Reduced confidence intervals and novel candidate genes for quantitative trait loci associated with apple scab resistance in Malus domestica

1 Reduced confidence intervals and novel candidate genes for quantitative trait 2 loci associated with apple scab resistance in Malus domestica 3 4 Short title: Fine mapping of apple scab resistance quantitative trait loci 5 6 Romane Lapous 1¶*, Camille Haquet 1¶#a, Caroline Denancé 1, Juliette Bénéjam 1#b, Laure 7 Perchepied 1, Kaat Hellyn1, Hélène Muranty1, Charles-Eric Durel1, Julie Ferreira de Carvalho1* 8 9 1 Univ Angers, Institut Agro, INRAE, IRHS, SFR QUASAV, F-49000 Angers, France 10 #a Current address: INRAE, UMR 1095 INRAE–UCA Genetics, Diversity & Ecophysiology of 11 Cereals, Clermont-Ferrand, France. 12 #b Current address: Univ. Bordeaux, INRAE, UMR BFP, F-33140, Villenave d’Ornon, France. 13 14 *Corresponding authors 15 E-mail: [email protected] (RL), [email protected] (JF) 16 17 ¶ These authors contributed equally to this work. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 18 Abstract 19 Apple scab, caused by Venturia inaequalis, remains one of the most damaging diseases in 20 apple orchards, driving intensive pesticide use worldwide. Reducing this dependence requires 21 the deployment of durable resistance, ideally through the combination of major resistance 22 genes (R genes) with quantitative trait loci (QTL) that confer partial and potentially 23 complementary protection. Yet, few apple scab QTLs have been functionally validated, and 24 their underlying mechanisms remain largely unresolved. 25 Here, we refined and functionally described, with transcriptomic data, five resistance QTLs in 26 a biparental population of 1,970 individuals derived from the cross ‘TN 10-8’ × ‘Fiesta’. Using 27 43 newly developed KASP markers, QTL locations were substantially precised through high- 28 resolution genotyping and phenotyping with two V. inaequalis isolates exhibiting contrasting 29 virulence. Four QTL (qT1, qF11, qF17, qT13) were validated, while qF3 was not confirmed. 30 Transcriptomic data comparison revealed the expression of candidate genes within the 31 narrowed intervals, including receptor-like proteins in qT1, and RNAi- and signaling-related 32 genes in qF11 and qF17, suggesting a diversified and complementary defense network. 33 These findings refine the genetic architecture of apple scab resistance and suppose the 34 existence of shared molecular pathways between major R gene, such as the well-described 35 Rvi6 gene, and quantitative resistance, with for instance the QTL qT1. The identified loci and 36 markers provide robust tools for marker-assisted and genomic breeding aimed at developing 37 apple cultivars with complementary and potentially durable resistance pathways. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 38 Introduction 39 Venturia inaequalis, the causal agent of apple scab, remains the most economically damaging 40 fungal disease of apple worldwide, causing substantial yield losses and requiring intensive 41 fungicide applications in temperate regions 1. The pathogen overwinters in fallen leaves, 42 producing ascospores that initiate primary infections on young leaves and fruits during spring. 43 Subsequent cycles of conidial dissemination lead to secondary infections, resulting in 44 defoliation, fruit deformation, and, in severe cases, tree weakening 2. To limit the 45 environmental and health impacts of repeated fungicide use, the development of genetically 46 resistant cultivars has become an essential strategy for sustainable apple production 3. 47 Historically, breeding for scab resistance has relied on monogenic resistance conferred by 48 resistance genes or R genes. Up to now, at least 18 such loci (Rvi1 –Rvi18) have been identified 49 from cultivated and wild Malus accessions 4–7. The first and most widely used, Rvi6 (previously 50 named Vf), originates from Malus floribunda and encodes a leucine-rich repeat receptor-like 51 protein (LRR-RLP) 8–10. Approximately 90% of resistant cultivars released since the 1980s carry 52 this Rvi6 locus 11. However, some V. inaequalis populations are able to overcome this 53 resistance, as firstly reported in the cultivar ‘Prima’ in Germany in 1988 12. The origin of such 54 populations could result from the migration of virulent isolates from non-European apples, 55 and their maintenance in environmental disease reservoirs, like ornamental crabapples 13. 56 This breakdown underscores the limitations of monogenic resistance, which imposes strong 57 selection pressure on pathogen populations. 58 In contrast, quantitative resistance, controlled by quantitative trait loci (QTL), typically 59 provides partial but presumably more durable protection, covering functions that might 60 complement R genes 14. Each locus contributes modestly to resistance, and the combined .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 61 effect of multiple QTLs may slow pathogen adaptation and enhance resistance durability if 62 they mobilize different mechanisms. To date, at least 14 apple scab QTLs have been reported 63 7, yet only a few have been precisely mapped or functionally characterized due to broad 64 confidence intervals (CIs) and limited marker density in earlier studies. Among these, three 65 QTL—qT1, qF11, and qF17—have been repeatedly detected and partially characterized in 66 biparental populations derived from the cultivars ‘TN10-8’ and ‘Fiesta’ 15–18. The locus qT1, 67 located on chromosome 1 and derived from ‘TN10-8’, co-localizes with Rvi6, which has been 68 shown to be a member of an extracellular leucine-rich repeat receptor family 8,9. In contrast, 69 qF11 and qF17, both originating from ‘Fiesta’, reside on chromosomes 11 and 17, respectively, 70 and exhibit a strong synergistic interaction, reducing disease severity only when both 71 resistance alleles are combined 18,19. Interestingly, these two QTLs were effective against a 72 broader range of isolates than qT1 16,18. Moreover, the pyramiding of qT1, qF11, and qF17 in 73 a ‘TN10-8 × Fiesta’ progeny has been shown to block V. inaequalis development at multiple 74 infection stages—from penetration to sporulation 20—although subsequent pathogen 75 adaptation has eroded the protection conferred by qF11 and qF17 19. More recently, two 76 additional QTLs—qT13 on chromosome 13 (from ‘TN10-8’) and qF3 on chromosome 3 (from 77 ‘Fiesta’)—were reported by Bénéjam et al. (2021) 18 but remained poorly resolved due to wide 78 CIs and low SNP density. 79 To elucidate their genetic basis and exploit them for breeding, it is essential to increase 80 mapping precision and identify putative causal genes. Fine-mapping approaches rely on larger 81 offspring populations (~500 to >10,000 individuals when available), thereby increasing the 82 number of recombination events 21. Coupled with SNP genotyping—the most common source 83 of DNA sequence variation—these approaches help reduce QTL CIs, provide new markers for .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 84 marker-assisted selection (MAS), limit the co-introgression of undesirable genes, and decrease 85 the number of candidate genes for functional validation 22–25. In addition to these approaches, 86 integrating ‘-omics’ data—such as genomics, transcriptomics, proteomics, and 87 metabolomics—can help prioritize candidate genes for functional validation. Such integration 88 may involve sequence alignment 21, gene expression analyses 25, or the identification of QTL 89 co-localizations 26–28. 90 In this study, we analyzed an extended ‘TN10-8 × Fiesta’ progeny of 1,970 individuals to refine 91 the genetic localization of five QTL (qT1, qF11, qF17, qT13, and qF3) involved in quantitative 92 resistance to V. inaequalis. Using newly developed KASP markers and phenotyping with two 93 V. inaequalis isolates of contrasting virulence, we improved the mapping resolution of the 94 QTLs. Then, the integration of transcriptomic data from Bénéjam et al. (2024) 29 enable us to 95 propose candidate genes potentially involved in apple scab resistance. These findings improve 96 our understanding of the genetic basis of quantitative resistance in apple and provide 97 molecular resources for marker-assisted and genomics-assisted breeding strategies aimed at 98 combining complementary resistance mechanisms for durable scab resistance. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 99 Materials and methods 100 Plant material 101 Seeds of the cross between apple genotype ‘TN10-8’ and cultivar ‘Fiesta’ (referred to as the 102 ‘TxF' progeny) were produced in the framework of the DARE European project (Durable Apple 103 Resistance in Europe, 1998-2002). A part of these seeds were firstly elevated in greenhouse 104 until their implantation in orchards, then allowing their phenotyping through grafting to 105 decipher the genetic control of scab resistance segregating in this material 16. This initial 106 population, hereafter referred as base population, is composed of 267 individuals and allowed 107 for the detection of three QTLs of ‘Fiesta’ and two of ‘TN10-8’ 15,16,18,30,31. ‘Fiesta’ is an offspring 108 from a cross between the cultivars ‘Cox’s Orange Pippin’ (resistant to apple scab) and ‘Idared’ 109 (susceptible to apple scab) and is carrying resistance and susceptibility alleles on each of 110 chromosomes 3, 11 and 17 that are subsequently referred as qF3, qF11 and qF17, respectively. 111 The hybrid ‘TN10-8’, is coming from a cross between the French heirloom cultivar ‘Reinette 112 Clochard’ and a descendent from the Russian cultivar ‘Antonovka’ (fairly resistant to apple 113 scab) 32. This genotype is carrying resistance and susceptibility alleles on each of chromosome 114 1 (qT1) and 13 (qT13). The MUNQ (Malus UNiQue genotype code), as described by Muranty 115 et al. (2020) 33 and Durel et al. (2023) 34, was 2396 for ‘TN10-8’, 763 for ‘Fiesta, 163 for ‘Cox’s 116 Orange Pippin', 717 for ‘Idared’, 118 for ‘Reinette Clochard’ and 4 for Antonovka. 117 The remaining seeds from the ‘TxF' progeny, stored at -20°C, were used in the present study 118 and each seed still corresponds to a unique genotype. This population, hereafter referred as 119 extended population, is composed of 1,970 seeds. In the fall 2022, seeds were germinated, 120 transplanted into soil, and subsequently cultivated in a greenhouse on their own roots. Young .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 121 leaf discs of each individual were sampled using the “BioArkTM Leaf collection kit” from LGC 122 Biosearch Technologies TM company (United Kingdom, UK) for subsequent genotyping. 123 Identification of QTL boundaries, development of SNP markers and 124 qPCR-based assays 125 In previous studies, QTL locations were assessed using 267 individuals 18, genotyped with SNPs 126 from the 20K SNP arrays 35,36. Context sequences of 43 SNPs localized upstream, within, and 127 downstream of the five QTL CIs, and heterozygous in ‘Fiesta’ or ‘TN10-8’, were retrieved from 128 both the 20K and 480K SNP arrays and blasted against the reference genome of ‘Golden 129 Delicious’ doubled haploid GDDH13 37 to then be visualized using JBrowse 38. This step allowed 130 the assignment of broader context sequence surrounding each candidate SNP (mean length 131 150 bp; range 105-216 bp). Additional SNPs, identified by aligning the sequencing data used 132 to develop the 480K array 36 to the GDDH13 genome 37, were also considered in order to 133 develop adapted primers for KASP genotyping. 134 Leaf material was sent to LGC Biosearch Technologies TM (Herts, UK) for DNA extraction and 135 SNP genotyping using KASP technology. For each selected SNP, two allele specific forward 136 primers and one common reverse primer were designed and validated by LGC Biosearch 137 Technologies TM (Herts, UK). The assay mix preparation and PCR amplifications were 138 performed according to the user’s guide and manual (LGC Biosearch Technologies TM, 139 Hoddesdon, Herts, UK). The concentration of DNA samples was measured using 140 NanoDrop2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). DNA 141 samples were diluted for genotyping, as required. KASP genotyping was carried out on a 142 StepOnePlusTM real-time PCR system (Thermo Fisher Scientific, Waltham, MA, USA). .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 143 Genotyping data were cleaned and analyzed using SNPviewer™ (LGC Biosearch Technologies, 144 UK) at INRAE (Angers), which visualizes KASP results as cluster plots. Marker quality was 145 evaluated based on clustering performance and genotype discrimination among seedlings 146 derived from two genotyped parents. Out of the 43 markers tested, 36 were validated. The 147 seven non validated markers exhibited poor clustering, likely resulting from suboptimal primer 148 hybridization due to imperfect primer design or additional polymorphisms near the target 149 SNP. 150 Using the same DNA extracts, additional SNP markers, were developed at INRAE to further 151 densify the CIs and peak of four QTLs (qT1, qF11, qF17, and qT13). A second set of seven SNP 152 markers was successfully designed and incorporated. KASP primers were designed at Eurofins 153 Genomics. These markers were analyzed on the ANAN genotyping platform (Univ. Angers, 154 France). Genotyping was then done using the PACE® mix (3crbio, Harlow, UK) following 155 manufacturer protocol. Thus, a total of 43 SNPs distributed along the five QTLs of interest 156 were used. Each SNP is presented with sequence context, SNP of interest position and 157 additional SNPs in S1 File. This file also contains virtual markers, added with computational 158 analyses described below. 159 Apple scab phenotypic scoring and analyses 160 Venturia inaequalis isolate inoculum and apple scab scoring 161 The seedlings described above were own-rooted and non-replicated; therefore, scab 162 inoculations were performed sequentially on the same plants, which were pruned between 163 each inoculation. Seedlings were randomly distributed into four blocks in the greenhouse. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 164 Two V. inaequalis isolates were used: the reference isolate ‘EU-B04’ (Origin: Belgium, host: 165 ‘Golden Delicious’) previously described in Caffier et al. (2015) 39 and Le Cam et al. (2019) 40 166 and the isolate ‘09BCZ014’ (Origin: France, host: ‘TN10-8’ × ‘Prima' progeny), referred to as 167 isolate ‘2557’ in Laloi et al. (2017) 20. The isolate ‘09BCZ014’ partially overcomes the resistance 168 conferred by the qT1 resistance QTL segregating in the ‘TxF’ progeny. The use of this isolate 169 allows to better study the effect of other QTLs segregating in the progeny. 170 Monoconidial suspensions were prepared from diseased dry leaves at a concentration of 2.5 171 × 10 5 conidia.ml−1 and sprayed on seedlings. Inoculated plants were then incubated for two 172 days at 17°C under a plastic sheet to maintain high humidity, according to the conditions 173 described by Caffier et al. (2010) 41. The percentage of leaf surface exhibiting sporulating 174 lesions was scored at 14 and 21 days post-inoculation (dpi) using the ordinal scale (0 to 7) 175 described in Calenge et al. (2004) 16. Three experiments were conducted from January to May 176 2022 (Table 1). One experiment was performed with isolate ‘EU-B04’ (coded ‘Vi-B04’) for 177 which additional scoring of resistance symptoms (leaf chlorosis and crispation) was recorded 178 at 14 dpi using the same ordinal scale (0 to 7 scores, reflecting the percentage of foliar surface 179 covered with chlorosis or crisped). Two other experiments were then conducted using the 180 isolate ‘09BCZ014’ (coded ’Vi-Z14_1’ and ’Vi-Z14_2’). ’Vi-Z14_1’ was performed on all the ‘TxF' 181 progeny seedlings (1,970) whereas ’Vi-Z14_2’ was restricted to individuals recombining in CIs 182 of QTLs of interest (1,101). For ‘Vi-Z14’ experiments, only leaf sporulation was recorded 183 (susceptibility symptoms). 184 Table 1. Summary of experiment characteristics. Experiment 1 Experiment 2 Experiment 3 Name Vi-B04 Vi-Z14_1 Vi-Z14_2 Month January March May .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint Seedling number 1,970 1,970 1,101 Vi isolate EU-B04 09BCZ014 09BCZ014 Phenotyping Sporulation at 14 and 21 dpi Chlorosis at 14 dpi Crispation at 14 dpi Sporulation at 14 and 21 dpi Sporulation at 14 and 21 dpi 185 Dpi, days post-inoculation. 186 Using sporulation scores at both dates i.e 14 dpi and 21 dpi, the Area Under Disease 187 Progression Curve (AUDPC) was calculated as a quantitative summary of disease severity over 188 the course of the infection using the following equation: 189 𝐴𝑘 = 𝑁𝑖―1 𝑖=1 (𝑦𝑖 + 𝑦𝑖+1) 2 (𝑡𝑖+1 ― 𝑡𝑖) 190 Where yi is the disease score at the ith day of observation and t is the number of days after V. 191 inaequalis inoculation at the ith observation. For AUDPC calculation, the initial time point (t0) 192 was set to 0. 193 All statistical analyses were performed using the software R V2022.07.2+576 (R Core Team, 194 2022). ANOVAs (function aov from the stats package) were conducted to test for assessor and 195 block effects and, when necessary, were followed by Tuckey post hoc tests (function 196 tukey_hsd from the rstatix package). Although each block consisted of genetically distinct 197 individuals, their mean genetic value was assumed to be equivalent due to the random 198 distribution of individuals among blocks, allowing adjustment of phenotypic variables for 199 block, assessor, and experiment effects, with the experiment effect included only for the Vi- 200 Z14 experiments. For the ‘Vi-B04’ experiment, QTL mapping was performed using the adjusted 201 AUDPC, chlorosis and crispation scores (hereafter ‘AUDPC_B04’, ‘CHLOROSIS_B04’ and .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 202 ‘CRISPATION_B04’). For the ‘Vi-Z14’ experiments, adjusted AUDPC additionally accounted for 203 variations between the two experiments (‘AUDPC Vi-Z14_adj’). Phenotyping data used for QTL 204 mapping are available in S2 File. 205 Distributions and correlations between studied variables were examined with ggplot function 206 (ggplot2 package) and ggpairs function (GGally package), respectively. In the ‘Vi-B04’ 207 experiment, the strong effect of the major QTL qT1, detected on LG1 of ‘TN10-8’ with the 208 reference isolate ‘EU-B04’ 16,18, is hiding the more moderate effects of other QTLs. Therefore, 209 the 'TxF' progeny was further subdivided into two subsets of individuals according to the 210 presence or absence of the resistance allele at qT1 predicted by SNP data at the QTL peak. 211 AUDPC distributions were then observed on two distinct subpopulations: one including the 212 entire extended progeny (1,970 individuals) and a second, named ‘qT1-’ including 721 213 individuals carrying the susceptible allele at qT1. 214 Mapping and genetic analyses of the targeted QTLs 215 Genetic maps: Genotyping data were analyzed using SNPviewer software provided by LGC 216 Biosearch Technologies TM with manual correction whenever necessary to improve dataset 217 quality. The order of SNP physical positions was checked against the latest apple reference 218 genetic map 42. A linkage map was constructed for each parent with JoinMap 4.1 software 43. 219 In total, over the five targeted regions, 43 SNP markers were used and the associated genetic 220 map is available in S3 File. 221 QTL mapping: 222 QTL analyses were conducted using the R/qtl package 44. In addition to SNP markers, ‘virtual’ 223 markers were added with calc.genoprob and sim.geno functions using 1,000 simulations in .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 224 order to have one marker per centiMorgan (cM). Simple interval mapping and composite 225 interval mapping were then performed using a multiple imputation method and a normal 226 distribution model (using cim function). The logarithm of the odd (LOD) score threshold to 227 identify the statistically significant QTLs was determined using 1,000 permutations (α = 0.05). 228 LOD thresholds were approximately 1.80 for all variables. 229 The LOD score, the 2-LOD support CI, and the contribution of each QTL to the overall 230 phenotypic variance (individual R 2) were extracted from R/qtl analyses, together with the 231 global QTL contribution (global R 2). Individual and global R² were calculated with the fitqtl 232 function (used for fitting a defined multiple-QTL model). Interactions between QTLs were 233 studied by variance analysis using the genotyping data of each SNP closest to the peak of each 234 QTL, and were detailed by the effectplot function. These results were used to define the model 235 for the calculation of the global R 2 with the fitqtl function. Graphics with LOD score and 236 positions of markers were constructed using MapChart 2.32 45. 237 Identification of candidate genes 238 Sequence contexts of SNPs localized at both extremities of QTL CIs were blasted against 239 reference genome GDDH13 v1.1 (using INRAE JBrowse tool) to retrieve QTL physical borders. 240 Physical positions of virtual markers were estimated according to the physical positions of 241 flanking SNP markers. The CI representing the shortest genetic region for each QTL was used 242 to look for candidate genes involved in plant-pathogen interaction. For all QTLs, lists of genes 243 along with their functional annotations and orthologs in Arabidopsis thaliana were extracted 244 from GDDH13 v1.1 gene annotation (https://iris.angers.inra.fr/gddh13/the-apple-genome- 245 downloads.html). .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 246 Transcriptomic datasets were retrieved from a previous experiment performed on offspring 247 of the base population of the ‘TxF' progeny 29. Four individuals for each of four classes were 248 pooled according to the presence/absence of favorable alleles at three QTLs: qT1, qF11 and 249 qF17. The four classes were constructed depending on the combination of QTLs present with 250 either ‘NoScabQTL’, ‘qT1’, both ‘qF11qF17’ and all three ‘qT1qF11qF17’. This design allowed 251 to test for the effects of QTLs on gene transcription before and after inoculation of V. 252 inaequalis with the ‘EU-B04’ isolate (see Bénéjam et al. (2024) 29 for additional information). 253 For all candidate genes underlying the three QTLs qT1, qF11 and qF17, we thus retrieved fold 254 changes and corrected p-values for genes showing differential transcription before and after 255 V. inaequalis inoculation according to the genotypic classes. For instance, for genes within the 256 qT1 range, we analyzed differential transcription between both qT1 and qT1qF11qF17 classes 257 against the 'NoScabQTL' class, but also between qT1qF11qF17 class against the qF11qF17 258 class, at 0 dpi (before V. inaequalis inoculation) and 5 dpi (after V. inaequalis inoculation). 259 Candidate genes with an absolute fold change (|FC|) > 1.5 and a Benjamini–Hochberg 260 adjusted p-value (BH method) < 0.05 were selected for further analysis 46. 261 For the qT13 and qF3 QTLs, no transcriptomic data were available to study the expression of 262 candidate genes contained in the CIs. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 263 Results 264 1. Broadening the mapping population and recombinant detection 265 Among the 1,970 genotyped seedlings, 1,101 displayed at least one recombination event 266 within the five target QTL regions, thereby increasing mapping precision relative to Bénéjam 267 et al. (2021) 18. Recombinant counts were notably higher for qT1 and qF11 (188 and 432 268 individuals, respectively; Table 2), substantially refining the genetic resolution. 269 Table 2. Summary of number of individuals presenting recombination events in confidence 270 intervals of resistance quantitative trait loci. Recombinant number from Bénéjam et al. (2021) Recombinant number in this study qT1 7 2% 188 9% qT13 16 6% 168 8% qF3 47 17% 364 18% qF11 43 16% 432 21% qF17 30 11% 229 11% Progeny size 266 1,970 271 2. Phenotypic variation and symptom characterization 272 Phenotypes revealed a clear segregation pattern within the extended ‘TxF’ progeny. In the ‘Vi- 273 B04’ experiment, out of the 1,970 seedlings, AUDPC values showed a distribution with a 274 bimodal tendency (Fig 1A), with a large subset of ~300 seedlings displaying near-zero values, 275 indicative of strong resistance. The distribution of AUDPC showed a maximum value of 109 276 and a median of 44.6. When individuals carrying the favourable qT1 allele (n = 1,249) were 277 removed, the remaining distribution became unimodal and continuous (Fig 1B). Median .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 278 AUDPC value rose to 63.5, demonstrating the effect of qT1 face to this isolate. In contrast, 279 adjusted AUDPC values obtained with ‘Vi-Z14’ experiments, induced a unimodal distribution 280 across the 1,101 recombinants ranging from 0 to 112 AUDPC values, with a median of 55.8 281 (Fig 1C). The increase of median AUDPC in ‘Vi-Z14’ experiments illustrates the ability of the 282 ‘09BCZ014’ to partially overcome the QTL qT1. 283 Fig 1. Distribution of susceptibility symptoms of the extended 'TxF' progeny. Adjusted Area 284 Under the Disease Progress Curves (AUDPC) variation for the experiment with ‘EU-B04’ isolate 285 (A) and experiment with the ’09BCZ014’ isolate (C) are shown. Distribution of AUDPC 286 estimated with ‘EU-B04’ isolate (B) is also represented in the subpopulation qT1- (i.e., 287 individuals selected as not carrying the resistance allele of the major quantitative trait locus 288 qT1). 289 Additionally, leaf resistance responses were scored in the ‘Vi-B04’ experiment through the 290 assessment of chlorosis and crispation, each ranging from 0 to 7. Over 80% of genotypes 291 exhibited both reactions (scores > 1). These two symptoms were strongly correlated (r = 0.649, 292 p < 0.001) and inversely associated with disease severity (AUDPC; Fig 2). However, a 293 substantial amount of genotypes was resistant (low AUDPC values) without presenting 294 resistance symptoms, as illustrated by the low correlation scores between AUDPC and 295 resistance phenotypes (r = -0.405 and -0.368). 296 Fig 2. Relationship between resistance and susceptibility symptoms of the extended 'TxF' 297 progeny after inoculation with isolate ‘EU-B04’. Pearson coefficient of correlation (Corr) are 298 indicated for each couple of variables. Three variables measured in ‘Vi-B04’ experiment are 299 presented: Area Under the Disease Progress Curve (AUDPC), chlorosis and crispation scoring. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 300 3. QTL validation and fine mapping 301 QTL mapping analyses with ‘Vi-B04’ experiment variables were conducted on the entire 302 progeny (1,970 seedlings) while only recombinants (1,101 seedlings) were considered for the 303 analysis of AUDPC from ‘Vi-Z14’ experiment. Four of the five targeted QTLs were validated 304 across experiments and phenotypic variables (qT1, qF11, qF17, and qT13) (Fig 3, S1 Table and 305 S1 Fig). For the QTL qF3, crispation variable reached the detection threshold (LOD > 1.8) with 306 a maximum value of 3.3, anchored by the virtual marker c3.loc15 (15 cM). Another signal was 307 detected at the border of the studied interval (20 cM) with a value of 1.88 obtained with 308 AUDPC from ‘Vi-Z14’ experiments. Despite dense marker coverage, this variable did not lead 309 to the refinement of qF3 region and other phenotypic variables (chlorosis and AUDPC from 310 ‘Vi-B04’ experiment) failed to reach the detection threshold (S1 Table and S1 Fig). 311 Fig 3. Genetic maps and LOD curves of the fine-mapped scab resistance quantitative trait 312 loci of the extended 'TxF' progeny. Genetic position (in centiMorgans) and marker name used 313 in this study are respectively indicated on the left and right of linkage groups (LG). Marker 314 name associated to maximum LOD value for each QTL are bolded. 2-LOD and 1-LOD support 315 QTL confidence interval are represented by vertical lines and solid rectangles, respectively. 316 The QTL qT1 was detected in ‘Vi-B04’ experiment for all phenotypic traits, displayed the 317 highest significance (LOD = 202) when using chlorosis as the trait. Its 2-LOD CI was narrowed 318 to 1.1 cM (~600 kb on the GDDH13 genome), anchored by marker AX-115385376. Similar 319 intervals were estimated with AUDPC and crispation variables from this experiment. 320 For qF11 and qF17, the shortest 2-LOD CIs were obtained with AUDPC from ‘Vi-Z14’ 321 experiments. The qF11 region was delimited to a 6 cM region (~2.5 Mb), while qF17 was .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 322 refined to 3.5 cM (~1.2 Mb). The most probable position of these QTLs (i.e. the positions with 323 highest LOD scores) were anchored by markers AX-115185079 and AX-115185087 for qF11 324 (5.2 cM), and by the virtual marker c17.loc12 (12 cM) for qF17. These two QTLs were also 325 identified in the ‘Vi-B04’ experiment using AUDPC and exhibited longer intervals. 326 Finally, qT13 was detected exclusively under ‘Vi-Z14’ inoculation, within a 7 cM (~2 Mb) region 327 on chromosome 13, anchored by the AX-115187347 marker. Although it did not reach the LOD 328 threshold, the AUDPC from the ‘Vi-B04’ experiment yielded a LOD score of 1.78 on this marker. 329 4. Epistatic interaction between qF11 and qF17 330 Consistent with prior findings, seedlings carrying favorable alleles at both qF11 and qF17 331 exhibited significantly lower AUDPC values than all other genotypic classes (p < 0.001; S2 Fig). 332 Under ‘Vi-B04’, median AUDPC dropped from 50–60 in single-allele carriers to ~15 in double 333 carriers; under ‘Vi-Z14’, values decreased from ~60 to ~30. Also, single-allele exhibited 334 susceptibility levels similar to those of non-carriers. These patterns confirm a synergistic effect 335 between qF11 and qF17. 336 5. Identification of candidate genes and defense mechanisms 337 To pinpoint plausible resistance determinants, we focused on genes located within the refined 338 CIs and showing differential expression in relevant genotypic classes (before and five days 339 after inoculation). 340 qT1 region 341 The narrow 590 kb qT1 interval encompassed 89 annotated genes, 10 of which were 342 differentially expressed (Table 2). Most of these were upregulated following inoculation in .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 343 resistant (qT1+) genotypes. Notably, five clustered genes (MD01G1178700, MD01G1179300, 344 MD01G1179700, MD01G1179800, and MD01G1180500) encode Leucine-Rich Repeat 345 Receptor-Like Proteins (LRR-RLPs), classical sensors of fungal or oomycete invasion. This 346 transcriptional pattern supports the hypothesis that qT1 could represent either a functional 347 variant or a paralog of Rvi6 with partial quantitative activity. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 348 Table 2. Annotation and relative expression of genes underlying the QTL qT1 confidence interval before (0 dpi) or five days (5 dpi) after 349 infection with Vi in different genetic background of pooled-samples from the extended ‘TxF’ progeny. 0 dpi 5 dpi 0 dpi versus 5 dpi QTL Gene ID Annotation Swissprot TAIR Name Biological process NoScabQTL/ qT1 NoScabQTL/ qT1qF11qF17 qF11qF17/ qT1qF11qF17 NoScabQTL/ qT1 NoScabQTL/ qT1qF11qF17 qF11qF17/ qT1qF11qF17 NoScabQTL qT1 qF11qF17 qT1qF11qF17 MD01G1173000 Phytosulfokine receptor 1 AT1G74190.1 Signal transduction 2,4 - - - - - - 2,2 - - MD01G1177000 Ethylene- responsive transcription factor 2 AT4G17500.1 Cellular process / response to nematode - - - - - - - 1,6 - - MD01G1177100 Ethylene- responsive transcription factor 5 AT5G51190.1 Response to heat - - - 1,67 - - - - - - MD01G1178700 LRR receptor-like serine/threonine -protein kinase FLS2 AT2G34930.1 Defense response to fungus and oomycetes - - - - - - - 2,7 - 2,6 MD01G1178900 Receptor-like protein 12 AT1G07390.3 Signal transduction - - - - - - - 2,7 - - MD01G1179300 LRR receptor-like serine/threonine -protein kinase GSO2 AT2G34930.1 Defense response to fungus and oomycetes - - - - - - 1,7 1,5 - - MD01G1179700 Probable LRR receptor-like serine/threonine -protein kinase IRK AT2G34930.1 Defense response to fungus and oomycetes - - - - - - - 2,8 - 5,4 MD01G1179800 Receptor-like protein 12 AT2G34930.1 Defense response to fungus and oomycetes - - - - - - - 2,3 - - MD01G1180500 Receptor-like protein 12 AT2G34930.1 Defense response to fungus and oomycetes - - -1,5 - - - - 2,3 - 2,4 qT1 MD01G1181100 Probable inactive leucine-rich repeat receptor kinase XIAO AT2G34930.1 Defense response to fungus and oomycetes - 2,3 - 3,19 3 3,3 - - - - .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 351 qF11 and qF17 regions 352 The refined qF11 (2.6 Mb) and qF17 (1 Mb) intervals contained 304 and 130 genes, 353 respectively. Fourteen were differentially expressed in each region (Table 3). In qF11 interval, 354 several DEGs (MD11G1025200–MD11G1033300) were constitutively upregulated in resistant 355 genotypes at both time points, suggesting baseline activation of signaling pathways rather 356 than induced defense. Their functions, which are linked to RNA interference and signal 357 transduction, suggest the existence of a regulatory layer that modulates defense gene 358 networks by inducing kinases, ligases, and helicases. 359 Conversely, two DEGs in qF17 interval ( MD17G110420, MD17G1105000) were constitutively 360 downregulated in resistant individuals. Although their annotations remain uncertain (one 361 glucosyltransferase and one uncharacterized protein), predicted enzymatic roles could cover 362 various functions. Interestingly, several DEGs in both regions were induced by inoculation, but 363 none of them was specific to the qF11qF17 genotypic class. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 364 Table 3. Annotation and relative expression of genes underlying the QTL qF11 and qF17 confidence intervals before (0 dpi) or five days (5 dpi) 365 infection with Vi in different genetic background of pooled-samples from the extended ‘TxF’ progeny. 0 dpi 5 dpi 0 dpi versus 5 dpi QTL Gene ID Annotation Swissprot TAIR Name Biological process NoScabQTL/ qF11qF17 NoScabQTL/ qT1qF11qF17 qT1/ qT1qF11qF17 NoScabQTL/ qF11qF17 NoScabQTL/ qT1qF11qF17 qT1/ qT1qF11qF17 NoScabQTL qT1 qF11qF17 qT1qF11qF17 MD11G1022100 - - - 2,4 2,2 - - - - - - - - MD11G1025200 G-type lectin S- receptor-like serine/threonine- protein kinase LECRK3 AT5G60900.1 - 3,4 3,3 2,2 3,4 4,2 2,1 - - - - MD11G1025300 DEAD-box ATP- dependent RNA helicase 20 AT1G55150.1 RNAi-mediated immune response 8,7 8,7 8,8 8,5 8,5 7,7 - - - - MD11G1025500 G-type lectin S- receptor-like serine/threonine- protein kinase LECRK3 AT5G60900.1 - 2,7 2,4 - 2,4 3,3 - - 1,8 - 1,5 MD11G1026100 DEAD-box ATP- dependent RNA helicase 20 AT1G55150.1 RNAi-mediated immune response 10,3 10,2 10,4 10,8 10,9 12,4 - - - - MD11G1027500 Beta-glucosidase 12 AT2G44480.1 Carbohydrates metabolic process - - - - - - 1,8 2,5 - - MD11G1027800 TIR domain- containing adapter molecule 1 AT4G12070.1 - - - - - - - - 2,2 - 2,7 MD11G1031100 F-box protein CPR30 AT4G12560.1 Regulation of defense response - - - - 2,6 - - - - - MD11G1031200 #N/A #N/A - - - - -2,0 - - - - - - MD11G1033300 Serine/threonine- protein kinase STY17 AT3G22750.1 Signal transduction 6,0 6,0 3,4 5,3 4,5 2,9 - - - - MD11G1038900 Fructose- bisphosphate aldolase 5, cytosolic AT4G26530.1 Carbohydrates metabolic process - - - -1,6 - - - - - - qF11 MD11G1039300 Heavy metal- associated isoprenylated plant protein 16 AT3G07600.1 - - -3,5 -3,8 - - - - - - - .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint MD11G1043900 C6 finger domain transcription factor tcpZ AT1G07745.1 Cellular process: DNA repair - - -3,1 - - - - - - - MD11G1049600 Cytochrome P450 714C2 AT5G24910.1 - - - - - - - - - - 2,8 MD17G1098400 Trophinin AT5G07570.1 - 1,5 - - - - - - - - - MD17G1099000 WRKY transcription factor 42 AT4G04450.1 Leaf senescence regulation - - - - - - - - - 1,7 MD17G1099500 Cellulose synthase- like protein B3 AT2G32530.1 Cellulose biosynthetic process - - - - - - - 2,0 - - MD17G1099600 Cellulose synthase- like protein H1 AT2G32530.1 Cellulose biosynthetic process - - - - - - - 1,7 - - MD17G1100100 Fatty acid oxidation complex subunit alpha AT4G20720.1 - - 1,6 - - - - - - - - MD17G1101300 Condensin complex subunit 2 AT2G32590.1 Mitotic chromosome condensation - 2,1 3,7 - - - - - - - MD17G1101700 UPF0753 protein YbcC AT1G77130.1 Plat secondary cell wall biogenesis - - - - - - - 1,8 - - MD17G1102000 Ribonuclease 3-like protein 3 AT4G15417.1 RNA processing - - - - - - - 2,5 - - MD17G1104200 UDP- glucose:glycoprotei n glucosyltransferase AT5G04700.1 - -12,2 -12,0 -11,8 -5,9 -8,4 -8,5 - - - - MD17G1104700 DEAD-box ATP- dependent RNA helicase 45 AT3G09620.1 mRNA splicing - - - - - -5,0 - - - - MD17G1105000 Uncharacterized protein DDB_G0275933 AT5G16060.1 - -2,0 -2,1 -1,8 -2,1 -2,2 -2,2 - - - - MD17G1105800 Palmitoyl- monogalactosyldiac ylglycerol delta-7 desaturase, chloroplastic AT3G15850.1 Fatty acid biosynthetic process - - - - - - - - - 1,6 MD17G1106300 1- aminocyclopropane -1-carboxylate oxidase AT1G05010.1 Ethylene biosynthetic process / response to fungus - - - - - - - 2,0 - 2,1 qF17 MD17G1109600 Chlorophyll a-b binding protein 40, chloroplastic AT1G29930.1 Photosynthesis / response to light stimulus - - 4,3 - - - - 3,6 - - .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 367 qT13 region 368 qT13 spanned a 2 Mb interval containing 289 genes, 22 of which lacked functional annotation 369 (S2 Table). .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 370 Discussion 371 The broadening of the biparental population and the analysis of nearly 2,000 seedlings 372 allowed the refinement and validation of several resistance QTLs previously reported in the 373 ‘TxF' progeny. Among the five QTLs studied, qT1, qF11, qF17, and qT13 were validated under 374 at least one experimental condition, while qF3 could not be confirmed despite an increased 375 number of individuals. The ‘Vi-B04’ experiment revealed a somewhat bimodal distribution of 376 resistance phenotypes mainly driven by the presence of the qT1 QTL. The joint analysis of 377 differential gene expression within refined CIs identified several candidate genes, including 378 RLPs in qT1 and constitutively expressed genes in qF11 and qF17. Epistatic interaction 379 between qF11 and qF17 was confirmed, supporting the functional association of their 380 underlying genes in resistance expression. 381 Experimental validation limits and interpretation of QTL stability 382 The absence of qF3 detection and the variability observed in QTL expression across 383 experiments underline the context-dependence of QTL detection. For qF3, QTL detection was 384 successful only for the crispation variable, whereas no QTL were identified for chlorosis, in 385 contrast to other detections (Fig 3) and despite their correlation (Fig 2). In addition, given the 386 low LOD score, this result should be interpreted with caution and confirmed in independent 387 experiments with crispation variable. Although the overall results are mixed, we observed an 388 increase in LOD values towards the end of the targeted genomic region, which was even more 389 pronounced when using AUDPC from the ‘Vi-Z14’ experiment (S1 Fig). This pattern may 390 suggest that the QTL is located slightly downstream of the initially targeted region. Such an 391 interpretation would be consistent with previous findings, in which the position of the highest 392 LOD scores varied among QTL detections performed with two different Vi isolates 18. Although, .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 393 Bénéjam et al. (2021) 18 studied 267 individuals, out of them 17% presented recombination 394 events in qF3 interval, which is similar to our study (Table 2). However, without biological 395 replicates, it is possible that phenotypic responses were less precise in this study, leading to 396 an absence of detection of low effect QTLs, such as qF3. Another possible explanation is that 397 qF3 may have been a false positive in the previous study, in which it was detected for the first 398 time 18. Further analyses using a higher number of markers and replicates would be required 399 to clarify this point. 400 The non-validation of qT13 in ‘Vi-B04’ experiment contrasts with its previous identification by 401 Bénéjam et al. (2021) 18 in both isolate contexts. However, the improved precision of its 402 mapping interval supports the robustness of the locus itself. In addition, QTL detection with 403 AUDPC from ‘Vi-B04' experiment almost reach the LOD threshold (Fig 3). Increasing the level 404 of replication in future experiments, notably with the isolate ‘EU-B04’ could help refine the 405 localization of the QTL by estimating more reliable AUDPC values. The current detection of 406 qT13 only in ‘Vi-Z14’ experiments may also indicate isolate-specific activation or 407 environmental modulation of gene expression. Despite the reduction of the confidence 408 interval, 283 genes still fall within the boundaries of qT13 (S2 Table). With only seven markers 409 on this QTL in this study, it would be relevant to further increase marker density in this region, 410 particularly around the marker AX-115187347 (5.89 cM), since the peak of LOD score was 411 located near it (S1 Table). Moreover, no transcriptomic data are available for this QTL, 412 preventing the initiation of gene number reduction and the investigation of potential gene 413 functions. Due to its efficacy against at least isolate ‘09BCZ014’, ‘TN10-8’ remains an attractive 414 parent for breeding. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 415 In addition, several non-genetic factors may have contributed to this instability. Differences in 416 environmental conditions between experiments, and particularly the use of rootstocks for 417 individual replication in earlier studies, could have modified physiological parameters and 418 biased disease symptom expression. Ontogenic effects and developmental stage of seedlings 419 at inoculation are known to strongly influence the expression of resistance loci, as reported in 420 apple and other perennial hosts 17,47,48. Here, scab inoculations were carried out sequentially 421 on the same plants, which were pruned between each inoculation. Similarly, Soufflet-Freslon 422 et al. (2008) 17 reported differences in AUDPC distributions of a progeny when screened 423 successively in two scab experiments using the same Vi isolate, highlighting differential 424 resistance expression depending on the physiological state of the plants. Collectively, these 425 results suggest that the variability in QTL validation partly reflects the ontogenic and polygenic 426 nature of apple scab resistance. 427 Functional interpretation and convergence of molecular mechanisms 428 The co-localization of differentially expressed genes (DEGs) with refined QTL intervals can 429 reveal potential functions of these loci. The working hypothesis underlying this analysis is that 430 the QTL effect results from allele-specific differences in gene expression, caused by sequence 431 variation in regulatory regions of the underlying gene. Under this model, differential 432 transcription between alleles would drive the observed phenotypic variation. Alternative 433 mechanisms cannot be excluded, including coding sequence variation leading to structural 434 differences in the encoded protein without expression differences 49, or epigenetic regulation 435 generating allele-specific expression in the absence of sequence polymorphism 50. In this 436 context, RNA-seq data provide a relevant framework to investigate the expression-based .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 437 hypothesis and to prioritize genes within the QTL CI, notably by excluding genes not expressed 438 under our experimental conditions. 439 For qT1, due to its specificity, large-effect and physical localization on the genome, it has 440 already been hypothesized to be an allelic variant or a paralog of the major/R gene Rvi6 16,18. 441 Moreover, qT1 often leads to a hypersensitive response 20 and transcriptomic datasets 442 highlighted a large amount of DEGs (~1,500) when comparing genotypes carrying 443 susceptibility and resistance alleles of qT1 infected with ‘EU-B04’ isolate 29. Here, we refined 444 the CI of qT1 to a ~600 kb region (Fig 3, S1 Table) and analyzed DEGs within this interval. More 445 precisely, several RLPs were upregulated upon Vi inoculation (Table 2), a pattern consistent 446 with canonical R-gene–mediated perception 51. This supports the hypothesis that qT1 might 447 be an allele or a paralog of the well-known Rvi6 locus, with different specificities 52. A 448 comparative genomic analysis of the potential allelic series in Rvi6 region, comprising Rvi6 449 cloned in 'Florina' 8,9,53 and also present in recently sequenced ‘Prima’ and ‘Priscilla’ 54, Rvi17 450 in 'Antonovka' 55, Vhc1 in 'Honeycrisp' 56 and qT1 in 'TN 10-8', would help confirm this co- 451 localization and clarify their possible redundancy or evolutionary divergence. A definitive test 452 would examine segregation in progeny from crosses between parents carrying different 453 putative alleles of these R genes/QTLs. Detection of both “alleles” in offspring would indicate 454 recombination, showing they are distinct loci rather than alternative alleles. In a perennial 455 species, such crosses are labor- and time-intensive, and testing all pairwise allele combinations 456 is practically difficult. 457 For qF11 and qF17, CIs were narrowed to 2.6 Mb and 1 Mb (Fig 3, S1 Table), respectively. Their 458 epistatic interaction was also confirmed (S2 Fig), consistent with other studies 18–20. 459 Integration of RNA-seq data provided additional insight into the biological mechanisms .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 460 potentially underlying the interaction between these QTLs, possibly through coordinated 461 transcriptional regulation. Within the qF11 interval, genes putatively involved in RNAi 462 signaling were expressed, whereas the qF17 interval showed repression of genes with 463 currently unknown functions (Table 3). In both regions, expression differences were primarily 464 associated with genotypic classes rather than with Vi inoculation, supporting the hypothesis 465 that these QTLs are constitutively expressed and may contribute to basal resistance 20. 466 Nevertheless, improved gene annotation will be necessary to better characterize the 467 molecular basis linking these two loci. The use of haplotype-resolved parental genomes would 468 provide an ideal framework for identifying genes underlying QTLs. The causal gene may be 469 absent from one haplotype in the case of hemizygous genes, or even missing from the 470 reference genome if the sequenced genotype does not carry the QTL allele. 471 Conclusions 472 The development of new markers and the screening of nearly 2,000 seedlings helped 473 consolidate our understanding of the genetic architecture of scab resistance in the ‘TxF' 474 progeny, validating four QTLs with a promising breadth of action. The improved precision of 475 QTL intervals, comprised between 2.6 Mb and 600 kb, and integration of transcriptomic data 476 strengthen the genetic framework for downstream applications. Future efforts should focus 477 on (i) confirming the co-localization between qT1 and Rvi6, (ii) functionally characterizing 478 qT13, and (iii) elucidating the biological basis of the epistatic interaction between qF11 and 479 qF17 integrating other phenotypic data. In breeding terms, the combination of these loci 480 through marker-assisted selection, supported by the development of new SNP markers, 481 represents a promising path toward durable resistance in apple. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 482 Author contributions 483 Romane Lapous: Investigation, Data Curation, Validation and Visualization of results, Writing 484 – Original Draft Preparation, Review & Editing. Camille Haquet: Investigation, Data Curation, 485 Formal analysis, Writing – Review & Editing. Caroline Denancé: Investigation, Data Curation. 486 Juliette Bénéjam: Validation and Visualization of results, Writing – Review & Editing. Laure 487 Perchepied: Validation and Visualization of results, Writing – Review & Editing. Kaat Hellyn: 488 Resources. Hélène Muranty: Conceptualization, Funding Acquisition, Writing – Review & 489 Editing. Charles-Eric Durel: Conceptualization, Funding Acquisition, Writing – Review & 490 Editing. Julie Ferreira de Carvalho: Investigation, Conceptualization, Funding Acquisition, 491 Project Administration, Supervision, Writing – Original Draft Preparation, Review & Editing. 492 Acknowledgements 493 This work was supported by (i) a grant to the METAdiVERSE project from the BAP (Plant biology 494 and breeding) division of INRAE and by (ii) a grant from the French government managed by 495 the Agence Nationale de la Recherche (ANR) as part of the Programme Prioritaire de 496 Recherche “Cultiver et Protéger Autrement” under the reference ANR-20-PCPA-0003 497 (CapZeroPhyto project). Romane Lapous was supported by a Ph.D. fellowship from the BAP 498 division of INRAE and the ‘Pays de la Loire’ region (France). Camille Haquet was supported by 499 a GIS Fruit grant. The authors greatly thank the PHENOTIC platform 500 (https://doi.org/10.17180/YKBZ-2V85) for carefully taking care of the plant material. The 501 authors would like to thank the Biological Resource Center “RosePom - Pome Fruits and 502 Roses” (https://eng-irhs.angers-nantes.hub.inrae.fr/shared-facilities/genetic-resources/crb- 503 pome-fruit-and-rose) and associated staff for maintaining the plant material and associated 504 datasets used in the present article. We are also grateful to the Horticole Experimental Unit .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 505 (https://doi.org/10.15454/1.5573931618268674E12) for maintaining the young trees derived 506 from seedlings. For generating the seeds used in this study, we thank present and former staff 507 of the VaDiPom team (IRHS). We thank Dr. Diego Micheletti for sharing the alignment of 508 sequencing data to the GGDH13 genome. For selecting and developing SNP markers, we thank 509 Aurélien Petiteau and the ANAN platform (SFR QuaSaV, Univ Angers). Finally, the authors wish 510 to thank M. Tiret for thorough reviewing of the original draft as well as all the people who 511 helped in the sampling of leaf material and apple scab symptom scoring: L. Vitteaut, A. 512 Daligault, A. Petiteau, B. Petit, R. Leclair, F. Lebreton. 513 References 514 1. Bowen JK, Mesarich CH, Bus VGM, Beresford RM, Plummer KM, Templeton MD. 515 Venturia inaequalis: the causal agent of apple scab. Molecular Plant Pathology. 516 2011;12(2):105-122. doi:10.1111/j.1364-3703.2010.00656.x 517 2. Jha G, Thakur K, Thakur P. The Venturia Apple Pathosystem: Pathogenicity 518 Mechanisms and Plant Defense Responses. BioMed Research International. 519 2009;2009(1):680160. doi:10.1155/2009/680160 520 3. Didelot F, Caffier V, Orain G, Lemarquand A, Parisi L. Sustainable management of scab 521 control through the integration of apple resistant cultivars in a low-fungicide input 522 system. Agriculture, Ecosystems & Environment. 2016;217:41-48. 523 doi:10.1016/j.agee.2015.10.023 524 4. Gessler C, Patocchi A, Sansavini S, Tartarini S, Gianfranceschi L. Venturia inaequalis 525 Resistance in Apple. Critical Reviews in Plant Sciences. 2006;25(6):473-503. 526 doi:10.1080/07352680601015975 527 5. Bus VGM, Rikkerink EHA, Caffier V, Durel CE, Plummer KM. Revision of the 528 Nomenclature of the Differential Host-Pathogen Interactions of Venturia inaequalis 529 and Malus. Annu Rev Phytopathol . 2011;49(1):391-413. doi:10.1146/annurev-phyto- 530 072910-095339 531 6. Khajuria YP, Kaul S, Wani AA, Dhar MK. Genetics of resistance in apple against Venturia 532 inaequalis (Wint.) Cke. Tree Genetics & Genomes. 2018;14(2):16. doi:10.1007/s11295- 533 018-1226-4 534 7. Švara A, De Storme N, Carpentier S, Keulemans W, De Coninck B. Phenotyping, genetics 535 and ‘-omics’ approaches to unravel and introgress enhanced resistance against apple 536 scab (Venturia inaequalis) in apple cultivars (Malus × domestica). Horticulture 537 Research. Published online January 10, 2024:uhae002. doi:10.1093/hr/uhae002 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 538 8. Vinatzer BA, Patocchi A, Gianfranceschi L, et al. Apple Contains Receptor-like Genes 539 Homologous to the Cladosporium fulvum Resistance Gene Family of Tomato with a 540 Cluster of Genes Cosegregating with Vf Apple Scab Resistance. MPMI. 2001;14(4):508- 541 515. doi:10.1094/MPMI.2001.14.4.508 542 9. Belfanti E, Silfverberg-Dilworth E, Tartarini S, et al. The HcrVf2 gene from a wild apple 543 confers scab resistance to a transgenic cultivated variety. Proc Natl Acad Sci USA. 544 2004;101(3):886-890. doi:10.1073/pnas.0304808101 545 10. Vinatzer BA, Patocchi A, Tartarini S, Gianfranceschi L, Sansavini S, Gessler C. Isolation 546 of two microsatellite markers from BAC clones of the Vf scab resistance region and 547 molecular characterization of scab-resistant accessions in Malus germplasm. Plant 548 Breeding. 2004;123(4):321-326. doi:10.1111/j.1439-0523.2004.00973.x 549 11. Brown SK, Maloney KE. An Update on Apple Cultivars, Brands and Club-Marketing. . 550 NUMBER. 2013;21(1). 551 12. Parisi L, Durel CE, Laurens F. A New Race of Venturia inaequalis Virulent to Apples with 552 Resistance due to the Vf Gene. Phytopathology. 1993;83(5):533. doi:10.1094/Phyto- 553 83-533 554 13. Lemaire C, De Gracia M, Leroy T, et al. Emergence of new virulent populations of apple 555 scab from nonagricultural disease reservoirs. New Phytologist. 2016;209(3):1220- 556 1229. doi:10.1111/nph.13658 557 14. Pilet-Nayel ML, Moury B, Caffier V, et al. Quantitative Resistance to Plant Pathogens in 558 Pyramiding Strategies for Durable Crop Protection. Front Plant Sci . 2017;8(1838). 559 doi:10.3389/fpls.2017.01838 560 15. Durel CE, Parisi L, Laurens F, Van De Weg WE, Liebhard R, Jourjon MF. Genetic 561 dissection of partial resistance to race 6 of Venturia inaequalis in apple. Genome. 562 2003;46(2):224-234. doi:10.1139/g02-127 563 16. Calenge F, Faure A, Goerre M, et al. Quantitative Trait Loci (QTL) Analysis Reveals Both 564 Broad-Spectrum and Isolate-Specific QTL for Scab Resistance in an Apple Progeny 565 Challenged with Eight Isolates of Venturia inaequalis . Phytopathology®. 566 2004;94(4):370-379. doi:10.1094/PHYTO.2004.94.4.370 567 17. Soufflet-Freslon V, Gianfranceschi L, Patocchi A, Durel CE. Inheritance studies of apple 568 scab resistance and identification of Rvi14 , a new major gene that acts together with 569 other broad-spectrum QTL. Francki M, ed. Genome. 2008;51(8):657-667. 570 doi:10.1139/G08-046 571 18. Bénéjam J, Ravon E, Gaucher M, Brisset MN, Durel CE, Perchepied L. Acibenzolar- S - 572 Methyl and Resistance Quantitative Trait Loci Complement Each Other to Control 573 Apple Scab and Fire Blight. Plant Disease. 2021;105(6):1702-1710. doi:10.1094/PDIS- 574 07-20-1439-RE .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 575 19. Caffier V, Lasserre-Zuber P, Giraud M, et al. Erosion of quantitative host resistance in 576 the apple ×Venturia inaequalis pathosystem. Infection, Genetics and Evolution. 577 2014;27:481-489. doi:10.1016/j.meegid.2014.02.003 578 20. Laloi G, Vergne E, Durel CE, Le Cam B, Caffier V. Efficiency of pyramiding of three 579 quantitative resistance loci to apple scab. Plant Pathology. 2017;66(3):412-422. 580 doi:10.1111/ppa.12581 581 21. Jaganathan D, Bohra A, Thudi M, Varshney RK. Fine mapping and gene cloning in the 582 post-NGS era: advances and prospects. Theor Appl Genet. 2020;133(5):1791-1810. 583 doi:10.1007/s00122-020-03560-w 584 22. Soriano JM, Madduri M, Schaart JG, et al. Fine mapping of the gene Rvi18 (V25) for 585 broad-spectrum resistance to apple scab, and development of a linked SSR marker 586 suitable for marker-assisted breeding. Mol Breeding. 2014;34(4):2021-2032. 587 doi:10.1007/s11032-014-0159-3 588 23. Pagliarani G, Dapena E, Miñarro M, et al. Fine mapping of the rosy apple aphid 589 resistance locus Dp-fl on linkage group 8 of the apple cultivar ‘Florina.’ Tree Genetics 590 & Genomes. 2016;12(3):56. doi:10.1007/s11295-016-1015-x 591 24. Emeriewen OF, Richter K, Flachowsky H, Malnoy M, Peil A. Genetic Analysis and Fine 592 Mapping of the Fire Blight Resistance Locus of Malus ×arnoldiana on Linkage Group 12 593 Reveal First Candidate Genes. Front Plant Sci . 2021;12. doi:10.3389/fpls.2021.667133 594 25. Švara A, Feulner H, Sakina A, et al. Fine-Mapping of the Vhc1 QTL for Apple Scab 595 Resistance on Linkage Group 1 of ‘Honeycrisp.’ Plant Breeding . Published online 596 September 28, 2025:pbr.70035. doi:10.1111/pbr.70035 597 26. Cubillos FA, Coustham V, Loudet O. Lessons from eQTL mapping studies: non-coding 598 regions and their role behind natural phenotypic variation in plants. Current Opinion 599 in Plant Biology. 2012;15(2):192-198. doi:10.1016/j.pbi.2012.01.005 600 27. Zhou Q, Fu Z, Liu H, et al. Mining novel kernel size-related genes by pQTL mapping and 601 multi-omics integrative analysis in developing maize kernels. Plant Biotechnol J. 602 2021;19(8):1489-1491. doi:10.1111/pbi.13634 603 28. Lapous R, Amegan KE, Caromel B, et al. When Metabolomics Meets Quantitative 604 Genetics: An Integrative Strategy to Elucidate Plant Resistance Mechanisms. Plant, Cell 605 & Environment. 2026;49(3):1712-1727. doi:10.1111/pce.70328 606 29. Bénéjam J, Ferreira De Carvalho J, Ravon E, et al. Phenotyping data coupled with RNA 607 sequencing of apple genotypes exhibiting contrasted quantitative trait loci 608 architecture for apple scab (Venturia inaequalis) resistance. Data in Brief. 609 2024;56:110778. doi:10.1016/j.dib.2024.110778 610 30. Durel CE, Calenge F, Parisi L, et al. AN OVERVIEW OF THE POSITION AND ROBUSTNESS 611 OF SCAB RESISTANCE QTLS AND MAJOR GENES BY ALIGNING GENETIC MAPS OF FIVE .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 612 APPLE PROGENIES. Acta Hortic. 2004;(663):135-140. 613 doi:10.17660/ActaHortic.2004.663.17 614 31. Lê Van A, Gladieux P, Lemaire C, et al. Evolution of pathogenicity traits in the apple 615 scab fungal pathogen in response to the domestication of its host: Pathogen evolution 616 under host domestication. Evolutionary Applications. 2012;5(7):694-704. 617 doi:10.1111/j.1752-4571.2012.00246.x 618 32. Bus VGM, van de Weg WE, Peil A, et al. The role of Schmidt ‘Antonovka’ in apple scab 619 resistance breeding. Tree Genetics & Genomes . 2012;8(4):627-642. 620 doi:10.1007/s11295-012-0470-2 621 33. Muranty H, Denancé C, Feugey L, et al. Using whole-genome SNP data to reconstruct 622 a large multi-generation pedigree in apple germplasm. BMC Plant Biol. 2020;20(1):2. 623 doi:10.1186/s12870-019-2171-6 624 34. Durel CE, Denancé C, Muranty H, Lateur M, Ordidge M. MUNQ and PUNQ – a European 625 and international apple and pear germplasm coding system. Acta Hortic. 626 2023;(1384):471-476. doi:10.17660/ActaHortic.2023.1384.59 627 35. Bianco L, Cestaro A, Sargent DJ, et al. Development and Validation of a 20K Single 628 Nucleotide Polymorphism (SNP) Whole Genome Genotyping Array for Apple (Malus × 629 domestica Borkh). PLOS ONE. 2014;9(10):e110377. 630 doi:10.1371/journal.pone.0110377 631 36. Bianco L, Cestaro A, Linsmith G, et al. Development and validation of the 632 Axiom®Apple480K SNP genotyping array. The Plant Journal. 2016;86(1):62-74. 633 doi:10.1111/tpj.13145 634 37. Daccord N, Celton JM, Linsmith G, et al. High-quality de novo assembly of the apple 635 genome and methylome dynamics of early fruit development. Nat Genet. 636 2017;49(7):1099-1106. doi:10.1038/ng.3886 637 38. Diesh C, Stevens GJ, Xie P, et al. JBrowse 2: a modular genome browser with views of 638 synteny and structural variation. Genome Biol. 2023;24(1):74. doi:10.1186/s13059- 639 023-02914-z 640 39. Caffier V, Patocchi A, Expert P, et al. Virulence Characterization of Venturia inaequalis 641 Reference Isolates on the Differential Set of Malus Hosts. Plant Disease. 642 2015;99(3):370-375. doi:10.1094/PDIS-07-14-0708-RE 643 40. Le Cam B, Sargent D, Gouzy J, et al. Population Genome Sequencing of the Scab Fungal 644 Species Venturia inaequalis, Venturia pirina, Venturia aucupariae and Venturia 645 asperata. G3 Genes|Genomes|Genetics. 2019;9(8):2405-2414. 646 doi:10.1534/g3.119.400047 647 41. Caffier V, Didelot F, Pumo B, Causeur D, Durel CE, Parisi L. Aggressiveness of eight 648 Venturia inaequalis isolates virulent or avirulent to the major resistance gene Rvi6 on .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 649 a non-Rvi6 apple cultivar: Aggressiveness of Venturia inaequalis. Plant Pathology . 650 2010;59(6):1072-1080. doi:10.1111/j.1365-3059.2010.02345.x 651 42. Howard NP, Troggio M, Durel CE, et al. Integration of Infinium and Axiom SNP array 652 data in the outcrossing species Malus × domestica and causes for seemingly 653 incompatible calls. BMC Genomics. 2021;22(1):246. doi:10.1186/s12864-021-07565-7 654 43. Van Ooijen JW. JoinMap® 4, Software for the Calculation of Genetic Linkage Maps in 655 Experimental Populations. Kyazma BV, Wageningen, Netherlands . Published online 656 2006. Accessed May 15, 2025. https://cir.nii.ac.jp/crid/1370287939162096384 657 44. Broman KW, Wu H, Sen S, Churchill GA. R/qtl: QTL mapping in experimental crosses. 658 Bioinformatics. 2003;19(7):889-890. doi:10.1093/bioinformatics/btg112 659 45. Voorrips RE. MapChart: Software for the Graphical Presentation of Linkage Maps and 660 QTLs. Journal of Heredity. 2002;93(1):77-78. doi:10.1093/jhered/93.1.77 661 46. Pelletier S. AnaDiff: A tool for differential analysis of microarrays and RNAseq. 662 Published online April 22, 2022. doi:10.5281/ZENODO.6477917 663 47. Develey-Rivière MP, Galiana E. Resistance to pathogens and host developmental stage: 664 a multifaceted relationship within the plant kingdom. New Phytologist. 665 2007;175(3):405-416. doi:10.1111/j.1469-8137.2007.02130.x 666 48. Gusberti M, Gessler C, Broggini GAL. RNA-Seq Analysis Reveals Candidate Genes for 667 Ontogenic Resistance in Malus-Venturia Pathosystem. PLOS ONE. 2013;8(11):e78457. 668 doi:10.1371/journal.pone.0078457 669 49. Cloutier S, Reimer E, Khadka B, McCallum BD. Variations in exons 11 and 12 of the 670 multi-pest resistance wheat gene Lr34 are independently additive for leaf rust 671 resistance. Front Plant Sci . 2023;13. doi:10.3389/fpls.2022.1061490 672 50. Gravot A, Liégard B, Quadrana L, et al. Two adjacent NLR genes conferring quantitative 673 resistance to clubroot disease in Arabidopsis are regulated by a stably inherited 674 epiallelic variation. Plant Comm. 2024;5(5). doi:10.1016/j.xplc.2024.100824 675 51. Kourelis J, van der Hoorn RAL. Defended to the Nines: 25 Years of Resistance Gene 676 Cloning Identifies Nine Mechanisms for R Protein Function. The Plant Cell . 677 2018;30(2):285-299. doi:10.1105/tpc.17.00579 678 52. Poland JA, Balint-Kurti PJ, Wisser RJ, Pratt RC, Nelson RJ. Shades of gray: the world of 679 quantitative disease resistance. Trends in Plant Science. 2009;14(1):21-29. 680 doi:10.1016/j.tplants.2008.10.006 681 53. Vinatzer BA, Zhang HB, Sansavini S. Construction and characterization of a bacterial 682 artificial chromosome library of apple. Theor Appl Genet. 1998;97(7):1183-1190. 683 doi:10.1007/s001220051008 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 684 54. Watts S, Yates S, Vanderzande S, et al. Haplotype-resolved chromosome-level genome 685 assemblies of nineteen apple (Malus domestica Borkh.) cultivars. Sci Data. 686 2026;13(1):258. doi:10.1038/s41597-026-06583-y 687 55. Švara A, Sun H, Fei Z, Khan A. Chromosome-level phased genome assembly of 688 “Antonovka” identified candidate apple scab-resistance genes highly homologous to 689 HcrVf2 and HcrVf1 on linkage group 1. Akhunov E, ed. G3: Genes, Genomes, Genetics. 690 2024;14(1):jkad253. doi:10.1093/g3journal/jkad253 691 56. Khan A, Carey SB, Serrano A, et al. A phased, chromosome-scale genome of 692 ‘Honeycrisp’ apple (Malus domestica). GigaByte. 2022;2022:gigabyte69. 693 doi:10.46471/gigabyte.69 694 695 Supporting information 696 S1 Fig. Genetic maps and LOD curves of minor variables used for scab resistance quantitative 697 trait loci mapping of the extended TxF progeny. Genetic position (in centiMorgans) and 698 marker name used in this study are respectively indicated on the left and right of linkage 699 groups (LG). Marker name associated to maximum LOD value for each QTL are bolded. 2-LOD 700 and 1-LOD support QTL confidence interval are represented by vertical lines and solid 701 rectangles, respectively. 702 S2 Fig. Distribution of AUDPC scores after scab infection with two isolates according to 703 favorable allele of two epistatic QTL segregating in the extended TxF progeny. Genotypic 704 classes are attributed according to allelic variant of marker associated to maximum LOD score 705 of each QTL. Two variables are presented (adjusted AUDPC from ‘Vi-B04’ and ‘Vi-Z14’ 706 experiments). Significant differences between classes have been evaluated through ANOVA 707 test. 708 S1 Table. Parameters associated with the fine-mapped quantitative trait loci (QTL) identified 709 for scab resistance of the extended TxF progeny. Virtual markers, for which physical positions .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint 710 were estimated, are also included (e.g., ‘c11.loc9’, which is placed on chromosome 11 at 9 711 centiMorgans). | Chr, chromosome, LOD, Logarithm of the odds; R², variance; CI, confidence 712 interval. 713 S2 Table. Gene list included in the refined confidence interval of qT13. Gene names (‘Gene 714 ID’) are extracted from the GDDH13 reference genome. | QTL, Quantitative Trait Loci. 715 S1 File. Markers comprising the genetic map of the extended TxF progeny. Virtual markers, 716 for which physical positions were estimated, are included (e.g., ‘c11.loc9’, which is placed on 717 chromosome 11 at 9 centiMorgans). Physical positions are extracted from the GDDH13 718 reference genome. 719 S2 File. Phenotyping data used for QTL mapping in the extended TxF progeny. Data were 720 acquired across three experiments. One includes the inoculation with the Vi isolate ‘EU-B04’ 721 and three phenotypic variables were scored. The two others experiments were conducted 722 with the inoculation of isolate ‘09BCZ14’ and only sporulation symptoms were measured. 723 Missing values are indicated with an asterisk (*). 724 S3 File. Genotyping data used for QTL mapping in the extended TxF progeny. For each SNP, 725 alleles are coded as 1 or 2. Virtual markers, further imputed in genetic analyses, are not 726 included. Missing values are indicated with an asterisk (*). .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 10, 2026. ; https://doi.org/10.64898/2026.04.08.717319doi: bioRxiv preprint