Wild barley (Hordeum spontaneum) and landraces (Hordeum vulgare) from Turkey contain an abundance of novel Rhynchosporium commune resistance loci

Rhynchosporium commune is a globally devastating pathogen of barley. Wild and landrace barley are underutilized, however, contain an abundance of loci that can be used as potential sources of resistance. Rhynchosporium commune, the causal agent of the disease scald or leaf blotch of barley, is a hemibiotrophic fungal pathogen of global importance, responsible for yield losses ranging from 30 to 40% on susceptible varieties. To date, over 150 resistance loci have been characterized in barley. However, due to the suspected location of the R. commune host jump in Europe, European germplasm has been the primary source used to screen for R. commune resistance leaving wild (Hordeum spontaneum) and landrace (H. vulgare) barley populations from the center of origin largely underutilized. A diverse population consisting of 94 wild and 188 barley landraces from Turkey were genotyped using PCR-GBS amplicon sequencing and screened with six Turkish R. commune isolates. The isolates were collected from distinct geographic regions of Turkey with two from the Aegean region, two from central Turkey and two from the Fertile Crescent region. The data set was utilized for association mapping analysis with a total of 21 loci identified, of which 12 were novel, indicating that these diverse primary barley gene pools contain an abundance of novel R. commune resistances that could be utilized for resistance breeding.


Introduction
Rhynchosporium commune, the causal agent of the barley (Hordeum vulgare) foliar disease known as scald or leaf blotch, is a hemibiotrophic fungal pathogen (Avrova and Knogge 2012;Zhang et al. 2020). The disease is characterized by pale gray lesions that form distinctive brown margins that often coalesce (Fig. 1). Scald is of global importance, as it is found in the majority of barley production regions, particularly those with cool and humid environments (Zhan et al. 2008). Yield losses due to scald typically range between 30 and 40%, and also diminishes grain quality reducing its acceptance by end users (Brown 1985;Zhan et al. 2008;Zhang et al. 2020). Originally isolated from rye, R. commune was given the Rhynchosporium secalis nomenclature, however Zaffarano et al. (2011) reclassified Rhynchosporium into three species with R. commune infecting Hordeum spp. and Bromus diandrus. However, R. commune has also been reported and isolated from H. murinum subsp. glaucum, Festuca perennis and Avena sativa (Seifollahi et al. 2018). Therefore, wild grass species can serve as a reservoir of potentially virulent isolates and primary hosts to complete its sexual cycle (Linde et al. 2016;Seifollahi et al. 2018), whereas the secondary host remains unidentified. In a recent publication, Crous et al. (2021) proposed renaming R. commune to R. graminicola.
An R. commune study of Turkish isolates found low levels of genetic diversity, supporting the hypothesis that R. commune did not coevolve with barley at its center of origin in Communicated by Andreas Graner.
the Fertile Crescent, but represented a founder population in the region (Çelik ). This hypothesis was also supported by the low genetic diversity of R. commune populations collected from Syria, Jordan and Iran (Kiros-Meles et al. 2011;Seifollahi et al. 2018). However, 14 Icelandic R. commune isolates phenotyped on a differential set of 15 near-isogenic lines all showed unique virulence profiles , implying that the Northwestern Europe R. commune populations contain considerable diversity. This study, along with other studies using the R. commune effector gene NIP1 (Rohe et al. 1995;Brunner et al. 2007), restriction fragment length polymorphisms (Zaffarano et al. 2009) and microsatellites ) implied that barley-R. commune host-parasite interactions co-evolved in Northwestern Europe. This high level of pathogen diversity also exemplifies the need to identify new sources of scald resistance in diverse barley populations. These efforts are important for durable resistance because of the genetic bottlenecks created in modern breeding programs (Clare et al. 2021).
Since R. commune was originally classified as R. secalis, gene designations follow the Rrs nomenclature for Reaction/Resistance to Rhynchosporium secalis (Bjørnstad et al. 2002). A total of eleven major Rrs genes have been mapped under the Rrs1, Rrs2,Rrs3,Rrs4,Rrs12,Rrs13,Rrs14,Rrs15,Rrs16,Rrs17, and Rrs18 gene nomenclatures (Bjørnstad et al. 2002;Zhang et al. 2020). The loci Rrs5, rrs6, rrs7, rrs8, Rrs9, Rrs10 and rrs11 (Chełkowski et al. 2003) were not mapped and may therefore represent alleles of previously mentioned Rrs loci. Therefore, the numbering of Rrs loci is not in strict sequential order due to nomenclature alterations, the unmapped Rrs loci mentioned, and the fact that Rrs3 remains unanchored to the barley reference genome of Morex (Zhang et al. 2020). Major loci have been designated on all chromosomes except for 5H, with Rrs14 on chromosome 1H, Rrs17 on chromosome 2H, Rrs1 and Rrs4 on chromosome 3H, Rrs3 and Rrs16 on chromosome 4H, Rrs13 on chromosome 6H and Rrs2, Rrs12 and Rrs15 on chromosome 7H (Fig. 2) (Bjørnstad et al. 2002;Zhang et al. 2020). Adding to the complexity are the 27 major or minor QTL reported within the Rrs1 locus at the centromeric region of chromosome 3HL that spans from 429 to 503 Mb on version 2 of the Morex genome assembly (Zhang et al. 2020), and two additional loci designated Rrs1 reported within cvs. Sultan and Halcyon at 383.9 and 573.5 Mb, respectively ( Fig. 2) (Genger et al. 2003;Zhang et al. 2020). Rrs1 is therefore a complex locus at a centromeric region (Büttner et al. 2020;Zhang et al. 2020), showing similarities to the Rpt5/Spt1 locus found at the centromeric region of chromosome 6HL conferring dominant resistance and susceptibility to Pyrenophora teres f. teres the causal agent of the disease net form net blotch of barley (Richards et al. 2016;Clare et al. 2020). The R. commune effector NIP1 is known to interact with Rrs1 (Schürch et al. 2004), and species-specific effectors within R. commune are believed to extend the length of its biotrophic phase and shorten its necrotrophic phase (Penselin et al. 2016).
Wild and landrace barley lines are excellent sources of novel disease resistance genes due to their diversity (Russell et al. 2016;Çelik Oğuz et al. 2017Çelik Oğuz et al. , 2019Azamparsa et al. 2019). A recent study used nested association mapping in the HEB-25 wild barley population crossed to the German line 'Barke' identifying eight loci (Büttner et al. 2020). In addition, association mapping within the Scottish 'Bere' landrace barley population, used due to the suspected Northwestern Europe origin of R. commune, also identified eight loci (Cope et al. 2021). The work by Clare et al. (2021) recently demonstrated that despite low marker density, novel resistance/susceptibility loci against fungal pathogens with complex host-pathogen genetic interactions can be identified by association mapping using natural wild and landrace barley populations. The work reported here relied on the previously reported genotyping data of Clare et al. (2021) and identified an abundance of novel resistance loci within the barley-R. commune interaction from wild and landrace barley of Turkish origin. This diversity of scald resistance genes is remarkable considering that the barley accessions are from the barley origin of domestication (Fertile Crescent). The Fertile Crescent is not considered a center of R. commune diversity, therefore some of the resistance may be representative of non-host resistance mechanisms that may be more prevalent in the wild population and may account for the higher level of resistance found in the wild barley population as compared to the domesticated barley population. However, it would be of interest to challenge these new sources of resistance with diverse pathogen isolates to determine the effectiveness of these novel resistances.

Biological materials
The barley populations utilized in this study consisted of 105 Hordeum spontaneum genotypes and 198 barley landraces previously described by Clare et al. (2021). These genotypes were obtained from the Central Research Institute for Field Crops located in Ankara, Turkey. Insufficient seeds were obtained from H. spontaneum genotypes numbered as 4, 15, and 41 and barley landraces numbered as 43 and 116 and were not used, thus a total of 303 genotypes were included in the phenotypic evaluation.

Phenotypic assay
To produce inoculum, each single spore isolate was grown on bean agar medium (140 g fresh bean boiled in 1 L distilled water for 30 min, followed by addition of 18 g agar, and 20 g dextrose) for approximately 14 days before distilled water was added to the Petri dish to allow spores to be released and collected. The harvested spore suspension was cleaned using a cheesecloth and the spore concentration was adjusted to 1 × 10 6 spores mL −1 (Mert and Karakaya 2003). One drop of Tween 20 was added to each 100 mL of inoculum. Five to ten seeds of each barley genotype were planted in plastic pots (7 cm diameter) and placed in the greenhouse. Three replications were inoculated at growth stage 11 (Zadoks et al. 1974) using a small volume (1-1.5 L) hand sprayers until runoff. Following inoculation, plants were transferred to a mist chamber with 100% relative humidity at 16-17 °C for 48 h. Two days post-inoculation, plants were moved to the greenhouse with a 20-25 °C temperature range (Azamparsa et al. 2019 Marker density is plotted as number of markers per 10 Mb window where green is low density and red is high density. LOD score is on the y-axis with thresholds indicated by the solid (α level 0.05) and dashed (α level 0.01) red lines made 18 days post inoculation using the 0-4 scale developed by El-Ahmed (1981). Evaluations were made based on a per pot basis in which with highly resistant (0), no visible symptoms; resistant (1), small lesions on the tips or on the base of the leaf; intermediate (2), one to two small lesions on the blade and/or a narrow band of lesions extending at the margin of the leaf; susceptible (3), well-developed lesions on the blade, but without collapse; highly susceptible (4), leaves collapsed. Scale values between 0 and 2.0 were considered resistant reactions and scale values between 2.1 and 4.0 were considered susceptible reactions. The disease phenotyping data were used in subsequent association mapping analysis.

Barley genotyping, filtering and relatedness
Genotyping of barley lines was performed by Clare et al. (2021). In brief, two independent PCR-GBS SNP marker panels consisting of 365 and 1,272 SNPs were used to genotype lines on the Ion Torrent sequencing platform (Sharma Poudel et al. 2018;Tamang et al. 2019;Ruff et al. 2020;Clare et al. 2021). Redundant markers were eliminated, and absolute marker positions were obtained from the cv. Morex reference genome (Mascher et al. 2017). Markers without absolute positions were estimated based on the iSelect consensus map (Muñoz-Amatriaín et al. 2014). Heterozygous calls were included in the analysis due to the fact that the population was considered natural. Barley genotypes and markers with more than 30% missing data were eliminated, resulting in 94 Hordeum spontaneum genotypes and 188 landraces for a total of 282 genotypes and 530 markers. Subsequently, missing data were imputed using LinkImpute (Money et al. 2015) within TASSEL 5.2.60 (Bradbury et al. 2007) and markers within linkage disequilibrium above an R 2 threshold of 0.8 were removed resulting in 522 markers. Markers with a minor allele frequency above 0.05 were included in the analysis based on GAPIT best practice. Population structure was accounted for with STRU CTU RE analysis (Pritchard et al. 2000) and principal component analysis within GAPIT 3.1.0 (Wang and Zhang 2021), whereas relatedness was accounted for with the efficient mixed model association (EMMA) kinship matrix (Kang et al. 2008).

Association mapping
For association mapping, a naïve model for each isolate was constructed using only genotypic and phenotypic data in a general linear model (GLM) in GAPIT version 3.1.0 (Wang and Zhang 2021) in R 4.2.1. Barley row type was used as a control phenotype. To account for population structure an additional four fixed effect models were generated using STRU CTU RE analysis (Q), and principal component analysis accounting for 25% (PC1), the plateau of the scree plot (PC4, Supplemental Fig. 1) and 50% of the phenotypic variation (PC9) using the GLM method. The kinship matrix was created with the EMMA algorithm (Kang et al. 2008) to use as the random effect component in random and mixed models based on Clare et al. (2021). In addition, up to 35 models using the random effect were generated using MLM (Yu et al. 2006), CMLM (Zhang et al. 2010), ECMLM , SUPER (Wang et al. 2014a), MLMM (Segura et al. 2012), FarmCPU (Liu et al. 2016) and BLINK (Huang et al. 2019) algorithms in conjunction with the EMMA kinship matrix or reconstructed kinship matrices of the respective algorithm. All combinations of models mentioned resulted in a total of 40 models per isolate. To determine the best model, mean-squared deviation (MSD) of each model was calculated (Mamidi et al. 2011) and used to inform the best model based on visual inspection of the quantile-quantile (QQ) plot (Fig. 3

Loci identification
Absolute marker positions for significant marker trait associations (MTAs) were extracted from the version one of the Morex reference genome (Mascher et al. 2017) due to the fact that Rrs loci are currently only anchored in this version of the genome assembly (Zhang et al. 2020). Loci were collapsed together if the nearest neighbor marker was not significant or if the gap to the next marker exceeded 10 Mb in physical distance if no marker was present in a similar strategy to Clare et al. (2021). Loci were deemed novel if no previously reported quantitative trait loci (QTL) were reported within approximately 10 Mb of the significant markers in a similar manner to Clare et al. (2021). This was performed as double recombination has been detected within as low as 10 Mb in barley (Dreissig et al. 2017), and therefore loci are potentially unlinked if double recombination goes undetected.

Phenotypic analysis
In all cases, the data were not normally distributed, and landrace barley was statistically (Wilcoxon rank sum test) more susceptible to the R. commune isolates than wild barley at a significance of less than 0.001 (Fig. 4). On average, approximately 18.4% of landraces were resistant to the six R. commune isolates, whereas 66.1% of wild barley lines were resistant ( Table 2). The mean phenotypic score for all isolates was 3.2 on the landraces and 1.6 on wild barley. The isolate GPS71-U was the most virulent overall with a mean phenotypic score of 3.4 on the barley landraces and 2.6 on wild barley (Table 2).

Association panel and population structure
As previously reported by Clare et al. (2021), a total of 282 barley genotypes were used in the association mapping analysis after removing genotypes that exceeded 30% missing data or not deemed unique using a similarity of individual matrix. In addition, 522 SNP markers were used after removing markers that exceeded 30% missing data or an LD R 2 threshold of 0.8. The population structure of the wild and landrace barley was sourced from Clare et al. (2021). In brief, two subpopulations were identified, with subpopulation one consisting of 89 wild genotypes and subpopulation two consisted of 118 landraces. A total of 75 genotypes consisting of landrace or wild barley were considered admixture due to their population membership probabilities of less than 0.8 (Richards et al. 2017;Clare et al. 2021). Principal component analysis was also used to infer population structure. The first four principal components explained 29.4, 6.4, 3.4 and 2.4% of the genotypic variation, respectively.

Association mapping
Across all models and isolates, a total of 151 MTA were identified, resulting in the identification of 62 unique QTL, of which 36 are suspected of being novel and 26 previously reported and consolidated by Zhang et al. (2020) and more recent publications (Büttner et al. 2020;Cope et al. 2021;Hautsalo et al. 2021). However, when selecting only the best model for each isolate, a total of 31 MTA (28 unique SNP makers) were identified, resulting in 21 unique QTL, of which 12 are novel and nine have been previously reported and consolidated by Zhang et al. (2020). The BLINK algorithm model was selected as the best model for four isolates (GPS71-U, 13GPS149, 13GPS203 and 13GPS207), whereas the FarmCPU algorithm model was selected as the best for the remaining two isolates (E4 and 13GPS109). Using only the best model, a total of one, five, three, three, five, one and two QTL were identified on barley chromosomes 1H, 2H, 3H, 4H, 5H, 6H and 7H, respectively ( Fig. 4; Table 3). There was one QTL identified on the "unassigned" chromosome. The most MTA (nine) were identified on chromosome 5H, which was surprising given that chromosome 5H is typically depleted for QTL in barley-R. commune interactions.
Barley row type was used as a control phenotype with most significant SNP marker (SSM) SCRI_RS_13565 on chromosome 2H at position 655,112,232 corresponding to the known row type vrs1 locus (Supplemental Fig. 2 & 3).

Isolate E4
The best model for isolate E4 was the mixed model using the FarmCPU algorithm and one principal component. Seven unique QTL were identified at the 0.01 significance threshold and an additional four at the 0.05 significance threshold, of which four are hypothesized to be novel (Table 3). Three QTL were identified on chromosome 2H, two QTL on chromosome 3H, two QTL on chromosome 5H and one QTL on chromosome 7H (Fig. 2)

Isolate GPS71-U
The best model for isolate GPS71-U was the random model using the BLINK algorithm and nine principal components. A total of five QTL were identified at the 0.01 significance threshold, with two additional QTL at the 0.05 significance threshold, with a total of three out of the seven being novel QTL. One QTL each was identified on chromosomes 2H, 3H, 4H, three QTL on chromosome 5H and one locus that has not been assigned to a chromosome (Fig. 2)

Isolate 13GPS109
The best model for isolate 13GPS109 was the mixed model using the FarmCPU algorithm and Q population structure. A total of three QTL were identified using isolate 13GPS109 at the 0.01 significance threshold and three additional QTL at the 0.05 significance threshold, with three being novel. One QTL each was identified on chromosomes 1H, 4H, 6H and 7H, and two QTL on chromosome 2H (Fig. 2) (Schweizer et al. 1995) and 0.2 Mb proximal to Rrs12 (Abbott et al. 1991;Genger et al. 2003).

Isolate 13GPS149
The best model for isolate 13GPS149 was the mixed model using the BLINK algorithm and four principal components. A total of two QTL were identified at the 0.01 significance threshold, with one additional QTL at the 0.05 significance threshold, with two of the three being novel QTL. These QTL were identified on chromosomes 2H, 3H and 4H (Fig. 2). The novel loci QRrs-2H.2 and QRrs-4H.2 were identified on chromosomes 2H and 4H at positions 567,016,838 (SSM 12_30724) and 27,327,299 (SSM 12_10860), respectively (Table 3). The previously identified locus QRrs-3H.2 was also identified using the isolate E4 on chromosome 3H

Isolate 13GPS203
The best model for isolate 13GPS203 was the mixed model using the BLINK algorithm and nine principal components. A total of three QTL were identified at the 0.01 significance threshold, of which two were novel and identified on chromosome 5H (Fig. 2). No additional QTL were detected at the 0.05 significance threshold. The novel loci QRrs-5H.2 and QRrs-5H.4, and QRrs-5H.5 were identified on chromosome 5H at positions 528,355,024 (SSM 12_31427), and 599,128,110 (SSM 12_30930), respectively (Table 3). The locus QRrs-5H.5 at position 638,951,179 (SSM 12_30162) was previously identified and embedded within the locus Qsc_5H_1 (Hautsalo et al. 2021).

Isolate 13GPS207
The best model for isolate 13GPS207 was the mixed model using the BLINK algorithm and four principal components. One QTL was identified at the 0.01 significance threshold and one additional QTL at the 0.05 significance threshold, with no novel loci. The first previously identified locus QRrs-5H.5 was identified on chromosome 5H at position 649,232,960 (SSM 11_10600) and embedded within the Qsc_5H_1 locus (Hautsalo et al. 2021). The previously identified locus QRrs-7H.1 was identified on chromosome 7H at position 22,774,581 (SSM 11_20495) approximately 6.7 Mb distal of Rrs12 (Abbott et al. 1991;Genger et al. 2003).

Enrichment analysis
The percentage of resistant alleles for all SSMs was calculated for wild and landrace barley (Fig. 5)

Discussion
Barley scald is regarded as one of the most economically important diseases of barley worldwide (Zhang et al. 2020). Yet, important information pertaining to hostpathogen interactions remain elusive in the pathosystem such as the identity of the sexual host/s (Penselin et al. 2016;Holtz 2021). Over 150 resistance loci have been described in barley against R. commune (Büttner et al. 2020;Zhang et al. 2020;Cope et al. 2021;Hautsalo et al. 2021), and an additional 12 novel loci were identified in this study out of a total of 21 significant loci detected. Of the major known scald resistance loci, only Rrs1, Rrs2, Rrs12, and Rrs16 were identified in our study indicating that these Turkish populations are a rich resource of diverse untapped scald resistance genes. In addition, our study identified five unique loci (three novel) on chromosome 5H, where there are no formally designated genes and only eight of over 150 previously described loci have been located (Looseley et al. 2012;Coulter et al. 2019;Zantinge et al. 2019;Daba et al. 2019;Büttner et al. 2020;Zhang et al. 2020;Cope et al. 2021;Hautsalo et al. 2021), increasing the number of loci described on chromosome 5H by over 50%. Interestingly, using the isolate 13GPS203, three QTL identified exclusively on barley chromosome 5H were found. Loci of particular interest include the locus QRrs-2H.2, identified using isolates E4 and 13GPS149 with marker 12_30724. However, the variation in virulence for isolate E4 also showed a significant marker trait association with SNP 11_21166 approximately 12 Mb from 12_30724, with no marker in between, suggesting that this locus is complex, with different haplotypes and possibly distinct genes providing resistance or susceptibility to the specific isolates. Other loci identified across multiple isolates include QRrs-2H.4 with isolate E4 and 13GPS109, QRrs-3H.2 with isolate E4 and 13GPS149, QRrs-5H.3 with isolate E4 and GPS71-U, and QRrs-7H.1 with isolates 13GPS109 and 13GPS207 (Fig. 1, Table 3). The QRrs-5H.5 locus, which was only previously reported once as Qsc_5H_1 (Hautsalo et al. 2021), was identified for all isolates except 13GPS109 and 13GPS149, suggesting this locus represents broader-spectrum resistance to the Turkish R. commune population. The locus Qsc_5H_1 was identified in two MAGIC populations with European founder lines possessing resistance alleles consisting of Iron, RGT Planet, SJ111998, Ylitornion, MBR-1012, Chevron, Fairytale, Middle Eastern JLB06-034 and the American founder line Nordic (Novakazi et al. 2020;Hautsalo et al. 2021  QRrs-5H.5 locus, SSMs SCRI_RS_155322 (landrace 96% vs. wild 62%), 11_10600 (95% vs. 83%), 12_30566 (90% vs. 46%) and 12_30162 (94% vs. 80%) were all more prevalent in barley landraces, despite in general landraces being more susceptible than wild accessions. We therefore hypothesize that a selective sweep at the QRrs-5H.5 locus occurred during domestication to account for the enrichment of all four markers when compared to the wild accessions. However, considering wild accessions are overall more resistant than landraces, additional sources of resistance must be present within wild accessions, and could supplement the QRrs-5H.5 resistance in barley breeding programs.
The locus QRrs-7H.1 is positioned between loci identified using isolates L2A and 13-13, which were postulated to be Rrs12 (Abbott et al. 1991) and Rrs2 (Hanemann et al. 2009). The isolates 13GPS109 and 13GPS207, which were used to identify QRrs-7H.1, are virulent on two and three of the three differential lines that contain Rrs2 (Rrs12 is not suspected to be in the differential set), respectively (Table 1). Therefore QRrs-7H.1 may represent Rrs12 or a novel Rrs2 allele. The locus QRrs-4H.1 may colocalize with a locus reported against isolate L2A, and correlates with Rrs16 (Pickering et al. 2006), however no marker position was provided (Cope et al. 2021). The remaining loci identified in the Scottish Bere landrace association mapping (Cope et al. 2021) and potentially all but one locus from HEB-25 wild barley population (Büttner et al. 2020) show no overlap with our study showing that geographically distinct barley populations and their interactions with local isolates harbor different loci involved in R. commune interactions.
The fact that the wild barley population was more resistant to all six R. commune isolates compared to the landrace population suggests that this wild population harbors more effective alleles or a greater diversity of resistance genes (Fig. 4). Although there are several possible explanations for these findings it is probable that this study may not have had the diversity within the isolates utilized to identify individuals with virulence on the wild barley genotypes. Thus, it would be prudent to screen these populations with a greater diversity of isolates to identify additional resistance sources and identify those with broad resistance that could be more valuable to breeding programs.
Interestingly, three loci colocalize with genes associated with different post-domestication barley market classes. The locus QRrs-2H.4 is approximately 3.1 Mb distal to vrs1, and the locus QRrs-5H.3 is approximately 4 Mb distal from vrs2. Both vrs1 and vrs2 control spikelet morphology and the switch from wildtype two-row barley to the mutant sixrow phenotype (Zwirek et al. 2019). The majority of six-row barley cultivars are controlled by the presence of natural recessive vrs1 alleles and accompanied by a natural vrs5 allele that improves lateral spikelet grain fill (Zwirek et al. 2019). Therefore, the fact that vrs1 is the most significant marker in our row type association and the vrs2 locus was not identified (Supplemental Fig. 2 & 3) is not surprising given that vrs2, vrs3 and vrs4 are not prevalent in six-row barley (Zwirek et al. 2019). The vrs1 and QRrs-2H.4 loci were identified with the same SSM SCRI_RS_13565, with six-row barley carrying the allele that increases resistance and thus is more prevalent in the landrace barley lines. This is surprising given that in general, wild barley is more resistant than landrace barley, however this provides the opportunity to break this linkage and stack resistance genes from both landrace and wild barley into breeding material. The QRrs-2H.4 locus was previously reported in three previous studies, one of which noted the close proximity to vrs1 (Li and Zhou 2011;Looseley et al. 2018;Fériani et al. 2020). The last locus QRrs-5H.4 colocalizes with vrn-H1, which is responsible for the spring and winter growth habit, with 99% of wild lines having the allele, whereas 49% of landrace barley having the same allele. The fact that 99% of wild barley have the same allele for QRrs-5H.4 may be due to the linkage drag from the close proximity to vrn-H1 and the fact that all wild barley are of winter growth habit (Fernández-Calleja et al. 2021).
Barley scald remains a global issue, being one of the most destructive diseases that hampers barley production (Zhan et al. 2008). Compared to other barley diseases such as net blotch or leaf rust, relatively few resistance loci have been identified that can be used in breeding programs (Dracatos et al. 2019;Clare et al. 2020;Zhang et al. 2020). Worldwide, barley is predominantly grown as a feed crop (IBGSC 2012), however, barley is the precursor to malt, a vital component within the brewing and distilling industry that cannot be replaced by other cereal grains (Olaniran et al. 2017). In addition, barley is gaining traction as a food crop because of its high nutritional value and heart healthy attributes (Zeng et al. 2020). In the USA, barley is primarily grown for malt, as malt commands a premium within the multi-billion dollar brewing and distilling industry. However, within the USA, barley has seen significant acreage drops and has been forced off premium agricultural land by more profitable crops such as corn and soybeans . Along with the impacts of climate change, there could be worldwide malt shortages in the near future that are beginning to be witnessed with stagnating yields seen in southern Europe (Dawson et al. 2015;Xie et al. 2018). Therefore, identification of sources of resistance that can be used to maintain yield increases and minimize inputs to barley production against important pathogens such as R. commune are paramount. Here we report the identification of 21 distinct loci encompassing 12 novel loci from wild and landrace barley from the origin of domestication, rather than the initial location of barley-R. commune coevolution, that could be introgressed into elite barley cultivars.