Workshops Friday 18 October
Process-based crop models: practical application and further development
Neil Budko, Corné Verburg, Tamara ten Den, Clemens Stolk
A central element in CropXR is the development of models that not only enhance our understanding of crop development and speed up breeding but can also be reliably applied in farming practice. This workshop, moderated by Neil Budko (Delft University of Technology), was aimed at exploring the calibration of current process-based models, their possible stochastic extension, and potential practical applications of such models.
The first speaker was Clemens Stolk on behalf of the four potato processing companies participating in CropXR Potato. CropXR Potato is the research package withing CropXR aimed at developing models that predict yield and quality of potatoes as a function of genotype, environment and management practices. The potato processing companies have the ambition to develop practical applications based on the scientific models that are expected as output of CropXR Potato. One such application could be a decision support system that provides real-time advice to growers on management interventions in potato cultivation. As a starting point, students of HAS Green Academy have been asking potato farmers how they take cultivation decisions, where they look for information, and what role decision support systems could play. The presentation led to several questions from the audience that will now be taken into account in a follow-up study among Belgian potato growers.
The second speaker was Tamara ten Den. She presented three examples illustrating the calibration needed for crop growth models. As new cultivars are introduced recalibration of crop growth models is required. Calibration has to be based on data collected in (experimental) fields. It is vital to design experimental treatments and measurements to align with the model’s data requirements. The first calibration example demonstrated calibrating a parameter that varied between cultivars, specifically linked to cultivar earliness. The second example focused on a parameter that changed throughout the season but was not cultivar specific. The final example highlighted the complexities in model set up. Some models can simulate a process in multiple ways; yet not all simulation methods are suitable for every situation. Additionally, model sensitivity to parameters may vary between simulation methods. In conclusion it is important to carefully consider and evaluate the model set up each time a model is applied or recalibrated.
Lastly, Corné Verburg (Delft University of Technology) introduced a new approach to modelling stochasticity in crop growth models to account for the realistic variation of crop traits within a genotype and across a field. This approach builds on the equations used in well-studied crop growth models, such as WOFOST, but introduces stochasticity by assuming that the model parameters are stochastic. The Liouville equation is used to propagate these distributions forward in time, illustrating how this leads to variation in observed crop traits. Several examples were discussed, including a modulated exponential growth model and an Arrhenius-equation-based model. Additionally, SINDy, a simple, data-driven method for estimating nonlinear growth rate equations in ODE form, was discussed. The presentation was followed by a discussion that led to new insights, including potential applications and possible collaborations on dataset collection and the further development of existing crop models.
Societal implications
Jonathan Arentoft, Víctor Betriu Yáñez, Koen Beumer, Aisha So, Eva Valencia Leñero, Elsenoor Wijlhuizen
The workshop leaders hosted a session to discuss research on the societal implications of data-driven plant breeding. Starting from the observation that it is important to research the interaction between society and technological development, the societal research in CropXR is structured via 4 themes: public values, regulations, ethics, and innovation system dynamics. During the workshop the participants discussed these different themes in two roundtable sessions.
In the first roundtable the group focused on the way different actors do and do not collaborate and how these collaborations are valued. Using a fishbowl-method to structure the conversation the group discussed two questions: (1) What makes a ‘good’ collaboration? (2) What would be the added value of such collaborations?
Drawing on their own experiences different actors from government, industry and academia all contributed to a rich discussion. A key reason for wanting to collaborate was that solving complex problems is best done by combining and integrating different types of knowledge. At the same time participants emphasizes that collaboration can sometimes be difficult to manage. Collaboration is, therefore, no silver-bullet. Instead, the value of collaboration depends on how we collaborate, why we collaborate, and on what issues.
The second roundtable focused on innovation system surrounding CropXR. The participants were asked to reflect for themselves on three different questions: 1) When would CropXR be a success for you? 2) What experiences from the past – both positive and negative – do you take with you to working in CropXR? 3) What opportunities or barriers do you experience with coming to that ‘success’ in CropXR?
A diversity of answers to the first question demonstrated that different actors bring their own definition of success to CropXR that does not always overlap with others’. These ranged from expanding the consortium by including smaller breeding companies to successfully executing fundamental research in molecular biology, and from planting resilient crops in the field to the integration of societal values in plant breeding. These different views on success were not experienced as a barrier though, but rather as an opportunity for knowledge integration. At the same time, participants were aware of the fact that in the end a project needs a focus. For the successful integration of different kind of goals and knowledge collaboration was deemed necessary, echoing what the particpants expressed during the first roundtable discussion. The outcomes of the roundtables serve as further input for the enhancement and refining of the focus of our research. The workshop leaders thank the participants very much for their input and hope to meet each other soon again!
CropXR growth under stress
By Anna Neeftjes (UvA), Xenja Ploeger (UvA) and Julia Ruiz Capella (UU)
In this workshop the workshop leaders aimed to introduce the three work packages (Wps) that focus on plant response under temperature stress. Furthermore, they tried to stimulate discussions on similarities and differences between their experimental approaches. C2 investigates the defence response under (a)biotic stress combinations, C4 focuses on studying the trade-offs between growth and resilience and C5 studies how plants maintain fruit and seed yield under temperature stress.
Apart from members from the involved WPs, people from other WPs (WP1 and WP3) and members from companies involved in the research done at Crop-XR participated in the workshop (Enza Zaden, Rijk Zwaan and Bejo, among others). After each WP and appointed members were introduced, ideas about points in common, differences and points to agree upon were discussed through a quiz. The results are the following:
Even though all WPs involved will include elevated temperature, the research questions of each of these work packages has a different focus. This led to a fruitful discussion regarding the points that were found important to agree upon in our experimental set-ups, and regarding those aspects that should be specific for every WP.
The main points that were discussed regarding the usage of different temperatures:
- The most relevant temperature that should be defined among WPs is the mild elevated temperature. This should be between 27 and 28 °C. What is to be defined as control and extreme temperatures might be dependent on the tissue that each group is working with and the research question.
- The choice for a specific temperature (regime) is fully dependent on the phenotype that it is pursued, as the temperature treatment should have an effect on the specific phenotypic output.
- The participants discussed the need to come to a consensus regarding other conditions such as day length or treatment length. They concluded that certain aspects might be aligned, but that this too might be dependent on the research question.
NB It is important to keep in mind that temperatures that we are referring to are those set at the growth chamber and might not be the temperature perceived at the different organs of the plant under study.
Regarding other aspects of the experimental set-up, the following was discussed:
- The main point that was highlighted during the discussion regarding the experimental set-up was that it is completely dependent on the research question that each WP pursues. It might not be so relevant to agree on a particular set-up but it definitely is when deciding method/molecular markers that we use to prove that the plant is suffering from a particular stress.
- In case of drought stress a combination of stop watering followed by providing less water to maintain lower water content in the soil in the most common. However, this will not be applicable in the research on seedling establishment.
- Genotype is considered to be the most important variable to have in common between WPs. Arabidopsis seeds from the same batch are being used among WPs. In a later phase this might change, as different WPs might introduce different mutants relevant to their own research question.
- When working with dual stress, it is important to keep in mind that previous reports have shown that the first stressor often is dominant in the plant response over the second stressor.
- A dynamic discussion followed from the subject of the relevance of considering the ecology aspects of the research and the further application into crops when defining our experimental set-ups. This is relevant as the work currently focuses on Arabidopsis. One example is the selection of the pathogens and length of temperature regimes.
- Further follow-up could be of interest between the WP leaders involved in the workshop in order to agree upon: (i) markers to define that the plant is suffering from the studied stresses, (ii) relevance of keeping in mind the translation of our findings into crops when defining experimental set-ups and (iii) defining what output generated by the people involved in modelling could be of direct use by the experimentalists.
CropXR in the soil
By Gerard van der Linden, Wageningen UR Plant Breeding, Wageningen University and Corné Pieterse, Plant Microbe Interactions at Utrecht University, Francisco Pinto Espinosa & Jochem Evers from the Centre of Systems Analysis, Wageningen University
In this workshop, the workshop leaders stressed the importance of studying the parts of crops that are in the soil: the roots. In particular, the communication of the roots with the root microbiome and with aboveground parts. The workshop addressed the challenges that the involved researchers have in studying the adaptation of root-to-microbiome and root-to-shoot signaling to abiotic stress in the soil. The workshop leaders presented results and insights from studies leading to the CropXR Potato program as well as from CropXR Potato itself.
The microbiome may play an integral part in the stress response of potato. The involved researchers have the tools to include this regulatory layer in their studies and possibly the models that will be developed in CropXR. Extensive aboveground (remote-sensing) phenotyping will play a pivotal role in understanding shoot contribution in connection with the roots to the stress response, and translating this into useful plant models.
The main takeaway message of this workshop is: We can only understand the stress response of crops and how this may contribute to more resilient crops if we not only study the stress response mechanisms but place this study in the context of the whole plant in its cultivated environment. Knowledge of the developmental stage, above and below ground architecture, and the interaction between different adaptive mechanisms at the physiological, morphological and ecological level, is imperative to yield insights in specific mechanisms that lead to more resilient crops. Similarly, big datasets (for instance extensive transcription datasets) can only give useful outcomes when placed in the context of the whole plant (crop) response.