The Snap Bean Association Panel (SnAP) is comprised of 378 historic-to-modern cultivars that span both fresh market and processing types, including large sieve, small sieve, wax, bush blue lake, Kentucky, Romano, pole, and others. The SnAP includes public and private industry-released snap bean cultivars predominantly from the United States, some landrace cultivars, and a few breeding lines. The SnAP has recently been used for Genome-Wide Association Studies (GWAS) of herbicide tolerance (Saballos et al., 2022) and resistance to white mold (Arkwazee et al., 2022). The SnAP expands on the 149 snap bean accessions for the Snap Bean Diversity Panel (SBDP) that composed part of the BeanCAP (Common Bean Coordinated Agricultural Project) panel, which was comprised of both dry and snap beans (http://arsftfbean.uprm.edu/beancap/). The SBDP was used in a diversity analysis of snap bean (Wallace et al., 2018) and for GWAS of total phenolic content and root rot resistance (Myers et al., 2019; Huster et al., 2021). In addition to the SnAP, a dry bean population (DBP) comprised of 88 advanced inbred breeding lines and cultivars representing different market classes was included from the USDA-ARS bean breeding program in Prosser, WA.

Disease resistance or susceptibility reactions to CTV were obtained for 83 of the SnAP accessions from published information, including from the Agricultural Marketing Service-USDA scanned Plant Variety Protection (PVP) database, cultivar registration articles, and the Germplasm Resources Information Network (https://npgsweb.ars-grin.gov/gringlobal/search). An additional 289 SnAP accessions were phenotyped for CTV disease reaction using an infectious CA/Logan clone of CTV and an agro-inoculation screening procedure in a controlled growth room environment at Harris Moran Seed Company in Sun Prairie, Wisconsin. The CTV CA/Logan strain was cloned into the binary vector pBIN19 in Agrobacterium tumefaciens strain GV3111 (Ti plasmid pTiB6S3SE) obtained originally from Dr. Drake Stenger. This unpublished infectious clone had increased infectivity for agro-inoculation of common bean compared to earlier-developed clones (Stenger et al., 1990; Stenger et al., 1992).

For each of the 289 accessions, two pots with a diameter of 15 cm were filled with MetroMix-510 growing medium (Premier Horticulture, Inc., Red Hill, PA), and each pot contained two plants. The plants were covered and placed in a dark growth chamber for 24 h after inoculation. After 24 h, they were maintained in a growth chamber with a 14-/10-h light cycle and a temperature range of 18°C–27°C. The seedlings were transplanted into 1-gallon pots (two plants/pot) 2–6 days post-inoculation. The accessions were tested in subsets of approximately 30 accessions, in 4-day intervals, in a growth room. Each subset included two pots with two plants, each of the susceptible ‘Caprice’ and resistant ‘Hystyle’ cultivars as controls. Six accessions did not germinate, including ‘Angers’, ‘Processor’, ‘RH13’, ‘Serin’, ‘Golden Rod’, and ‘Sundial’. Disease reactions were recorded 17–24 days post-inoculation (dpi) as resistant (no visible symptoms), susceptible (with assorted symptoms: leaf rugosity, leaf cupping, leaf curling, plant stunting), or intermediate resistant (with fewer and less severe symptoms) (Larsen et al., 2010).

The DBP accessions were sown during the summer of 2022 in two replications at the Washington State University (WSU) Research Station—Roza Unit (46°17′43.6ʺN 119°43′48.0ʺW) in Prosser, WA, USA, where CTV is endemic. The experimental design was a randomized complete block, and each plot consisted of four rows of 3-m row length, with a spacing of 0.56 m between rows. To determine the CTV strain(s) causing infection in the trial, leaf tissue was collected from 1 asymptomatic plant and 16 symptomatic plants during the early pod-fill R3 growth stage. The tissue was extracted using a modified Dellaporta protocol (Crosslin et al., 2006) and analyzed by conventional PCR using the strain-specific primers previously described in Strausbaugh et al. (2008) to detect three strains of CTV, namely, the CFH strain [also referred to as beet severe curly top virus (BSCTV)], Wor strain [beet mild curly top virus (BMCTV)], and CA/Logan strain. All PCR reactions consisted of 5 µl of 5× GoTaq Green PCR buffer (Promega Corp, USA), 0.25 µl 10 mM dNTP (each), 0.5 µl of 20 µM forward and reverse primers, 17.65 µl H₂O, 1 µl of total nucleic acid extract, and 0.1 µl GoTaq DNA polymerase (Promega Corp., USA). Thermal cycler conditions consisted of 95°C for 2 min, followed by 40 cycles of 95°C for 1 min, 54°C for 1 min, and 72°C for 1 min, and a final extension step of 72°C for 5 min. Amplicons were visualized on a 1.5% agarose gel with ethidium bromide staining.

Disease reaction was rated 60 days after planting on a scale from 1 to 9, where 1 = no symptoms and 0% of plants infected, 5 = moderate disease symptoms and up to 30% of plants infected, and 9 = severe disease symptoms with ≥80% of the plants dead or dying (Larsen and Miklas, 2004). Trait data were spatially corrected, and best linear unbiased predictors (BLUPs) were calculated using a two-dimensional P-spline mixed model with the Mr.Bean web application (Aparicio Arce, 2018) and the R package SpATS v1.0-11 (Rodríguez-Álvarez et al., 2018).

DNA was extracted from each SnAP accession using the Qiagen Plant DNeasy Mini Kit (Hilden, Germany) and diluted to 10 ng/μl with nuclease-free water. Multiplexed genotyping-by-sequencing (GBS) libraries were generated through digestion with ApeKI following the protocol and with barcodes detailed in Elshire et al. (2011). Single-end sequencing of multiplex GBS libraries was performed using the Illumina HiSeq2000 (Illumina, San Diego, CA, USA). The barcodes of raw Illumina sequence were trimmed, and reads were aligned to the reference Andean G19833 Phaseolus vulgaris v2.1 genome sequence (https://phytozome-next.jgi.doe.gov/info/Pvulgaris_v2_1) using Bowtie2 v2.4.1 (Langmead and Salzberg, 2012). The alignments in BAM format were sorted by SAMtools v1.9 (Li et al., 2009).

SNP discovery and genotype calling were conducted using NGSEP v3.3.0 (Next Generation Sequencing Experience Platform) pipeline (Perea et al., 2016) with G19833 Phaseolus vulgaris v2.1 as reference genome (Schmutz et al., 2014). Maximum base quality score was set to 30, and the quality minimum was set to 40 for reporting a variant. All SNP markers detected with <40% missing values, and a minor allele frequency (MAF) of 0.01 were retained to perform imputation with the ImputeVCF module in NGSEP, which is a reimplementation of the Hidden Markov Model (HMM) implemented in the package fastPHASE (Scheet et al., 2006). Annotation of variants was performed using the command Annotate by NGSEP.

For the DBP accessions, the leaf tissue was collected from individual plants grown in the USDA-ARS greenhouses at Prosser, WA. Genomic DNA was isolated from 20 mg of the leaf tissue using a Qiagen DNeasy 96 Plant Kit (Hilden, Germany) and sent to the US Department of Agriculture Agricultural Research Service (USDA-ARS), Soybean Genetics and Improvement Laboratory, Beltsville, MD, for the 11,292 SNPs BARC Illumina chip assay (Miklas et al., 2020). The SNP genotyping was conducted on the Illumina platform following the Infinium HD Assay Ultra protocol (Illumina Inc., San Diego, CA) (Song et al., 2015). The SNP positions were updated by alignment to the v2.1 reference genome assembly of G19833 and were filtered based on a minimum allele frequency (MAF) of 0.01 and a missing data rate of <20%.

The SNP data sets used for population structure analysis in SnAP and DBP were adjusted by linkage disequilibrium (LD) pruning using a 0.5 r² threshold and the –indep-pairwise function in Plink v2.0 (https://www.cog-genomics.org/plink/2.0/), with a sliding window of 50 kb shifting by five bases. In addition, two dry bean lines, ‘BAT 93’ (representing the Middle American gene pool) and ‘Jalo EPP 558’ (representing the Andean gene pool), were added to assess the structure and diversity in SnAP. The sequence data for ‘BAT 93’ and ‘Jalo EPP 558’ was obtained by GBS, as described above.

The pruned set of SNPs with <20% missing values and a MAF > 0.01 were retained. Population structure was estimated using a Bayesian Markov chain Monte Carlo model (MCMC) implemented in STRUCTURE 2.3 software (Pritchard et al., 2000). Ten runs (k) were performed for each number of populations set from 1 to 10. Burn-in time and MCMC replication number were set to 100,000 and 200,000, respectively. The most likely number of populations (K) in the SnAP was calculated with the Evanno method (Evanno et al., 2005) using Pophelper v2.3.1 (Francis, 2017). Finally, the merge Q-matrix was obtained using the CLUMPP v1.1 software (Jakobsson and Rosenberg, 2007) to determine the robustness of the population assignments at each K.

GWAS analysis was conducted by mixed linear model (MLM) using the genome association and prediction integrated tool (GAPIT) (Lipka et al., 2012). An identity-by-state kinship matrix was created using the Efficient Mixed Model Association (EMMA) algorithm implemented in GAPIT R package with corrections for kinship and population structure. Five principal components (PCs) generated from GAPIT were included as covariates, and the Bonferroni procedure was applied to control the experiment-wise type I error rate at 0.05. The GWAS results were visualized using CMPlot v3.62 (Yin, 2016) and IntAssoPlot v3 (He et al., 2020) for a graphical representation of regional-based marker–trait associations.

Candidate genes annotated within the Bct region were identified using the gene models from G19833 v2.1 common bean reference genome. Selected candidate genes were amplified and sequenced in a subset of 14 snap and 14 dry bean genotypes from Andean and Middle American gene pools.

A set of primers were designed to obtain full-length candidate genes using Primer3 software (Koressaar and Remm, 2007) (http://bioinfo.ut.ee/primer3-0.4.0/). The PCR volume was 25 µl, containing 1× Taq buffer, 1.8 mM of MgCl₂, 0.4 mM of dNTPs mix (Promega^®, Madison, WI), 0.25 µM of each primer (forward and reverse), 1 U of Taq polymerase (Promega^®, Madison, WI), and 25 ng of genomic DNA. The amplification conditions were 2 min at 95°C, followed by 38 cycles at 94°C for 20 s, 57°C for 30 s, and 72°C for 90 s, and a final extension at 72°C for 5 min, performed in a PCR Eppendorf Mastercycler (Eppendorf AG, Hamburg, Germany). PCR products were visualized by gel electrophoresis on 2% (w/v) agarose. Sanger sequencing from PCR products obtained was performed using the services of Eurofins MWG Operon (Louisville, KY, USA). Sequencing data were analyzed using Geneious 9.1.8 (Kearse et al., 2012), which was used for sequence trimming, alignment, and variant detection.

A set of five SNPs and one InDel marker were developed from a variants catalog obtained from Lobaton et al. (2018) and used to refine the Bct interval identified by GWAS in SnAP ( Supplementary Table S1 ). From the refined interval, three nonsense variants (S07_2966197, S07_ 2970276, and S07_ 2970381) were selected in candidate gene sequences based on polymorphism among 14 snap and 14 dry beans genotypes and converted to markers using the allele-specific method (Wang et al., 2005). These three variants (S07_2966197, S07_ 2970276, and S07_ 2970381), and the original ‘SAS8.₁₅₅₀’ SCAR marker, using the protocol described by Larsen and Miklas (2004), were assayed in the SnAP and DBP. Lastly, three additional SNP markers (Pv07_3044767, Pv07_331344, and Pv07_3644749), identified by Ariza-Suarez et al. (2023) to be linked with resistance to bean leaf crumple virus (BLCrV) in proximity to the Bct region, were converted to Tm-shift assays for genotyping the DBP ( Table S1 ). BLCrV is a single-stranded DNA Begomovirus in family Geminiviridae.

Tm-shift SNP markers were amplified by PCR on an Eppendorf Mastercycler (Eppendorf AG, Hamburg, Germany). The PCR volume was 20 μl containing 20 ng of genomic DNA, 1× Taq buffer, 1.5 mM of MgCl₂, 0.2 mM of dNTPs mix (Promega), 0.15 μM of each primer (two allele-specific forward primers and the common reverse primer), 1× EvaGreen^® dye, and 0.1 μl of Taq polymerase (Promega) under the following thermal profile: 94°C for 2 min, then 38 cycles of denaturation at 92°C for 20 s, annealing for 20 s (the temperature was specific to each primer trio), and extension at 72°C for 20 s, and final extension at 72°C for 5 min. Melting point analysis for allele determination of the PCR products was performed with a fluorescence-detecting thermocycler (LightCycler^® 4890 Instrument II, Roche, Basal Switzerland). EvaGreen Fluorescent detection was performed for 1 min at 95°C, and the melting curve step ramped from 65°C to 95°C in increments of 1°C every 20 s.

The S07_2966197, S07_ 2970276, and S07_ 2970381 markers were used to identify haplotypes correlated with quantitative field reaction to CTV in the DBP. An analysis of variance between haplotypes was carried out by F_{trimmed-means} one-way ANOVA using the ggbetweenstats function in the R-Package ggstatsplot. Yuen’s trimmed means test was implemented using Robust p-value fitting method using the ggbetweenstats function in the R-Package ggstatsplot (Patil, 2021). Overall, the F_{trimmed-means} method provides a robust approach for analyzing data that may contain outliers and can help identify statistically significant differences between groups.

Out of the 289 SnAP accessions screened by agro-inoculation with the CTV CA/Logan strain, 166 accessions were resistant, 66 were susceptible, 4 showed intermediate resistance, and 53 were segregating with plants exhibiting either resistant or susceptible (20 accessions); resistant or intermediate resistance (18 accessions); or resistant, intermediate resistant, or susceptible responses (15 accessions). The disease reactions for 83 SnAP accessions, compiled from various public information sources, were categorized as 42 resistant, 32 susceptible, and 9 intermediate resistant to CTV. For upstream analysis, the qualitative reactions of each accession were transformed into a quantitative score, with resistant being coded as 1, susceptible as 0, and intermediate resistant as 2 ( Supplementary Table S2 ).

In addition, 88 DBP accessions were evaluated for CTV reaction (scored from 1 to 9) under natural field infection. Based on the BLUP values for CTV reaction, 43 accessions were resistant (disease score between 1 to 4), 33 were tolerant (between 4.1 to 6.9), and 14 were susceptible (between 7 to 9) to the Wor strain 60 days after planting ( Supplementary Table S3 ). PCR analysis with strain-specific primers detected the Wor strain in leaf tissue collected from this naturally infected field. Specifically, one asymptomatic plant and 11 of 16 symptomatic plants amplified a product of the expected size (397 bp) with the Wor strain primers.

According to STRUCTURE analysis of the SnAP (378 accessions), using a pruned set of 1,833 SNPs, a likelihood of K=2 was estimated, indicating the presence of two sub-populations belonging to the Andean and Middle American gene pools, as confirmed by gene-pool checks ‘Jalo EEP 558’ and ‘BAT 93’ ( Supplementary Figure S1 ). Accessions were further categorized based on a membership coefficient of Q ≥ 0.85. The Andean population comprised 166 accessions (43.9%), and the Middle American, 27 accessions (7.1%). The remaining accessions were grouped as admixtures according to gene-pool predominancy, using a membership coefficient Q ≥ 0.5, with 108 accessions as Andean admixtures (28.6%) and 77 as Middle American admixtures (20.4%) ( Supplementary Table S4A ).

A similar STRUCTURE analysis was performed in the DBP (88 accessions), using an LD-pruned set of 974 SNP markers, and K=2 was estimated ( Supplementary Figure S2 ), with 59 Andean (67.0%) and 29 Middle American (33.0%) accessions ( Supplementary Table S4B ). Gene pool classification was based on previous diversity studies (Cichy et al., 2015; Moghaddam et al., 2016).

A filtered set of 25,472 biallelic SNPs from 378 SnAP accessions were obtained by GBS. Multiple GWAS analyses were conducted to identify the genomic regions associated with resistance to CTV in the SnAP. The first GWAS compared 208 resistant, 13 intermediate-resistant, and 98 susceptible SnAP accessions, and 53 accessions segregating for CTV symptoms were excluded. A significant interval from 2.79 to 4.32 Mb on chromosome Pv07 was identified, exceeding the Bonferroni-corrected significance threshold of α = 0.05 (p = 1.7E−36), which was associated with resistance and intermediate resistance to CTV ( Supplementary Figure S3A ). This interval corresponds with the location of the Bct gene that was identified by Larsen and Miklas (2004) in a snap bean RIL population and a QTL (BCT7.1^GT) that conditions resistance to CTV found in the dry bean landrace G122 named ‘Jatu Rong’ (PI 163120) from India (Larsen et al., 2010).

A separate GWAS comparing 208 resistant with 98 susceptible accessions and 13 intermediate-resistant with 98 susceptible accessions revealed similar significant peak intervals of 2.66–4.32 Mb (p = 7.81E−44) and 2.79–3.60 Mb (p = 1.89.17) on Pv07, respectively ( Supplementary Figures S3B, C ). No significant SNPs were detected by the GWAS comparing 208 resistant and 13 intermediate-resistant SnAP accessions. Therefore, a GWAS was conducted comparing 239 resistant and intermediate-resistant accessions coded as 1 with 98 susceptible accessions coded as 0. This analysis showed a genomic interval (containing 129 significant SNPs) from 2.66 to 4.32 Mb on chromosome Pv07 with significantly increased significance (p =1.12E−45), as expected for a single major allele ( Figure 1A ). One LD block within the 2.66–4.32 interval, from 2,794,761 to 3,199,136 bases ( Supplementary Table S5 ), consisting of 50 significant SNPs with the highest p-values ranging from p = 1.08E−34 to 1.12E−45, became the target region for the Bct allele for CTV resistance on Pv07 ( Figure 1B ). Other LD blocks with CTV resistance associations less than p = 2.43E−28 were excluded as targets within the 2.66 – 4.32 Mb interval.

GWAS with the 88 DBP accessions, using 7,983 SNP markers after filtering, did not detect any significant peaks for resistance to CTV. To improve the analysis, we incorporated six additional SNP markers, including three identified from sequencing candidate genes in the Bct region described below and three associated with resistance to bean leaf crumple virus (BLCrV) from Ariza-Suarez et al. (2023). Using these additional SNPs, the GWAS analysis identified a missense SNP, S07_2970381 (G19833v2.1 reference genome), within a candidate gene in Bct region with a high p-value (p =1.84E−07) ( Figure 2A ). The three SNP markers linked to BLCrV resistance ( Figure 2B ) had low LD with the sBct region and did not associate with CTV resistance.

Six resistant SnAP accessions displayed recombination within the targeted LD block (2,794,761–3,199,136 bp) for the Bct region on Pv07 ( Supplementary Table S6 ). These accessions were ‘Goldcrop’ (SnAP152) and ‘Keygold’ (SnAP191) from the Andean gene pool; ‘Blue Mountain’ (SnAP037) from the Middle American gene pool; and Andean admixtures ‘Early Bird’ (SnAP107), ‘FR 266’ (SnAP136), and ‘Labrador’ (SnAP194). Haplotypes for the six developed markers spanning the LD block in the recombinant SnAP accessions narrowed the interval for the Bct region to 58.0 kb from 2,943,470 to 3,001,466 bases ( Figure 3 ; Supplementary Table S6 ).

Six candidate genes were annotated in the refined 58.0 kb Bct region based on the Andean G19833 common bean v2.1 gene models ( Table 1 ). However, one of these gene models, Phvul.L001619, was missing from a gap segment (Pv07: 2,988,789–2,998,791 bases) in the Bct region of G19833 v2.1. A BLASTn search in the previous version of the G19833 v1.0 reference genome revealed that this gene model had 100% similarity with the former Phvul.007G036700 in G19833 v1.0.

Interestingly, two candidate genes in the Bct region, Phvul.007G036300 encoding a 5′ deoxyribonuclease domain-containing protein (Exonuclease V) and Phvul.L001619, which functions as a mediator of RNA polymerase II transcription subunit, showed homology with gene models from other plant species involved in the replication of geminiviruses (Hanley-Bowdoin et al., 2013). Another candidate gene, Phvul.007G036200 (transmembrane protein), may play a role in defense response to other organisms based on its homology to the Arabidopsis thaliana gene AT1G71110. Target open reading frame (ORF) sequencing of the three candidate genes revealed 29 variants, 28 SNPs, and 1 InDel among a subset of 28 snap and dry bean genotypes from diverse backgrounds ( Supplementary Table S7 ). Additionally, Phvul.007G036400 and Phvul.007G036500 gene models within the Bct region were sequenced, with two nonsense InDel variants and two synonymous SNP variants identified, respectively. Overall, for the 33 variants in the five candidate genes assayed across the 28 genotypes, only one variant, a missense SNP on Pv07: 2,970,381 bases in candidate gene Phvul.007G036300, was completely associated with BCTV resistance.

The in silico analysis of the Exonuclease V (EXO5) protein of the sequenced ORF for gene Phvul.007G036300 did not uncover any disruptions in the EXO5 domain. Nevertheless, the missense SNP variant located in Pv07: 2,970,381 bases involves a transition of Adenine (A) to Thymine (T) base, resulting in radical replacement of a polar, uncharged amino acid, glutamine (Q), with a non-polar, hydrophobic amino acid, leucine (L), at position 279 in the EXO5 domain (i.e., p.Q279L) ( Table 2 ).

The dendrogram based on the genetic distance of the translated protein for the Bct candidate gene Phvul.007G036300 in the 28 sequenced genotypes showed one cluster for susceptibility and two distinct clusters for resistance/intermediate resistance to CTV ( Figure 4 ). The first cluster consisted of susceptible snap bean genotypes from the Andean gene pool. The second cluster included the resistant Andean snap bean genotypes ‘Shannon’, ‘Amity’, ‘Fortex’, and the Andean dry bean ‘G122’. The third cluster included the resistant Andean snap bean genotypes ‘Early Bird’, ‘Astun’, ‘Profit’, ‘Tenderlake’, the Middle American snap bean ‘Dubbele Witte’, and all the sequenced dry beans from the Middle American gene pool. The Bct EXO5 candidate gene protein sequence exhibits 99.75% identity between Andean-resistant and Andean-susceptible genotypes, with only one amino acid change observed for the codon containing SNP S07_2970381. In contrast, resistant and tolerant Middle American genotypes display 100% identity for the EXO5 protein.

A Tm-shift assay named S07_2970381 was specifically designed to detect the missense SNP Pv07: 2,970,381 in the Bct candidate gene Phvul.007G036300 (EXO5), which exhibited polymorphism between resistant and susceptible genotypes to CTV. Additionally, two Tm-shift assays were developed to distinguish introgression haplotypes of the Bct region from Andean and Middle American gene pools. The first assay, named S07_2970276, targeted the missense SNP variant Pv07: 2,970,276 bases within Phvul.007G036300 gene and differentiated genotypes from Andean and Middle American gene pools. The second assay, named S07_2966197, focuses on another missense SNP variant, Pv07: 2,966,197 bases, found in the Phvul.007G36200 gene model by Soler-Garzón et al. (2021) in a race Durango dry bean diversity panel. This variant differentiated genotypes within the Middle American genepool. The primer sequences and Tm-shift melting curves for each SNP marker can be found in Table 3 and Supplementry Figure S4 , respectively.

The S07_2970381 marker exhibited a 100% correlation with the CTV phenotype observed in SnAP. For the 378 SnAP accessions, 238 (63.0%) showed resistance or intermediate resistance to the CTV CA/Logan strain and possessed the resistance allele of the S07_2970381 marker, and 100 (26.5%) were susceptible and possessed the susceptible allele of the S07_2970381 marker, while 33 (8.7%) were segregating for CTV reaction, and 7 (1.8%) had missing data ( Supplementary Table S8 ).

For the DBP (88 accessions) infected with the Wor strain, a correlation of 99% was observed between the S07_2970381 SNP marker and the disease score. There were 42 accessions with resistant disease scores ≤ 4.0, which possessed the resistant allele of the S07_2970381 SNP marker, and 13 with susceptible disease scores from 7.0 to 9.0, which possessed the susceptible allele ( Supplementary Table S9 ). The 31 tolerant accessions, with disease scores from 4.1 to 6.9, possessed the resistant marker allele with two exceptions: ‘SVS-0683’ (DBP42) with a disease score of 6.3 and ‘Beluga’ (DBP45) with a disease score of 6.2, possessed the susceptible marker allele. Note that the ‘Beluga’ genotype exhibited a susceptible reaction when agro-infiltrated with Logan strain, while SVS-0863 was untested.

A comparison of genotypes for the S07_2970381 SNP marker and the ‘SAS8_.1500’ SCAR marker developed by Larsen and Miklas (2004) showed a 97.1% similarity in the SnAP and 47.4% in the DBP. The ‘SAS8_.1550’ SCAR marker, Pv07: 2,975,432–2,977,000 bases, is 4.1 kb downstream of the EXO5 candidate gene Phvul.007G036300 and in high linkage disequilibrium (LD) with the S07_2970381 and S07_2970276 SNP markers in the SnAP and DBP. In fact, ‘SAS8_.1550’ and S07_2970276 markers were 100% identical in both bean panels.

The SnAP and DBP were assayed for the S07_2970381, S07_2970276, and S07_2966197 markers to identify haplotypes for the Bct region. Four haplotypes were identified: Haplo1 with AGT sequence for the S07_2966197 (A/G), S07_2970276 (G/C), and S07_2970381 (T/A) SNPs, respectively; Haplo2 with AGA sequence; and Haplo3 divided into two sub-haplotypes, Haplo3-1 with ACA sequence and Haplo3-2 with GCA sequence, based on polymorphism for S07_2966197 in Middle American accessions.

In the SnAP ( Table 4 ), Haplo1 was found in all 98 susceptible accessions: 97% Andean and Andean admixtures and 3% Middle American and Middle American admixtures. Haplo2 was a rare haplotype found in only five resistant accessions, four Andean/Andean admixtures, and one Middle American admixture. Haplo3-1 was observed in 162 resistant/intermediate resistant accessions: 76% Andean and Andean admixtures and 24% Middle American and Middle American admixtures. Haplo3-2 was found in 71 resistant/intermediate resistance accessions: 68%) Middle American and Middle American admixtures and 32% Andean and Andean admixtures.

For the DBP ( Table 5 ), Haplo1 was found in 15 Andean accessions with a mean disease score of 8.0, 13 with susceptible reactions and 2 tolerant accessions, ‘SVS-0683’ and ‘Beluga’ with scores of 6.3 and 6.2, respectively. Haplo2 was observed in 27 resistant/tolerant Andean accessions (3.8 mean score). There were 30 resistant/tolerant accessions (3.5 mean score) with Haplo3-1, 63% Middle American and 37% Andean. For Haplo3-2, 16 resistant/tolerant accessions (4.4 mean score), 62% Middle American and 38% Andean, were observed. CTV mean score comparisons among haplotypes in the DBP, using p _Holm-adjusted p-values, showed that the susceptible Haplo1 was significantly different from the three resistant haplotypes (p < 0.0001), while the only significant difference among resistant haplotypes mean scores was between Haplo3-1, which was more resistant than Haplo3-2 (p = 0.05) ( Figure 5 ).

The haplotype survey using the same three markers was expanded to 57 dry bean genotypes, 30 Andean and 27 Middle American, which were agro-inoculated with the CA/Logan strain in Sun Prairie, WI ( Supplementary Table S10 ). Both Haplo1 and Haplo2 were composed of only Andean genotypes, 7 with susceptible and 12 with resistant reactions to CTV, respectively. All Middle American genotypes were resistant to the CA/Logan strain and exhibited either Haplo3-1 (19 genotypes) or Haplo3-2 (8 genotypes). In addition, there were seven resistant Andean accessions with Haplo3-1 and 4 with Haplo3-2. Qualitative reactions to agro-inoculation with the CA/Logan strain matched the quantitative field reactions to Wor strain, for the 12 genotypes included in both the DBP and this additional survey, except for Beluga, which was susceptible to agro-inoculation but tolerant in the field.