A genetic analysis of seed and berry weight in grapevine

A genetic analysis of seed and berry weight in grapevine
of 14
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
   A genetic analysis of seed and berry weight ingrapevine J.A.Cabezas,M.T.Cervera,L.Ruiz-Garcı´ a,J.Carren ˜ o,andJ.M.Martı´ nez-Zapater Abstract:  Fruit size and seedlessness are highly relevant traits in many fruit crop species, and both are primary targets of breeding programs for table grapes. In this work we performed a quantitative genetic analysis of size and seedlessness inan F 1  segregating population derived from the cross between a classical seeded ( Vitis vinifera  L. ‘Dominga’) and a newlybred seedless (‘Autumn Seedless’) cultivar. Fruit size was scored as berry weight (BW), and for seedlessness we consid-ered both seed fresh weight (SFW) and the number of seeds and seed traces (SN) per berry. Quantitative trait loci (QTL)analysis of BW detected 3 QTLs affecting this trait and accounting for up to 67% of the total phenotypic variance. QTLanalysis for seedlessness detected 3 QTLs affecting SN (explaining up to 35% of total variance) and 6 affecting SFW (ex-plaining up to 90% of total variance). Among them, a major effect QTL explained almost half of the phenotypic variationfor SFW. Comparative analysis of QTLs for these traits reduced the number of grapevine genomic regions involved, oneof them being a major effect QTL for seedlessness. Association analyses showed that microsatellite locus VMC7F2,closely linked to this QTL, is a useful marker for selection of seedlessnes. Key words: Vitis , table grape, genetic mapping, QTL mapping, berry size, seedlessness, seed number. Re´sume´ :  La taille du fruit et l’absence de pe´pins sont des caracte`res reveˆtant un grand inte´reˆt chez plusieurs espe`ces frui-tie`res et les deux font l’objet d’efforts dans le cadre de programmes de se´lection du raisin de table. Dans ce travail, les au-teurs ont re´alise´ une e´tude ge´ne´tique quantitative de la taille de fruits et de l’absence de pe´pins chez une population F 1  ense´gre´gation de´rive´e du croisement entre un cultivar classique a` pe´pins (‘Dominga’) et un nouveau cultivar sans pe´pins(‘Autumn Seedless’). La taille des fruits a e´te´ mesure´e en de´terminant la masse des baies, tandis que pour les pe´pins, lesauteurs ont mesure´ tant la masse fraıˆche des pe´pins (SFW) que le nombre de pe´pins ou de traces de pe´pins (SN) par baie.Une analyse QTL de la masse des baies a de´tecte´ 3 QTL affectant ce caracte`re et expliquant jusqu’a` 67 % de la variancephe´notypique totale. L’analyse QTL pour les pe´pins a de´cele´ 3 QTL influenc¸ant le nombre de pe´pins (expliquant jusqu’a`35 % de la variance totale) et 6 QTL affectant la masse fraıˆche des pe´pins (expliquant jusqu’a` 90 % de la variance totale).Parmi ceux-ci, un QTL majeur expliquait presque la moitie´ de la variation phe´notypique pour le SFW. Une analyse com-pare´e des QTL pour ces caracte`res a permis de re´duire le nombre de re´gions ge´nomiques implique´es chez la vigne, l’unece celles-ci e´tant un QTL majeur pour l’absence de pe´pins. Des analyses d’association ont montre´ que le locus microsatel-lite VMC7F2, e´troitement lie´ a` ce QTL, constitue un marqueur utile pour se´lectionner l’absence de pe´pins.  Mots cle´ s : Vitis , raisin de table, cartographie ge´ne´tique, cartographie QTL, taille des baies, absence de pe´pins, nombre degraines.[Traduit par la Re´daction] Introduction Large fruit size and reduced seed number and size aremajor goals in the genetic improvement of fruit crop species(Varoquaux et al. 2000). In table grape, seedlessness is oneof the most appreciated quality traits. Most breeding pro-grams have focused on the generation of new table grapecultivars combining seedlessness with other berry qualitytraits, such as large size, muscat flavour, or crispiness(Loomis and Weinberger 1979). Two different types of seedlessness have been observed among grape genetic re-sources: parthenocarpy and stenospermocarpy (Stout 1936).Parthenocarpy, or fruit development in the absence of polli-nation, yields small berries that completely lack seeds, suchas ‘Corinto’ and related cultivars, mainly used for the pro-duction of seedless raisins (Ledbetter and Ramming 1989).In stenospermocarpy, pollination takes place, but seed devel- Received 9 June 2006. Accepted 14 September 2006. Publishedon the NRC Research Press Web site at on27 February 2007.Corresponding Editor: J.P. Gustafson. J.A. Cabezas and J.M. Martı´nez-Zapater. 1 Departamento deGene´tica Molecular de Plantas, Centro Nacional deBiotecnologı´a, Consejo Superior de Investigaciones Cientı´ficas(CSIC), C/ Darwin 3, 28049 Madrid, Spain. M.T. Cervera.  Centro de Investigacio´n Forestal, InsititutoNacional de Investigacio´n y Tecnologı´a Agraria y Alimentaria(INIA), Ctra. de A Corun˜a Km 7, 28040 Madrid, Spain. L. Ruiz-Garcı´a and J. Carren˜o.  Instituto Murciano deInvestigacio´n y Desarrollo Agrario, Instituto Murciano deInvestigacio´n y Desarrollo Agrario y Alimentario (IMIDA), LaAlberca, 30150 Murcia, Spain. 1 Corresponding author (e-mail: 1572 Genome  49 : 1572–1585 (2006) doi:10.1139/G06-122   2006 NRC Canada  opment fails because of early degeneration of endosperm(Ledbetter and Ramming 1989). Thus, stenospermocarpiccultivars are not strictly seedless but contain seminal rudi-ments or seed traces of different sizes bearing viable em-bryos that can be germinated by in vitro culture techniques(Cain et al. 1983). Several grape cultivars, such as ‘Muscatof Alexandria’, occasionally produce stenospermocarpic ber-ries (Winkler et al. 1974). However, a stable stenospermo-carpy phenotype is found in only a few ancient orientalcultivars known as ‘Kishmish’ and derived cultivars. Amongthem, the white ‘Kishmish’, also known as ‘Sultanina’ or‘Thomson Seedless’ (Dangl et al. 2001), has been the mainsource of seedlessness in table grape breeding programs(Adam-Blondon et al. 2001). Unfortunately, as in other fruitcrop species (Crane 1965), in grapevine there is a direct re-lation between seed number and berry size (Coombe 1973)that results in an undesired negative correlation betweenseedlessness and berry size (Fanizza et al. 2005; Wei et al.2002). This correlation likely results from the fact that gib-berellins produced by seeds are required to promote berrygrowth during late berry developmental stages (Coombe1960; Perez et al. 2000). Identification of the genetic deter-minants responsible for seedlessness and berry size couldprovide alternatives to dissociate this negative correlation,as well as closely linked molecular markers to assist breed-ing programs.Stenospermocarpy is a quantitative trait. All degrees of seed development can be observed in seedless cultivarsand in progenies segregating for this trait. Qualitative anal-yses of stenospermocarpy in different progenies have ledto different hypothesis explaining its genetic control(Bouquet and Danglot 1996). Analysis of the progeny of 2partially seedless grape genotypes led to a genetic modelin which seedlessness would be controlled by recessive al-leles at 3 independent loci regulated by a dominant alleleat a single locus inhibiting the development of the seed(Bouquet and Danglot 1996). This locus was later namedas  SdI   for seed development inhibitor (Lahogue et al.1998). Studies focused on the identification of markerslinked to seedlessness using bulked segregant analysis(BSA; Michelmore et al. 1991) have identified markerslinked to this putative  SdI   locus in specific segregatingprogenies (Adam-Blondon et al. 2001; Lahogue et al.1998; Mejı´a and Hinrichsen 2003; Striem et al. 1996).However, their usefulness was reduced when tested in ad-ditional segregating progenies (Adam-Blondon et al. 2001;Mejı´a and Hinrichsen 2003).The large availability of molecular markers for grape ge-nome analysis has opened up the possibility of performingquantitative genetic analyses of traits such as berry weightand seedlessness. Nevertheless, the reduced population sizesand the limitations posed by the 2-way pseudo-testcrossmapping strategy (Grattapaglia and Sederoff 1994), com-monly used to build maps in highly heterozygous species us-ing F 1  progenies, only allow detection of quantitative traitloci (QTLs) with large phenotypic effects. Several geneticmaps of different grape cultivars and  Vitis  species havebeen built in recent years using these strategies (Adam-Blondon et al. 2004; Dalbo´ et al. 2000; Doligez et al. 2002;Doucleff et al. 2004; Fischer et al. 2004; Grando et al. 2003;Lodhi et al. 1995; Riaz et al. 2003), and preliminary resultsof QTL detection for berry size and seedlessness have beengenerated (Doligez et al. 2002; Fanizza et al. 2005; Fischeret al. 2004). Doligez et al. (2002) showed the existence of 1major effect QTL for seedlessness, explaining up to 49% of the total phenotypic variance for seed fresh weight (SFW) ina controlled cross between 2 partially seedless varieties.This QTL likely corresponds to the  SdI   locus in the Bouquetand Danglot (1996) model. Unfortunately, in this work thelinkage group (LG) involved did not include any marker incommon with other available  Vitis  genetic maps allowing itsidentification. QTLs for berry size and seed number weredetected colocalizing with  SdI   in the same study. AdditionalQTLs with small effects were also found for berry weight,seed number, and seed dry weight but only in one of the 3analyzed seasons. The genetic control of berry size has alsobeen analyzed in 2 additional segregating progenies devel-oped for the genetic analysis of disease resistances in winecultivars (Fischer et al. 2004) and fruit yield components intable grapes (Fanizza et al. 2005). These studies reported theidentification of additional QTLs for berry size on differentlinkage groups.To further investigate the complex genetic control of seedlessness and berry size in table grape, we have analyzeda different progeny derived from a seeded (‘Dominga’) anda seedless (‘Autumn Seedless’) cultivar. For this purpose weconstructed parental and consensus genetic maps, mainlybased on amplified fragment length polymorphism (AFLP)and microsatellite markers, and performed QTL analyses of seed weight, seed number, and berry weight. Our resultsconfirm the existence of a major effect QTL affecting bothseed and berry weight on LG18. In addition, we identifynew QTLs accounting for berry size, seed weight, and seednumber in grape. Finally, we show the usefulness of anLG18 microsatellite locus, VMC7F2, as a marker for seed-lessness breeding in table grape. Materials and methods Mapping population The F 1  segregating population used for the genetic analy-ses was derived from controlled crosses between cultivars‘Dominga’ (female progenitor) and ‘Autumn Seedless’(male progenitor) performed in 1988 and 1989. ‘Dominga’is a seeded cultivar of unknown srcin cultivated in the re-gion of Murcia (Spain). ‘Autumn Seedless’ is a seedlessbred cultivar derived from the cross between cultivar‘Calmeria’ and a hybrid progeny of ‘Sultanina’ and ‘Muscatof Alexandria’ ( Vitis  International Variety Catalogue; http:// Cultivar ‘Calmeria’, pre-viously considered to be the result of open pollination of cultivar ‘Aledo’ ( Vitis  International Variety Catalogue; is likely an F 1  prog-eny of ‘Aledo’ and ‘Sultanina’ according to microsatellitegenotyping (Iban˜ez et al. personal communication, 2005).The mapping population consisted of 118 hybrid plants andrepresentative plants of both progenitors. All of them weregrown in the same field and on their own roots. Vine androw spacing were 2.5 and 1 m, respectively, with an east–west orientation. The vines were supported on overhead ten-done trellis (parral) and pruned to 2 canes with 10–12 nodes.Irrigation (5500 m 3  /ha per year) and fertilization regimes Cabezas et al. 1573   2006 NRC Canada  were determined largely by grower experience. Fertilizationwas uniformly performed across all blocks and included120 U of N per hectare and year (U/ha yearly), 80 U/hayearly of P 2 O 5 , 180 U/ha yearly of K 2 O, 30 U/ha yearly of MgO, and 50 U/ha yearly of CaO. Oligoelements were sup-plied at 1.5, 2.5, and 5 g/plant yearly of Mn, Zn, and Fechelates with 16%, 16%, and 6% richness, respectively.This population was genotyped and phenotyped for seedand berry traits. Marker analysis Total DNA was isolated from young frozen leaves usingthe procedure described by Dellaporta et al. (1983) with ex-traction buffer supplemented with 1% polyvinylpirrolidoneto reduce polyphenols (Lodhi et al. 1994). Parents and map-ping population, all 118 progeny individuals, were geno-typed using AFLP, selective amplification of microsatellitepolymorphic loci (SAMPL), sequence-specific amplifiedpolymorphisms (S-SAP), and microsatellite and sequence-characterized amplified region (SCAR) markers.AFLP procedures were performed following the protocoldescribed by Vos et al. (1995) with slight modifications(Cervera et al. 1998). Primer combinations were selected onthe basis of the total number of polymorphic fragments am-plified on a sample with both parents and 10 offspring indi-viduals. Eighteen primer combinations were used forgenotyping the parents and the whole progeny. SAMPLmarkers (Witsenboer et al. 1997) were generated using thesame AFLP protocol but with selective primer combinationsof 1 AFLP primer and 1 microsatellite-based primer (basedon tandem repeats and including a 3 ’  or 5 ’  anchor sequence).Sixteen primer combinations were selected for genotypingparents and progeny. Finally, for 2 S-SAP reactions (Waughet al. 1997) a modified AFLP protocol was used in whichselective primer combinations included an AFLP primerand a retrotransposon-based primer. Primer sequences andprimer combinations for all markers are shown in Tables S1and S2 of Supplementary materials. 2 For microsatellite analyses both progenitors and 6 prog-eny individuals were first genotyped at 167 microsatelliteloci (primer sequences for most of them are available at theUniSTS database of GeneBank ( to select useful polymorphisms (Costantini et al., inpress). Ninety-three of them (markers starting with ‘‘VMC’’,‘‘VrZAG’’, ‘‘VVMD’’, and ‘‘VVI’’ in Fig. 1) were selectedfor linkage mapping according to their segregation type. Mi-crosatellite genotyping was carried out using radioactive la-belling of the forward primer with [ g 33 P] followed byseparation of the amplified products on acrylamide gels (6%acrylamide:bisacrylamide 19:1, 1   Tris–borate–EDTA(TBE), and 7.5 mol/L urea) or fluorescence labelling of theforward primers followed by separation of the amplifiedproducts on an ABI-Prism 3700 sequencer (Applied Biosys-tems, Foster City, Calif.). In all cases polymerase chain re-action (PCR) was conducted in a final volume of 10  m Lcontaining 20 ng of genomic DNA, 10 mmol/L Tris–HCl(pH 8.3), 2 mmol/L MgCl 2 , 50 mmol/L KCl, 5% DMSO,0.2 mmol/L of each deoxynucleoside triphosphate (dNTP),2  m mol of forward primer, 2  m mol of reverse primer, and0.2 U of   Taq  DNA polymerase (Perkin Elmer, Waltham,Mass.). The amplification program consisted of 1 cycle of 30 s at 94  8 C, 30 s at ( T  m –3)  8 C, and 45 s at 72  8 C, followedby 14 cycles in which the annealing temperature was de-creased by 0.2  8 C/cycle, and then 20 cycles of 30 s at94  8 C, 30 s at ( T  m –6  8 C), and 45 s at 72  8 C; there was afinal extension of 5 min at 72  8 C. In some cases, microsatel-lite reactions amplified additional fragments. When thesefragments showed clear patterns of Mendelian segregationthey were considered as additional markers in the mappingdata set and named using the microsatellite locus name fol-lowed by ‘‘-FA’’ (fragment associated) and a number.The mapping population was also genotyped for theSCAR marker SCF27 (Mejı´a and Hinrichsen 2003), identi-fied using BSA as being linked to berry seedlessness. PCRwas performed in a final volume of 20  m L containing 20 ngof genomic DNA, 10 mmol/L Tris–HCl (pH 8.3), 2 mmol/LMgCl 2 , 50 mmol/L KCl, 5% DMSO, 0.2 mmol/L of eachdNTP, 2  m mol of forward primer (5 ’ -CAGGTGGGAGTA-GTGGAATG-3 ’ ), 2  m mol of reverse primer (5 ’ -CAGGTG-GGAGTAAGATTTGT-3 ’ ), and 0.4 U of   Taq  DNApolymerase (Perkin Elmer). The amplification program con-sisted of 35 cycles of 30 s at 94  8 C, 30s at 55  8 C, and 45 sat 72  8 C. Amplified fragments were visualized after electro-phoresis on 1.5% agarose gels and 1   TBE buffer. Phenotypic evaluation Ninety of the 118 progeny individuals of the mappingpopulation bore fruit in at least one of the 3 studied seasons:2002, 2003, and 2004 (from now on, harvests 02, 03, and04, respectively). These plants were evaluated for parame-ters related to berry size and berry seedlessness, but becauseof environmental conditions and disease incidence the num-bers of individuals scored were 89 in 02, 66 in 03, and 65 in04. For each plant we randomly sampled 100 berries from 2or 3 representative ripened clusters. Berry size was scored asthe average weight (BW) and volume (calculated as dis-placed water volume) per berry using the whole set of 100berries. Afterwards, 25 berries of the cluster mix were ran-domly collected to be evaluated for seedlessness. The degreeof seedlessness was scored as the fresh and dry weight of total seed content per berry, including completely developedseeds and seed traces. The average number of seeds andseed traces per berry (SN) was also scored. Statistical analy-ses were carried out using the software SPSS v. 12.0 (SASCorporation, Cary, S.C.). Linkage mapping Genetic maps for ‘Dominga’ and ‘Autumn Seedless’ culti-vars and a consensus linkage map for the cross were inde-pendently generated using all 118 progeny individuals andthe 2-way pseudo-testcross strategy (Grattapaglia andSederoff 1994). The mapping software Joinmap v. 3.0 (VanOoijen and Voorrips 2001) was used with a cross-pollinationpopulation type. Logarithm of the odds (LOD) and recombi- 2 Supplementary data for this article are available on the Web site or may be purchased from the Depository of Unpublished Data, Docu-ment Delivery, CISTI, National Research Council Canada, Building M-55, 1200 Montreal Rd., Ottawa, ON K1A 0R6, Canada. DUD5121. For more information on obtaining material refer to 1574 Genome Vol. 49, 2006   2006 NRC Canada  nation frequency thresholds were fixed at 3.5 and 0.3, respec-tively, to assign markers to LGs and establish marker order.Map distances were estimated using the Kosambi map func-tion (Kosambi 1944). For QTL mapping, framework cultivarmaps (not shown) and a cross consensus map were also devel-oped using only markers that were evenly distributed alongLGs and fully informative when possible (segregation typesabxac, abxcd, and abxaa for ‘Dominga’ and abxac, abxcd,and aaxab for ‘Autumn Seedless’). Partially informativemarkers were only included in gaps longer than 15 cM whenno other marker was available. Framework cultivar mapswere developed with Joinmap v. 3.0 using the double haploidpopulation type (Van Ooijen and Voorrips 2001). Numberingof LG was performed according to Riaz et al. (2003) andAdam-Blondon et al. (2004). Observed genome size ( G ob )was calculated for each linkage map as the sum of all LGsizes. Estimated genome size ( G e ) was calculated accordingto method 3 of Chakravarti (Hulbert et al. 1988). Observedmap coverage is the quotient of the observed and estimatedgenome sizes ( G ob  /  G e ). QTL analysis QTL detection was carried out using the framework mapsof each progenitor and the consensus frame map. QTL map-ping was performed using MapQTL v. 4.0 (Van Ooijen et al.2002) software and 3 different methods as described in thereference manual ( Kruskal Wallisnonparametric test (Lehmann 1975), interval mapping(Lander and Botstein 1989), and multiple QTL mapping(Jansen and Stam 1994). The results derived from the finalmultiple QTL mapping (MQM) model are shown. Bothgenome-wide and linkage-group-wide LOD thresholds corre-sponding to  a  = 0.05 were used for QTL detection. LODthresholds were estimated with the permutation test imple-mented in MapQTL using 10 000 permutations. Only QTLswith LOD values higher than genome-wide thresholds orLOD values higher than LG thresholds in more than 1 grow-ing season (Lander and Kruglyak 1995) were considered.Two LOD support intervals were established as approximate95% QTL confidence intervals (Van Ooijen 1992). Additiveeffect (the effect of substituting 1 parental allele for the other)and percentage of phenotypic variance explained by eachQTL, as well as total phenotypic variance explained by allthe QTLs, were estimated from the MQM model. Dominancerelations among alleles was estimated by studying the pheno-typic averages of genotypic classes. Two-way analysis of var-iance (ANOVA) was performed using the statistics softwareSPSS v. 12.0 (SAS Corp.) to test possible interactions amongthe identified QTLs; for this analysis the markers selected ascofactors in the QTL mapping were used as fixed factors. Results Mapping data set The final mapping data set for ‘Dominga’ and ‘AutumnSeedless’ included a total of 595 molecular markers, of which 75% allowed discrimination between maternal(‘Dominga’) and paternal (‘Autumn Seedless’) inherited al-leles (completely informative markers: segregation types ab-xaa, aaxab, abxcd, and abxac) (Table 1). Fifty-one markersof the mapping data set (9%) showed deviations from theexpected Mendelian segregations at  P  < 0.01 (28 in ‘Dom-inga’ (6.8%) and 40 in ‘Autumn Seedless’ (10.1%)). The ba-sic linkage maps were constructed using 93 microsatelliteloci (Table 1) selected to cover all linkage groups accordingto their position in other  Vitis  genetic maps (Adam-Blondonet al. 2004; Doligez et al. 2002; Doucleff et al. 2004;Grando et al. 2003) and their segregation in this population.Amplification of 8 of the microsatellite markers generatedcomplex patterns. The primer pair used for microsatelliteVMC1E11 amplified 2 different loci, named VMC1E11-Iand VMC1E11-II in Fig. 1. Furthermore, amplification of 6microsatellite loci yielded 17 additional segregating frag-ments (fragments associated with the amplification of micro-satellites (FA-SSRs)). Six microsatellites amplified a total of 14 FA-SSR fragments that were used as dominant markersin the linkage analysis (Table 1). Eight of those markersmapped in positions linked to the srcinal microsatellite lo-cus. Finally, amplification of VVMD30 generated a complexprofile that could not be interpreted, and the 3 segregatingfragments were handled as 3 independent FA-SSR markers.Basic microsatellite maps were saturated with 484 segre-gating AFLP, SAMPL, and S-SAP markers and 1 SCARmarker (Table 1). Of these techniques, SAMPL was themost efficient in the detection of polymorphic segregatingmarkers, with an average of 17 markers per reaction (vs 11for AFLPs and 9 for S-SAPs). Only S-SAP markers follow-ing Mendelian segregation were considered in the linkageanalyses. Many additional S-SAP-amplified polymorphicfragments did not fit the expected segregation ratios, and Table 1.  Markers used in the linkage analysis.‘Dominga’ ( , ) ‘Autumn Seedless’ ( < )abxaa abxab abxac abxcd Total aaxab abxab abxac abxcd Total TotalAFLP 72  55  127. 76  55.  131. 203. SAMPL 96  77   173. 90  77 .  167. 263. S-SAP 9  6   15. 3  6 .  9. 18. SSR 17  9  37 21 84. 9  9.  37 21 76. 93. FA-SSR 5  4  9. 8  4.  12. 17. SCAR  1  1.  1.  1. 1. Total 199  152  37 21 409. 186  152.  37 21 396. 595. Note:  Markers segregating 1:1 allowed the study of linkage relations only in ‘Dominga’ (abxaa) or ‘Autumn Seedless’ (aaxab).Markers segregating 3:1 (abxab dominant markers), 1:2:1 (abxab codominant markers), and 1:1:1:1 (abxcd and abxac markers) al-lowed the study of linkage relations in both progenitors. Partially informative markers (abxab) are indicated in italics. Cabezas et al. 1575   2006 NRC Canada  Fig. 1.  Genetic maps. Linkage maps of ‘Dominga’ (D), ‘Autumn Seedless’ (AS) and the cross consensus (C) are shown with empty, grey, and black bars, respectively. Partiallyinformative (segregation type abxab) and completely informative markers (the remaining ones) are indicated in normal and bold type, respectively. Marker names ending in ‘‘-D’’or ‘‘-AS’’ could be mapped only in ‘Dominga’ (segregation type abxaa) or ‘Autumn Seedless’ (aaxab), respectively. Markers showing distorted segregation ratios ( P  > 0.01) areindicated in italics. Relevant QTLs (in bold in Table 2) are shown for berry weight (BW; black background) seed number (SN; grey background), and seed fresh weight (SFW;white background). One logarithim of the odds (LOD) and 2 LOD confidence intervals of QTLs are indicated with boxes and lines, respectively. 1  5 7  6  G e n om e V  ol    .4  9  ,2  0  0  6    2   0   0   6   NR C  C  a n a  d   a 
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks