Dissection of the multigenic wheat stem rust resistance present in the Montenegrin spring wheat accession PI 362698

Background Research to identify and characterize stem rust resistance genes in common wheat, Triticum aestivum, has been stimulated by the emergence of Ug99-lineage races of the wheat stem rust pathogen, Puccinia graminis f. sp. tritici (Pgt), in Eastern Africa. The Montenegrin spring wheat landrace PI 362698 was identified as a source of Pgt resistance. This accession exhibits resistance to multiple Ug99-lineage and North American Pgt races at seedling and adult-plant stages. A recombinant inbred population was developed by crossing the susceptible line LMPG-6 with a single plant selection of PI 362698. A genetic map was constructed using the Illumina iSelect 90 K wheat assay and the markers csLv34, NB-LRR3, and wMAS000003 and quantitative trait locus (QTL) analysis was performed. Results QTL analysis identified five significant QTLs (α = 0.05) on chromosomes 2B, 3B, 6A, 6D, and 7A associated with wheat stem rust resistance. The QTL on chromosome 3B was identified using both field data from Kenya (Pgt Ug99-lineage races) and seedling data from Pgt race MCCF. This QTL potentially corresponds to Sr12 or a new allele of Sr12. The multi-pathogen resistance gene Sr57 located on chromosome 7D is present in PI 362698 according to the diagnostic markers csLv34 and wMAS000003, however a significant QTL was not detected at this locus. The QTLs on chromosomes 2B, 6A, and 6D were identified during seedling trials and are thought to correspond to Sr16, Sr8a, and Sr5, respectively. The QTL identified on chromosome 7A was detected using MCCF seedling data and may be Sr15 or a potentially novel allele of recently detected Ug99 resistance QTLs. Conclusions The combination of resistance QTLs found in PI 362698 is like the resistance gene combination present in the broadly resistant cultivar Thatcher. As such, PI 362698 may not be a landrace as previously thought. PI 362698 has been crossed with North Dakota wheat germplasm for future breeding efforts. Additional work is needed to fully understand why the combination of genes present in PI 362698 and ‘Thatcher’ provide such durable resistance. Electronic supplementary material The online version of this article (10.1186/s12864-018-4438-y) contains supplementary material, which is available to authorized users.


Background
The emergence of new races of the wheat stem rust pathogen, Puccinia graminis Pers.:Pers. f. sp. tritici (Pgt), has historically been a problem in many wheat (Triticum aestivum L.) producing regions [1][2][3][4][5]. Pgt is a macrocyclic rust that mainly persists in its asexual uredinial state and is primarily controlled using resistance genes [1][2][3][4]. The use of resistance genes has been particularly effective in the United States due to low pathogen diversity caused by the removal of the alternate host of Pgt, the common barberry (Berberis vulgaris L.), which is required for sexual recombination [2,6]. Despite the reduction of Pgt's sexual cycle in most of North America, new Pgt races occasionally emerge due to a sexual population in the Pacific Northwest and selection pressures imposed upon the Pgt population by deployed resistance genes [6]. In Eastern Africa new races also emerge due to selection pressures imposed by deployed resistance genes; however, the rate of emergence is enhanced due to multiple cropping seasons in a year causing a "green bridge" [3,7].
The emergence of Ug99-lineage races and new virulent races in the durum [Triticum turgidum subsp. durum (Desf ) Husn] producing regions of Ethiopia has instilled new concerns regarding global food security [3-5, 8, 9]. In 1998 the Pgt race TTKSK (Ug99), which is virulent on Sr31, was identified in Uganda [10]. Two years later a second Pgt race, TTKSF, was identified in South Africa with shared ancestry to Ug99 [7]. Selection pressure imposed by deployed resistance genes has resulted in Ug99-lineage races that are virulent on Sr24, Sr36, Sr9h, and SrTmp [3,[11][12][13][14]. In Ethiopia, cultivars with resistance to Ug99-lineage races were deployed [8,9]. These cultivars led to selection against Ug99-lineage races and a predominance of the race TKTTF in these regions, which lead to major epidemics in 2013 and 2014 [9].
Over 60 wheat stem rust resistance genes have been found in wheat, at least 31 of which confer resistance to at least one race in the Ug99-lineage [3-5, 14, 15]. Approximately one third of all stem rust resistance genes and half of the resistance genes effective to Ug99-lineage races were introgressed from wild wheat relatives [5,14]. Many of these genes suffer from linkage drag and low efficacy when deployed alone [14]. Due to the continued emergence of virulent Pgt races it is important to identify and characterize new sources of resistance.
Landraces can serve as unexploited sources of resistance to many diseases. The deployment of previously used resistance sources in modern breeding programs has created diversity bottlenecks for many crop species, creating vulnerability to disease [16]. As such, the increased diversity found in landraces due to their association with traditional farming systems make them useful tools for many wheat breeding programs [17][18][19][20]. Resistance genes identified in landraces are easier to incorporate into adapted material than those from wild relatives [18]. The identification and subsequent mapping of disease resistance loci from landraces has been effectively applied to wheat [21][22][23][24], including wheat stem rust resistance loci [25][26][27][28].
During a recent study to identify Pgt resistance in spring habit T. aestivum landraces, Newcomb et al. [29] [29][30][31]. A mesothetic infection type is unusual but a similar mesothetic seedling infection type has been observed for 'Thatcher' , which is postulated to have the resistance genes Sr5, Sr9g, Sr12, and Sr16 [32][33][34]. PI 362698 is postulated to have Sr57 based on DNA markers, however any additional resistance genes present within the line were unknown [29]. To evaluate the mode of inheritance for TTKSK resistance a single plant selection of PI 362698, PI 362698-1, was crossed to the susceptible Canadian line LMPG-6 (Little Club//Prelude/8*Marquis/3/Gabo) [30,35]. Mendelian ratios corresponding for one to three genes were not observed during seedling resistance evaluations at the F 2 generation, suggesting the stem rust resistance to TTKSK found in PI 362698-1 is complex [30]. Consequently, the LMPG-6/PI 362698-1 population was advanced via single seed descent to identify the genetic regions associated with stem rust resistance in PI 362698-1.

Population development
The susceptible Canadian line LMPG-6 [34] was used as a female parent and crossed to PI 362698-1, a single plant selection [30]. A recombinant inbred population was created by advancing the F 2 progeny via single seed descent to the F 6 generation. Seeds from the 151 F 7 generation individuals were pooled and advanced to produce enough seeds for phenotyping.

Phenotypic evaluation
The LMPG-6/PI 362698-1 population, PI 362698, and PI 362698-1 were tested for seedling resistance to North American (NA) and African Pgt races. PI 362698 was tested against the NA Pgt races MCCFC (isolate 59KS19), QFCSC (06ND76C), QTHJC (75D717C), RCRSC (77ND82A), RKQQC (99KS76A), TPMKC (74MN1409), and TTTTF (01MN84A-1-2) and the African Pgt races TRTTF (06YEM34-1), TTKSK (04KEN156/04), and TTKST (06KEN19v3) to determine the spectrum of seedling resistance present in PI 362698. PI 362698-1 was tested against the NA Pgt races HHBJ, HKHJ, HKQJ, HPCJ, HPLB, QFCQC, QKCS, QKMS, QTHJ and the African Pgt races TTKSF and TTKSF+. Races MCCFC, QFCSC, QTHJC, RCRSC, RKQQC, TPMKC, TTTTF, TRTTF, TTKSK, and TTKST were evaluated at the Cereal Disease Laboratory. Races TTKSF and TTKSF+ were evaluated at the University of the Free State in Bloemfontein, South Africa and races HHBJ, HKHJ, HKQJ, HPCJ, HPLB, QFCQC, QKCS, and QKMS were evaluated at North Dakota State University in Fargo, ND, U.S.A. The LMPG-6/PI 362698-1 population was tested for seedling resistance against Pgt race HHBJ (isolate R29J), HPLB (A-15), MCCF (A-5), and QFCQC at North Dakota State University; TRTTF at the Cereal Disease Laboratory; and TTKSF+ at the University of the Free State. Two replicates of the population were planted with five seeds per experimental unit. Seedling evaluations at the Cereal Disease Laboratory were conducted by inoculating seedlings 7 to 10 days after planting with urediniospores retrieved from −80°C storage. Urediniospores were revitalized with a 15 min 45°C heat shock followed by 2 to 4 h rehydration under 80% relative humidity created with a KOH solution [36]. A spray inoculator was used to distribute a urediniospore and mineral oil suspension (Sotrol 170, Philips Petroleum, Borger, TX, U.S.A.). The oil was allowed to evaporate from the plants in a fume hood for 30 min and then plants were placed in a dark dew chamber for 14 h at 18°C after which the plants were exposed to fluorescent light for 3 to 4 h and transferred to a 18 ± 2°C greenhouse with a 16 h photo period [36,37]. Inoculations at North Dakota State University and University of the Free State were similar to those conducted at the Cereal Disease Laboratory; however freshly collected Pgt urediniospores were used and greenhouse temperatures ranged from 20 to 24°C. The plants were grown for 12 to14 days before evaluating infection types using the 0 to 4 scale developed by Stakman et al. [31].
The LMPG-6/PI 362698-1 population was tested for field response to Ug99-lineage races during the 2014 main season, July to October, and the 2015 off season, February to May, at the International Stem Rust Nursery at the Kenyan Agricultural Research and Livestock Organization in Njoro, Kenya. Hill plots with 10 to 15 seeds were planted between spreader rows of several wheat lines, including 'Cacuke' , bearing Sr31 and Sr24 to select for Pgt race TTKST. The population was replicated twice with the parental lines planted every 20 entries to evaluate the distribution of disease pressure. A mixture of talc powder and urediniospores was used to inoculate the spreader rows. The plants were rated using categorical scores to evaluate infection response and the area of infection using a modified Cobb Scale ranging from 0 to 100% [38,39]. Infection response categories could be used individually or in combination to describe multiple responses on a stem. The categories included immune (I), resistant (R), moderately resistant (MR), moderately susceptible (MS), and susceptible (S) [39].

Linkage mapping
Leaf tissue was collected from each recombinant inbred line (RIL) at the F 6 generation, lyophilized for 24 h, and then ground using a Retsch mm301 mixer mill (Retsch GmbH, Haan, Germany) as described in Rouse et al. [40]. DNA was extracted in 96-welled plates using a modified cetyltrimethylammonium bromide (CTAB) method [41]. The F 6 RILs and parental lines were genotyped using the Illumina Infinium iSelect 90 K wheat SNP assay [42], an Illumina BeadStation, and an Illumina iScan according to the manufacturer's protocol (Illumina Inc., San Diego, CA, U.S.A.). The markers were scored using version 1.0 of the polyploid clustering module for Illumina GenomeStudio ® version 1.9.4 (Illumina Inc.).
The diagnostic STS marker csLv34 was used to map the cloned resistance gene Sr57 [43]. Reactions were completed at a volume of 20 μl containing 1X PCR buffer, 2.5 mM MgCl 2 , 187.5 μM dNTPs, 500 nM primer, 60 ng DNA, and 1 unit of GoTaq (Promega Corporation, Madison, WI, U.S.A.). Amplification was performed under the following conditions: initial denaturation at 94°C for 5 min, followed by 40 cycles of a 30 s denaturation at 94°C, a 57°C annealing step for 30 s, and an extension step of 72°C for 45 s, followed by a final extension at 72°C for 7 min. Amplicons were visualized via 2% agarose gel electrophoresis.
The KASP markers NB-LRR3 [32] and wMAS000003 [44] were included for mapping due to their associations with Sr12 and Sr57, respectively. Sr12 often displays a mesothetic response to many races at seedling stages [32]. wMAS000003 is designed to detect the functional polymorphism for the resistant allele of Sr57 [44]. The KASP assays were performed as 5 μL reactions consisting of 2.5 μL of KASP 2X reaction mix (LGC Ltd., Teddington, United Kingdom), 0.07 μL of KASP primer mix (containing 12 μM of each allele-specific forward primer and 30 μM reverse primer diluted in 10 mM TRIS, pH 8.3), and 37.5 ng of DNA. Amplifications were performed under the following conditions: initial denaturation at 94°C for 15 min, followed by 10 cycles of a 20 s denaturation at 94°C and a combined annealing and extension touch-down step for 1 min where the initial temperature began at 61°C and decreased by 0.6°C per cycle to reach a final temperature of 55.6°C, followed by 26 cycles of denaturation at 94°C for 1 min and annealing and extension at 55°C for 1 min. Reaction fluorescence was measured using an Omega Fluorostar plate reader (BMG LABTECH GmbH, Ortenberg, Germany). Allelic data were analyzed using KlusterCaller version 2.21 (LGC Ltd).
ISelect markers with less than 5% missing data, NB-LRR3, wMAS000003, and csLV34 were used for mapping. Mapping was conducted using Mapdisto version 1.7.7.0.1.1 using a LOD of 5 and maximum recombination frequency of 0.1 [45]. The robustness of the map was evaluated using a ripple command with a window size of five markers. Problematic loci were identified using the drop command and all markers which increased the map length more than 3 centimorgans (cM) were removed. The ripple command was applied after each problematic locus was removed to revalidate the robustness of the map.
Linkage groups were assigned to chromosomes using the 90 K wheat consensus map [42] and combined into chromosomal linkage groups with gaps less than 35 cM. The genetic distance between markers was calculated using the Kosambi mapping function [46].

QTL analysis
To utilize the data for quantitative trait locus (QTL) analysis, adult-plant infection responses were converted to a numerical value modified from Yu et al. [47].  [47]. Seedling infection types were converted from the 0 to 4 Stakman scale to a 0 to 9 linear scale as described by Zhang et al. [48]. Median values were calculated for replicates and used for QTL analysis. The Shapiro-Wilk test was used to determine the normality of the population data. All statistical tests were performed using R version 3.3.2 [49]. QTL analysis was conducted using QGene version 4.4.0 [50]. Traits were analyzed using a multiple interval mapping algorithm based on a general linearized framework (MIM-GLZ) [51]. Significant QTLs (α = 0.05) were identified using a resampling analysis of 1000 permutations [52]. The 95% confidence intervals for significant QTLs were estimated using the 2-LOD drop method [53]. The Kruskal-Wallis test was used to test for epistatic interactions. Post-hoc analysis was conducted using Student's T-Test with the Benjamini-Hochberg correction for multiple comparisons.

Phenotypic evaluation
PI 362698 was resistant to the races tested at seedling stages and was immune to many of the NA Pgt races ( Table 1). The races TPMKC, TTKSK, TTKSF, and TTKSF+ exhibited mesothetic reactions at seedling stages (Fig. 1). Infection types varied greatly from race to race when evaluating the LMPG-6/PI 362698-1 population with races HHBJ, HPLB, MCCF, QFCQC, TRTTF, and TTKSF+ at the seedling stages (Table 2). Missing data was observed in each of the trials. In the seedling trials 13 (8.

Linkage mapping
The STS marker csLv34, KASP markers NB-LRR3 and wMAS000003, and 6863 SNP markers from the 90 K Infinium iSelect wheat assay were used for mapping. A genomic map with a total length of 2544.3 cM was produced consisting of 2153 unique loci (bins) distributed  Table S1). Co-segregating markers were found at 44.3% of the bins, ranging from 2 to 65 co-segregating markers per bin. Severe segregation distortion was observed for the long arm of chromosome 3B, approximately a 3:1 ratio of PI 362698-1 to LMPG-6 alleles for each locus, and linkage group 5A2, approximately a 2:1 ratio of PI 362698-1 to LMPG-6 alleles at each locus. Markers at these genetic locations were homozygous in each parent.
QSr.ace-3B, which was detected during the MCCF seedling trial, was detected during both field trials in Kenya during 2014 and 2015 (Table 3). This QTL explained 47% and 50% of the phenotypic variation for the 2014 and 2015 Kenyan trials, respectively. The position of this QTL mapped near the Sr12 associated marker NB-LRR3 but shifted slightly between each of the trials, likely due to missing data and environmental variation.
Sr57 is thought to be present in PI 362698-1 according to the diagnostic markers csLv34 and wMAS000003, however no significant QTLs (α = 0.05) were detected at the Sr57 locus.
The MCCF seedling trial was the only trial where multiple QTLs were detected ( Table 3). The individuals within the population were divided into groups based on the presence or absence of the PI 362698-1 allele at markers in the middle of the QTL region defined by the 2-LOD drop method [53]. The Kruskal-Wallis test was performed on these groups and found differences between different QTL combinations. Post-hoc analysis with Student's T-test identified that the PI 362698-1 allele was required for all three QTLs; QSr.ace-3B, QSr.ace-6A, and QSr.ace-7A2; for the highest level of resistance (Fig. 3). The PI 362698-1 allele is required for at least two QTLs to provide resistance (Fig. 3). If the PI 362698-1 allele is present for only one QTL, the individual is less susceptible than individuals without a PI 362698-1 allele for any of the QTLs (Fig. 3).

Discussion
Composite interval mapping (CIM) is one of the most popular algorithms used in plant QTL mapping experiments, however its robustness and power is less than that of multiple interval mapping (MIM) [51,54]. The increased power of MIM stems from the algorithm considering multiple chromosomal regions simultaneously allowing for the detection of multiple QTLs and improved detection of epistatic interactions [54]. Both of these algorithms, as well as simple interval mapping and Bayesian interval mapping, assume that trait data are continuous and normally distributed [51,[53][54][55][56][57]. As such, these algorithms experience reduced power when applied to categorical or non-normally distributed data, which is commonly found in agricultural studies [51,58,59]. Data normality is an important factor which is often overlooked for many QTL studies and it can influence the ability to detect minor effect QTLs and partial disease resistance, especially in the presence of large effect QTLs [51,59]. Transformation of data is an option to meet normality assumptions, but transformation often introduces unwanted biases that skew QTL effect estimates [51]. Algorithms based on a general linearized framework, such as the MIM-GLZ algorithm, are an excellent option to overcome model assumption violations in trait data and provide more power than traditional algorithms used on transformed data [51,59]. In the present study, the MIM-GLZ algorithm was used to dissect the complex stem rust resistance found in the highly resistant Montenegrin spring wheat landrace PI QSr.ace-3B was identified using MCCF seedling data and the 2014 and 2015 Kenyan field data and explained 45%, 47%, and 50% of the phenotypic variation, respectively. QSr.ace-3B was mapped near the marker NB-LRR3, which has been shown to co-segregate with the recessively inherited Sr12 [32]. Sr12 has been shown to be ineffective to many races of the Ug99-lineage races at seedling stages [32,34,60]. Rouse et al. [34] observed that the appearance of Sr12 mesothetic phenotype was inconsistent, possibly due to environmental influences. A large degree of variation was observed when mapping QSr.ace-3B, which is likely due to environmental variation and the effects of other QTLs in the genome. Hiebert et al. [32] suggested two resistance genes may be present near the Sr12 locus and the variation observed may support the hypothesis of two resistance genes. At adult-plant stage, Sr12 provides some resistance to Ug99-lineage races in 'Thatcher' , especially in the presence of Sr57 [32,34,61]. 'Thatcher' and the line RL6058, a 'Thatcher' background near isogenic line containing Sr57, were included for evaluation during the 2015 Kenyan adult-plant trial. Both 'Thatcher' and RL6058 were more resistant than the susceptible line LMPG-6 with median scores of 35MSS and 5MS, respectively. RL6058 has a similar level of severity to PI 362698-1, which is thought to have Sr57 based on diagnostic markers csLv34 and wMAS000003. However, PI 362698-1 consistently had a lower disease response than RL6058, ranging from R to MR. As such, QSr.ace-3B is likely Sr12 or an allele of Sr12.
Marker density near the centromere of 7D is low resulting in two linkage groups. The pleiotropic disease resistance gene Sr57 mapped to the centromeric end of the 7D2 linkage group. However, no significant QTLs (α = 0.05) near the Sr57 locus were observed during any of the trials. Sr57 has been shown to be greatly influenced by environmental variation, which can explain why no effects were detected [32,34,47,62]. Sr57 is often considered as an adult-plant resistance gene, however it has been shown to enhance the effects of other stem rust resistance genes at the seedling stage [32,61,63]. To determine if Sr57 was confounding the detection of QTLs for mesothetic resistance, a QTL analysis was performed with the TTKSF+ data using a subset of 68 individuals lacking Sr57. QSr.ace-3B was detected as significant (α = 0.05) and explained 60.8% of the phenotypic variation. The combination of environmental influences on QSr.ace-3B and Sr57, recessive inheritance, and the segregation distortion observed on the long arm of chromosome 3B may explain the abnormal segregation ratios observed at the F 2 generation [30]. PI 362698-1 was not an aneuploid based on a chromosome count (data not shown). PI 362698-1 and the RIL population produced some tillers with sterile heads, indicating a degree of infertility. Additionally, RILs often displayed reduced vigor. Therefore, a gene or set of epistatic genes affecting fertility and linked to the areas of segregation distortion may be present in PI 362698-1.
QSr.ace-2B was detected with QFCQC seedling data and explained 34% of the phenotypic variation. The T. aestivum-derived stem rust resistance genes Sr9, Sr16, Sr28, and SrWLR are present on the long arm of chromosome 2B [14,28,40]. Comparative mapping suggests QSr.ace-2B is likely Sr16, as the QTL maps approximately 30 cM distal of the postulated location of Sr28 on the LMPG-6/PI 362698-1 map [27,28,40]. Both Sr9 and SrWLR map approximately 12 cM proximal to Sr28, suggesting they cannot be QSr.ace-2B [14,28,40]. Moreover, Sr28 provides resistance to the Ug99 lineage races TTKST [27]. As QSr.ace-2B was not detected during either of the adult plant trials in Kenya it is unlikely to be Sr28 and is more likely to be Sr16. Sr16 is thought to be in the background of many wheat accessions and virulence to Sr16 is relatively common [33].
QSr.ace-7A2 was detected with MCCF seedling data and explained 45% of the phenotypic variation. Linkage group 7A2 represents the long arm of chromosome 7A. A non-significant (α = 0.10) QTL in this location was observed with the TTKSF+ data when using the subset population. Two resistance genes, Sr15 and Sr22, and two resistance QTL have been reported on the long arm of chromosome 7A [26,73,74]. Sr22 was introgressed into T. aestivum from T. monococcum L. and provides resistance to Ug99 and TRTTF [8,26,60,73,74]. As QSr.ace-7A2 was not detected with the Kenyan or TRTTF data, it is unlikely to be Sr22 (Table 3). Moreover, QSr.ace-7A2 maps distally to the postulated location of Sr22 on the LMPG-6/PI 362698-1 map [26,42,75]. QSr.ace-7A2 maps to the same location as QSr.abr-7AL [26], the QTL detected by Pujol et al. [74], and Sr15 (E. Babiker, personal communication). The Pujol et al. QTL [74] provided resistance to TTKSK and is thought to either be an allele of Sr15 or a homoeologous allele of a stem rust suppressor thought to be present in 'Thatcher' derived cultivars. As such, QSr.ace-7A2 is likely an allele of Sr15, however we are unable to conclude if the QTL is different than the QSr.abr-7AL [26] or the Pujol et al. QTL [74].
Newcomb et al. [29] assessed PI 362698 with diagnostic markers associated with the stem rust resistance genes Sr2, Sr24, Sr31, Sr36, and Sr57 and modern breeding associated genes Rht-B1b, Rht-D1b, and Ppd-D1a. The genes Rht-B1b and Rht-D1b are associated with reduced plant height and Ppd-D1a is associated with photoperiod insensitivity which has a large effect on flowering time and heading date [29]. PI 362698 was positive for Sr57 and Ppd-D1a, but negative for Rht-B1b and Rht-D1b [29]. Despite its short stature, PI 362698 was thought to be a landrace and not the product of modern breeding because Sr57 and Ppd-D1a are thought to have originated multiple times [29]. 'Thatcher' is postulated to have Sr5, Sr12, and Sr16 [32,34,61]. As QTLs thought to be associated with Sr5, Sr12, and Sr16 were detected in PI 362698-1 it is possible that PI 362698-1 may not be a landrace and may be derived from 'Thatcher' (Table 3). To determine the similarity of 'Thatcher' and PI 362698-1, the alleles for each were compared at the 6866 markers mapped in the LMPG-6/ PI 362698-1 population (Additional file 1: Table S1). 'Thatcher' and PI 362698-1 were found to be 76.2% similar at the mapped loci. As such, it is highly likely that 'Thatcher' was used as a parent multiple times in the pedigree of PI 362698 and PI 362698 is not a landrace as previously reported [29]. Despite the high similarity 'Thatcher' and PI 362698-1 appeared morphologically distinct when grown to maturity in the NDSU green houses.

Conclusions
The identification of stem rust resistance QTLs at genomic locations with known resistance genes makes PI 362698 a useful accession for future breeding efforts. Five significant QTLs (α = 0.05) were detected on chromosomes 2B, 3B, 6A, 6D, and 7A. The QTL on chromosomes 2B, 3B, 6A, and 6D are thought to be Sr16, Sr12, Sr8a, and Sr5, respectively, with Ug99-lineage race resistance being inherited from Sr12. Interestingly, Sr5, Sr12, and Sr16 are present in the cultivar Thatcher which displays broad resistance to stem rust [32,34,61]. In 'Thatcher' , Sr12 was inherited from the durum line Iumillo [32]. The stem rust resistance present in PI 362698-1 and 'Thatcher' is similar and the two were 76.2% similar genetically suggesting PI 362698 is not a landrace and may be a product of modern breeding efforts with 'Thatcher' as an ancestor. The QTL detected on chromosome 7A, QSr.ace-7A2, is possibly Sr15, QSr.abr-7AL, or the QTL detected by Pujol et al. [74]. If QSr.ace-7A2 is the QTL detected by Pujol et al. [74], it could support the hypothesis that 'Thatcher' is an ancestor of PI 362698. Additional work is needed to determine the range of resistance QSr.ace-7A2. Despite Sr57 being present in PI 362698-1, according to diagnostic markers, no QTLs were detected during any of the trials at the Sr57 locus. Crosses have been made with PI 362698-1 to begin introgressing its stem rust resistance into adapted North Dakota wheat germplasm.

Additional file
Additional file 1: Table S1. The LMPG-6/PI 362698-1 linkage map and the positions of significant QTLs detected using seedling and adult data.