Introduction

Over 85% of spring wheat in Canada is produced in the western prairie provinces of Manitoba, Saskatchewan, and Alberta (McCallum and DePauw 2008). The primary target traits of spring wheat breeders in western Canada is developing early maturing and short stature cultivars with high-yield potential and grain protein content, combined with other end-use qualities, and resistance to diseases. Currently, cultivars to be released in the region must be at least intermediately resistant to stem rust (Puccinia graminis f. sp. tritici), leaf rust (Puccinia triticina), yellow (stripe) rust (Puccinia striiformis f. sp. tritici), common bunt (caused by two very closely related fungi, Tilletia tritici and Tilletia laevis), and Fusarium head blight (caused mainly by Fusarium graminearum) (http://www.pgdc.ca). Breeding for resistance to diseases involves (i) identification of sources of resistance, (ii) introgressing the new sources of resistance from the resistant parents into the genetic background of other parents, and (iii) selecting progenies showing acceptable combinations of resistance to diseases and other agronomic and quality traits. Breeding for disease resistance is often challenging for at least two reasons. First, breeders often need to pyramid diverse sources of resistance to multiple diseases in to the same genetic background. Second, the inheritance of each disease is often both qualitative and quantitative (Faris et al. 1996, 1997; Faris and Friesen 2005; Singh et al. 2007; Chu et al. 2008, 2010; Singh et al. 2016b), which complicates the selection process. Qualitative resistance is controlled by a single gene with a major effect, but resistance regulated by single genes often loses their effectiveness over time due to changes in pathogen populations. Quantitative resistance is controlled by multiple minor effect genes or quantitative trait loci (QTLs) with small additive effects, which are more durable (Singh et al. 2008), but require the introgression of multiple genes or QTLs that confer resistance for a given disease.

Currently, over 73 leaf rust (Lr) and 65 stripe rust (Yr) resistance genes have been reported in the literature on almost every chromosome (McIntosh et al. 2012; Dakouri et al. 2013; Park et al. 2014). Several gene combinations have been reported to contribute some level of resistance to rusts in many of the cultivars grown in western Canada (McCallum et al. 2007, 2012; Randhawa et al. 2012). However, only few gene combinations (pyramids) have provided good levels of durable resistance in global spring wheat cultivars released since the mid-1980s. Such combinations include the slow-rusting Lr34/Yr18 gene on chromosome 7DS (Suenaga et al. 2003; Spielmeyer et al. 2005; Lagudah et al. 2009), Lr46/Yr29 on 1BL (Singh et al. 1998; William et al. 2003), Sr2/Yr30 on 3BS (Singh et al. 2005), Yr17/Lr37/Sr38 cluster on 2AS (Helguera et al. 2003; Milus et al. 2015), and Lr67/Yr46 on 4DL (Hiebert et al. 2010; Herrera-Foessel et al. 2011).

At least 15 monogenic and race-specific genes (named from Bt1 to Bt15) conferring resistance to common bunt (Goates 1996) (also known as stinking smut and covered smut (Gaudet and Puchalski 1989)), have also been reported in wheat. Both incidence and severity of common bunt have been controlled largely by introgressing Bt10 and Bt8 (Menzies et al. 2006; McCallum and DePauw 2008; Hiebert et al. 2011), but the vulnerability of such major genes through intense selection pressure on the pathogen is a concern (Wang et al. 2009). QTLs associated with resistance to common bunt have also been reported on some chromosomes, including 1B and 7A (Galaev et al. 2006; Fofana et al. 2008; Wang et al. 2009). Tan spot, caused by Pyrenophora tritici-repentis, is the most destructive leaf spotting disease of wheat in Canada and other major wheat-growing countries (Faris et al. 1997; Friesen and Faris 2004). The virulence of P. tritici-repentis (Ptr) depends on the production of Ptr ToxA, Ptr ToxB, and/or Ptr ToxC host-specific toxins by the different races of the fungus (Ciuffetti et al. 1998, 2010; Lamari et al. 2003). Fungal isolates of races 2, 3, and 5 produce Ptr ToxA, Ptr ToxC, and Ptr ToxB, respectively (Strelkov and Lamari 2003; Lamari and Strelkov 2010). Isolates of races 1, 6, and 7 each produce two toxins, with race 1 producing Ptr ToxA and Ptr ToxC, race 6 producing Ptr ToxB and Ptr ToxC, and race 7 producing Ptr ToxA and Ptr ToxB. Race 8 isolates produce all the three toxins (Faris et al. 2013). Fungal isolates producing both Ptr ToxA and Ptr ToxC are highly abundant in the Canadian prairies, while those producing Ptr ToxB are extremely rare in this region (Lamari et al. 1998, 2003; Aboukhaddour et al. 2013).

We have been evaluating the performance of wheat cultivars and mapping populations at the University of Alberta, Edmonton, Canada for a wide range of agronomic traits and diseases. One of the mapping populations was a recombinant inbred line (RIL) population derived from a cross between two spring wheat cultivars, ‘Attila’ and ‘CDC Go’. ‘Attila’ is an awned, medium-yielding, semi-dwarf, and early-maturing cultivar (in South Asiatic regions) developed by the International Maize and Wheat Improvement Center from CM85836-50Y-0M-0Y-3M-0Y (Tadesse et al. 2010). It has been released in several countries with different local names and grown on millions of hectares throughout the world. ‘Attila’ carries at least two additive genes for slow-rusting resistance to leaf rust and three for stripe rust, such as Yr27 (Rosewarne et al. 2008) and Lr46/Yr29 (Rosewarne et al. 2006; Datta et al. 2009), showed moderate to high levels of field resistance to both leaf and stripe rusts, and has been frequently used as a slow-rusting donor parent in international spring wheat breeding programs (Rosewarne et al. 2008). ‘CDC Go’ is a Canadian Western Red Spring wheat cultivar characterized by strong straw, medium height, relatively late maturity (in western Canada), high yield, and high test weight and thousand kernel weight (Asif et al. 2015); it is also resistant to common bunt, and moderately resistant to resistant to both leaf and stripe rust (Randhawa et al. 2012; Perez-Lara et al. 2017). ‘CDC Go’, however, does not have the Lr34/Yr18 resistance allele (Randhawa et al. 2012). We previously used the ‘Attila’ × ‘CDC Go’ RIL population to map QTLs associated with agronomic traits both under conventional and organic management systems (Asif et al. 2015; Zou et al. 2017a, b). This RIL population has also shown good segregation for different wheat diseases, but information on the genetics of disease resistance in this population has not been previously reported. Here, we present QTLs associated with common bunt, tan spot, leaf rust, and stripe rust resistance in the ‘Attila’ × ‘CDC Go’ RIL population using the Wheat 90K SNP array.

Materials and methods

Phenotyping

The present study was conducted on a biparental mapping population of 167 RILs developed from a cross between two spring wheat cultivars—‘Attila’ (CM85836-50Y-0M-0Y-3M-0Y) and ‘CDC Go’. The RIL population was developed as described in one of our previous papers (Asif et al. 2015). The RIL population and the two parents, along with two susceptible (‘AC Barrie’ and ‘AC Crystal’) and two resistant (‘Lillian’ and ‘Carberry’) checks were evaluated eight times for their reaction to stripe rust in disease screening nurseries at Creston, British Columbia (49° 6′ 28.32″ N, 116° 34′ 4.67″ W) in 2011, 2013, and 2014; at Lethbridge Research Centre, Alberta (49.6989° N, 112.7636° W) from 2012 to 2015, and at Ellerslie Research Station, Edmonton, Alberta (53° 25′ 35.30″ N, 113° 32′ 59.03″ W) in 2015.

The RIL population, the two parents, and checks also were evaluated for 3 years (2012–2014) for reaction to leaf rust, common bunt, and tan spot at the Crop Research Facility of the University of Alberta, South Campus (53° 19′ N, 113° 35′ W), Alberta, Canada. The following cultivars were used as susceptible/moderately susceptible and moderately resistant/resistant checks: (i) leaf rust nurseries: ‘AC Barrie’ and ‘Park’ as moderately susceptible to susceptible checks, and ‘Peace’ and ‘Carberry’ as moderately resistant to resistant checks; (ii) tan spot nurseries: ‘AC Barrie’, ‘Unity’, and ‘Glenlea’ as moderately susceptible checks, and ‘Neepawa’ as moderately resistant check; and (iii) for common bunt nurseries: ‘Glenlea’ and ‘Neepawa’ as moderately susceptible checks; ‘AC Barrie’ and ‘Unity’ as moderately resistant and resistant checks, respectively. In Creston and Lethbridge, plots consisted of a single 1-m long row per genotype, arranged in a randomized complete block (RCB) design with two to three replicates per trial, depending on seed and space availability. In Edmonton, we grew hill-plots of ten seeds per genotype spaced 25 cm apart, with a similar RCB design. Details on disease evaluation methodology have been described in our recent paper (Perez-Lara et al. 2017). Visual disease assessments (disease scores) for leaf rust and stripe rust were done on a scale of 1 (no visible sign or symptom = resistant) to 9 (leaf area totally covered with spores = highly susceptible) on a plot basis when the spreader rows reached maximum infection. Tan spot reaction was recorded on a scale of 1 to 9 at milk-stage as described above. For common bunt, all heads of each genotype in a hill plot were examined for infection at the dough stage and recorded in percentages as the ratio of the number of infected heads to the total number of head per hill plot.

Genotyping and linkage maps

DNA extraction, genotyping, and linkage analyses were done as described in our previous study (Perez-Lara et al. 2016), and the linkage maps are summarized by Zou et al. (2017a). Briefly, a total of 5665 of the 81,587 SNPs and 3 functional markers initially used for genotyping the RIL population were incorporated into 27 linkage groups that covered all chromosomes except chromosomes 3D and 4D (Supplemental Table 1); none of the SNPs were polymorphic on chromosomes 3D and 4D. Markers that cosegregated in the linkage maps were excluded from the final dataset, which reduced the final number of markers retained for QTL analyses to 1200 SNPs and 3 functional markers (Ppd-D1, Vrn-A1a, and Rht-B1). The total map length for the 19 chromosomes, excluding 3D and 4D, was 3442 cM, with each chromosome varying between 14 cM on 1D and 325 cM on 5B (Zou et al. 2017a).

Statistical analyses

In each environment (year), outlier disease scores were checked on row basis using box plot and frequency distribution histograms in MiniTab v14, and some extreme values were deleted prior to any statistical analyses. Least square means and F statistics were computed for combined disease scores across all environments using PROC MIXED, whereas plot-based broad-sense heritability was computed using PROC ILM in SAS version 9.3 (SAS Institute Inc. Cary, USA). Genotypes (entries) were considered fixed, while environments, replications, replication within environment, and entry × environment interactions were considered random effects. Both test for normality and the frequency distribution were computed using MiniTab v14. Inclusive composite interval mapping (ICIM) was performed on disease scores averaged across all environments with QTL IciMapping v4.1 (Li et al. 2007; Meng et al. 2015) using a mean replacement for missing data, 1 cM walking distance, a minimum logarithm of odds (LOD) score of 2.5, and an additive model to determine the effect of individual QTL. QTL names were designated following the International Rules of Genetic Nomenclature (http://wheat.pw.usda.gov/ggpages/wgc/98/Intro.htm), which consisted of three letters for trait acronym, lab designation (dms = Dean Michael Spaner), and chromosome. Genetic maps and QTL graphs were drawn using MapChart v2.1 (Voorrips 2002).

Results

Disease evaluation

Tan spot, leaf rust, and stripe rust mean disease scores among the checks varied from 1.1 to 3.8 (in the 1 to 9 scale) for the resistant checks and from 5.3 to 7.4 for the susceptible checks. For common bunt, mean scores varied from 1.8 to 3.8% for the resistant checks and from 7.4 to 34.3% for susceptible checks (data not shown). Table 1 provides a summary of descriptive and F statistics. For tan spot, stripe rust, and leaf rust, RILs with mean disease scores ≤3.0 were considered resistant, 3.1–5.0 moderately resistant, 5.1–7.0 moderately susceptible, and 7.1–9.0 susceptible. Based on such categorization, both parents (‘Attila’ and ‘CDC Go’) exhibited moderate resistance to leaf rust (4.2 for both parents) and stripe rust (3.3–3.5), but moderately susceptible to tan spot (5.4–5.8). However, common bunt infection in ‘CDC Go’ was 9%, which was half of the 18% observed for ‘Attila’.

Table 1 Summary of descriptive and F statistics for 167 recombinant inbred lines (RILs) evaluated for common bunt, tan spot, leaf rust and stripe rust under field conditions between 2011 and 2015 in Canada
Full size table

The 167 RILs showed highly variable reactions to the four diseases, varying from 4.0 to 8.5 for tan spot, from 2.5 to 5.8 for leaf rust, from 1.8 to 7.2 for stripe rust, and from 1.8 to 37.2% for common bunt (Table 1). RILs differed (p < 0.0001) for their reactions to all four diseases. Broad-sense heritability varied from 0.25 for leaf rust to 0.48 for common bunt (Table 2). The distribution of disease scores averaged across all environments was normal (P > 0.050) for both tan spot and leaf rust, but it deviated from normality (P ≤ 0.018) both for common bunt and stripe rust. Overall, the RIL population exhibited transgressive segregation for all four diseases and several RILs that were superior or inferior to the parents were observed (Fig. 1). A total of 11 and 33 RILs were found to be resistant with a disease score rating of ≤3.0 for leaf rust and stripe rust, respectively, but only 6 RILs were resistant to both rusts.

Fig. 1
figure 1

Frequency distribution of least squares means of 167 recombinant inbred lines (RILs) evaluated at three environments for tan spot, common bunt, and leaf rust and eight environments for stripe rust. The arrows indicate values of the two parents: ‘Attila’ (A) and ‘CDC Go’ (C)

Full size image

QTL mapping

Using inclusive composite interval mapping and disease scores averaged across all environments, we identified a total of 11 QTLs (Table 3; Fig. 2) associated with resistances to tan spot (3), common bunt (2), leaf rust (3), and stripe rust (3). The three QTLs for resistance to tan spot mapped at 264 cM on 2B (QTs.dms-2B), at 3 cM on 2D (QTs.dms-2D), and at 27 cM on 6B (QTs.dms-6B), which individually explained between 7 and 10% (Table 3) and altogether accounted for 24.0% of the phenotypic and 52.2% of the genetic variance for tan spot (Table 2). RILs with ‘CDC Go’ alleles at the two flanking markers of all three QTLs for tan spot showed between 0.6 and 0.9 (on 1 to 9 scale) less tan spot scores than those with ‘Attila’ alleles.

Table 2 Summary of the number of QTLs identified, plot-based broad-sense heritability, and total phenotypic and genetic variance
Full size table
Fig. 2
figure 2

Chromosomal position of QTLs associated with common bunt (black), tan spot (pink), leaf rust (red), and stripe rust (blue) across combined environments. Map position in centiMorgans (cM) is shown on the left side of the chromosomes, with each horizontal line representing a marker. QTLs are shown on the right side of each linkage group, with bars indicating their 95% genetic confidence interval. Details of each QTL is given in Table 3, while the markers and linkage map is given in Supplemental Table 1

Full size image
Table3 Summary of QTLs associated with four wheat diseases based on 167 recombinant inbred lines (RIL) evaluated from three to eight environments between 2011 and 2015 in field nurseries
Full size table

ICIM uncovered two QTLs associated with common bunt resistance (Table 3) that mapped at 52 cM on the second linkage group of chromosome 1B (QCbt.dms-1B.2) and at 202 cM on chromosome 3A (QCbt.dms-3A) that altogether accounted for 26.5% of the phenotypic and 55.2% of the genetic variance across all environments (Table 2). QCbt.dms-1B.2 is a moderate effect QTL that mapped between BS00086854_51 and wsnp_Ex_c5679_9976893 on chromosome 1B, had a LOD score of 7.2 and individually explained 18.7% the phenotypic variance. QCbt.dms-3A was a minor effect QTL that mapped between RAC875_c17453_896 and RAC875_c57584_240 on 3A and explained 7.9% of the phenotypic variance across all combined environments. On average, RILs with the ‘CDC Go’ alleles at the two flanking markers for QCbt.dms-1B.2 and QCbt.dms-3A had 7.1% and 2.9%, respectively, less common bunt disease score than those with ‘Attila’ alleles.

Three QTLs associated with resistance to leaf rust were mapped at 39 cM on 2D (QLr.dms-2D), at 2 cM on 2D second linkage group (QLr.dms-2D.2), and at 17 cM on 3A (QLr.dms-3A) that altogether explained 21.5% of phenotypic and 86.0% of genetic variance across all environments (Table 2). Each QTL for resistance to leaf rust individually explained between 5.9 and 8.6% of phenotypic variance across combined environments (Table 3). RILs with ‘CDC Go’ alleles at the flanking markers of all three QTLs for leaf rust scored from 0.3 to 0.5 less leaf rust disease score than those with ‘Attila’ alleles. The three QTLs associated with resistance to stripe rust were mapped at 300 cM on 3A (QYr.dms-3A), at 123 cM on 4A (QYr.dms-4A), and at 191 cM on 5B (QLr.dms-5B). Each QTL for stripe rust resistance individually explained between 6.7 and 8.5% (Table 3) and altogether accounted for 23.1% of the phenotypic and 82.5% of the genetic variance across all combined environments (Table 2). The resistant alleles for QYr.dms-3A and QYr.dms-4A originated from ‘CDC Go’, while that of QYr.dms-5B from ‘Attila’. RILs with the resistance alleles at the two flanking markers of every QTL for stripe rust had between 0.5 and 0.7 less stripe rust disease scores than those with susceptible alleles.

Discussion

Resistance to common bunt and tan spot

Here, we report three and two QTLs associated with resistance to tan spot and common bunt accounting for 52.2 and 55.2% of the total genetic variance, respectively (Table 2). Given our medium heritability estimates for tan spot (0.46) and common bunt (0.48), nearly 45–48% of the genetic variation of these two wheat diseases remained unexplained, which may be partly due to limitations of the current linkage map (wider gaps on marker distribution in some chromosomes; lack of SNP polymorphism on chromosomes 3D and 4D, and very low genome coverage on the D genome as whole) (Zou et al. 2017a) and/or presence of other minor effect QTLs that remained undetected in the current phenotypic data. Furthermore, tan spot infection in wheat fields occurs in association with Stagonospora nodorum and Zymoseptoria tritici (Singh et al. 2006), which may have partly contributed to elevated tan spot disease scoring in the current study.

One of the QTLs associated with common bunt was a minor effect QTL that mapped on chromosome 3A (QCbt.dms-3A) and accounted nearly for 8% of the phenotypic variation (Table 3). We are not aware of any mapping studies that reported either single major effect gene(s) or QTL on chromosome 3A that confers resistance to common bunt in wheat. It is possible that QCbt.dms-3A may be a novel minor effect QTL that has not been reported elsewhere. In addition, we also identified a moderate effect QTL on chromosome 1B (QCbt.dms-1B.2), which accounted for nearly 19% of the phenotypic variation for common bunt disease reaction. Several previous studies have reported genes and QTLs associated with resistance to common bunt on chromosome 1B (Fofana et al. 2008; Wang et al. 2009; Dumalasová et al. 2012; Singh et al. 2016a). One of the previous studies used a doubled haploid (DH) spring wheat population derived from a cross between ‘RL4452’ × ‘AC Domain’ and reported three QTLs associated with resistance to common bunt (Fofana et al. 2008). In that study, the authors observed a continuous frequency distribution and transgressive segregation for common bunt and reported two QTLs on chromosome 1B (QCbt.crc-1B.1 and QCbt.crc-1B.2) that together explained 29%, and another QTL on chromosome 7A that accounted for 3% of the phenotypic variation. QCbt.crc-1B.1 mapped on the short arm of chromosome 1B between XGwm374.1 and XWmc818, had a LOD score of 9.0 and accounted for 21% of the phenotypic variance, while QCbt.crc-1B.2 mapped between GluB1 and XGwm274 on the long arm of chromosome 1B, had a LOD score of 3.6 and explained 8% of the phenotypic variance. Our results, together with others suggest the presence of a moderate to major effect QTL associated with common bunt resistance on chromosome 1B irrespective of the genetic background of the mapping populations, the type of markers and marker density. However, directly comparing the two studies is difficult due to the use of diverse marker types and lack of physical position for all flanking markers.

At least 15 monogenic and race specific genes (named from Bt1 to Bt15) conferring resistance to common bunt have also been reported in wheat (Goates 1996), of which Bt4, Bt5 and Bt6 mapped on chromosome 1B (Sears et al. 1960; Scmidt et al. 1969). In a winter wheat DH population derived from a cross between ‘Blizzard’ × ‘8405-JC3C’, a single gene associated with common bunt resistance has been reported on the short arm of chromosome 1B (Wang et al. 2009). Using a DH population derived from a cross between ‘Trintella’ and ‘Piko’, another group mapped a single major effect gene conferring resistance to common bunt around the centromere region on chromosome 1B, flanked by Xgwm273 (Dumalasová et al. 2012). That gene had a LOD score of 38 and explained up to 30% of the phenotypic variance for common bunt disease severity. Common bunt disease severity in the ‘Trintella’ × ‘Piko’ DH population showed bimodal distribution, which clearly support the presence of a major effect gene or QTL segregating in that population. In the present study, common bunt disease scores in the ‘Attila’ × ‘CDC Go’ RIL population showed somewhat a bimodal distribution (Fig. 1) that may partly explain the moderate effect QCbt.dms-1B.2 QTL identified on chromosome 1B (Table 3). Again, direct comparisons of the position of genes and QTLs reported in the current study with previous studies were not possible due to differences in the types of markers used and lack of physical positions of the flanking markers reported in the different studies. For tan spot, we identified three minor effect QTLs on chromosomes 2B (QTs.dms-2B), 2D (QTs.dms-2D), and 6B (QTs.dms-6B), which individually explained between 6.9 and 10.0% of the phenotypic variance for tan spot disease reaction. Previous studies have reported genes and QTLs associated with tan spot resistance on several chromosomes, including 2B (Friesen and Faris 2004; Gurung et al. 2011), 2D (Gurung et al. 2011), and 6B (Singh et al. 2016b). Ptr ToxB is a proteinaceous host-selective toxin that induces chlorosis in wheat lines harboring the dominant Tsc2 gene (Strelkov et al. 1999; Lamari et al. 2003), which mapped on the short arm of chromosome 2B (Friesen and Faris 2004; Abeysekara et al. 2010). Recently (Perez-Lara et al. 2017), our group evaluated 81 spring wheat cultivars registered (released) in western Canada for reaction to tan spot isolates in field nurseries and for sensitivity to the three Ptr toxins in a greenhouse, and genotyped them with a subset of 19,919 of the Wheat 90K single-nucleotide polymorphic (SNP) array and 11 gene-specific markers. Using genome-wide association analysis, we uncovered clusters of 30 SNPs on chromosome 2B (from 17 to 39 cM interval) that were significantly (p value ranging from 4.5 × 10−5 to 2.9 × 10−10) associated with Ptr ToxB; each SNP individually explained from 21.3 to 44.9% of the phenotypic variation of Ptr ToxB, with an overall average of 28.9%. In the current study, the two flanking markers (Kukri_c148_1346 and GENE-1343_556) for QTs.dms-2B fell within the confidence interval that we reported for Ptr ToxB in the association mapping panel, which suggests that the two regions are the same. However, the genetic position of QTs.dms-2B (262–270 cM) is different from Ptr ToxB QTL position that we reported using association analysis (Perez-Lara et al. 2017), because the linkage map used in the current study was developed using the 167 RILs derived from ‘Attila’ × ‘CDC Go’, while the linkage map used in our GWAS study was a consensus map developed based on multiple mapping populations available to our programs.

At least three minor effect QTLs associated with Ptr ToxB have also been reported on the short arm of chromosomes 2A, the long arms of both 2B and 4A (Friesen and Faris 2004), which suggests that the sensitivity of wheat genotypes to Ptr ToxB depends not only on the Tsc2 gene on chromosome 2B, but also on additional minor effect QTLs that are located on different chromosomes. Using a DH population derived from a cross between ‘CPI133872’ and ‘Janz’, Zwart and colleagues (Zwart et al. 2010) reported a major effect QTL on chromosome 3D, the recessive tsn1 gene on 5BL that confers insensitivity to Ptr ToxA, and five environment specific QTLs on chromosomes 2B, 2D, 3A, 4B, and 5A (Zwart et al. 2010). Based on an association mapping study conducted on 567 spring wheat landraces from the United States Department of Agriculture National Small Grains Collection, several genomic regions, including chromosomes 2B and 2D, that individually explained between 1.3 and 5.9% of the phenotypic variance for race 1 and/or race 5 isolates have been reported (Gurung et al. 2011). An associating mapping study conducted on a set of bread wheat germplasm from CIMMYT has also reported 9 genomic regions associated with tan spot resistance, including chromosome 6B (Singh et al. 2016b). Therefore, the QTLs that we identified for tan spot on chromosomes 2B, 2D, and 6B may be on the same genomic regions with those genes and QTLs reported in previous studies. However, we were not able to make direct comparisons among the different studies for diverse reasons: (i) most studies were conducted using different types of markers, mapping populations, and linkage mapping software’s, which makes direct comparison based on genetic position unreliable; (ii) physical positions for most flanking markers reported in multiple studies, including most SNPs used in the present study, not yet available; (iii) there is even some level of discrepancy (disagreement) on loci order and genetic positions of several SNPs from the Wheat 90K SNP array depending on the mapping population; and (iv) for SNPs where physical positions are available, there are also some level of disagreement with physical positions given in publicly available database, such as Triticeae toolbox (https://triticeaetoolbox.org/wheat/) and Ensembl Plants (http://plants.ensembl.org/Triticum_aestivum/Info/Index).

Resistance to leaf and stripe rust

‘Attila’ was reported to possess moderate to high level of field resistance to both leaf and stripe rusts (Rosewarne et al. 2006; Datta et al. 2009) and has been frequently used as a slow-rusting donor parent in international spring wheat breeding programs (Rosewarne et al. 2008; Datta et al. 2009). ‘CDC Go’ has also been reported to be moderately resistant to leaf rust, and resistant or moderately resistant to stem rust (http://www.agric.gov.ab.ca) and stripe rust (McCallum et al. 2012; Randhawa et al. 2012). In our disease evaluation nurseries, both ‘Attila’ and ‘CDC Go’ showed moderate level of resistance to both leaf and stripe rusts, which agrees with previous studies. Using ICIM, we identified 3 QTLs associated with resistance to leaf rust on the short arms of both chromosomes 2D (QLr.dms-2D.1 and QLr.dms-2D.2) and 3A (QLr.dms-3A), which individually explained between 5.9 and 8.6% of the phenotypic variance (Table 3). Previous mapping studies have reported several single genes associated with leaf rust resistance on chromosome 2D, which includes Lr2a, Lr2b and Lr2c (Dyck and Samborski 1974), Lr15 (Luig and McIntosh 1968), Lr22a, Lr22b and Lr22c (Rowland and Kerber 1974; Dyck 1979), Lr39 and Lr41 (Singh et al. 2004), and Lr54 (Marais et al. 2005). However, the observed continuous leaf rust disease scores distribution with a single peak in the ‘Attila’ × ‘CDC Go’ RIL population does not support the presence of major effect single genes. Several QTLs associated with leaf rust resistance have also been reported on chromosome 2D. One of the QTLs located on chromosome 2D was QLrid.osu-2D, which explained between 21.5 and 26.4% of the phenotypic variance for leaf rust infection in a RIL population derived from ‘CI13227’ × ‘Suwon92’ (Xu et al. 2005). A major effect QTL on the short arm of chromosome 2A (QYr.ufs-2A), along with three minor effect QTLs on chromosomes 2D (QYr.ufs-2D ), 5B (QYr.ufs-5B), and 6D (QYr.ufs-6D) have been reported (Agenbag et al. 2012). QYr.ufs-2D was located on the short arm of chromosome 2D that is believed to be the position of Yr16 and explained between 4.7 and 10.3% of the phenotypic variance for stripe rust on individual experiments. An environment specific QTL associated with leaf rust has also been reported on the short arm of chromosome 2D distal to Xwmc25.2 (Buerstmayr et al. 2014). Three other minor effect QTLs for leaf rust resistance were reported on chromosomes 1B, 2A, and 2D (Rosewarne et al. 2012). A recent meta-analysis compiled 144 QTLs reported in 19 studies conducted between 1999 and 2015 using 20 mapping populations (Soriano and Royo 2015). That study reported a total of 35 meta-QTLs associated with leaf rust resistance, of which three meta-QTLs mapped on chromosome 2D (MQTL9, MQTL10, and MQTL1) and two meta-QTLs mapped on 3A (MQTL12, MQTL13, and MQTL14), with each meta-QTL consisting of clusters of two to six QTLs (Soriano and Royo 2015). MQTL14 mapped around Xmwg570 on chromosome 3A, and the individual QTLs explained between 19 and 30% of leaf rust resistance (Maccaferri et al. 2008). Overall, results from the various studies, together with ours, clearly revealed the presence of several genomic regions on chromosome 2D conferring resistance to leaf rust, which does not seem the case for chromosome 3A.

For stripe rust resistance, we identified three QTLs at 300 cM on 3A (QYr.dms-3A), at 123 cM on 4A (QYr.dms-4A), and at 191 cM on 5B (QLr.dms-5B (Table 3). Each QTL associated with stripe rust resistance individually explained between 6.7 and 8.5% of the phenotypic variance. Although we are not aware of stripe rust resistance gene assigned to chromosome 3A, a few minor effect QTLs for stripe rust have been reported on chromosome 3A. Using composite interval mapping in the Avocet × Saar population, for example, a QTL associated with resistance to stripe rust was reported on the short arm of 3A, which colocalized with a QTL for powdery mildew resistance (Lillemo et al. 2008). In another study, a minor effect QTL (QYr.ifa-3AS) that explained 5.6% of the phenotypic variance for stripe rust severity was reported on the short arm of chromosome 3A in three of five experiments in the ‘Capo’ × ‘Arina’ population (Buerstmayr et al. 2014). Another study (Rosewarne et al. 2012) has also reported minor effect QTLs for stripe rust resistance on several chromosomes, including 3A.

Stripe rust resistance genes Yr51 (Randhawa et al. 2014) and Yr60 (Herrera-Foessel et al. 2015) are two genes reported on chromosome 4A. Yr60 is located on the long arm of 4A and conferred moderate levels of resistance at both seedling and adult plant stages against two Mexican races of P. striiformis (Herrera-Foessel et al. 2015). The distribution of stripe rust disease score in the present study, however, was continuous with a single peak, which did not support the segregation of single gene with major effect. Recently, two QTLs on chromosomes 4A (QYrel.wak-4A) and 6B (QYrfi.wak-6B) that explained between 15 and 16% of the phenotypic variance for stripe rust disease severity in a RIL population derived from ‘Eltan’ and ‘Finch’ have been reported (Klarquist et al. 2016). Yr19 (Chen et al. 1995) and Yr47 that mapped distal to Lr52 (Bansal et al. 2011) are the single genes that confer resistance to stripe rust resistance on chromosome 5B. Some QTLs that confer an adult stage stripe rust resistance have also been reported on chromosome 5B, which includes Yrco.wpg-5B and QYrbr.wpg-5B (Case et al. 2014).

Conclusions

Using combined disease scores across three to eight environments, we identified a total of 11 QTLs associated with resistance to four wheat diseases, which included two for common bunt and three each for tan spot, leaf rust, and stripe rust. Each QTL showed either minor or moderate effect and individually explained between 5.9 and 18.7% of phenotypic variance, and altogether accounted from 21.5 to 26.5% of phenotypic and from 52.2 to 86.0% of genetic variance. Even though the identified QTLs are of minor to moderate effect, they provide useful information to spring wheat breeders aiming to pyramid such types of genomic regions for developing wheat cultivars with durable levels of disease resistance. Some of the QTLs identified in the present study were novel, while others were located on the same genomic regions as previously reported QTLs. Direct comparisons on QTL positions across multiple studies was difficult due to differences in marker platforms, lack of common set of markers, and/or physical positions of flanking markers reported in different studies.