These conference reports are written by the beneficiaries of our travel fund.
Click here to read more about the fund and apply yourself
The 10th International Epidemiology Workshop, Geneva, New York, USA 6-12 June 2009
The first of these workshops was held in 1963, and just preceded the publication of Van der Plank’s influential textbook. It had a self-conscious aim of defining the discipline, raising its status and creating a network that would improve the management of plant diseases.
Food shortages were within hungry memory in Europe and regular or looming in large parts of Asia and Africa; Borlaug was busy shuttling wheat between central Mexico and the coast, the green revolution still a hope; and the structure of DNA had been known for 10 years. Since then, the workshops have been held irregularly, but always with small numbers and the aim of effective discussion and networking. Just 1 person was present at both the 1st and the 10th workshops: the inspiring figure of Jan Zadoks.
I was enabled to go to this workshop through the generosity of my University and the British Society for Plant Pathology. All who went presented either a poster or a short talk, and all the sessions were as one group.
Session titles were roughly erected around groups of similar talks; panel discussions on groups of posters occupied the second day. The social programme and accommodation was thoroughly well designed to ensure that the group stayed as one group, and mixed freely and productively; most people understood the need to allow part of their slot for discussion. I particularly enjoyed the heavy steel doors of the Cornell Food Science section, built in the early 1960s and dramatically labelled “Fallout Shelter á “.
Written versions of the talks and posters will be published as a booklet by the organisers, so I will try to give a personal view of some emerging themes of the workshop rather than a listing in any form.
Genomics has not yet really hit the subject. Metagenomics will do so, but ways of phrasing questions that require the answers these techniques offer remain scarce. DNA marker-based methods, by contrast, have become routine and dramatically lit up much that was obscure: the curious confirmation by Phytophthora infestans that there is more sex in Sweden than Scotland, for example. Sampling remains a problem, with the marshland between the diagnosticians (“all I need is a picogram, but I can manage very large samples if you like, as much as a gram” – 1 % of 1 ha to a depth of 10 cm is around 1012 g: which one would you choose?), the statisticians (“take a 1% sample”) and the field pathologists (“remember to turn the leaves over, squint and judge it”) still trackless. Bold people have constructed platforms a few metres out from each side of the marsh, but mostly it is still necessary to don large flat shoes, run very fast and hope for the best.
Disease forecasting remains a topic as important as ever , with many contributions on the old favourites of potato blight, apple scab, and the mildews of grape, downy and powdery. Understanding the risk structure of growers, and the sort of information they require, emerged as a key theme, with frequent invocation of Bayesianism as a way in which both sides could understand what they were doing.
Joseph Russo, a meteorologist from the company ZedX, talked about “cyber infrastructure”, and the promise it could give of relatively painless construction of systems integrating weather forecasts, disease intelligence and whatever models we liked across a continent but with local data and output. This has worked well for soybean rust; whether it has promise for a divided Europe with increasingly scarce public advisory input is less clear.
Methodological advances in description and modelling seemed to me largely incremental, but the inclusion of multiple scales is beginning to show real expansion of our understanding in areas where intuition is very poor (until you’ve absorbed the lesson of the model, when it becomes “obvious”).
Some fascinating successes, with important policy implications, were described. Evolution was in evidence at three scales. The population structure and history of pathogens is a still expanding subject (but is it epidemiology, some asked? I don’t care!). The trade-offs and compromises constraining evolution and adaptation to rare or common hosts are not always what one might have expected. And Jan Zadoks’s opening lecture was a challenging description not of an invasion but of a disappearance: yellow rust of rye. This is the inverse of invasion and the first case study of its kind I have heard. The spectres of Huanglongbing, Ug99, new Phytophthora spp. and soybean rust mostly, though not entirely, skulked outside the workshop. At the moment, it seems, we can forecast the way disaster will develop, but not where it will come from or why.
The delight of the workshop for me was Dani Shtienberg’s saving of 14/15 of the sprays on capsicum crops in Israel by careful attention to epidemic dynamics. The modelling was simple but vital; it worked because it was coupled to a good understanding of the system; and the change was adopted because the research was closely linked to smallish growers. Much of the applied work discussed in the workshop was public sector research on problems of smallish growers: the classic research – extension – farmer sequence. But increasing volumes of agricultural production, probably the majority of traded commodities, is entirely privately managed, with very large areas overseen by skilled advisers drawing on trials organisations working in private.
This is as true in the US as in Europe or Indonesia, yet almost everyone at the workshop was public sector in one way or another. It seems a worrisome split.
Once again, I thank the society for its support.
Mike Shaw, University of Reading
About 60 researchers from 15 countries attended the 10th epidemiology workshop, held at Geneva, New York, 7-12 June 2009. The workshop was hosted by the Departments of Plant Pathology and Plant Microbe-Biology at Ithaca (Cornell University campus) and Geneva (The New York State Agricultural Experiment Station). In total, there were 10 oral presentation sessions, covering a range of topics; in addition, there were 6 poster sessions.
Two pathogen systems dominated the workshop: potato late blight and grape powdery mildew, particularly the former. For late blight, several presentations focused on its population variability to infer dispersal characteristics, sexual reproduction, and forecasting. Prof Ristaino demonstrated the importance of applying modern molecular techniques to historical samples to infer population changes over time, speciation and dispersal at a continental scale. This illustrates the importance of detailed records and maintenance of experimental samples/materials. Prof Yuen showed that two mating types exist in Sweden; molecular data suggested that sexual reproduction among two mating types did take place. Several presentations dealt with finetuning disease forecasting models for late blight and delivering forecasts to users. For grapevine powdery mildew, presentations focused on hostphenology-related susceptibility and disease forecasting.
A few presentations from French researchers focused on the trade-off or balanced selection in pathogen virulence traits. Detailed research of fungal traits related to virulence, aggressiveness and transmission on potato late blight (Dr Andrivon) suggested the existence of compromises in the fungal evolution, such as aggressiveness versus transmission. Similarly, aggressiveness of cereal rust was decomposed into different components, which were further characterized for a few selected isolates on several cultivars. Results again suggested a trade-off between the latent period and spore production. Another modeling study suggested that partial resistance may favour interseasonal disease transmission.
Understanding dynamics of interacting populations was dealt with in several presentations. Prof Jeger described a theoretical approach to model a generic biocontrol system. This approach links dynamics of a biocontrol agent with the pathogen dynamics via a specified biocontrol mechanism. Through detailed model analysis, it is possible to identify key characteristics that a successful biocontrol agent should possess. Observed population dynamics of oak powdery mildew can be interpreted in terms of the interactions between the mildew fungus and resident microorganisms (Prof Shaw). Survival analysis of a citrus disease suggested that infection of one isolate may inhibit infection of another isolate (Prof Gottwald).
Several modeling tools were described for their potential uses in epidemiological research. Dr Pautasso described a network-based approach to model disease development, and illustrated the effects of various network characteristics on pathogen dynamics. A new statistical method is being developed to estimate spatial disease dispersal characteristics with a set of sparsely mapped data (Dr Parnell). Prof Madden described an approach (spline smoothing) to analyse temporal epidemic data when there is no definable functional relationship between the variable and time.
Parallel to the modeling techniques, several presentations were devoted to introduce new molecular techniques to solving epidemiological problems. Prof Seems described several promising sensors and nanotechnologies, in particular new development in rapid sequencing of fungal genomes. Multiple soil and foliar pathogens may be simultaneously detected with a microarray (Dr Smart). Dr Garrett described future opportunities for plant disease epidemiologists with anticipated availability of massive microbial community molecular data. Such data may be used in studying taxonomic and functional diversities as well as the relationship of their relative abundances with their functions.
As expected, understanding spatial disease dynamics received considerable attention, including aforementioned work on dispersal of potato late blight on a large scale using molecular techniques. Prof Nutter described how spatial disease development can be described by combining GPS, GIS and remote sensing technologies. Statistical analysis of HLB in citrus suggested a lack of local disease spread but rather disease spread on a regional scale; this may place greater difficulties in controlling this disease. Extensive 3D mapping of healthy and diseased cherry fruit is being undertaking to understand spatial dynamics of cherry brown rot (Prof Scherm). As always, sampling came up as an important topic, not only in the context of obtaining samples for molecular analysis but also for describing spatiotemporal dynamics accurately. Prof Ferrandino discussed in detail how sampling strategy is dependent on plant phenology and sampling objectives, i. e. needing to estimate disease means, variability and/or spatiotemporal distribution. He further described a sampling strategy based on a variable sample size (fractal sampling). Prof Aylor discussed spore dispersal at various spatial scales and the resulting disease patterns.
There were many oral and poster presentations on various disease management strategies, including use of disease forecasting systems. Several papers dealt with development of disease forecasting models on several pathosystems at various stages: development, validation, evaluation and commercial use, including apple Nectria fruit rot, grapevine powdery mildew, cucumber downy mildew and apple scab. Dr Shtienberg presented a study on pepper powdery mildew that illustrated how careful field epidemiological research can result in better disease control but with significantly less input of fungicides.
I am grateful to the British Society for Plant Pathology and the Worshipful Company of Fruiterers for travel funds. Finally, Beijing was chosen as the venue for the next workshop, which will be held in August 2013, the week before the ISPP congress in Beijing.
Xiangming Xu, East Malling Research
