henhaobio
Sign in

Horticulture ResearchPlant BiologyPeer-reviewed journal article

Investigating the genetic architecture of biotic stress response in stone fruit tree orchards under natural infections with a multi-environment GWAS approach

Marie SerrieiD, Vincent SeguraiD, Alain Blanc, Laurent BruniD, Naïma DlalahiD, Frédéric Gilles, Laure HeurteviniD, Mathilde Le-Pans, Véronique Signoret, Sabrina Viret, Jean-Marc AudergoniD, Bénédicte Quilot‐TurioniD, Morgane RothiD

PDFDOIPreprintCodeData
In-depth full-text digest available · Jump to digest ↓

Abstract

The mapping and introduction of sustainable plant immunity to pests and diseases in fruit tree is still a major challenge in modern breeding. This study aims at deciphering the genetic architecture underlying resistance or tolerance across environments for major pests and diseases in peach (P. persica) and apricot (P. armeniaca). We set up a multi-environment trial (MET) approach by studying two core collections of 206 peach and 150 apricot accessions deployed under low phytosanitary conditions in respectively three and two environmentally contrasted locations in South-East of France. To capture the complex dynamics of pest and disease spread in naturally infected orchards, visual scoring of symptoms was repeated within and between 3 years, for five and two pests and diseases respectively for peach and apricot, resulting in the maximum of damage score and the AUDPC. These traits were used as phenotypic inputs in our genome-wide association studies (GWAS) strategy, and leading to the identification of: i) non-additive genotype-phenotype associations, ii) environment-shared QTLs iii) environment-specific QTLs, and iv) interactive QTLs which changes in direction ( antagonist) or intensity ( differential) according to the environment. By conducting GWAS with multiple methods, we successfully identified a total of 60 high confidence QTLs, leading to the identification of 87 candidate genes, the majority belonging to the Leucine-rich repeat containing receptors (LRR-CRs) family gene. Finally, we provided a comparative analysis of our results on peach and apricot, two closely related species. The present results contribute to the development of genomics-assisted breeding to improve biotic resilience in Prunus varieties.

In-depth summary

Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 1 (17 Oct 2024) · CC BY 4.0

Multi-environment GWAS of 206 peach and 150 apricot accessions under natural infection finds mostly environment-specific or interactive loci, 60 high confidence QTLs and 87 candidate genes.

Durable resistance in long-lived fruit trees likely depends on polygenic traits, yet in Prunus genotype-by-environment (G×E) effects and the genetics of partial resistance to diverse biotic stresses were little studied, and most of the stresses examined had never been mapped. Combining additive, non-additive, single-environment, multi-environment and meta-analysis GWAS on naturally infected orchards separates stable from environment-dependent loci in peach and apricot.

Key results

  1. ShownG×E interactions strongly shaped biotic stress responses in both collections. Location contributed more interaction variance than year for most stresses, but year dominated for blossom blight. Leaf curl was the most stable response across environments and blossom blight the least stable.

    Genotype variance 5.1 to 66.7% of total; genotype-by-location (G×L) 6.3 to 26.5%, reported above genotype-by-year (G×Y), though blossom blight G×Y reached 39.7%; broad-sense heritability 0.07 to 0.78; mean between-environment genetic correlations 0.68 (Max) and 0.73 (AUDPC); blossom blight r = 0.41 (Max) vs leaf curl r = 0.87 (Max) and 0.91 (AUDPC). · A large variability in the phenotypic response…; Biotic stress responses are modulated by G and…; A focus on peach leaf curl and apricot blossom…

  2. ShownAdding dominance, recessive and overdominance codings to the additive model found many loci that additive GWAS missed. Non-additive QTLs made up a larger share in peach than in apricot.

    Across-environment GWAS: 341 QTLs (188 additive, 153 non-additive, 30 found by both); 96 (Max) and 45 (AUDPC) QTLs only from non-additive models; non-additive share 56% in peach vs 39% in apricot; dominance variance up to 16.4% of total variance, highest for apricot rust. · Biotic stress responses are modulated by G and…; Complementarity between additive and…; Similar, yet divergent: difference in the…

  3. ShownMost loci depended on the environment. Single-environment GWAS gave mainly environment-specific QTLs. MTMM and meta-analysis GWAS found interactive loci whose effect size changed across environments (and for blossom blight, often the sign), plus loci the single-environment scans missed.

    Single-environment GWAS: 801 QTLs, 62.5% to 100% environment-specific per trait; only a leaf curl chromosome 4 QTL found in every environment for both variables. MTMM 209 and Meta-GWAS 226 QTLs (107 shared); 29 Meta-GWAS QTLs absent from single-environment results; 61% of significant blossom blight SNPs showed z-score sign inversion. · Identification of environment specific QTLs with…; Modelling QTLs with G×E interaction effects…

  4. ShownKeeping only QTLs detected by at least 50% of the additive GWAS models left 60 high confidence QTLs, which captured a sizeable share of genetic and location-interaction variance. Blossom blight gave none, and some were detectable only with AUDPC.

    High confidence QTLs explained 9.6 to 64.6% of additive genetic variance (average 36.8%), up to 54.4% of G×L and up to 29.3% of G×Y variance; for apricot rust they explained 53.6% of G×L variance; 32% of high confidence QTLs were detected only with AUDPC. · Identification of high confidence QTLs and…; Usefulness of AUDPC and Max variables to study…; High confidence QTLs reveal the role of basal…

  5. SuggestedAnnotation of high confidence QTLs (intervals LD-refined in peach) nominated 87 candidate genes, mostly leucine-rich repeat (LRR) receptors (RLKs, RLPs, NBS-LRR). Three apricot genes each overlapped several rust QTLs; in two, the main haplotypes differed in susceptibility.

    Candidates included seven TMV resistance protein, seven LRR receptor-like kinase, four cysteine-rich receptor-like kinase, four TIR-NBS-LRR and two MLO-like genes; up to 9 for the peach rust chromosome 2 QTL. The two commonest top-SNP haplotypes in PruarM.6G017900 and PruarM.4G204800 differed in Max G-BLUPs (Wilcoxon P < 0.05). · Local LD and annotation analyses within high…; High confidence QTLs reveal the role of basal…; Similar, yet divergent: difference in the…

Methods, data and reproducibility

Trees were scored visually under natural infection, repeatedly within seasons except for blossom blight (scored once), giving per-tree maximum damage (Max) and area under the disease progress curve (AUDPC). Spatially corrected BLUPs were analysed with MM4LMM mixed-model GWAS (additive and non-additive codings), per-environment GWAS, multi-trait mixed model (MTMM) GWAS and MetaGE meta-analysis.

  • Peach core collection (P. persica plus P. davidiana, P. mira, P. kansuensis) in three South-East France orchards, scored 2021 to 2023 for five pests and diseases · Prunus persica and wild relatives · n = 206 accessions (192 genotyped) · doi:10.57745/2HNRN0
  • Apricot core collection in two South-East France orchards, scored 2020 to 2023 for blossom blight and rust · Prunus armeniaca · n = 150 accessions · doi:10.57745/2HNRN0
  • Peach IRSC 16K SNP array genotypes after quality filtering (10,082 markers kept for non-additive GWAS) · Prunus persica · n = 13,680 SNPs
  • Apricot Illumina HiSeq 2000 resequencing SNPs aligned to the third version of the Marouch genome (25,508 markers kept for non-additive GWAS) · Prunus armeniaca · n = 207,000 SNPs
  • Reference genomes for candidate gene search: Prunus persica Whole Genome v2.0 (annotation v2.1) and Prunus armeniaca Marouch v1.0, Genome Database for Rosaceae · Prunus persica; Prunus armeniaca
Methods & tools
MM4LMM (single-locus linear mixed model GWAS), Multi-trait mixed model (MTMM), MetaGE (meta-analysis GWAS), SNMF (population structure), poolR (Meff effective SNP number for Bonferroni), VCFTools v0.1.16, LDcorSV (structure- and kinship-corrected LD), IRSC 16K SNP array, Illumina HiSeq 2000, JBrowse, Genome Database for Rosaceae, Field Book Android application, R 4.1.2
Software & environment
R 4.1.2; VCFTools v0.1.16; R packages MM4LMM, SNMF, poolR, MetaGE and LDcorSV (versions not given); field data captured with the Field Book Android app.
Reproducibility gaps
Versions of most R packages are not stated. Model details are left to Supplementary Methods, which are not in the text. No sequence archive accession is cited for the apricot resequencing reads.
Limitations (6)
  • Apricot blossom blight fluctuated strongly between years, so no high confidence QTL was found despite a heritability of 0.51. The authors say the epidemiological context (bloom-time weather, flower density) needs to be modelled better. · stated by the authors
  • Allelic effects and candidate genes are not functionally validated. The authors say they must be validated before use in marker-assisted selection, and that validating G×E QTLs is rare and needs demanding work such as gene cloning. · stated by the authors
  • Results overlap little with earlier biparental QTL studies (leaf curl on linkage groups 3 and 6; powdery mildew Vr2/Vr3 on chromosomes 8 and 2). Major alleles that are rare in a core collection may go undetected, and markers, models and phenotyping also differ. · stated by the authors
  • Environments were not characterized by envirotyping, so the climatic variables that modulate QTL effects remain unidentified. The authors propose envirotyping to find them and to group locations. · stated by the authors
  • Long-range LD in peach (decay 508 to 1620 Kbp) left QTL intervals up to 3240 Kbp holding 52 to 528 genes. Peach candidates were chosen by functional annotation inside LD-trimmed intervals, so they remain poorly resolved and untested. · noted in this summary
  • Coverage was uneven across traits: MTMM could not be fitted for 4 variables (both powdery mildew variables, AUDPC for peach rust and leafhopper), and 19.6% of the data were dropped for weak symptom expression, unbalancing environments per trait. · noted in this summary

Read this if you breed or map disease resistance in perennial crops and want a worked design combining additive, non-additive, multi-environment and meta-analysis GWAS under natural infection.

Citations
20
FWCI
10.40
Field percentile
top 1%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISGao '26Labbancz '26Osuna-Cabal '24Vanalli '23Li '21Santos '22
60% semantic matchPlant Biology
Yuhao Gao, W B Wang, Y H Liu+16 more·bioRxiv·May 12, 2026