New tools for predicting spread of Fusarium wilt in banana
Lead Research Organisation:
University of Warwick
Department Name: School of Life Sciences
Abstract
Banana (Musa spp) is the 5th most important global food staple, but as a consequence of its clonal nature, banana production suffers significantly from pathogens. In the last banana apocalypse, Fusarium oxysporum var. cubense Race 1 (Foc1) destroyed global production of "Gros Michel". India, the biggest producer of bananas in the world (~25% global production), consumes nearly all its production. The highly virulent Fusarium Wilt Tropical Race 4 (Foc4) arrived in India in 2018 and in 2019 it was found in South America, threatening both food security (India is the world's largest banana producer) and the trade in banana commodities.
This project has two key objectives.
(i) Generating robust molecular markers that can be used in field for identification of TR4 and inform management strategies.
(ii) Developing satellite-based remote sensing approaches will be developed for automatic demarcation of banana plantations, and identification of plant stress.
Methods:
Next generation sequencing
Big data analysis - satellite radar
Hyperspectral imaging
Development of diagnostic molecular markers.
Development of LAMP assays for in field detection of Fusarium.
This will involve training in handling large datasets, both genomic (Fusarium; comparative genomics across different races and geographical/population structures) and geo-sattelite.
to develop intra-specific genomic markers that can be used to track spread of individual pathogen genotypes across space and time. These TR4-specific sequences, will be used to develop PCR- and LAMP-based diagnostic assays, the latter of which could be deployed in the field using a battery-powered portable device.
The remote sensing studies (from month 18) will use the European Space Agency's Sentinel Synthetic Aperature Radar (SAR) to develop radar signatures for banana plantations, using Google Earth imagery and plantation maps from partner organizations. In the first instance Costa Rica data will be used for training and statistical clustering algorithms such as Random Forest used for classification. Open-access, spectral data from LANDSAT will be used to develop vegetation indices for plantations, to understand spatial and temporal variation in plant health. Higher resolution imagery will be acquired using UAVs with multispectral scanners, both to validate satellite imagery and to avoid problems with image acquisition due to cloud cover. For ground truthing drone data collection, we aim to include some glasshouse experimentation to monitor effects of diseases on reflectance, due to changes in leaf chemistry or morphology.
This project has two key objectives.
(i) Generating robust molecular markers that can be used in field for identification of TR4 and inform management strategies.
(ii) Developing satellite-based remote sensing approaches will be developed for automatic demarcation of banana plantations, and identification of plant stress.
Methods:
Next generation sequencing
Big data analysis - satellite radar
Hyperspectral imaging
Development of diagnostic molecular markers.
Development of LAMP assays for in field detection of Fusarium.
This will involve training in handling large datasets, both genomic (Fusarium; comparative genomics across different races and geographical/population structures) and geo-sattelite.
to develop intra-specific genomic markers that can be used to track spread of individual pathogen genotypes across space and time. These TR4-specific sequences, will be used to develop PCR- and LAMP-based diagnostic assays, the latter of which could be deployed in the field using a battery-powered portable device.
The remote sensing studies (from month 18) will use the European Space Agency's Sentinel Synthetic Aperature Radar (SAR) to develop radar signatures for banana plantations, using Google Earth imagery and plantation maps from partner organizations. In the first instance Costa Rica data will be used for training and statistical clustering algorithms such as Random Forest used for classification. Open-access, spectral data from LANDSAT will be used to develop vegetation indices for plantations, to understand spatial and temporal variation in plant health. Higher resolution imagery will be acquired using UAVs with multispectral scanners, both to validate satellite imagery and to avoid problems with image acquisition due to cloud cover. For ground truthing drone data collection, we aim to include some glasshouse experimentation to monitor effects of diseases on reflectance, due to changes in leaf chemistry or morphology.