A team at the University of Cambridge have developed new tools to predict disease risk in cassava and help protect a vital food security crop in sub-Saharan Africa.
Cassava, a staple food grown widely across sub-Saharan Africa, is severely threatened by diseases such as Cassava Mosaic Disease (CMD) and Cassava Brown Streak Disease (CBSD).
Whiteflies are vectors of the viral pathogens which cause these diseases. Whiteflies feed on the leaves of cassava and transfer the viruses between plants in the field and across neighbouring farmland.
The diseases reduce the quality and quantity of crop harvested, triggering food insecurity and significant economic hardship, with estimated annual crop losses exceeding $1.25 billion in sub-Saharan Africa.
To address this, researchers from the University of Cambridge’s Epidemiology and Modelling group have developed two innovative tools, the EpiPvr R package and the CropMix web application, designed to improve understanding, prediction, and management of cassava disease epidemics across diverse farming landscapes.
Disease transmitters across time and space
The EpiPvr package provides a new model-driven tool that helps researchers understand how plant diseases spread through insects. It works by using data compiled from experiments where insects are monitored when feeding on healthy and infected plants.
Researchers can upload their data, and the tool estimates key factors that influence disease spread, such as how easily insects pick up and pass on infections and how long they stay infectious. It can also predict the risk of local outbreaks from pathogen introductions in the field.
When tested on two cassava viruses in a recent study led by Dr Ruairi Donnelly, Postdoctoral Research Associate at the Department of Plant Sciences, Cambridge, the team found striking differences.
One virus (CBSI) spreads poorly from insects to plants but is dangerous because infections are hard to spot. The other (CMB) spreads easily and causes obvious symptoms, while insects stay infectious for a long time.
As a result, CMB can trigger outbreaks even when insect numbers are low, whereas CBSI poses less risk unless insect numbers are moderate or high.
Cassava disease management scenarios for optimising yield
Investigations led by Dr Israël Tankam, former member of the Department of Plant Sciences, Cambridge and now a Postdoctoral Researcher at the Institut Agro Rennes-Angers in France, complimented these studies.
The team produced ‘CropMix’, a web-based decision support tool built around a model that helps optimise crop variety mixtures to maintain high yields when disease-risk is high. The model compares different cassava types, ‘phytotypes’, that are varieties known to be susceptible, resistant or disease-tolerant, and ‘decoy’ plants that do not host the disease.
CropMix takes into account how the disease spreads, how easily infections can be detected, and how each type performs when healthy or infected. Using these factors, the model identifies the best mix of varieties to manage disease and protect harvests.
When estimates from EpiPvr for the two cassava viruses (CMB and CBSI) were added to CropMix, the model suggested that mixing susceptible cassava with varieties resistant to CBSI works well when whitefly numbers are moderate or low. This approach can significantly protect yields from CBSI outbreaks. Planting non-host ‘decoy’ crops alongside cassava can also help by attracting and cleaning whiteflies, as long as these decoy plants have comparable appeal to whiteflies as cassava.
In contrast, for CMB, the model found that no mixture of varieties could beat planting only tolerant or resistant cassava when it comes to protecting crops from epidemics.
Practical strategies for managing plant diseases
These studies show how modelling tools can guide practical strategies for managing plant diseases.
EpiPvr is now being expanded to predict local CBSI outbreak risks across sub-Saharan Africa by combining its estimates with whitefly population data. CropMix will also be upgraded to handle more realistic planting patterns, moving beyond random mixtures to spatially structured layouts.
Together, these improvements will allow more accurate forecasts of disease risk and management options under future climate conditions. Overall, this suite of tools helps design targeted, sustainable interventions that balance effectiveness with farmer needs—an essential step toward long-term crop protection.
Funding: This work was supported by Bill and Melinda Gates Foundation; UK’s Foreign, Commonwealth and Development Office (FCDO). Dr Donnelly was additionally awarded a BBSRC Flexible Talent Mobility Grant from the University of Cambridge for engagement: ‘Translating cassava virus modelling into guidance for cassava growers and policy-makers’.
References:
Donnelly R., Tankam I., Gilligan C.A., ‘Plant pathogen profiling with the EpiPvr package’. Accepted at Methods in Ecology and Evolution (19 Nov 2025), preprint DOI: 10.32942/X29K9P
Donnelly, R., ‘EpiPvr: Estimating Plant Pathogen Epidemiology Parameters from Laboratory Assays (R package version 0.0.1). Comprehensive R Archive Network (CRAN)’. DOI: 10.32614/CRAN.package.EpiPvr
Tankam, I., Donnelly R., Gilligan, C.A., ‘Optimizing crop varietal mixtures for viral disease management: A case study on cassava virus epidemics’. PLOS Computational Biology, DOI: 10.1371/journal.pcbi.1012842
Image: Cassava root infected by Cassava Brown Streak Disease (CBSD). Credit: Anna Szyniszewska, Epidemiology and Modelling group, University of Cambridge.
Text by Alison Scott-Brown and Ruairi Donnelly.