Robigalia Roundup LXXXV: Machine learning can spot scab resistance before the fungus even shows up

Hello {{first_name | Robigalia readers}},

A new Horizon Europe project is about to get underway, and if your work touches plant health systems, pest risk, or quarantine policy, it's worth knowing about.

PhytoPRISM, short for Platform for supporting a systems approach in Pest RISk Management, is a four-year, €6 million consortium project launching in September 2026. The project is led from the University of Cambridge and brings together fifteen partner organisations across the UK, Europe, and beyond, including Wageningen University, INRAE, Rothamsted Research, and the European and Mediterranean Plant Protection Organisation (EPPO).

The core of PhytoPRISM is a co-designed, data-driven platform that integrates epidemiological modelling with economic and behavioural tools to help National Plant Protection Organisations make better decisions about quarantine pest management. Rather than producing a one-pathogen solution, the platform is designed to be pest-generic and interoperable, capable of real-time scenario simulation and strategy design for different local contexts. The project focuses initially on six high-profile EU quarantine pests and will produce ready-to-use tools, training materials, and policy guidance.

It's also one of the first visible outputs of the UK's renewed association to the Horizon Europe programme, with UK institutions including Cambridge, Warwick, Rothamsted, and Forest Research all part of the consortium.

If you're working on plant biosecurity, IPM systems, or epidemiological modelling and think your research could connect to what PhytoPRISM is building, it's worth watching the project as it gets off the ground.

Read more about PhytoPRISM

Now, onto this week’s edition:

Let’s dive in!

Pecan scab remains one of the most persistent management challenges in commercial pecan production across the southeastern United States, and identifying resistant genotypes faster and more objectively is a standing bottleneck in breeding. An article in press by Kang et al., in BMC Plant Biology, investigates the early metabolic signatures that distinguish scab-resistant from scab-susceptible reactions in pecan leaves, with an eye toward developing biomarkers that could complement or replace microscopy-based phenotyping.

The researchers inoculated the commercially important cultivar 'Desirable' with two isolates of Venturia effusa, one virulent and one avirulent, then sampled leaf tissue from 0 to 7 days post-inoculation. They measured 155 metabolites spanning plant hormones, flavonoids, organic acids, amino acids, and sugars. Rather than testing each metabolite individually, they used a machine learning algorithm called logistic regression with L1 regularisation, which automatically selects the smallest subset of variables that best predicts a given outcome while penalising redundant ones. In plain terms, the model was trained to distinguish resistant from susceptible leaves based on their chemical profiles, and then asked to identify which specific metabolites were doing most of the work.

Three sequential metabolic phases emerged in the resistant reaction. Within the first two days, salicylic acid and HMG-CoA were rapidly upregulated, pointing to early immune activation prioritised over energy metabolism. By days 3 to 4, the resistant reaction shifted toward accumulating flavonoids including astragalin, genistein, biochanin A, apigenin, and luteolin. The susceptible reaction, by contrast, showed delayed and broadly elevated primary metabolism throughout, with salicylic acid eventually surging to levels 7.5 times those in resistant leaves late in infection. That late-stage increase appears to reflect reactive stress rather than proactive defence. Crucially, the machine learning models distinguished resistant from susceptible reactions at 76 to 94% accuracy as early as 1 to 4 days post-inoculation, well ahead of the 5-day minimum required for conventional microscopy to detect fungal structures in stained tissue.

The study is a proof of concept within a single cultivar, so the biomarkers identified reflect this specific host-isolate combination rather than universal resistance markers. The approach itself, pairing metabolite profiling with machine learning to achieve earlier and more objective resistance classification, has clear potential for the wider field of host resistance screening.

Cut a bracket of Ganoderma resinaceum and something remarkable happens: a thick yellow resin oozes from the wound and hardens rapidly into a glossy surface, as though the fungus is sealing itself shut. This distinctive behaviour gave the species its name. First described in 1889 by the French mycologist Jean Louis Émile Boudier, G. resinaceum is the only European Ganoderma that produces annual rather than perennial fruiting bodies, a trait that makes it easy to overlook and harder to track.

Ganoderma resinaceum (family Ganodermataceae, order Polyporales) is a laccate basidiomycete, meaning its bracket-shaped fruiting bodies display that characteristic lacquered, shiny surface. Phylogenetically, it sits within a well-supported "resinaceum subclade" of laccate Ganoderma that spans Europe and parts of Asia, distinct from the closely related North American G. sessile and the widely used medicinal species G. lucidum. Like other members of the genus, it is a white rot fungus, deploying ligninolytic enzymes to selectively break down lignin in both heartwood and sapwood, leaving behind pale, stringy, structurally weakened wood.

The primary hosts are deciduous hardwoods, with Quercus species particularly frequently recorded, alongside willow, hornbeam, hackberry and poplar. The fungus attacks the roots and lower trunk, causing what foresters and arborists call butt rot. Externally, infected trees may show little for years. Above ground, canopy thinning, reduced growth and dieback signal that internal decay is already advanced. The most visible indicator is the bracket conk itself, emerging from the base of the trunk or root flares. Because G. resinaceum produces only annual fruiting bodies rather than the persistent perennial conks typical of relatives like G. applanatum, it can be missed entirely during surveys conducted outside its fruiting window.

The species has a cosmopolitan distribution, recorded across Europe, northern Africa, parts of Asia, Australia and the Americas, though its taxonomy has historically been entangled with misapplied names in the G. lucidum complex. New host records continue to emerge, including recent reports from Iran on Gleditsia caspica, Populus deltoides, Morus alba and Manilkara zapota. Phytopathogenic Ganoderma species collectively account for estimated annual losses exceeding USD 68 billion across agriculture and forestry worldwide, though attributing specific losses to individual species remains difficult.

There is no curative treatment once infection is established. Management relies on prevention: protecting root systems and trunk bases from wounding, maintaining tree vigour, and removing infected stumps and root debris that serve as inoculum reservoirs. Basidiospores are the primary long-range dispersal agent, entering trees through wounds and compromised root bark. Where infected trees pose a structural hazard, removal is the only safe outcome. Risk assessment by a qualified arborist, including sonic tomography to map internal decay, remains the most reliable tool for managing standing trees that have already fruited.

This week, we meet Tonjock Rosemary Kinge, an Associate Professor of Mycology and Phytopathology and Head of the Department of Plant Sciences at the University of Bamenda, Cameroon.

Rosemary grew up accompanying her parents to the farm. Over the seasons she noticed that some harvests came in well while others did not. When she asked her father why yields sometimes failed, he explained that plant diseases were to blame, but when she pressed him for how they might cure those diseases, he had no answer. She went on to study plant pathology to find answers, and to make sure that understanding reached farmers like her family.

At Bamenda, she leads the Fungal Biodiversity, Ecology, Ethnomycology and Phytopathology Research Group (the FBEEP Group), and her current projects focus on the identification and management of fungal pathogens affecting indigenous vegetables, including Talinum triangulare, as well as vegetable seeds and cucurbits.

Alongside this diagnostic work, she investigates the use of medicinal mushrooms as biocontrol agents against fungal crop diseases, and examines their antimicrobial properties more broadly. Fungal diversity and conservation remain a thread running through all of it.

Alongside her own research, Rosemary has invested heavily in training postgraduate students and early-career researchers in molecular diagnostics of fungal pathogens and in scientific and grant writing, with support from the British Society for Plant Pathology, the Society for Molecular Biology and Evolution, the British Mycological Society, the JRS Biodiversity Foundation, and Inqaba Biotec Central Africa.

Students who came through her workshops have gone on to complete Master's and PhD theses with skills and grants they would not otherwise have had. In 2024, the Alexander von Humboldt Foundation recognised this work with a Humboldt Alumni Networking Award for innovative mentoring.

You can connect with Rosemary on LinkedIn and follow the work of the FBEEP Group from there.

Want to be featured in Robigalia? Submit your details in the form.

Have a job, scholarship, or event to advertise? List it in for FREE Robigalia. I’ll help promote your opportunity or event to a global network of over 10,000 plant pathologists.

See you next Monday!

P.S. Why Robigalia? The name originates from the Ancient Roman festival dedicated to crop protection. You can read all about the history here: