Accurate prediction of huanglongbing occurrence in citrus plants by machine learning-based analysis of symbiotic bacteria
Huanglongbing (HLB), the most prevalent citrus disease worldwide, is responsible for substantial yield and economic losses. Phytobiomes, which have critical effects on plant health, are associated with HLB outcomes. The development of a refined model for predicting HLB outbreaks based on phytobiome...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2023-05-01
|
| Series: | Frontiers in Plant Science |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fpls.2023.1129508/full |
| _version_ | 1827939454020485120 |
|---|---|
| author | Hao-Qiang Liu Ze-long Zhao Hong-Jun Li Shi-Jiang Yu Lin Cong Li-Li Ding Chun Ran Xue-Feng Wang |
| author_facet | Hao-Qiang Liu Ze-long Zhao Hong-Jun Li Shi-Jiang Yu Lin Cong Li-Li Ding Chun Ran Xue-Feng Wang |
| author_sort | Hao-Qiang Liu |
| collection | DOAJ |
| description | Huanglongbing (HLB), the most prevalent citrus disease worldwide, is responsible for substantial yield and economic losses. Phytobiomes, which have critical effects on plant health, are associated with HLB outcomes. The development of a refined model for predicting HLB outbreaks based on phytobiome markers may facilitate early disease detection, thus enabling growers to minimize damages. Although some investigations have focused on differences in the phytobiomes of HLB-infected citrus plants and healthy ones, individual studies are inappropriate for generating common biomarkers useful for detecting HLB on a global scale. In this study, we therefore obtained bacterial information from several independent datasets representing hundreds of citrus samples from six continents and used these data to construct HLB prediction models based on 10 machine learning algorithms. We detected clear differences in the phyllosphere and rhizosphere microbiomes of HLB-infected and healthy citrus samples. Moreover, phytobiome alpha diversity indices were consistently higher for healthy samples. Furthermore, the contribution of stochastic processes to citrus rhizosphere and phyllosphere microbiome assemblies decreased in response to HLB. Comparison of all constructed models indicated that a random forest model based on 28 bacterial genera in the rhizosphere and a bagging model based on 17 bacterial species in the phyllosphere predicted the health status of citrus plants with almost 100% accuracy. Our results thus demonstrate that machine learning models and phytobiome biomarkers may be applied to evaluate the health status of citrus plants. |
| first_indexed | 2024-03-13T08:54:44Z |
| format | Article |
| id | doaj.art-e1db68636cbc407da80cd10fac9f7ae2 |
| institution | Directory Open Access Journal |
| issn | 1664-462X |
| language | English |
| last_indexed | 2024-03-13T08:54:44Z |
| publishDate | 2023-05-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Plant Science |
| spelling | doaj.art-e1db68636cbc407da80cd10fac9f7ae22023-05-29T04:21:26ZengFrontiers Media S.A.Frontiers in Plant Science1664-462X2023-05-011410.3389/fpls.2023.11295081129508Accurate prediction of huanglongbing occurrence in citrus plants by machine learning-based analysis of symbiotic bacteriaHao-Qiang Liu0Ze-long Zhao1Hong-Jun Li2Shi-Jiang Yu3Lin Cong4Li-Li Ding5Chun Ran6Xue-Feng Wang7Citrus Research Institute, Southwest University/Chinese Academy of Agricultural Sciences, National Engineering Research Center for Citrus, Chongqing, ChinaShanghai BIOZERON Biotechnology Co., Ltd., Shanghai, ChinaCitrus Research Institute, Southwest University/Chinese Academy of Agricultural Sciences, National Engineering Research Center for Citrus, Chongqing, ChinaCitrus Research Institute, Southwest University/Chinese Academy of Agricultural Sciences, National Engineering Research Center for Citrus, Chongqing, ChinaCitrus Research Institute, Southwest University/Chinese Academy of Agricultural Sciences, National Engineering Research Center for Citrus, Chongqing, ChinaCitrus Research Institute, Southwest University/Chinese Academy of Agricultural Sciences, National Engineering Research Center for Citrus, Chongqing, ChinaCitrus Research Institute, Southwest University/Chinese Academy of Agricultural Sciences, National Engineering Research Center for Citrus, Chongqing, ChinaCitrus Research Institute, Southwest University/Chinese Academy of Agricultural Sciences, National Engineering Research Center for Citrus, Chongqing, ChinaHuanglongbing (HLB), the most prevalent citrus disease worldwide, is responsible for substantial yield and economic losses. Phytobiomes, which have critical effects on plant health, are associated with HLB outcomes. The development of a refined model for predicting HLB outbreaks based on phytobiome markers may facilitate early disease detection, thus enabling growers to minimize damages. Although some investigations have focused on differences in the phytobiomes of HLB-infected citrus plants and healthy ones, individual studies are inappropriate for generating common biomarkers useful for detecting HLB on a global scale. In this study, we therefore obtained bacterial information from several independent datasets representing hundreds of citrus samples from six continents and used these data to construct HLB prediction models based on 10 machine learning algorithms. We detected clear differences in the phyllosphere and rhizosphere microbiomes of HLB-infected and healthy citrus samples. Moreover, phytobiome alpha diversity indices were consistently higher for healthy samples. Furthermore, the contribution of stochastic processes to citrus rhizosphere and phyllosphere microbiome assemblies decreased in response to HLB. Comparison of all constructed models indicated that a random forest model based on 28 bacterial genera in the rhizosphere and a bagging model based on 17 bacterial species in the phyllosphere predicted the health status of citrus plants with almost 100% accuracy. Our results thus demonstrate that machine learning models and phytobiome biomarkers may be applied to evaluate the health status of citrus plants.https://www.frontiersin.org/articles/10.3389/fpls.2023.1129508/fullcitrus microbiomeHuanglongbingmachine learningmeta-analysiscommunity assembly |
| spellingShingle | Hao-Qiang Liu Ze-long Zhao Hong-Jun Li Shi-Jiang Yu Lin Cong Li-Li Ding Chun Ran Xue-Feng Wang Accurate prediction of huanglongbing occurrence in citrus plants by machine learning-based analysis of symbiotic bacteria Frontiers in Plant Science citrus microbiome Huanglongbing machine learning meta-analysis community assembly |
| title | Accurate prediction of huanglongbing occurrence in citrus plants by machine learning-based analysis of symbiotic bacteria |
| title_full | Accurate prediction of huanglongbing occurrence in citrus plants by machine learning-based analysis of symbiotic bacteria |
| title_fullStr | Accurate prediction of huanglongbing occurrence in citrus plants by machine learning-based analysis of symbiotic bacteria |
| title_full_unstemmed | Accurate prediction of huanglongbing occurrence in citrus plants by machine learning-based analysis of symbiotic bacteria |
| title_short | Accurate prediction of huanglongbing occurrence in citrus plants by machine learning-based analysis of symbiotic bacteria |
| title_sort | accurate prediction of huanglongbing occurrence in citrus plants by machine learning based analysis of symbiotic bacteria |
| topic | citrus microbiome Huanglongbing machine learning meta-analysis community assembly |
| url | https://www.frontiersin.org/articles/10.3389/fpls.2023.1129508/full |
| work_keys_str_mv | AT haoqiangliu accuratepredictionofhuanglongbingoccurrenceincitrusplantsbymachinelearningbasedanalysisofsymbioticbacteria AT zelongzhao accuratepredictionofhuanglongbingoccurrenceincitrusplantsbymachinelearningbasedanalysisofsymbioticbacteria AT hongjunli accuratepredictionofhuanglongbingoccurrenceincitrusplantsbymachinelearningbasedanalysisofsymbioticbacteria AT shijiangyu accuratepredictionofhuanglongbingoccurrenceincitrusplantsbymachinelearningbasedanalysisofsymbioticbacteria AT lincong accuratepredictionofhuanglongbingoccurrenceincitrusplantsbymachinelearningbasedanalysisofsymbioticbacteria AT liliding accuratepredictionofhuanglongbingoccurrenceincitrusplantsbymachinelearningbasedanalysisofsymbioticbacteria AT chunran accuratepredictionofhuanglongbingoccurrenceincitrusplantsbymachinelearningbasedanalysisofsymbioticbacteria AT xuefengwang accuratepredictionofhuanglongbingoccurrenceincitrusplantsbymachinelearningbasedanalysisofsymbioticbacteria |