Abstract
The disease scald of barley is caused by the pathogen Rhynchosporium commune and can cause up to 30-40% yield loss in susceptible varieties. In this study, the Australian barley cultivar Yerong was demonstrated to have resistance that differed from Turk (Rrs1) based on seedling tests with 11 R. commune isolates. A doubled haploid population with 177 lines derived from a cross between Yerong and Franklin was used to identify quantitative trait loci (QTL) for scald resistance. Scald resistance against four pathogen isolates was assessed at the seedling growth stage in a glasshouse experiment and at the adult growth stage in field experiments with natural infection over three consecutive years. A QTL on chromosome 3H was identified with large effect, consistent with a major gene conferring scald resistance at the seedling stage. Under field conditions, scald percentage was negatively correlated with early relative maturity. A bivariate analysis was used to model scald percentage and relative maturity together, residuals from the regression of scald percentage on relative maturity were used as our phenotype for QTL analysis. This analysis identified one major QTL on chromosome 3H, which mapped to the same position as the QTL identified for scald resistance at seedling stage. The identified QTL on 3H is proposed to be different from the Rrs1 on the basis of seedling resistance against different R. commune isolates and physical map position. The analysis also identified an additional novel QTL on chromosome 7H. This study increases the current understanding of scald resistance and identifies genetic material possessing QTLs useful for the marker-assisted selection of scald resistance in barley breeding programs.
Introduction
Scald is a serious foliar disease in barley (Hordeum vulgare) that is caused by Rhynchosporium commune. The pathogen can cause up to 30-40% yield loss in susceptible varieties and is found in all barley-growing regions worldwide 1. Control of scald disease requires a multi-facetted approach, including application of fungicides, cultural disease management, manipulation of sowing date and the use of resistant cultivars 2. R. commune populations have changed rapidly in response to newly-developed fungicides and resistant plant varieties 3–6. One of the most sustainable strategies for R. commune management is to develop and deploy disease-resistant barley cultivars through the introgression and pyramiding of different resistance genes (major or minor). Traditional methods of phenotypic selection for complex patho-systems such as scald can be improved through detailed genetic studies, which allow the implementation of marker-assisted selection (MAS) in breeding programs.
Scald resistance is governed by both major and minor genes. Major resistance genes provide high levels of resistance at all plant growth stages, while minor resistance genes generally provide partial levels of resistance at the adult plant stage 7,8. Reduced scald symptoms in adult stage plants under field conditions might also result from disease escape through physical barriers to infection 9. In terms of scald resistance in barley, flowering time, plant height and canopy structure can affect scald symptoms by physically limiting the upward spread of the splash-dispersed pathogen 8. The major scald resistance genes discovered so far have been mainly identified through experiments with seedlings via inoculation with specific isolates 8. The problem with using major scald resistance genes in breeding programs is a lack of durability. Quantitative genes are thought to be more durable, and it has been suggested that pyramiding these genes could reduce the ability of R. communes to rapidly acquire new virulence combinations 10.
A large number of QTLs for scald resistance have been discovered in barley. Two genomic regions in particular have been frequently associated with resistance; the Rrs1 locus on chromosome 3H and the Rrs2 locus on chromosome 7H 8,11. These two loci have been detected in different mapping populations, across different environments, under glasshouse conditions after inoculation with specific isolates, and under field conditions 8,11. Progress of our understanding of these loci has been made through fine mapping studies 12,13. It remains unknown if the QTLs detected at each of these two loci are alleles of the same gene, or if they are part of a closely linked gene cluster at each locus 14,15.
For the Rrs1 locus, multiple major and minor scald resistance genes or QTLs have been identified, supporting its importance in barley germplasm worldwide 12,16–21. Rrs1 was the first scald resistance gene reported in barley 22 with an associated RFLP marker of cMWG680, positioned at 455.3 Mb on chromosome 3H on the barley physical map 18. The location of the Rrs1 locus was further fine mapped including all known markers to an interval of less than 9 Mbp at 448.4 Mb from Spanish barley varieties 12.
In this study, we performed screening against 11 R. commune isolates at seedling stage to demonstrate that resistance in the variety Turk (international source of Rrs1) differs from that resistance in the variety Yerong. These results contradicted previous findings that suggested Yerong carries a scald resistance QTL at the Rrs1 locus on chromosome 3H 23. To further investigate this finding, screening of a Yerong/Franklin doubled haploid population was conducted for seedling resistance against four R. commune isolates and adult plant resistance under natural field conditions across three years. To overcome potential confounding effects of differences in the onset of flowering (determined as relative maturity) on scald resistance QTL detection within the population, a linear mixed model with a bivariate approach was used to analyse the field data. Using this approach, a major QTL on chromosome 3H from Yerong was detected for scald resistance at both seedling and adult plant stage. The QTL detected did not map to the Rrs1 locus, suggesting the presence of a new and useful scald resistance QTL in Yerong.
Materials and Methods
Differential varieties screening
The seedling scald resistance of different barley varieties was tested against 11 R. commune isolates from the Wagga Wagga Agricultural Institute fungal isolate collection. These isolates were collected from southern New South Wales between 2013 and 2016 (Table 1). The R. commune isolates were grown on lima bean agar (LBA) at 20°C under 24 hour light condition 7. After 2 to 3 weeks, spores were harvested with a scalpel blade and the spore solutions for spray inoculation were adjusted to 2×106 spores ml−1 in distilled water using a haemocytometer. Seeds of varieties were sown in 6-cm pots arrayed in a randomized complete block design consisting of 4 columns by 30 rows with three replicates of each genotype in each experiment for a total of 120 pots. A total of 40 varieties were included in the differential varieties screening, however, only the results for six key varieties (Yerong, Franklin, Turk, Atlas, Atlas46 and Litmus) were reported in this study (Table 2). Each variety by isolate combination was evaluated in at least two experiments. The responses of Turk (Rrs1), Atlas (Rrs2) and Atlas46 (both Rrs1 and Rrs2) were compared with that of Yerong, while Litmus was used as the susceptible check variety. The barley varieties were uniformly sprayed with spore solutions at the 3-leaf-stage, and kept in a dark chamber for 48 hours at 18°C with 100% humidity. The seedlings were scored 14 days post-inoculation with a scoring scale from 1 to 5 (1 = resistant and 5 = very susceptible) following the methods of Jackson and Webster 24.
Plant material for QTL analysis
A doubled haploid (DH) population of 177 lines from a cross between varieties Yerong and Franklin was used in this study. Seeds for each genotype were obtained from the University of Tasmania. Franklin is an Australian two-rowed malting quality variety, and Yerong is an Australian six-rowed feed quality variety. The genetic linkage map of the population comprised 28 microsatellites and 196 diversity arrays technology (DArT) markers assembled by Li and Zhou 23.
Evaluation of seedling resistance to scald
Four different R. commune isolates, WAI2466, WAI2470, WAI2471 and WAI2473, were used in the seedling resistance screening in the glasshouse experiments (Table 1). All four isolates were collected from southern New South Wales and are held in the Wagga Wagga Agricultural Institute fungal isolate collection (Table 1). Inoculum preparation and inoculations were as outlined for the differential screening. Spores were diluted to 1×106 spores ml−1 in distilled water using a haemocytometer. Each of the four isolates was tested in a separate experiment. For each experiment, the 177 DH lines as well as the parental varieties Franklin and Yerong and an additional 21 check varieties (for a total of 200 genotypes) were sown in 6-cm pots arrayed in a 24 column by 25 row randomized complete block design with three replicates of each genotype for a total of 600 pots. Each variety by isolate combination was tested with three different experiments with three technical replicates per experiment. Seedlings were scored 14 days post-inoculation with a scoring system from 1 to 4 (1 = resistant and 4 = very susceptible) following the methods of Wallwork and Grcic 7.
Field screening
Field screening for scald resistance was conducted at Wagga Wagga Agricultural Institute (Wagga Wagga, New South Wales) in 2015, 2016 and 2017. All field trials were sown in May with a randomised complete block design with two replicates for each genotype. Each genotype was sown in 1.2 m rows with 0.4 m spacing between each row. The primary inoculum for R. commune infection was residual barley crop debris from the previous harvest. Overhead irrigation was used regularly to supplement rainfall throughout the growing season to enhance the development of disease. Experiments were subject to a strict weed control and crop nutrition regime to maximize yield potential. Assessment of disease was based on leaf symptoms using the percentage of infected leaf area 6. Relative maturity at the time of disease assessment was determined using the Zadoks decimal score for plant development 25. Final plant height was also measured at physiological maturity in the 2017 experiment.
Statistical analysis
All data was analysed using the software package ASReml-R version 3 26 in the R environment 27. A linear mixed model following the approach of Gilmour, et al. 28 was used to analyse the data for the differential variety screening experiments as follows (Model 1): where y is the n × 1 vector of the response variable (scald score) across p = 22 experiments with each of the 11 R. commune isolates tested in a separate experiment, and that experiment repeated once. n = 2640 for the differential variety screening experiments as only selected varieties were included. τ is a t × 1 vector of fixed effects, including the overall mean scald score, corresponding to the n × t design matrix X. The term g is the vector of genotypic random effects with associated design matrix Z used to model the genotype by experiment effects. The term u is the vector of random effects corresponding to the experimental design matrix Z, which contains experiment-specific terms to capture extraneous variation including the experiment level blocking structure including replicate, row and column. The n × 1 residual vector e was modelled for each experiment.
A model similar to Model 1 above was also used to model scald scores for the seedling inoculation experiments with n = 7200 for the p = 12 experiments conducted.
For the field screening experiments measuring scald resistance and relative maturity, each of the three field experiments was modelled separately using a bivariate approach as follows: where y is a vector of length n = 2 × 479 containing stacked vectors for the two traits, S: scald resistance and R: relative maturity). τ is a vector of fixed effects including trait means and the trait by genotype effects for the design matrix X. The term u is the vector of replicate, column and row effects for each trait corresponding to the experimental design structure Z. The vector e of length n containing the residuals of the two traits S and R was modelled with a separable autoregressive process of order one (AR1 ⊗ AR1) and an unstructured variance-covariance matrix between traits. This structure permits the fitting of a linear relationship at the residual level between the two traits. In all models, the significance of fixed effects was assessed using the techniques of Kenward and Roger 29 and the significance of random effects other than ‘replicate’ was determined using log-likelihood ratio tests 30.
The linear relationship between scald resistance and relative maturity was determined from the trait:genotype covariance modelling after the general approach of Van Beuningen and Kohli 31 and the paired case-control study example detailed in Butler, et al. 26 as follows: where the slope of the regression is calculated as: and the intercept β0 was determined from the overall BLUPs for the two traits S and R from the bivariate mixed model. The difference (residual) between the BLUP from the mixed model for scald resistance of a genotype and the predicted value on the trend line was calculated. These values are referred to hereafter as the deviation from the regression of scald resistance on relative maturity (DRSRRM).
QTL analysis and positioning of identified QTL on the barley physical map
The DRSRRM values for each individual field experiment were used as the phenotype to detect QTL for scald resistance adjusted for the relationship between scald resistance and relative maturity. For the phenotype analyses of scald percentage, relative maturity and plant height BLUPs were obtained from the model for each experiment. Phenotypes were used for QTL analysis using the MapQTL6.0 software package 32. QTLs were first analysed with interval mapping (IM). The closest marker to each QTL was selected as a cofactor for multiple QTL mapping (MQM). A logarithm of the odds (LOD) threshold value of 3 was used to identify QTL.
The primer sequences of markers associated with the identified QTL for scald resistance and fine mapped Rrs1 12 were used to do BLAST searches by using the IPK Barley BLAST Server (http://webblast.ipk-gatersleben.de/barley). The barley pseudomolecules Morex V 2.0 2019, was used for the BLASTn search. Default settings were used to do the BLASTn search and the best hit was used to decide the physical position of the detected QTL.
Results
Differential varieties screening
Atlas46 (Rrs1 + Rrs2) was resistant to all the isolates used in this study, with scores under 2.3, except isolate WAI2840, which was virulent on all varieties with a score of more than 3.5 (Table 2). Turk (Rrs1) was resistant to all the isolates used in this study with scores under 2.4, except two isolates WAI2439 and WAI2840. Varieties carrying Rrs2 (Atlas and Atlas46) were resistant against isolate WAI2439 (Table 2). Varieties conferring Rrs1 (Turk and Atlas46) showed higher level of resistance than Atlas and Yerong against two isolates, WAI1245 and WAI2464. Varieties carrying either Rrs1 or Rrs2 were resistant against isolates WAI453, WAI2466, WAI2471 and WAI2636.
Turk was significantly more resistant than Yerong against five isolates and equivalent against the remaining six isolates tested. Together, these results suggested that Rrs1 from Turk was not present in Yerong. Compared to Yerong, Atlas showed a higher level of resistance against isolates WAI2439, WAI2463, WAI2466, WAI2470 and WAI2636, suggesting Rrs2 is absent in Yerong. Yerong was resistant against WAI2471, and displayed moderate resistance against isolates WAI453, WAI1245, WAI2439 and WAI2473 with the scores between 3.0 and 3.4. Yerong showed a better resistance than Franklin against isolates WAI453, WAI1245 and WAI2471with the overall scores of Yerong being lower than those of Franklin. Litmus was susceptible to all the isolates in this study.
QTLs for scald resistance at the seedling stage
There were no significant differences (using 95% confidence intervals) in resistance to the four isolates WAI2466, WAI2470, WAI2471 and WAI2473 among the two parent cultivars in the seedling stage screening experiments. However, there was phenotypic variation in resistance to the different isolates among the DH population lines (Fig. 1). The disease scores among DH population lines against WAI2466 showed a bimodal distribution of scores between resistant and susceptible lines. The variation of disease scores against WAI2470 was the lowest among all four isolates. The distribution of disease scores against WAI2471 was skewed towards higher levels of disease. In contrast, the distribution of disease scores against WAI2473 was skewed towards the lower end of the scoring scale.
One major QTL on chromosome 3H, designated as QTL-WAIYerong-3H, was identified for seedling resistance to all four different R. commune isolates (WAI2466, WAI2470, WAI2471 and WAI2473) at the same position (Table 3) at 178.7 Mb. Flanking markers of this QTL, bPb-0068 and Bmag0006, are mapped at 120.3 and 178.7 Mb on the physical map (Fig. 2). The LOD scores varied from 23.8 to 38.7, explaining more than 46% of the phenotypic variation (Table 3).
Scald resistance under field conditions
The parent varieties Franklin and Yerong showed similar levels of scald resistance and relative maturity across the different years (Table 4). There were no significant differences between Franklin and Yerong for any of the traits measured across the different years (p > 0.01). However, the DH lines showed substantial phenotypic variation in scald resistance and relative maturity (Fig. 3). Average disease incidence was more severe in 2016, reaching 43.0%, higher than the average scald percentages in 2015 (29.0%) and 2017 (22.1%) (Table 4).
A significant correlation (Pearson’s correlation coefficient) between scald percentage and relative maturity was observed among the DH lines (Fig. 3) in 2015 (r = 0.50, p < 0.01), 2016 (r = 0.51, p < 0.01). Later flowering lines (with lower relative maturity scores) tended to have a lower scald percentage. In 2017, although Pearson correlation coefficient between relative maturity and scald percentage was significant (r = 0.34, P < 0.01), no correlation was indicated between relative maturity and scald percentage from bivariate analysis (Fig. 3). A phenotype value was calculated to account for the linear relationship between scald percentage and relative maturity by determining the deviation from the regression of scald percentage on relative maturity (DRSRRM) and this was used as a phenotype for QTL analysis in our field trials.
QTLs for relative maturity and plant height
Four QTLs for relative maturity were identified in the Yerong/Franklin population consistently across all three years (Fig. 4 and Table 3): two on chromosome 2H, one on chromosome 5H and another one on chromosome 7H. These four QTLs explained more than 40% of the total phenotypic variance for relative maturity. One QTL for plant height in 2017 was identified on chromosome 3H that explained 14.1% of the total phenotypic variance, with a LOD value of 5.7 (Table 3).
QTL for scald resistance in the field
In 2015, one QTL for scald percentage on chromosome 2H (QTLSP-2H-2015) was located at the same position as the 2H QTL for relative maturity (QTLZ-2H-2015). In 2016, QTLSP-2H-2016 and QTLZ-2H-2016 were also mapped at the same position on chromosome 2H (Fig. 4 and Table 3). Markers associated with QTLSP-7H-2016 and QTLZ-7H-2016 were also located close to each other (Fig. 4 and Table 3). QTLs for scald percentage were not mapped at the same position as QTLs for relative maturity in 2017.
The QTLs for DRSRRM in 2016 (QTLSR-3H-2016) and 2017 (QTLSR-3H-2017) were located at 178.7 Mb in the same position as QTL-WAIYerong-3H, and the resistant allele was also derived from the Yerong parent. The QTL for DRSRRM (QTL-WAIYerong-3H) explained 20.7% of the phenotypic variance in 2016 (LOD value 8.8; Fig. 4 and Table 3). In 2015, the QTL for DRSRRM on chromosome 3H (QTLSR-3H-2015) was 15 cM away from this major QTL on the linkage map, at 503.4 Mb on the barley physical map. This QTL was still more than 50 Mb away from Rrs1, between Rrs1 (448.4 Mb) and Rrs4 (523.0 Mb) on chromosome 3H, explaining 14.8% of phenotypic variance. QTLs were identified for DRSRRM on chromosome 3H across the three years at the same location as the QTLs detected for scald percentage. In 2017, a novel QTL for scald resistance (both scald percentage and DRSRRM) was identified on chromosome 7H from Franklin, locating at 69.9 Mb on the barley physical map. This QTL explained 12.9 % phenotypic variance with a LOD value of 5.8.
Discussion
Scald resistance at seedling stage
Differential screening conducted in this study showed that the combination of major resistance genes Rrs1 and Rrs2 in the variety Atlas46 provides resistance to 10 out of 11 R. commune isolates from southern NSW. These results confirmed the effectiveness of pyramiding major resistance genes into one variety in providing broad protection against R. commune 7,33. However, resistance gene combinations need to be found that do not rapidly select for the corresponding virulence gene combinations in the pathogen population. This remains a challenge with only a limited understanding of the gene for gene interaction between R. commune and its host. We have identified specific R. commune isolates to distinguish the presence of Rrs1 or Rrs2. While isolates such as WAI2840 posed a threat to reliance on just these two major genes, they are valuable in detecting seedling resistances other than Rrs1 and Rrs2 for use in resistance breeding.
Five of the R. commune isolates screened in this study were avirulent against Turk and virulent against Yerong, indicating that the Rrs1 allele from Turk is not present in Yerong. Yerong was resistant to isolate WAI2471, and had moderate levels of resistance against WAI453, WAI1245, WAI2439 and WAI2473. Further QTL analysis located a major QTL for scald resistance against four different isolates at the seedling stage to chromosome 3H, which originated from Yerong. This QTL-WAIYerong-3H explained more than 46% of the phenotypic variation and is located at the same position of a QTL (QSc.YeFr-3H) identified using adult plant data in a previous study from the same DH population in Tasmania, Australia 23. However, Li and Zhou (2011) were unable to distinguish this QTL from Yerong from that of Rrs1.
The complete reference barley genome sequence enabled the projection of QTLs from different populations onto one barley physical map 34. In our study, flanking markers of the QTL-WAIYerong-3H, bPb-0068 and Bmag0006, are mapped at 120.3 and 178.7 Mb on the physical map. The position is some distance away from previously fine mapped Rrs1 which is located at a position of 448.4 Mb on chromosome 3H (Fig. 2) 12. This also suggested the scald resistance QTL-WAIYerong-3H is different from Rrs1. Multiple major and minor scald resistance genes or QTLs have been identified on chromosome 3H, close to the Rrs1 locus 12,16–20. Among these QTL, the QTL-IA3H from cultivar Abyssinian was identified in studies of seedling resistance to four R. commune isolates 19. A QTL Rrs1BC240 for R. commune resistance from wild barley H. spontaneum was also identified at the position of Rrs1 17. While ICARDA4 and ICARDA9 showed resistance to all isolates at seedling stage in one study 7 and QTLs identified from them were mapped at 383.9 Mb on physical map 17, the identity of these resistances remain unknown.
A scald resistance QTL in a similar position to the QTL-WAIYerong-3H (Fig. 2) was published by Coulter, et al. 20. QTL qSUK7_3, contributed by the variety Steptoe based on detached leaf assays, was located 3 Mb away from QTL-WAIYerong-3H. The authors suggested this resistance to be different from Rrs1 based on differential isolate reactions, but were unable to differentiate its map position from that of Rrs1 20. Further experiments are required to resolve whether the QTL-WAIYerong-3H is the same or different to the gene found in Steptoe on chromosome 3H.
Disease escape under field conditions
Our study illustrates the potentially confounding effects that relative maturity can have when phenotyping for disease resistance. The observed differences between genotypes may be conferred by both host resistance and mechanisms that lead to disease escape, such as plant maturity, plant height and canopy structure 8,15. For example, Zhan, et al. 8 postulated that later flowering time and taller plant height can physically slow the upward spread of splash-dispersed R. commune and contribute to disease escape. While disease escape traits are important, we sought to identify sources of resistance that are independent of relative maturity, as these QTL are more likely to be useful in breeding programs selecting for a constrained window of flowering time for their target environments.
In this study the positions of all four QTLs for relative maturity were co-located with known phenology QTLs. QTLZ-2H.1 was co-located with a QTL for heading date detected in a TX9425/Franklin population 35 and this relative maturity allele was derived from Franklin. This QTL is close to another flowering-time QTL, pseudo-response regulator Ppd-H1 36, which is located 7 cM away on the consensus map. QTLZ-2H.2 was located at the same position as another QTL for heading date detected by Nduulu, et al. 37. QTLZ-5H and QTLZ-7H were located at the same positions as the flowering time genes eam5 and Eps-7S, respectively 38.
All of these loci have also been implicated to be associated with resistance at seedling or adult plant stages to five biotrophic or necrotrophic pathogens. von Korff, et al. 39 identified QTLs for scald, leaf rust and powdery mildew resistance at a position similar to that of Ppd-H1. QTLZ-2H.2 was located at the same position as a QTL for Fusarium head blight resistance detected by Nduulu, et al. 37. The results of Nduulu, et al. 37 also indicated that the QTLs for heading date and Fusarium head blight resistance were tightly linked rather than pleiotropic. A new gene conferring adult plant resistance to leaf rust, Rph23, was identified in the same Yerong/Franklin population used in this study on chromosome 7H, co-located with the QTL QTLZ-7H detected in this study 40. Another QTL for stem rust resistance at the seedling stage was also mapped to the same position as QTLZ-7H from Yerong/Franklin population 41.
A QTL for plant height was mapped to chromosome 2H at the same position as QTLZ-2H.2 in the Yerong/Franklin population by Xue, et al. 42. Interestingly, the QTL for plant height identified in the Yerong/Franklin population is at the same position as a QTL for plant height in a CM72/Gairdner population 43. Mature plant height was measured in only one of the experiments in this study (in 2017), and a correlation between mature plant height and scald resistance was not observed (data not shown), and the QTL identified for plant height did not co-locate with scald resistance QTL (Table 3).
Modelling relative maturity together with disease resistance
The deviation from the regression of infection on important confounding traits has been employed previously as an effective phenotype for disease resistance 9,31. Van Beuningen and Kohli 31 in particular included linear functions for both heading date and plant height against Septoria tritici blotch (STB) infection in wheat. They aimed to calculate a phenotype that captured components of resistance that did not depend on those two traits, and represented a better approximation of genetic resistance from the experiments in their study. The residuals from a generalized linear model in which STB percentages were fitted to all escape related traits, including heading date, plant height, leaf spacing and leaf morphology, were used as the indicators of disease resistance to analyse the STB resistance in a set of wheat lines 44. Chartrain, et al. 45 identified a QTL for partial resistance to STB in wheat by using residuals from multiple regression on relative maturity and plant height as their phenotype. QTLs for spot blotch resistance in wheat were detected by using residuals when fitting disease severity as a dependent variable and plant height and days to heading as independent variables in multiple regression to exclude the effects of these traits 46. Most of the reported uses of maturity regression residuals phenotype pertain to STB resistance in wheat. This could be due to the well-characterised relationship between STB infection rates and both flowering time and plant height 47 and recognition that these confounding traits needed to be accounted for in ascertaining true genetic resistance for STB that would be useful for variety improvement 48. In general terms, for phenotypes based on deviation from the regression of a correlated trait (like our DRSRRM measure of resistance) and the approaches described above, the most resistant lines are those with the largest negative values of residuals, also allowing breeders to select resistant lines with desirable relative maturity and plant height 44,49. Further, as noted by Van Beuningen and Kohli 31, this approach is especially useful where disease resistance is evaluated in experiments at a single date, rather than at critical development stages for each line, especially in genetic material with large variation for confounding traits such as maturity.
In our study, a multiplicative mixed model was used to analyse the field data of scald resistance and relative maturity. A QTL for scald percentage on chromosome 2H were identified at the same position as QTLs for relative maturity; and the major QTL on chromosome 3H was also detected for scald percentage. When DRSRRM was utilised as a trait for QTL analysis, a major QTL was identified on chromosome 3H. This major QTL is co-located with a major QTL detected for scald resistance at seedling stage QTL-WAIYerong-3H. This indicates that by using DRSRRM, QTLs for disease resistance were identified as the confounding effects of relative maturity were removed.
In conclusion, a major QTL QTL-WAIYerong-3H providing scald resistance at both seedling and adult plant growth stages was identified. Results from this study indicate that this QTL-WAIYerong-3H is not Rrs1, as differential variety screening indicates the Rrs1 allele from Turk is not present in Yerong. The flanking markers for this QTL-WAIYerong-3H are located distantly (approx. 270Mb) from the Rrs1 locus based on the barley physical map. To overcome the confounding effects of relative maturity on adult plant disease resistance, a bivariate approach was used to model the field data of scald resistance and relative maturity. The phenotype derived from the bivariate analysis (DRSRRM) is a more effective trait to detect disease resistance QTLs as it removes the confounding effects of relative maturity. The identified new QTL identified in this study are a useful resource for pyramiding different resistance genes (major or minor) in breeding programs.
Author contributions
AM conceived and designed the experiments. MZ, RP and DS provided seeds of the Yerong-Franklin population. XZ performed the phenotype screening. BO, XZ, BAO and AM analysed the data. XZ, BO, BAO, MZ, RP, DS and AM prepared and edited the manuscript.
Funding
This work was financially supported by the Grains Research and Development Corporation (GRDC) of Australia under project DAQ00187.
Compliance with ethical standards Disclaimer
Conflict of interest
The authors declare that they have no conflict of interest.
Ethical approval
This article does not contain any studies with human participants or animals performed by the authors.
Acknowledgements
The authors thank Tony Goldthorpe, Michael McCaig and Brad Baxter for their expert technical contributions and data collection.