Deep learning empowers genomic selection of pest-resistant grapevine

Abstract

Crop pests significantly reduce crop yield and threaten global food security. Conventional pest control relies heavily on insecticides, leading to pesticide resistance and ecological concerns. However, crops and their wild relatives exhibit varied levels of pest resistance, suggesting the potential for breeding pest-resistant varieties. This study integrates deep learning (DL)/machine learning (ML) algorithms, plant phenomics, quantitative genetics, and transcriptomics to conduct genomic selection (GS) of pest resistance in grapevine. Building deep convolutional neural networks (DCNNs), we accurately assess pest damage on grape leaves, achieving 95.3% classification accuracy (VGG16) and a 0.94 correlation in regression analysis (DCNN-PDS). The pest damage was phenotyped as binary and continuous traits, and genome resequencing data from 231 grapevine accessions were combined in a Genome-Wide Association Studies, which maps 69 quantitative trait locus (QTLs) and 139 candidate genes involved in pest resistance pathways, including jasmonic acid, salicylic acid, and ethylene. Combining this with transcriptome data, we pinpoint specific pest-resistant genes such as ACA12 and CRK3, which are crucial in herbivore responses. ML-based GS demonstrates a high accuracy (95.7%) and a strong correlation (0.90) in predicting pest resistance as binary and continuous traits in grapevine, respectively. In general, our study highlights the power of DL/ML in plant phenomics and GS, facilitating genomic breeding of pest-resistant grapevine.

Acknowledgements

We thank members of the Zhou lab at AGIS for discussion and comments on the project. This work was supported by the National Key Research and Development Program of China (No. 2023YFD2200702), the project of National Key Laboratory for Tropical Crop Breeding (No. NKLTCB202325), the National Natural Science Foundation of China (No. 32372662), and the Science Fund Program for Distinguished Young Scholars of the National Natural Science Foundation of China (Overseas) to Yongfeng Zhou.

Author contributions

Y.Z. designed the project. Y.Z., Z.L., and Y.W. supervised the project. Y.G. and Z.L. designed and trained image recognition and object detection models. Y.G., F.Z., and Q.X. performed the bioinformatic analyses. X.W., H.X., W.L., L.S., Y.Z., and C.L. assisted in bioinformatics analyses. Y.G., X.S., and W.M. provided essential technology support. Q.L. and A.M. provided the plant material for genome analyses. Y.G., Z.L., and Y.W. wrote the draft. Y.L., X.F., C.Z., P.Z., C.L., and Z.Z. revised the manuscript. All authors contributed to manuscript preparation and read, commented on, and approved the manuscript.

Data availability

All resequencing data generated have been deposited in the NCBI database under the BioProject: PRJNA994294.

Code availability

All scripts performed in this study are available on Github: https://github.com/zhouyflab/Pest-Resistance.

Conflict of interest statement

The authors declare no competing interests.

Supplementary data

Supplementary data is available at Horticulture Research online.

References

[1] Oerke EC. Crop losses to pests. J Agric Sci. 2006; 144:31-43

[2] Simon J-C, Peccoud J. Rapid evolution of aphid pests in agricul-tural environments. Curr Opin Insect Sci. 2018; 26:17-24

[3] He Z, Guo JF, Reitz SR. et al. A global invasion by the thrip, Frankliniella occidentalis: current virus vector status and its man-agement. Insect Sci. 2020; 27:626-45

[4] Beddington J. Food security: contributions from science to a new and greener revolution. Philos Trans R Soc B Biol Sci. 2010; 365:61-71

[5] Mahmood I, Imadi SR, Shazadi K. et al. Effects of pesticides on environment. Plant Soil Microbes. 2016;1:253- 69 implications in crop science

[6] Forgash AJ. History, evolution, and consequences of insecticide resistance. Pestic Biochem Physiol. 1984; 22:178-86

[7] Gatehouse JA. Plant resistance towards insect herbivores: a dynamic interaction. New Phytol. 2002; 156:145-69

[8] War AR, Paulraj MG, War MY. et al. Herbivore-and elicitor-induced resistance in groundnut to Asian armyworm, Spodoptera litura (Fab.) (Lepidoptera: Noctuidae). Plant Signal Behav. 2011; 6: 1769-77

[9] Kundu A, Mishra S, Vadassery J. Spodoptera litura-mediated chem-ical defense is differentially modulated in older and younger systemic leaves of Solanum lycopersicum. Planta. 2018; 248:981-97

[10] War AR, Paulraj MG, Ahmad T. et al. Mechanisms of plant defense against insect herbivores. Plant Signal Behav. 2012; 7: 1306-20

[11] Verhage A, van Wees SC, Pieterse CM. Plant immunity: it’s the hormones talking, but what do they say? Plant Physiol. 2010; 154: 536-40

[12] Ahmad M, Ghaffar A, Rafiq M. Host plants of leaf worm, Spodoptera litura (Fabricius) (Lepidoptera: Noctuidae) in Pakistan. Asian J Agric Biol. 2013; 1:23-8

[13] EFSA Panel on Plant Health (PLH), Bragard C, Dehnen-Schmutz

[14] K. et al. Pest categorisation of Spodoptera litura. EFSA J. 2019; 17:e05765

[15] Balikai R, Kotikal Y. Pest status of grapevine in northern Kar-nataka. Agric Sci Dig. 2003; 23:276-8

[16] Jagginavar S, Biradar A, Bagali A. Menace of Spodoptera litura Fab. on grape. Insect Environ. 1999; 5:76-7

[17] Wang X, Liu Z, Zhang F. et al. Integrative genomics reveals the polygenic basis of seedlessness in grapevine. Curr Biol. 2024; 34:3763-3777.e5

[18] Liu Z, Wang N, Su Y. et al. Grapevine pangenome facili-tates trait genetics and genomic breeding. Nat Genet. 2024; 56: 2804-14

[19] Xiao H, Wang Y, Liu W. et al. Impacts of reproductive systems on grapevine genome and breeding. Nat Commun. 2025; 16:2031

[20] Bock CH, Chiang K-S, Del Ponte EM. Plant disease severity estimated visually: a century of research, best practices, and opportunities for improving methods and practices to maximize accuracy. Trop Plant Pathol. 2022; 47:25-42

[21] Laxa M, Liebthal M, Telman W. et al. The role of the plant antioxidant system in drought tolerance. Antioxidants. 2019; 8:94

[22] Buffara CRS, Angelotti F, Vieira RA. et al. Elaboration and valida-tion of a diagrammatic scale to assess downy mildew severity in grapevine. Ciência Rural. 2014; 44:1384-91

[23] Varshney RK, Bohra A, Roorkiwal M. et al. Fast-forward breeding for a food-secure world. Trends Genet. 2021; 37:1124-36

[24] LeCun Y, Bengio Y, Hinton G. Deep learning. Nature. 2015; 521: 436-44

[25] Mohanty SP, Hughes DP, Salathé M. Using deep learning for image-based plant disease detection. Front Plant Sci. 2016; 7:1419

[26] Sladojevic S, Arsenovic M, Anderla A. et al. Deep neural networks based recognition of plant diseases by leaf image classification. Comput Intell Neurosci. 2016; 2016:1-11

[27] Samanta A, Das G, Das SK. Roles of flavonoids in plants. Carbon. 2011; 100:12-35

[28] Ecker JR, Davis RW. Plant defense genes are regulated by ethy-lene. Proc Natl Acad Sci. 1987; 84:5202-6

[29] Rodriguez PL. Protein phosphatase 2C (PP2C) function in higher plants. Plant Mol Biol. 1998; 38:919-27

[30] Schweighofer A, Kazanaviciute V, Scheikl E. et al. The PP2C-type phosphatase AP2C1, which negatively regulates MPK4 and MPK6, modulates innate immunity, jasmonic acid, and ethylene levels in Arabidopsis. Plant Cell. 2007; 19:2213-24

[31] Jeong JC, Shin D, Lee J. et al. Isolation and characterization of a novel calcium/calmodulin-dependent protein kinase, AtCK, from Arabidopsis. Mol Cell ( Springer Science & Business Media BV). 2007; 24:276-82

[32] Miyamoto T, Uemura T, Nemoto K. et al. Tyrosine kinase-dependent defense responses against herbivory in Arabidopsis. Front Plant Sci. 2019; 10:776

[33] Díaz-Riquelme J, Zhurov V, Rioja C. et al. Comparative genome-wide transcriptome analysis of Vitis vinifera responses to adapted and non-adapted strains of two-spotted spider mite, Tetranyhus urticae. BMC Genomics. 2016; 17:1-15

[34] Van Damme M, Huibers RP, Elberse J. et al. Arabidopsis DMR6 encodes a putative 2OG-Fe (II) oxygenase that is defense-associated but required for susceptibility to downy mildew. Plant J. 2008; 54:785-93

[35] Rajavel A, Klees S, Schlüter JS. et al. Unravelling the complex interplay of transcription factors orchestrating seed oil content in Brassica napus L. Int J Mol Sci. 2021; 22:1033

[36] Fotouhi N, Fischer-Stettler M, Lenzoni G. et al. ACA pumps maintain leaf excitability during herbivore onslaught. Curr Biol. 2022; 32:2517-2528.e6

[37] Karn A, Zou C, Brooks S. et al. Discovery of the REN11 locus from Vitis aestivalis for stable resistance to grapevine powdery mildew in a family segregating for several unstable and tissue-specific quantitative resistance loci. Front Plant Sci. 2021; 12: 733899

[38] Possamai T, Wiedemann-Merdinoglu S. Phenotyping for QTL identification: a case study of resistance to Plasmopara viticola and Erysiphe necator in grapevine. Front Plant Sci. 2022; 13:930954

[39] Zhu Q, Zhang J, Gao X. et al. The Arabidopsis AP2/ERF transcrip-tion factor RAP2. 6 participates in ABA, salt and osmotic stress responses. Gene. 2010; 457:1-12

[40] Klimecka M, Muszy ńska G. Structure and functions of plant calcium-dependent protein kinases. Acta Biochim Pol. 2007; 54: 219-33

[41] Goverse A, Smant G, Bouwman L. et al. The BIOEXPLOIT project. Potato Res. 2009; 52:209-13

[42] Xiao H, Liu Z, Wang N. et al. Adaptive and maladaptive introgression in grapevine domestication. Proc Natl Acad Sci. 2023; 120:e2222041120

[43] Pound MP, Atkinson JA, Townsend AJ. et al. Deep machine learn-ing provides state-of-the-art performance in image-based plant phenotyping. Gigascience. 2017;6:gix083

[44] Arya S, Sandhu KS, Singh J. et al. Deep learning: As the new frontier in high-throughput plant phenotyping. Euphytica. 2022; 218:47

[45] Tang Z, Wang M, Schirrmann M. et al. Affordable high through-put field detection of wheat stripe rust using deep learning with semi-automated image labeling. Comput Electron Agric. 2023; 207:107709

[46] Guan H, Fu C, Zhang G. et al. A lightweight model for efficient identification of plant diseases and pests based on deep learn-ing. Front Plant Sci. 2023; 14:1227011

[47] Dray FA Jr, Goldstein SL. A novel method for quantifying insect damage to plant storage organs. Entomol Exp Appl. 2020; 168: 113-7

[48] Stotz HU, Pittendrigh BR, Kroymann J¨l. et al. Induced plant defense responses against chewing insects. Ethylene signaling reduces resistance of Arabidopsis against Egyptian cotton worm but not diamondback moth. Plant Physiol. 2000; 124:1007-18

[49] Horgan FG, Romena AM, Bernal CC. et al. Stem borers revisited: host resistance, tolerance, and vulnerability determine levels of field damage from a complex of Asian rice stemborers. Crop Prot. 2021; 142:105513

[50] Johnson SN, Riegler M. Root damage by insects reverses the effects of elevated atmospheric CO2 on eucalypt seedlings. PLoS One. 2013; 8:e79479

[51] Clohessy JW, Sanjel S, O’Brien GK. et al. Development of a high-throughput plant disease symptom severity assessment tool using machine learning image analysis and integrated geoloca-tion. Comput Electron Agric. 2021; 184:106089

[52] Tong H, Nikoloski Z. Machine learning approaches for crop improvement: leveraging phenotypic and genotypic big data. J Plant Physiol. 2021; 257:153354

[53] Redmon J, Divvala S, Girshick R. et al. You only look once: uni-fied, real-time object detection. arXiv e-prints arXiv:1506.02640. 2015

[54] Simard PY, Steinkraus D, Platt JC.Best practices for convolu-tional neural networks applied to visual document analysis. In: Seventh International Conference on Document Analysis and Recogni-tion, 2003. Proceedings. IEEE Computer Society: Washington, DC, United States, 2003, 958-63

[55] Abadi M, Agarwal A, Barham P. et al. Tensorflow: large-scale machine learning on heterogeneous distributed systems. arXiv preprint arXiv:1603.04467. 2016

[56] Krizhevsky A, Sutskever I, Hinton G. ImageNet classification with deep convolutional neural networks. Adv Neural Inf Process Syst. 2012;25:

[57] Simonyan K,Zisserman A. Very deep convolutional networks for large-scale image recognition. arXiv e-prints arXiv:1409.1556. 2014

[58] He K, Zhang X, Ren S. et al. Deep residual learning for image recognition. arXiv e-prints arXiv:1512.03385. 2015

[59] Szegedy C, Vanhoucke V, Ioffe S. et al. Rethinking the inception architecture for computer vision. arXiv e-prints arXiv:1512.00567. 2015

[60] Huang G, Liu Z, van der Maaten L. et al. Densely con-nected convolutional networks. arXiv e-prints arXiv:1608.06993. 2016

[61] Srivastava N, Hinton G, Krizhevsky A. et al. Dropout: a simple way to prevent neural networks from overfitting. JMachLearn Res. 2014; 15:1929-58

[62] Kingma DP,Ba J. Adam: a method for stochastic optimization. arXiv e-prints arXiv:1412.6980. 2014

[63] Ruder S.An overview of gradient descent optimization algo-rithms. arXiv e-prints arXiv:1609.04747. 2016

[64] Weiss K, Khoshgoftaar TM, Wang D. A survey of transfer learn-ing. J Big Data. 2016; 3:1-40

[65] Ketkar N, Ketkar N. Introduction to Keras. In: Ketkar N (ed.), Deep Learning with Python:A Hands-on Introduction. Apress: Berkeley, CA, 2017,97-111

[66] Lathuilière S, Mesejo P, Alameda-Pineda X. et al. A comprehen-sive analysis of deep regression. arXiv e-prints arXiv:1803.08450. 2018

[67] Krizhevsky A, Sutskever I, Hinton GE. ImageNet classifica-tion with deep convolutional neural networks. Commun ACM. 2017; 60:84-90

[68] Liu X, Chi M, Zhang Y. et al. Classifying high resolution remote sensing images by fine-tuned VGG deep networks. In:IGARSS 2018-2018 IEEE International Geoscience and Remote Sensing Sympo-sium. IEEE: New York, United States, 2018,7137-40

[69] Pearson K. Note on regression and inheritance in the case of two parents. Proc R Soc London Ser I. 1895; 58:240-2

[70] Shi X, Cao S, Wang X. et al. The complete reference genome for grapevine (Vitis vinifera L.) genetics and breeding. Hortic Res. 2023;10:uhad061

[71] Zhou X, Stephens M. Genome-wide efficient mixed-model anal-ysis for association studies. Nat Genet. 2012; 44:821-4

[72] Leinonen R, Sugawara H, Shumway M. et al. The sequence read archive. Nucleic Acids Res. 2010;39:D19-21

[73] Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics. 2014; 30:2114-20

[74] Martin M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet J. 2011; 17:10-2

[75] Dobin A, Davis CA, Schlesinger F. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics. 2013; 29:15-21

[76] Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014; 15:1-21