Advanced backcross QTL analysis of tomato. II. Evaluation of near-isogenic lines carrying single-donor introgressions for desirable wild QTL-alleles derived from Lycopersicon hirsutum and L. pimpinellifolium

 Improved-processing tomato lines were produced by the molecular breeding strategy of advanced backcross QTL (AB-QTL) analysis. These near-isogenic lines (NILs) contained unique introgressions of wild alleles originating from two donor wild species, Lycopersicon hirsutum (LA1777) and L. pimpinellifolium (LA1589). Wild alleles targeted for trait improvement were selected on the basis of previously published replicated QTL data obtained from advanced backcross populations for a battery of important agronomic traits. Twenty three NILs were developed for 15 genomic regions which were predicted to contain 25 quantitative trait factors for the improvement of seven agronomic traits: total yield, red yield, soluble solids, brix×red yield, viscosity, fruit color, and fruit firmness. An evaluation of the agronomic performance of the NILs in five locations worldwide revealed that 22 out of the 25 (88%) quantitative factors showed the phenotypic improvement predicted by QTL analysis of the BC₃ populations, as NILs in at least one location. Per-location gains over the elite control ranged from 9% to 59% for brix×red yield; 14% to 33% for fruit color; 17% to 34% for fruit firmness; 6% to 22% for soluble-solids content; 7% to 22% for viscosity; 15% to 48% for red yield, and 20% to 28% for total yield. The inheritance of QTLs, the implementation of the AB-QTL methodology for characterizing unadapted germplasm and the applicability of this method to other crops are discussed. Received: 27 October 1997 / Accepted: 25 November 1997     

Advanced backcross QTL analysis in tomato. I. Identification of QTLs for traits of agronomic importance from Lycopersicon hirsutum

 Advanced backcross QTL (AB-QTL) analysis is a new strategy for studying the effect of unadapted alleles on the agronomic performance of elite cultivated lines. In this paper we report results from the application of the AB-QTL strategy to cultivated tomato using the wild species Lycopersicon hirsutum LA1777 as the donor parent. RFLP genomic fingerprints were determined for 315 BC₂ plants and phenotypic data were collected for 19 agronomic traits from approximately 200 derived BC₃ lines which were grown in replicated field trials in three locations worldwide. Between 1 and 12 significant QTLs were identified for each of the 19 traits evaluated, with a total of 121 QTLs identified for all traits. For 25 of the QTLs (20%) corresponding to 12 traits (60%), the L. hirsutum allele was associated with an improvement of the trait from a horticultural perspective, despite the fact that L. hirsutum is overall phenotypically inferior to the elite parent. For example, L. hirsutum has fruit that remains green when ripe (lack of red pigment) yet alleles were found in this species that significantly increase red color when transferred into cultivated tomatoes. Wild alleles were also associated with increases in total yield and soluble solids (up to 15%) and brix×red yield (up to 41%). These results support the idea that one cannot predict the genetic potential of exotic germplasm based on phenotype alone and that marker-based methods, such as the AB-QTL strategy, should be applied to fully exploit exotic germplasm. Received: 27 October 1997 / Accepted: 25 November 1997     

QTL analysis of an advanced backcross of Lycopersicon peruvianum to the cultivated tomato and comparisons with QTLs found in other wild species

 A BC₃ population previously developed from a backcross of Lycopersicon peruvianum, a wild relative of tomato, into the cultivated variety L. esculentum was analyzed for QTLs. Approximately 200 BC₄ families were scored for 35 traits in four locations worldwide. One hundred and sixty-six QTLs were detected for 29 of those traits. For more than half of those 29 traits at least 1 QTL was detected for which the presence of the wild allele was associated with an agronomically beneficial effect despite the inferior phenotype of the wild parent. Eight QTLs for fruit weight could be followed through the BC₂, BC₃, and BC₄, generations, supporting the authenticity of these QTLs. Comparisons were made between the QTLs found in this study and those found in studies involving two other wild species; the results showed that while some of these QTLs can be presumed to be allelic, most of the QTLs detected in this study are ones not previously discovered. Received: 9 April 1997 / Accepted: 20 May 1997     

Advanced backcross QTL analysis in a cross between an elite processing line of tomato and its wild relative L. pimpinellifolium

Approximately 170 BC2 plants from a cross between an elite processing inbred (recurrent parent) and the wild species Lycopersicon pimpinellifolium LA1589 (donor parent) were analyzed with segregating molecular markers covering the entire tomato genome. Marker data were used to identify QTLs controlling a battery of horticultural traits measured on BC2F1 and BC3 families derived from the BC2 individuals. Despite its overall inferior appearance, L. pimpinellifolium was shown to possess QTL alleles capable of enhancing most traits important in processing tomato production. QTL-NIL lines, containing specific QTLs modifying fruit size and shape, were subsequently constructed and shown to display the transgressive phenotypes predicted from the original BC2 QTL analysis. The potential of exploiting unadapted and wild germplasm via advanced backcross QTL analysis for the enhancement of elite crop varieties is discussed.     

QTL analysis of horticultural traits differentiating the cultivated tomato from the closely related species Lycopersicon pimpinellifolium

Molecular markers were used to map and characterize quantitative trait loci (QTLs) for several characters of agronomic and biological importance in an interspecific backcross of tomato. The parents of the cross were an elite processing inbred Lycopersicon esculentum cv M82-1-7 and the closely related red-fruited wild species L. pimpinellifolium (LA1589). A total of 257 BC₁ plants were grown under field conditions in Ithaca, New York and scored for 19 quantitative traits. A genetic linkage map was constructed for the same population using 115 RFLP, 3 RAPD and 2 morphological markers that spanned 1,279 cM of the tomato genome with an average interval length of 10.7 cM. A minimum of 54 putatively significant QTLs (P<0.001; LOD> 2.4) were detected for all characters with a range of 1–7 QTLs detected per character. Of the total 54 QTLs 11% had alleles with effects opposite to those predicted by the parental phenotypes. The percentage of phenotypic variation associated with single QTLs ranged from 4% to 47%. Multilocus analysis showed that the cumulative action of all QTLs detected for each trait accounted for 12–59% of the phenotypic variation. The difference in fruit weight was controlled largely by a single major QTL (fw2.2). Digenic epistasis was not evident. Several regions of the genome (including the region near sp on chromosome 6) showed effects on more than one trait. Implications for variety improvement and inferences about the domestication of the cultivated tomato are discussed.     

Introgression and DNA marker analysis of Lycopersicon peruvianum, a wild relative of the cultivated tomato, into Lycopersicon esculentum, followed through three successive backcross generations

 Segregation of the Lycopersicon peruvianum genome was followed through three generations of backcrossing to the cultivated tomato L. esculentum cv ‘E6203’ using molecular markers. Thirteen BC₁ plants were genotyped with 113 markers, 67 BC₂ plants with 84 markers, and finally 241 BC₃ plants were genotyped with 177 markers covering the entire genome and a BC₃ map constructed. Several segments of the genome, including parts of chromosomes 3, 4, 6, and 10, quickly became fixed for esculentum alleles, possibly due to sterility problems encountered in the BC₁. Observed overall heterozygosity and chromosome segment lengths at each generation were very near the expected theoretical values. Markers located near the top telomeric region of chromosome 9 showed segregation highly skewed towards the wild allele through all generations, suggesting the presence of a gamete promoter gene. One markers, TG9, mapped to a new position on chromosome 9, implying an intrachromosomal translocation event. Despite the great genetic distance between the two parents, overall recombination was only 25% less than that observed in a previous tomato cross, indicating that L. peruvianum genes may be more readily introgressed into cultivated germplasm than originally believed. Received: 9 April 1997 / Accepted : 20 May 1997     

Description and analysis of genetic diversity between commercial bean lines (Phaseolus vulgaris L.)

The effectiveness of RFLP, DAMD-PCR, ISSR and RAPD markers in assessing polymorphism and relationships between 24 commercial lines of Phaseolus vulgaris L.was evaluated. We have used a Phaseolus-specific minisatellite sequence as a probe, which enabled 23 of the bean lines tested to be fingerprinted. Based on the sequence information obtained, primers corresponding to the bean-specific minisatellite core sequence were used in subsequent PCR amplifications. Our observations indicated that while the DAMD-PCR was sensitive in detecting genetic variation between bean species and between accessions of P. vulgaris, when used alone it may be limited in its ability to detect genetic variation among cultivated bean lines due to the low number of loci amplified. Only one out of the five ISSR primers tested was efficient in generating multiple band profiles, which was insufficient to distinguish all the different bean lines. Reproducible RAPD profiles were obtained, and these allowed us to differentiate all the genotypes tested with seven primers. We ultimately used only results from RFLP and RAPD markers to explore the genetic diversity among commercial bean lines. Both analyses led to the same clustering of the bean lines according to their geographical origins (United States or Europe). With respect to the European lines, the results obtained from RAPD data also enable the lines to be clustered according to their creators. Received: 15 January 2000 / Accepted: 21 March 2000     

Advanced backcross QTL analysis: a method for the simultaneous discovery and transfer of valuable QTLs from unadapted germplasm into elite breeding lines

Advanced backcross QTL analysis is proposed as a method of combining QTL analysis with variety development. It is tailored for the discovery and transfer of valuable QTL alleles from unadapted donor lines (e.g., land races, wild species) into established elite inbred lines. Following this strategy, QTL analysis is delayed until the BC2 or BC3 generation and, during the development of these populations, negative selection is exercised to reduce the frequency of deleterious donor alleles. Simulations suggest that advanced backcross QTL analysis will be effective in detecting additive, dominant, partially dominant, or overdominant QTLs. Epistatic QTLs or QTLs with gene actions ranging from recessive to additive will be detected with less power than in selfing generations. QTL-NILs can be derived from advanced backcross populations in one or two additional generations and utilized to verify QTL activity. These same QTL-NILs also represent commercial inbreds improved (over the original recurrent inbred line) for one or more quantitative traits. The time lapse from QTL discovery to construction and testing of improved QTL-NILs is minimal (1–2 years). If successfully employed, advanced backcross QTL analysis can open the door to exploiting unadapted and exotic germplasm for the quantitative trait improvement of a number of crop plants.     

Evaluation of breeding utility of a chromosomal segment from Lycopersicon chmielewskii that enhances cultivated tomato soluble solids

A chromosomal segment from the wild tomato species, Lycopersicon chmielewskii, when backcrossed into the tomato cultivar VF145B-7879, increases fruit soluble solids concentration. In the investigation reported here, the near-isoline (VF145B-7M) homozygous for the chmielewskii (chm) chromosomal segment was crossed to the cultivar to study allelic interactions and then hybridized to three commercial cultivars (UC204c, E6203, and ChicoIII) to investigate the effect of this segment when heterozygous in different genetic backgrounds. Parents, isogenic hybrids, and unmodified hybrid controls were evaluated in three consecutive years of field study. The chm segment, when either homozygous or heterozygous, significantly increased soluble solids concentration by 13% and 12%, respectively, over VF145B-7879 (esc/esc), suggesting dominant gene action. Averaged across the three isohybrids and one isoline, one dose of this segment significantly increased soluble solids concentration by 6.0% compared to the unmodified hybrids and VF145B-7879. Other than increasing fruit total and soluble solids in mature-green and ripe-red tomatoes, no consistent negative effects of the chmielewskii segment were observed in the various genetic backgrounds on plant yield, fruit weight, or fruit pH. These results favor the use of this segment in breeding programs to develop fresh market and processing tomato cultivars with enhanced quality and reduced processed product dehydration costs. Received: 15 April 2000 / Accepted: 13 January 2001     

QTL mapping of fruit-related traits in pepper (Capsicum annuum)

QTL analysis of pepper fruit characters was performed in an F₃ population derived from a cross between two Capsicum annuum genotypes, the bell-type cultivar Maor and the Indian small-fruited line Perennial. RFLP, AFLP^®1, RAPD and morphological markers (a total of 177) were used to construct a comparative pepper-tomato genetic map for this cross, and 14 quantitatively inherited traits were evaluated in 180 F₃ families. A total of 55 QTL were identified by interval analysis using LOD 3.0 as the threshold for QTL detection. QTL for several traits including fruit diameter and weight, pericarp thickness and pedicel diameter were often located in similar chromosomal regions, thus reflecting high genetic correlations among these traits. A major QTL that accounts for more than 60% of the phenotypic variation for fruit shape (ratio of fruit length to fruit diameter) was detected in chromosome 3. This chromosome also contained QTL for most of the traits scored in the population. Markers in linkage groups 2, 3, 8 and 10 were associated with QTL for multiple traits, thereby suggesting their importance as loci that control developmental processes in pepper. Several QTL in pepper appeared to correspond to positions in tomato for loci controlling the same traits, suggesting the hypothesis that these QTL may be orthologous in the two species.     

Simple sequence repeats for the genetic analysis of apple

 The development of highly informative markers, such as simple sequence repeats, for tagging genes controlling agronomic characters is essential for apple breeding. Furthermore the use of these markers is fundamental both for variety identification and for the characterisation and management of genetic resources. We have developed 16 reliable simple sequence repeat (SSR) markers that amplify all alleles from a panel of 19 Malus x domestica (Borkh.) cultivars or breeding selections and from Malus floribunda 821. Those markers show a high level of genetic polymorphism, with on average 8.2 alleles per locus and an average heterozygosity of 0.78. Due to this high level of polymorphism, it was possible using two selected SSRs to distinguish all cultivars except Starking and Red Delicious. Ten of the markers we developed have been mapped on a RAPD linkage map, proving their Mendelian segregation as well as their random distribution in the apple genome. Finally, we discuss the importance of using co-dominant markers in outbreeding species. Received: 8 October 1997 / Accepted: 9 December 1997     

Comparative microsatellite linkage analysis and genetic structure of two populations of F6 lines derived from Lycopersicon pimpinellifolium and L. cheesmanii

A population of recombinant inbred lines (RILs) has several advantages over its F₂ population counterpart with respect to quantitative trait loci (QTLs) and genomic studies. The objective of the investigation reported here was the comparative characterization by simple sequence repeat (SSR) and sequence characterized amplified region (SCAR) markers of two populations of F₆ lines derived from Lycopersicon pimpinellifolium (P population, consisting of 142 lines) and L. cheesmanii (C population, consisting of 115 lines) and sharing the female parent, L. esculentum var. cerasiforme. Almost the same percentage of polymorphic markers was found for each population although a different set of markers was involved. The proportion of SSR primer pairs (93 in total) that resulted in polymorphism for the main band was larger (55–56%) than for SCAR ones (13–16%). The C population showed the largest proportion of markers with zygotic and gametic segregation distortion, which is in agreement with the larger genetic distance reported between L. esculentum and L. cheesmanii than with the former and L. pimpinellifolium. Zygotic distortion corresponded primarily to an excess of heterozygotes in both populations, suggesting that the increment of homozygosity was the main factor limiting viability/self-fertility of the lines. Despite both populations sharing the female parent, P alleles were slightly favored in the P population while E alleles were the most frequently fixed in the C population. A linkage map for each population was obtained, with the average distances between consecutive markers being 3.8 cM or 3.4 cM depending on the population. Discrepancy between the maps for the location of only four markers on chromosomes 3, 6 and 10 was observed. Two possible causes of this discrepancy were investigated and can not be discarded: (1) the presence of duplicated markers and (2) segregation distortion caused by the selective advantage of gametes carrying one of the two alleles. This marker characterization of both populations will continue and will enable the comparative QTLs and candidate gene analysis of complex traits towards a more efficient utilization of genetic resources and breeding strategies.     

Marker-assisted introgression of blackmold resistance QTL alleles from wild Lycopersicon cheesmanii to cultivated tomato (L. esculentum) and evaluation of QTL phenotypic effects

Blackmold, caused by the fungus Alternaria alternata, is a major ripe fruit disease of processing tomatoes. Previously, we found blackmold resistance in a wild tomato (Lycopersicon cheesmanii) and quantitative trait loci (QTL) for resistance were mapped in an interspecific population. Five QTLs were selected for introgression from L. cheesmanii into cultivated tomato using marker-assisted selection (MAS). Restriction fragment length polymorphism and PCR-based markers flanking, and within, the chromosomal regions containing QTLs were used for MAS during backcross and selfing generations. BC₁ plants heterozygous at the QTLs, and subsequent BC₁S₁ and BC₁S₂ lines possessing different homozygous combinations of alleles at the target QTLs, were identified using DNA markers. Field experiments were conducted in 1998 (with 80 marker-selected BC₁S₂ lines) and 1999 (with 151 marker-selected BC₁S₂ and BC₁S₃ lines) at three California locations. Blackmold resistance was assessed during both years, and horticultural traits were evaluated in 1999. The BC₁S₂ and BC₁S₃ lines containing L. cheesmanii alleles at the QTLs were associated with a large genetic variance for resistance to blackmold and moderate heritability, suggesting that significant genetic gain may be achieved by selection in this genetic material. L. cheesmanii alleles at three of the five introgressed QTLs showed a significant, positive effect on blackmold resistance. A QTL on chromosome 2 had the largest positive effect on blackmold resistance, alone and in combination with other QTLs, and was also associated with earliness, a positive horticultural trait. The other four QTLs were associated primarily with negative horticultural traits. Fine mapping QTLs using near isogenic lines could help determine if such trait associations are due to linkage drag or pleiotropy.     

QTL analysis of yield and seed number in Citrus

Amount, regularity and low seed content of the crop are important properties of scion citrus cultivars. The genetic control of these traits was studied in a progeny derived from the cross Citrus volkameriana×Poncirus trifoliata using molecular marker analysis. Since the traits were not normally distributed, the Kruskal-Wallis non-parametric test was used for quantitative trait loci (QTLs) detection. Most of the QTLs detected correspond to the trait ”number of fruits per tree”, in agreement with its known physiological complexity. Related traits (fruit number, fruit size and seed number) are controlled by QTLs some of which are located in the same genomic regions, suggesting that undesired associations could be broken to some degree by recombination. QTL analysis over years revealed important effects of genotype-by-environment interaction on QTL detection. This result agrees with the differences found for the trait means among years, which was found to be related, among other causes, to the alternate bearing of some genotypes and the amount of rain before harvest. Received: 1 October 1999 / Accepted: 2 December 1999     

Quantitative trait locus mapping using sets of near-isogenic lines: relative power comparisons and technical considerations

 The power to detect QTL using near-isogenic line (NIL) comparisons versus recombinant inbred (RI) populations was assessed. The power to detect QTL was found to be generally greater when using RI populations than when using NIL contrasts. Power to detect QTL with NIL contrasts never exceeded that of RI populations when the number of RI lines is maximized relative to replication of lines for a given number of experimental units. The relative power of NIL contrasts is highest for traits with high heritability and when a gain in precision is realized due to increased replication of entries. Although NIL populations are generally less powerful than RI populations of similar size, some practical considerations may enhance the value of these materials. Availability of NILs allows the genetic effect of a specific chromosome region to be determined by comparing two lines; all RI lines in a population need to be rescored for each new trait even if the effect of a specific chromosome region is suspected. NIL comparisons may allow genetic differences to be detected by visual inspection; genetic effects can only be expressed as means and variances with recombinant inbred populations. In summary, RI populations generally, and in some cases, substantially, provide better power for QTL detection than NIL comparisons. Practical considerations, however, indicate that many factors need to be considered when choosing a population structure to meet an experimental objective. Received: 12 December 1996 / Accepted: 21 March 1997     

QTL analysis of leaf morphology in tetraploid Gossypium (cotton)

Molecular markers were used to map and characterize quantitative trait loci (QTLs) determining cotton leaf morphology and other traits, in 180 F₂ plants from an interspecific cross between a Gossypium hirsutum genotype carrying four morphological mutants, and a wild-type Gossypium barbadense. The prominent effects of a single region of chromosome 15, presumably the classical ”Okra-leaf” locus, were modified by QTLs on several other chromosomes affecting leaf size and shape. For most traits, each parent contained some alleles with positive effects and others with negative effects, suggesting a large potential for adapting leaf size and shape to the needs of particular production regimes. Twenty one QTLs/loci were found for the morphological traits at LOD≥3.0 and P≤0.001, among which 14 (63.6%) mapped to D-subgenome chromosomes. Forty one more possible QTLs/loci were suggested with 2.0≤LOD<3.0 and 0.001<P≤0.01. Among all of the 62 possible QTLs (found at LOD≥2.0 and P≤0.01) for the 14 morphological traits in this study, 38 (61.3%) mapped to D-subgenome chromosomes. This reinforces the findings of several other studies in suggesting that the D-subgenome of tetraploid cotton has been subject to a relatively greater rate of evolution than the A-subgenome, subsequent to polyploid formation. Received: 26 April 1999 / Accepted: 30 July 1999     

QTL analysis of malting quality in barley based on the doubled-haploid progeny of two elite North American varieties representing different germplasm groups

Characterization of the determinants of economically important phenotypes showing complex inheritance should lead to the more effective use of genetic resources. This study was conducted to determine the number, genome location and effects of QTLs determining malting quality in the two North American barley quality standards. Using a doubled-haploid population of 140 lines from the cross of Harrington×Morex, malting quality phenotype data sets from eight environments, and a 107-marker linkage map, QTL analyses were performed using simple interval mapping and simplified composite interval mapping procedures. Seventeen QTLs were associated with seven grain and malting quality traits (percentage of plump kernels, test weight, grain protein percentage, soluble/total protein ratio, α-amylase activity, diastatic power and malt-extract percentage). QTLs for multiple traits were coincident. The loci controlling inflorescence type [vrs1 on chromosome 2(2H) and int-c on chromosome 4(4H)] were coincident with QTLs affecting all traits except malt-extract percentage. The largest effect QTLs, for the percentage of plump kernels, test weight protein percentage, S/T ratio and diastatic power, were coincident with the vrs1 locus. QTL analyses were conducted separately for each sub-population (six-rowed and two-rowed). Eleven new QTLs were detected in the subpopulations. There were significant interactions between the vrs1 and int-c loci for grain-protein percentage and S/T protein ratio. Results suggest that this mating of two different germplasm groups caused a disruption of the balance of traits. Information on the number, position and effects of QTLs determining components of malting quality may be useful for maintaining specific allele configurations that determine target quality profiles. Received: 28 May 1999 / Accepted: 9 November 1999     

QTL analysis of agronomic traits in barley based on the doubled haploid progeny of two elite North American varieties representing different germplasm groups

A better understanding of the genetics of complex traits, such as yield, may be achieved by using molecular tools. This study was conducted to estimate the number, genome location, effect and allele phase of QTLs determining agronomic traits in the two North American malting barley (Hordeum vulgare L.) quality variety standards. Using a doubled haploid population of 140 lines from the cross of two-rowed Harrington×six-rowed Morex, agronomic phenotypic data sets from nine environments, and a 107-marker linkage map, we performed QTL analyses using simple interval mapping and simplified composite interval mapping procedures. Thirty-five QTLs were associated, either across environments or in individual environments, with five grain and agronomic traits (yield, kernel plumpness, test weight, heading date, and plant height). Significant QTL×environment interaction was detected for all traits. These interactions resulted from both changes in the magnitude of response and changes in the sign of the allelic effect. QTLs for multiple traits were coincident. The vrs1 locus on chromosome 2 (2H), which determines inflorescence row type, was coincident with the largest-effect QTL determining four traits (yield, kernel plumpness, test weight, and plant height). QTL analyses were also conducted separately for each sub-population (six-rowed and two-rowed). Seven new QTLs were detected in the sub-populations. Positive transgressive segregants were found for all traits, but they were more prevalent in the six-rowed sub-population.QTL analysis should be useful for identifying candidate genes and introgressing favorable alleles between germplasm groups. Received: 18 August 2000 / Accepted: 15 December 2000     

Evaluation of near-isogenic lines of rice introgressed with QTLs for root depth through marker-aided selection

Drought is one of the main abiotic constraints in rice. A deep root system contributes efficiently to maintaining the water status of the crop through a stress period. After identifying QTLs affecting root parameters in a doubled-haploid (DH) population of rice derived from the cross IR64/Azucena, we started a marker-assisted backcross program to transfer the Azucena allele at four QTLs for deeper roots (on chromosomes 1, 2, 7 and 9) from selected DH lines into IR64. We selected the backcross progenies strictly on the basis of their genotypes at the marker loci in the target regions up to the BC₃F₂. We assessed the proportion of alleles remaining from Azucena in the non-target areas of the BC₃F₂ plants, which was in the range expected for the backcross stage reached. Twenty nine selected BC₃F₃ near-isogenic lines (NILs) were developed and compared to IR64 for the target root traits and three non-target traits in replicated experiments. Of the three tested NILs carrying target 1, one had significantly improved root traits over IR64. Three of the seven NILs carrying target 7 alone, as well as three of the eigth NILs carrying both targets 1 and 7, showed significantly improved root mass at depth. Four of the six NILs carrying target 9 had significantly improved maximum root length. Five NILs carrying target 2 were phenotyped, but none had a root phenotype significantly different from that of IR64. A re-analysis of the initial data with the composite interval mapping technique revealed two linked QTLs with opposite effects in this area. Some NILs were taller than IR64 and all had a decreased tiller number because of a likely co-introgression of linked QTLs. The usefulness of NILs, the efficiency of marker-aided selection for QTLs and the relationship between root traits are discussed. The NILs with an improved root system will permit testing the importance of root depth for water-limited environments. Received: 17 July 2000 / Accepted: 20 October 2000     

Fine mapping of a quantitative trait locus (QTL) from Lycopersicon hirsutum chromosome 1 affecting fruit characteristics and agronomic traits: breaking linkage among QTLs affecting different traits and dissection of heterosis for yield

The near-isogenic Line TA523, containing a 40-cM introgression at the bottom of chromosome 1 from Lycopersicon hirsutum acc. LA1777, affects several agronomically important traits. A set of recombinant lines (subNILs) derived from the original NIL TA523 were developed in order to fine-map, by substitution mapping, the genetic factors included within the original introgression. In the current experiment, TA523 showed redder, rounded, less pigmented shoulder, lower-weighted fruits and higher brix, whereas higher yield and brix*yield was observed only in the hybrid TA253×TA209 suggesting heterosis for these traits. By substitution mapping we mapped independent genetic loci affecting brix, yield and fruit shape, whereas fruit weight, shoulder pigmentation and external color mapped to a position coincident with the brix locus. Analysis of the subNILs revealed that the gene action of most of the QTLs was additive or nearly additive. The exception was for the yield QTL which was dominant (d/a=0.7), eliminating the possibility that yield increase is due to true overdominance at a single gene locus. However, no negative yield effects were detected in other regions of the introgressed segment, as would be predicted by a dominance complementation model. Therefore, epistatic interactions among genetic factors along the introgressed segment are suggested as the cause of yield heterosis. Results from this study, combined with previous experiments involving different tomato wild species, demonstrate that the base of chromosome 1 of tomato contains multiple QTLs affecting various agronomic and fruit traits and that these effects can not be attributed to the pleiotropic effects of a single locus. Received: 21 April 1999 / Accepted: 17 June 1999