons) and its twocomponent maps, GV–BC1 and GV–RIL,

– CRI36 £ Hai7124–F2 (CH–F2) reported by Yu et al.(2007),

– Handan 208 £ Pima 90–F2 (HP-F2) reported by He et al.(2006),

– TM1 £ 3-79–RIL (T3–RIL) reported by Frelichowskiet al. (2007),

– TM1 £ Hai 7124–BC1 (TH–BC1) reported by Guo et al.(2004),

– Race palmeri £ K101–F2 (PK–F2) reported by Ronget al. (2004).

The 7,397 loci identiWed on the six maps represent 6,046diVerent loci: 5,254 are unique to one speciWc map and 792(13%) are shared between two and six maps (6 loci shared

Table 2 Distribution of shared loci between 8 whole genome cotton genetic maps: Guazuncho £ VH8-consensus (GV-cons) and its two compo-nent (GV–BC1 and GV–RIL) maps, and Wve interspeciWc G. hirsutum £ G. barbadense maps, CH–F2, HP–F2, T3–RIL, TH–BC1, and PK–F2

a Populations compared are: Guazuncho 2 £ VH8-consensus (GV-cons) in this article; Guazuncho 2 £ VH8–BC1 (GV–BC1) reported by Nguyenet al. (2004); Guazuncho 2 £ VH8–RIL (GV–RIL) in this article; CRI36 £ Hai7124–F2 (CH–F2) by Yu et al. (2007); Handan 208 £ Pima 90–F2(HP-F2) by He et al. (2007); TM1 £ 3-79–RIL (T3–RIL) by Frelichowski et al. (2006); TM1 £ Hai 7124–BC1 (TH–BC1) by Guo et al. (2007);palmeri £ K101 F2 (PK–F2) by Rong et al. (2004)b The number of loci is the one reported in referenced publication

A locus is common between two maps when it is (1) derived from the same marker and (2) mapped on the same chromosome on two maps (onlyone locus considered per chromosome/linkage group in case of paralogous duplication on the same LG)

Populationa No locib GV-cons GV–BC1 GV–RIL CH–F2 HP–F2 T3–RIL TH–BC1 PK–F2

GV-cons 1,745

GV-BC1 1,271 1,226

GV-RIL 800 546 255

CH-F2 1,097 296 260 184

HP-F2 1,029 243 235 136 195

T3-RIL 433 116 75 83 89 60

TH-BC1 1,790 313 249 205 236 180 85

PK-F2 2,584 203 202 50 68 68 28 64

Total common 657 – – 399 302 162 400 221

123

Theor Appl Genet (2009) 119:281–292 291

between all 6 maps, 49 between 5 maps, 104 between 4maps, 148 between 3 maps, and 477 between 2 maps).Except for the comparison of PK–F2 with GV-cons whichis partially based on RFLP markers, all other cross-popula-tion comparisons rely exclusively on SSR loci. SSR mark-ers from the BNL collection are shared between the highestnumbers of maps. This is easy to understand, as this was theWrst SSR library developed in cotton. The GV-consensusmap (by far) contains the highest numbers of bridge loci ina single map (Table 2), with 657 loci (38% of its loci)shared with at least one of the Wve other framework maps.This establishes the GV-consensus map as an excellent toolfor future map integration.

Acknowledgments The French National Research Agency, ANR,through its program specialized in plant genomics, Génoplante, hassponsored this research (project nr ANR-06-GPLA-018) that wasotherwise supported by Bayer CS and CIRAD funding.

References

Armstead I, Turner L, King I, Cairns A, Humphreys M (2002) Com-parison and integration of genetic maps generated from F2 andBC1-type mapping populations in perennial ryegrass. PlantBreeding 121:501–507

Bradeen JM, Staub JE, Wye C, Antonise R, Peleman J (2001) Towardsan expanded and integrated linkage map of cucumber (Cucumissativus L.). Genome 44:111–119

Burr B, Burr F, Thompson KH, Alberston MC, Stuber CW (1988)Gene mapping with recombinant inbreds in maize. Genetics118:519–526

Chee P, Draye X, Jiang CX, Decanini L, Delmonte TA, Bredhauer R,Smith CW, Paterson AH (2005) Molecular dissection of interspe-ciWc variation between Gossypium hirsutum and Gossypiumbarbadense (cotton) by a backcross-self approach: I. Fiber elon-gation. Theor Appl Genet 111:757–763

Cobos MJ, Rubio J, Strange RN, Moreno MT, Gil J, Millan T (2006)A new QTL for Ascochyta blight resistance in an RIL populationderived from an interspeciWc cross in chickpea. Euphytica149:105–111

De Vicente MC, Tanksley SD (1993) QTL analysis of transgressivesegregation in an interspeciWc tomato cross. Genetics 134:586–596

Dellaporta SL, Wood J, Hicks JN (1983) A plant DNA miniprepara-tion: Version II. Plant Mol Biol Rep 1(4):19–21

Frelichowski JE, Palmer MB, Main D, Tomkins JP, Cantrell RG, StellyDM, Yu J, Kohel RJ, Ulloa M (2006) Cotton genome mappingwith new microsatellites from Acala ‘Maxxa’ BAC-ends. MolGenet Genomics 275:479–491

Fu Y, Wen TJ, Ronin YI, Chen HD, Guo L, Mester DI, Yang Y, LeeM, Korol AB, Ashlock DA, Schnable PS (2006) Genetic dissec-tion of intermated recombinant inbred lines using a new geneticmap of maize. Genetics 174:1671–1683

Guo W, Cai C, Wang C, Han ZG, Song XL, Wang K, Niu X, Wang C,Lu K, Shi B, Zhang T (2007) A microsatellite-based, gene-en-riched linkage map reveals genome structure, function, and evo-lution in Gossypium. Genetics 176:527–541

Guo W, Cai C, Wang C, Zhao L, Wang L, Zhang T (2008) A prelimi-nary analysis of genome structure and composition in Gossypiumhirsutum. BMC Genomics 9:314

Hackett CA, Broadfoot LB (2003) EVects of genotyping errors, miss-ing values and segregation distortion in molecular marker data onthe construction of linkage maps. Heredity 90:33–38

He P, Li JZ, Zheng XW, Shen lS, Lu CF, Chen Y, Zhu LH (2001)Comparison of molecular linkage maps and agronomic trait locibetween DH and RIL populations derived from the same ricecross. Crop Sci 41:1240–1246

He DH, Lin Z, Zhang X, Nie YC, Guo XP, Zhang Y, Li W (2007) QTLmapping for economic traits based on a dense genetic map ofcotton with PCR-based markers using the interspeciWc cross ofGossypium hirsutum £ Gossypium barbadense. Euphytica153:181–197

He DH, Lin ZX, Zhang XL, Zhang YX, Li W, Nie YC, Guo XP (2008)Dissection of genetic variance of Wbre quality in advancedgeneration from an interspeciWc cross of Gossypium hirsutum andG. barbadense. Plant Breed 127:286–294

Jiang C, Chee P, Draye X, Morrell P, Smith CW, Paterson AH (2000)Multilocus interactions restrict gene introgression in interspeciWcpopulations of polyploid Gossypium (Cotton). Evolution54(3):798–814

Jie C, Bughio H, Chen DZ, Liu G, Zheng K, Zhuang J (2006) Devel-opment of chromosomal segment substitution lines from a back-cross recombinant inbred population of interspeciWc rice cross.Rice Sci 13:15–21

Knox MR, Ellis THN (2002) Excess heterozygosity contributes togenetic map expansion in pea recombinant inbred populations.Genetics 162:861–873

Lacape J-M, Nguyen TB (2005) Mapping quantitative trait loci associ-ated with leaf and stem pubescence in cotton. J Hered 96(4):441–444

Lacape J-M, Nguyen TB, Thibivilliers S, Courtois B, Bojinov BM,Cantrell RG, Burr B, Hau B (2003) A combined RFLP-SSR-AFLP map of tetraploid cotton based on a Gossypiumhirsutum £ Gossypium barbadense backcross population.Genome 46:612–626

Lacape J-M, Nguyen TB, Courtois B, Belot J-L, Giband M, Gourlot J-P, Gawryziak G, Roques S, Hau B (2005) QTL analysis of cottonWber quality using multiple G. hirsutum £ G. barbadense back-cross generations. Crop Sci 45:123–140

Lacape J-M, Dessauw D, Rajab M, Noyer JL, Hau B (2007) Microsat-ellite diversity in tetraploid Gossypium gene pool: assembling ahighly informative genotyping set of cotton SSRs. Mol Breed19:45–58

Lee SJ, Oh CS, Suh JP, McCouch S, Ahn SN (2005) IdentiWcation ofQTLs for domestication-related and agronomic traits in an Oryzasativa £ O. ruWpugon BC1F7 population. Plant Breed 124:20–219

Loudet O, Chaillou S, Camilleri C, Bouchez D, Daniel-Vedele F(2002) Bay-O £ Shahdara recombinant inbred line population: apowerful tool for the genetic dissection of complex traits inArabidopsis. Theor Appl Genet 104:1173–1184

Mei M, Syed NH, Gao W, Thaxton PM, Smith CW, Stelly DM, ChenZJ (2004) Genetic mapping and QTL analysis of Wber-relatedtraits in cotton (Gossypium). Theor Appl Genet 108:280–291

Nguyen TB, Giband M, Brottier P, Risterucci A-M, Lacape J-M (2004)Wide coverage of tetraploid cotton genome using newly devel-oped microsatellite markers. Theor Appl Genet 109:167–175

Paran I, Goldman S, Tanksley SD, Zamir D (1995) Recombinationinbred lines for genetic mapping in tomato. Theor Appl Genet90:542–548

Park Y-H, Alabady MS, Sickler B, Wilkins TA, Yu J, Stelly DM,Kohel RJ, El-Shihy OM, Cantrell RG, Ulloa M (2005) Geneticmapping of new cotton Wber loci using EST-derived microsatel-lites in an interspeciWc recombinant inbred line (RIL) cottonpopulation. Mol Genet Genomics 274:428–441

123

292 Theor Appl Genet (2009) 119:281–292

Paterson AH, Saranga Y, Menz M, Jiang C, Wright RJ (2003) QTLanalysis of genotype £ environment interactions aVecting cottonWber quality. Theor Appl Genet 106:384–396

Qi XP, Stam P, Lindhout P (1996) Comparison and integration of fourbarley linkage maps. Genome 39:379–394

Risterucci A-M, Grivet L, N’Goran JAK, Pieretti I, Flament MH,Lanaud C (2000) A high density linkage map of Theobroma ca-cao L. Theor Appl Genet 101:948–955

Rong J, Abbey C, Bowers JE, Brubaker CL, Chang C, Chee P, Del-monte TA, Ding X, Garza JJ, Marler BS, Park C-h, Pierce GJ,Rainey KM, Rastogi VK, Schulze SR, Trolinder N, Wendel JF,Wilkins TA, Williams-Coplin TD, Wing RA, Wright RJ, Zhao X,Zhu L, Paterson AH (2004) A 3,347 locus genetic recombinationmap of sequence-tagged sites reveals features of genome organi-zation, transmission and evolution of cotton (Gossypium). Genet-ics 166:389–417

Rong J, Feltus FA, Waghmare VN, Pierce GJ, Chee P, Draye X,Saranga Y, Wright RJ, Wilkins TA, May LO, Smith CW, Gann-away JR, Wendel JF, Paterson AH (2007) Meta-analysis of poly-ploid cotton QTL shows unequal contributions of subgenomes toa complex network of genes and gene clusters implicated in lintWber development. Genetics 176:2577–2588

Shen X, Guo W, Lu Q, Zhu X, Yuan Y, Zhang T (2007) Genetic map-ping of quantitative trait loci for Wber quality and yield trait byRIL approach in Upland cotton. Euphytica 155:371–380

Stam P (1993) Construction of integrated genetic linkage maps bymeans of a new computer package: JoinMap. Plant J 3:739–744

Stam P, Van Ooijen JW (1995) JOINMAP version 2.0: software for thecalculation of genetic linkage maps

Törjék O, Witucka-Wall H, Meyer RC, von KorV M, Kusterer B,Rautengarten C, Altmann T (2006) Segregation distortion inArabidopsis C24/Col-0 and Col-0/C24 recombinant inbred linepopulations is due to reduced fertility caused by epistatic interac-tion of two loci. Theor Appl Genet 113:1551–1561

Truco MJ, Antonise R, Lavelle D, Ochoa O, Kosik A, Witsenboer H,Fort SB, Jeuken MJW, Kesseli RV, Lindhout P, Michelmore RW,Peleman J (2007) A high-density, integrated genetic linkage mapof lettuce (Lactuca spp.). Theor Appl Genet 115:735–746

Van Berloo R (2008) GGT 2.0: versatile software for visualization andanalysis of genetic data. J Hered 99:232–236

Van Ooijen JW (2006) JoinMap® 4.0, Software for the calculation ofgenetic linkage maps in experimental populations. Kyazma B.V.Wageningen, Netherlands

Voorrips R (2002) MapChart: software for the graphical presentationof linkage maps and QTLs. J Hered 93(1):77–78

Vos P, Hogers R, Bleeker M, Reijans M, van de Lee T, Hornes M,Fritjers A, Pot J, Peleman J, Kuiper M, Zabeau M (1995) AFLP:a new technique for DNA Wngerprinting. Nucleic Acids Res23(21):4407–4414

Wang G-L, Mackill DJ, Bonman JM, McCouch S, Champoux MC,Nelson RJ (1994) RFLP mapping of genes conferring completeand partial resistance to blast in a durable resistant rice cultivar.Genetics 136:1421–1434

Wang B-H, Guo W, Zhu X, Wu Y, Huang N, Zhang T (2006a) QTLmapping of Wber quality in an elite hybrid derived-RIL populationof upland cotton. Euphytica 152:367–378

Wang K, Song XL, Han ZG, Guo W, Yu J, Sun J, Pan J, Kohel RJ,Zhang TZ (2006b) Complete assignment of the chromosomes ofGossypium hirsutum L. by translocation and Xuorescence in situhybrisation mapping. Theor Appl Genet 113:73–80

Wang B, Wu Y, Guo W, Zhu X, Huang N, Zhang T (2007) QTLanalysis and epistasis eVects dissection of Wber qualities in an elitecotton hybrid grown in second generation. Crop Sci 47:1384–1392

Wendel JF, Olson PD, Stewart JM (1989) Genetic diversity, introgres-sion, and independent domestication of old world cultivatedcottons. Am J Bot 76(12):1795–1806

Xu Z, Kohel RJ, Song G, Cho J, Yu J, Yu S, Tomkins JP, Yu J (2008)An integrated genetic and physical map of homologous chromo-somes 12 and 26 in Upland cotton (G hirsutum L.). BMCGenomics

Young WP, Schupp JM, Keim P (1999) DNA methylation and AFLPmarker distribution in the soybean genome. Theor Appl Genet99:785–790

Yu J, Yu S, Lu C, Wang W, Fan S, Song M, Lin Z, Zhang X, Zhang J(2007) High-density linkage map of cultivated allotetraploidcotton based on SSR, TRAP, SRAP and AFLP markers. J IntegrPlant Biol 49(5):716–724

Zamir D, Tadmor Y (1986) Unequal segregation of nuclear genes inplants. Bot Gaz 147:355–358

Zhang YX, Lin ZX, Xia QZ, Zhang MJ, Zhang XL (2008) Character-istics and analysis of simple sequence repeats in the cottongenome based on a linkage map constructed from a BC1 popula-tion between Gossypium hirsutum and G. barbadense. Genome51:534–546

123