The BGEM population is a BC₁F₁ exotic-derived DH population. The population was developed at Iowa State University in collaboration with the USDA-ARS Germplasm Enhancement of Maize project (Pollak and Salhuana, 2001). Briefly, BC₁F₁ lines were created by crossing exotic accessions to one of two expired Plant Variety Protection (PVP) lines, namely PHZ51 or PHB47, backcrossing the resulting F₁ accessions were backcrossed to their respective ex-PVP parent, and then DH lines were created and self-pollinated (Brenner et al., 2012). Any lines that were infertile, lacked uniformity, or with poor agronomic traits were discarded. Genetically, the BGEM population is comprised of roughly 25% donor parent and 75% recurrent parent (Sanchez et al., 2018). The BGEM population represents a broad diversity of exotic derived maize. In total 67 landraces from 13 different countries are represented (Sanchez et al., 2018).

The BGEM population was screened for resistance to E. turcicum in three environments from 2019 to 2021. In total, all lines were evaluated four times in the field between 2019 and 2021 and twice in the greenhouse in 2020. All experiments were designed as augmented randomized incomplete block designs using the agricolae package (de Mendiburu, 2021) in the statistical software R, version 3.6.0 (R Core Team, 2021). Lines from BGEM population were screened at the Crop Sciences Research and Education Centers in Urbana, IL in 2019 (n = 252) and 2021 (n = 240). Differences in the number of lines used in 2019 and the other years were due to seed availability. Field plots were planted in rows that were 3.65 m long with 0.91 m alleys. The 2019 field was irrigated to encourage disease development. Lines were replicated twice in 2019 and 2021. In 2019, incomplete blocks (n = 14) were augmented with Oh7B and NC344 as susceptible and resistant checks, respectively. In 2021, Oh7B, NC344, PHB47, and PHZ51 served as check lines in each incomplete block (n = 16). In 2021 the following NCLB differential lines were included in each replication: A619-Ht1, A619-Ht2, A619-Ht3, A619, B37-HtN1, and B37.

In 2020, BGEM lines (n=240) were screened at the Plant Care Facility in Urbana, IL. The 2020 greenhouse experiment was replicated twice with 13 incomplete blocks in each replication. Due to space constraints, replications had to be run separately. One plant per line was grown in a one-gallon pot filled with 1:1:1 general purpose potting mix. Greenhouse conditions were set to 12/12 hr. light-dark cycle. Ambient temperatures were set to 24-28°C during the day and 20-22°C during the night. Lines were organized in an augmented randomized incomplete block design. Oh7B and NC344 were susceptible and resistant check lines in each block. PHB47 and PHZ51 were randomized once in each replication. The same qualitative check lines used in 2021 were also used in the greenhouse in each replication.

In 2019 natural inoculum was relied upon for infection, as the growing season was conducive for disease development and evaluation. For all other experiments, inoculum was prepared using five fungal isolates (19StM06, 19StM07, 19StM09, 19StM11, 19StM12) that were isolated from diseased maize tissue collected from the 2019 field experiment at Crop Science Research and Education Center in Urbana, IL. A mixture of isolates was used to best reproduce disease pressure observed in 2019. For the 2020 greenhouse experiment, plants were inoculated with a spore suspension at the V3 stage (Adhikari et al., 2021). Fungal isolates were grown on lactose-casein hydrolysate agar (LCA) media for 14 days under 12/12 hr light/dark photoperiod at room temperature. Spores were harvested from 14-day old plates and the concentration adjusted to 4 × 10³ spores/mL in a 0.02% Tween 20 solution. For inoculations, 0.5 mL of the spore suspension was pipetted into the whorl of the plants. Following inoculation, plants were maintained in high humidity conditions for 24 hours to facilitate disease development. In 2021 plants were inoculated at the V4 to V5 stage using infested sorghum grains (Zhang et al., 2020). Fungal isolates were cultured on LCA for 14 days as described above. After 14 days of culturing, agar pieces cut directly from the plates were added to mushroom bags containing 1000 mL of soaked, autoclaved sorghum grains. The mushroom bags were cultured for two to three weeks at room temperature under 12/12 hr light/dark photoperiod. The infested grains were examined under a dissecting microscope to confirm the presence of conidia then dried and stored at room temperature. Each plant was inoculated with 1.3 g of the prepared inoculum in the whorl at the V3 growth stage.

Disease ratings for the 2019 and 2021 field experiments were taken on a whole plot basis using a 0-100% scale in 5% increments (Poland and Nelson, 2011) where no disease present was represented by 0%, and a rating of 100% indicated that the total leaf area of the plants was necrotic due to disease. In 2019 two field ratings were taken after lines had started flowering, and ratings were taken 7 days apart. In 2021 a total of four field ratings were taken. Ratings were conducted every 7-14 days starting two weeks before the onset of flowering in the earliest maturing plants.

Lines in the 2020 greenhouse experiment were evaluated for incubation period and disease severity. Incubation period is defined as the number of days after inoculation (DAI) that the first lesions were visible. Lines were evaluated for incubation period every 48 hours until all plants showed lesions. In the greenhouse experiment, a total of four diseased leaf area ratings were taken. Disease leaf area ratings were taken every 7-14 days starting the day when all plants had lesions present. Lines were rated on a single leaf per plant basis using the same 0-100% rating scale mentioned above.

Disease severity data was examined using the standardized area under the disease progress curve (sAUDPC). sAUDPC is used to compare disease progression between environments or experiments (Simko and Piepho, 2012). sAUDPC is calculated by first measuring the area under the disease progress curve (AUDPC), which is the area of a trapezoid between two or more time points on a progression curve and accounts for disease severity over time (Jeger and Viljanen-Rollinson, 2001; Madden et al., 2007). The sAUDPC value is then obtained by dividing AUDPC by the weighted total of the number of days of disease evaluation. sAUDPC was calculated using the R package agricolae (de Mendiburu, 2021). Both days to anthesis (DTA) and days to silking (DTS) were recorded for the 2019 field experiment. DTA is defined as the number of days after planting when 50% or more of the plot had visible anthers on the tassel. DTS is defined as the number of days after planting when 50% or more of the plot had visible silks emerging from the ears. Pearson correlation coefficients for sAUDPC across all environments were calculated using the rcorr() function in the R package Hmisc v4.5-0 (Harrell, 2021). Flowering data was only collected in 2019. Using AUDPC, DTS, and DTA data from 2019, Pearson correlation coefficients were calculated between each trait.

Both sAUDPC and incubation period values were used to fit mixed models and estimate least squared means (LS Means). The following model was fit to estimate LS Means for sAUDPC in the 2019 + 2021 dataset:

The factors are defined as follows: Y_ijkl is the sAUDPC value from genotype i in replicate j in block k and environment l; G_i is the fixed effect of genotype I; R_j is the random effect of replicate j; B_k(j) is the random effect of block k nested in replicate j nested in environment l; E_l the random effect of the environment l; GE_il the random effect of the interaction between genotype i and environment l; and ε_ijkl the error term. From this model, LS Means were calculated from the estimates of G_i from the fitted model. For the 2020 incubation period dataset the following mixed model was fit:

Terms in the model are described as follows Y_ij is the incubation period value in days of genotype i in block j; G_i is the fixed effect of genotype i; B_j is the random effect of block j; and ε_ij the error term associated with Y_ij. The fixed effects for genotype in each model were extracted and used for further analysis. As with the previous model, LS Means were calculated from the estimates of G_i from the fitted model.

The BGEM population was previously genotyped (Sanchez et al., 2018). A dataset consisting of 62,077 single nucleotide polymorphisms (SNPs) was obtained from Sanchez, Liu (Sanchez et al., 2018). This dataset was generated for the DH BGEM lines via genotyping-by-sequencing (Elshire et al., 2011) by the Cornell University Genomic Diversity Facility with data analysis by the Buckler Lab for Maize Genetics and Diversity. The 62,077 SNP dataset was generated by filtering SNPs with a large amounts of missing data and low allele frequencies, followed by SNPs in the same genetic position (Sanchez et al., 2018). Within the BGEM population, the average number of recombination events was higher than expected. A Bayes theorem described by Sanchez, Liu (Sanchez et al., 2018) was implemented to correct for monomorphic markers that were flanked by markers with donor parent genotypes. After correction, the number of recombination events was reduced, and the donor genome composition was closer to 25%, as expected. All edits to the genotypic dataset were conducted prior to this study.

To identify candidate regions associated with resistance to E. turcicum two mapping strategies were employed using TASSEL v5.2.81 (Bradbury et al., 2007). First, we used a generalized linear model (GLM) in TASSEL to identify significant markers, as well as a stepwise regression approach to identify markers that were associated with disease severity and incubation period. The Bayesian corrected genotypic dataset was filtered to remove markers with a minor allele frequency less than 0.05. Two principal components, calculated in TASSEL using the standard PCA plugin, were used to control for population structure. The phenotypic datasets used were the LS Means estimated using the 2019 + 2021 combined disease severity and the 2020 incubation period datasets ( File S1 ). The same genotypic dataset was used for both phenotypes. The GLM model was ran for both phenotypes with 1000 permutations. A significance threshold of α = 0.10 was applied using permutation corrected p-values. For the second approach, stepwise regression was implemented in TASSEL independently of the GLM association analysis. The two principal components calculated in TASSEL were included as numeric covariates along with the estimated LS Means value for each line. The stepwise regression analysis was implemented for both the 2019 + 2021 combined disease severity and the 2020 incubation period datasets. The p-value model was used with 1000 permutations with entry and exit limits of 1 × 10^-5 and 2 × 10^-5, respectively. A significance threshold of α = 0.10 was used to identify significant SNPs.

The Pathway Association Study Tool (PAST) approach (Thrash et al., 2020) was implemented via the MaizeGDB online platform (Woodhouse et al., 2021). PAST is designed to take outputs generated by TASSEL and provide biological insight into association study results. The software interprets the results from association analysis to identify metabolic pathways associated with genes that are strongly associated with the trait through significant markers or when genes are moderately associated with a trait, but the markers themselves may not have been significant in the original association study. PAST uses the output from association analysis, allelic effects files, and a linkage disequilibrium file to assign SNPs to genes based on LD and genomic distance between SNPs and genes before then identifying significant metabolic pathways. PAST analysis was conducted for the 2019 + 2021 combined disease severity dataset and the 2020 incubation period dataset. The association and effect files generated from the GLM were used as the input for the PAST analysis. A gene assignment window size of 1000 base pairs was used. Pathways with at least five genes were considered and significance was determined based on 1000 permutations. A Type I error rate of α = 0.05 was used to determine statistical significance for each pathway.

A recent study examined differences in gene expression in maize and sorghum plants inoculated with different strains of E. turicum (Adhikari et al., 2021) ( File S2 ). Our hypothesis was that genes involved in the synthesis of ethylene were differentially expressed in mock-inoculated maize plants versus E. turcicum inoculated maize plants. A list of genes associated with ethylene biosynthesis was downloaded from CornCyc7.5 available through the MaizeGDB (corncyc-b73-v3.maizegdb.org) ( File S3 ).

Transcriptome data were generated by Adhikari et al. (2021). Briefly the maize line B73 was grown in the greenhouse and either inoculated with E. turcicum or mock inoculated with sterile deionized H₂O (DI H₂O). RNA was extracted from plant tissue collected at 24 and 72 hours after inoculation (hai) and sequenced at the Roy J. Carver Biotechnology Center at the University of Illinois at Urbana-Champaign. Quality control, alignment, and normalization were conducted using standard procedures (Adhikari et al., 2021). We used the final dataset with annotated differentially expressed genes (DEGs) and calculated false discovery rates (FDR) (Adhikari et al., 2021). ( File S2 ). Global FDR adjusted p-values were calculated based on the Benjamini & Hochberg procedure (Benjamini and Hochberg, 1995).

Our objective with the RNA data was to examine the expression of genes associated with ethylene biosynthesis in maize during the E. turcicum infection stage. Therefore, we are interested in three contrasts: DEG of maize 24 and 72 hai with E. turcicum, DEG of maize 24 hai with E. turcicum and DI H₂O, and DEG of maize 72 hai with E. turcicum and DI H₂O. Using a list of known genes involved in ethylene biosynthesis ( File S3 ), we filtered the normalized list of DEGs for those only involved in ethylene biosynthesis. Then, for each contrast we only examined DEGs with an FDR value greater than or equal to 0.05. We hypothesized that ethylene biosynthesis could be involved in maize early defense against E. turcicum.

A subset of DH lines was chosen to evaluate the role of ethylene in resistance against E. turcicum. In maize, ethylene is inhibited by the foliar application of AgNO₃ (Kumar et al., 2009). Using LS Means data from the 2021 field experiment, we examined disease phenotypes within exotic donor families. Only the ex-PVP recurrent parents have been genotyped, and there is no genotypic information for the landrace donor parents. We chose a resistant and susceptible line, as we wanted to see if inhibiting ethylene affected phenotypes of resistant or susceptible lines differently ( Table 1 ). We selected pairs of lines categorized as resistant or susceptible within five exotic landrace groups. Both ex-PVP recurrent parents, PHB47 and PHZ51, were included in the assay. Our goal was to assess the role of ethylene between resistant and susceptible lines, within different exotic families, and within different recurrent parent backgrounds.

The experiment was arranged in a split-plot design. The whole-plot was a complete block of plants, and the subplot was each BGEM line randomly assigned to each treatment within each complete block. Lines were either treated with AgNO₃ followed by inoculation with E. turcicum, or treated with DI H₂O followed by inoculation with E. turcicum. There were three whole-plots (complete blocks) per treatment and each line occurred once in each whole-plot. Greenhouse growing conditions were the same as the conditions described previously. The E. turcicum spore suspension was prepared as explained above. At the V3 growth stage, plants were treated with either 20 mM AgNO₃ in 0.001% Tween 20 (Wang et al., 2017) or with sterile water in 0.001% Tween 20 at a rate of 187 L ha^-1 at 207 kPa using a spray chamber (Technical Machinery Inc., Sacramento, CA). The chamber was equipped with an even flat-fan nozzle 8002E (TeeJet Technologies, Wheaton, IL) and plants were sprayed 45 cm above the tallest leaf. After treatment, plants were allowed to dry completely before inoculation with 4 × 10³ spores/mL in a 0.02% Tween 20 solution prepared as mentioned previously. Plants were inoculated by pipetting 0.5 mL of spore suspension into the whorl. After inoculation, a high-humidity environment was maintained overnight to encourage disease development. All plants were evaluated for incubation period and disease severity on a per plant basis as described previously. AUDPC was calculated using the agricolae package (de Mendiburu, 2021) in the statistical software R, version 3.6.0 (R Core Team, 2021). We ran an ANOVA to appropriately account for the differences in the whole-plots and the subplot replicates.

The BGEM population was evaluated for DTA and DTS in 2019 and for resistance to E. turcicum in 2019 and 2021 in the field at the Crop Sciences Research and Education Center in Urbana, IL. A field trial of the BGEM population was lost during the 2020 growing season due to poor emergence, followed by a hailstorm, which destroyed the field. Thus, the BGEM population was grown a second time and evaluated for resistance in the greenhouse in 2020 at the Plant Care Facility in Urbana, IL. All checklines performed as expected. In the 2019 and 2021 field environments, Oh7B was more susceptible than NC344. The BGEM recurrent parents, PHB47 and PHZ51, were moderately resistant in the 2021 field season with PHZ51 being the more resistant recurrent parent. The observed range of phenotypes suggest that there was sufficient disease pressure in each experiment to evaluate for disease severity and incubation period, even in the uninoculated field trial ( Table 2 ).

Due to the high evolutionary potential of E. turcicum, we examined whether isolates collected from Champaign County could overcome major genes associated with resistance. In 2020 and 2021 A619-Ht1, A619-Ht2, A619-Ht3, A619, B37-HtN1, and B37 were included to evaluate whether the mixture of 19StM06, 19StM07, 19StM09, 19StM11, 19StM12 were able to overcome major genes associated with resistance. In the 2020 greenhouse experiment we did not observe a significant difference in AUDPC or incubation period among lines containing major genes and their respective background based on a Dunnett’s test (α = 0.05), but we observed typical resistant responses for Ht2, Ht3, and HtN1. In the 2021 field experiment the AUDPC for A619-Ht2 and A619-Ht3 were significantly different from A619 based on a Dunnett’s test (p< 0.001), indicating that Ht2 and Ht3 conferred resistance. Additionally, B37-HtN1 was significantly different from B37 (α = 0.01), indicating that HtN conferred resistance. Discrepancies between genotype performance in the greenhouse and the field environment could be due to other E. turcicum genotypes present in the field environment. The effectiveness of qualitative and quantitative under greenhouse conditions differs from the field environment (Raymundo and Hooker, 1981b; Chung et al., 2010). Thus, it is possible that the observed discrepancy in phenotypes between the field and greenhouse environments could be attributed to environmental effects on the plants, as well as additional strains present in the field environments.

We observed a difference in resistance responses between the greenhouse and field environments. The 2019 and 2021 field environments were strongly and significantly correlated, (r= 0.64, α = 0.001) while the 2020 greenhouse environment had a weaker correlation with the 2019 field environment (r = -0.21, α = 0.01) and the 2021 field environment (r = -0.29, α = 0.01, Table 3 ). The strong correlation between the 2019 and 2021 field data underscores that similar levels of resistance were observed even though one field was artificially inoculated while the other relied on natural inoculum. In the mixed model for the 2019 + 2021 combined dataset, environment contributed the most variance, followed by the genotype × environment interaction ( Table 4 ). The model used to estimate LS Means in the 2020 incubation period dataset only included block but not replication. Replication was not included as it had a covariance estimate of zero, indicating that block and replication were redundant. By including the blocking factor in the model, we are still able to account for variation in the greenhouse environment. The 2020 greenhouse diseased leaf area was analyzed in the same manner as the 2020 greenhouse incubation period. The estimated LS Means for the 2019 + 2021 combined dataset and the 2020 incubation period dataset showed a range of phenotypic responses suitable for association mapping ( Figure 1 , Table 1 ).

In order to identify candidate markers associated with resistance to NCLB, we employed genome-wide association analysis using TASSEL (Bradbury et al., 2007). To control for population structure, two principal components were included in the model, as most of the variation is explained by the first two components ( Figure 2 ). We then employed a GLM model using TASSEL to examine the genetic architecture of resistance to NCLB. For both the 2019 + 2021 field dataset and the 2020 greenhouse incubation period dataset log quantile-quantile plots were constructed to assess the effectiveness of two principal components to control population structure ( Figures 3B, D ). Given the genetic design of the BGEM population, this approach is appropriate to control for population structure while controlling for any false positives. The same mapping approach was used for the 2019 + 2021 field dataset, the 2020 greenhouse incubation period dataset, and the 2020 greenhouse diseased leaf area dataset. We did not identify any significant markers using the 2020 greenhouse diseased leaf area dataset. In total, four significant markers were identified using the GLM model in TASSEL ( Figure 3 ).

The BGEM-DH lines have a high percentage of recurrent parent genome, and as such, linkage disequilibrium (LD) decays at a much slower rate than expected. In some regions, the LD blocks are large and span 100,000,000 bp or beyond (Sanchez et al., 2018). Given the LD structure in this population, we considered genes within 1 Mbp of a significant marker when examining candidate genes. Three markers on chromosome 3 were associated with delayed incubation period ( Table 5 , Figure 3A ). A stepwise regression approach using the same marker dataset identified a single significant marker for incubation period on chromosome three (S3_8266879; p = 5 × 10^-5), which was also significant in the GLM model (p = 0.033, Table 5 , Figure 3A ). The marker S3_8266879 is within GRMZ2G108898, plant cysteine oxidase 1 (PCO1). PCO1 spans from 8,263,929 bp to 8,269,675 bp on chromosome 3. The other two significant markers identified in the GLM model for incubation period, S3_15712704 (p = 0.53) and S3_18008517 (p = 0.55) were proximal to GRMZM2G445261 a probable carboxylesterase 15 and GRMZM2G380195 a pentatricopeptide repeat-containing protein chloroplastic, respectively. While these genes were the closest to the most significant marker, it is important to keep in mind the extensive LD structure in this population.

Resistance to diseased leaf area (DLA) is genetically distinct from incubation period; a single marker was significant on chromosome 2 ( Figure 3C ). The significant marker, S2_10777410 (p = 0.053), is proximal to GRMZM2G068982, a methionine aminopeptidase. Using the same marker dataset, a stepwise regression approach did not reveal any significant markers for the field DLA dataset ( Table 5 ). Stepwise regression is more conservative than GLM, and our findings are consistent in that no significant markers were identified using stepwise regression.

There was no overlap in significant markers for incubation period and disease severity. We observed a weak negative correlation between disease severity and incubation period ( Table 3 ). Although we observed some chlorotic resistant responses in our 2019 and 2021 field experiments and 2020 greenhouse experiment, we did not identify any significant markers near any of the major genes for NCLB resistance, Ht1, Ht2, Ht3, HtN1 ( Figure S1 ). If the qualitative resistance genes were present in the population, they may have been at too low of a frequency to detect any significant markers in these regions. It is also possible that these responses were due to novel resistance genes.

To better understand the molecular mechanisms associated with resistance to NCLB we conducted a metabolic pathway analysis and examined a previously published RNA-seq dataset. We identified metabolic pathways associated with disease severity and incubation period using the results from the GLM model and the PAST software (Thrash et al., 2020). In total, 24 metabolic pathways (α = 0.05) were associated with resistance to E. turcicum ( Table 6 ), including nine associated with incubation period and 15 associated with disease severity. Several pathways associated with plant hormones were significant including jasmonic acid biosynthesis (PWY-735, p = 0.02) and cytokine-O-glucosides biosynthesis (PWY-2902, p = 0.04) in the combined disease severity dataset. Additionally, gibberellin inactivation I (PWY-102, p = 0.01) and brassinosteroids inactivation (PWY-6546, p = 0.03) were significant in the incubation period dataset. A pathway of particular interest is the ethylene biosynthesis pathway (ETHYL-PWY) which was significant for both incubation period (p = 0.006) and disease severity (p = 0.01). Ethylene synthesis has previously been associated with disease resistance in plants (Dong, 1987; van Loon et al., 2006). Additionally, S-adenosyl-L-methionine cycle II (PWY-5041) was significant for disease severity. S-adenosyl-L-methionine (SAM) is the starting substrate in ethylene synthesis. It is converted to 1-aminocyclopropane-1-carboxylic acid (ACC) by ACC synthase (ACS), which is subsequently converted into ethylene via oxidation by ACC oxidase (ACO) (Xu and Zhang, 2015).

Pathogens are known to mimic phytohormones and interfere with signaling (Denance et al., 2013), as well as to target regulatory components involved with in planta hormone production (Robert-Seilaniantz et al., 2011). Thus, we examined an RNA-seq dataset (Adhikari et al., 2021) for DEGs associated with ethylene biosynthesis during the early stages of infection. We examined three contrasts: inoculated versus uninoculated maize 24 hai, inoculated versus uninoculated maize 72 hai, and inoculated maize 24 versus 72 hai. Only one gene associated with ethylene biosynthesis, GRMZM2G07529 (FDR = 0.03), was upregulated, with a fold change of 16.87, in maize 24 hai with E. turcicum compared to maize 24 hai with DI H₂O. GRMZM2G07529 is annotated as ACC oxidase ACO31. ACO is the final oxidation step in the synthesis of ethylene (Xu and Zhang, 2015). Thus, during the initial phases of infection ethylene production may be increased, indicating that it may play a role in resistance to NCLB.

We hypothesized that inhibiting ethylene would alter NCLB disease severity. We used a foliar treatment of AgNO₃ to examine the effect of ethylene on disease resistance. AgNO₃ inhibits ethylene action in planta by reducing the ability of ethylene receptors to bind to ethylene, which results in decreased ethylene sensitivity and the inhibition of continuous ethylene production (Kumar et al., 2009). We hypothesized that inhibition of ethylene by applying AgNO₃ would decrease disease severity and increase incubation period. Furthermore, we postulated that host resistant level and genetic background might alter the magnitude by which inhibiting ethylene improves resistance and increases incubation period.

To examine the role of ethylene in E. turcicum infection, a total of 10 DH lines and both recurrent parents were selected. We selected DH lines in pairs within each exotic family, where the pair had contrasting disease phenotypes. Plants were either treated with AgNO₃ or DI H₂O prior to a whorl inoculation with E. turcicum. From our ethylene inhibition assay, disease severity was significantly impacted by genotype ( Table 7 , p = 0.012) and by treatment with AgNO₃ or deionized H₂O ( Table 7 , p = 1.756 × 10^-5). Similarly, incubation period was significantly impacted by genotype ( Table 8 , p = 0.004) and by treatment with AgNO₃ or deionized H₂O Table 8 , p = 4.281 × 10^-5). In plants treated with AgNO₃, incubation period increased and AUDPC decreased ( Figure 4 ). We did not observe a significant impact of the line’s resistance level, exotic family, or recurrent parent on AUDPC or incubation period (α = 0.05). In this experiment, application of AgNO₃ affected all genotypes similarly meaning that, of the lines screened, no line was more or less resistant relative to the other lines after AgNO₃ application; thus, inhibiting ethylene using AgNO₃ decreases disease severity similarly in the lines screened.