European Plant Phenotyping Network

Final Report Summary - EPPN (European Plant Phenotyping Network)

Executive Summary:
1. Summary
Phenotypic analysis has become a major limiting factor in genetic and physiological analyses in plant sciences as well as plant breeding. Molecular plant biology and molecular-based breeding techniques have developed rapidly within the last decade. In contrast, the understanding of the link between genotype and phenotype has progressed more slowly. Faster progress is currently hampered by insufficient technical and conceptual capacity in the plant science community to analyse the interaction between phenotypes of existing genetic resources and the environment, considering the timescale and practical challenges to actually grow plants in sufficiently large experiments in (semi) controlled environments. Therefore, improvement in phenotyping is a key factor for success in modern breeding as well as for advancement in basic plant research. Europe has reached a leading role in plant phenotyping and several institutions participating in EPPN are globally recognized as the hotspots for the development and implementation of plant phenotyping approaches such as development and implementation of novel technologies as well as data management and analysis tools. The European research area has a high potential to maintain and expand its leading role in this key technology by coherently integrating the expertise and capacities while, at the same time, reducing the fragmented landscape within this research area in Europe. The aim of the EPPN project was thus to create structural and functional collaborations between the leading plant phenotyping institutions in Europe and integrating the plant phenotyping community across Europe. EPPN followed the vision that the network of leading phenotyping infrastructures form the nucleus that provides a structured and efficient development of a persistently competitive plant phenotyping community in Europe.
EPPN successfully manged to integrate the plant phenotyping community within and beyond Europe by addressing a wide stakeholder community from academia and industry in different levels of interaction. Joint Research Activities developed, adapted and benchmarked novel sensors and established experimental as well as IT standards for application in plant phenotyping. The standards were made available for the wider plant phenotyping community on the EPPN website and by publications in scientific journals. Networking Activities provided a link between phenotyping experts, user communities, and technology developers within Europe and beyond. EPPN realized communication, networking, and education throughout the duration of the project at different levels: i) between existing and newly developing phenotyping platforms; ii) between phenotyping platforms and users from academia and industry; iii) between platforms, developers, and users; iv) with other leading international phenotyping centres. This effort represented the basis for novel scientific approaches in the utilisation of the existing facilities through Transnational Access. The access was based on demand driven, transparent access procedure which included independent reviewers from outside of EPPN. High demand from users across Europe for access to plant phenotyping facilities resulted in 66 experiments mostly from young scientists and new first-time users of phenotyping facilities.
EPPN became an important nucleus for the integration of the plant phenotyping community by the establishment of cooperation with the user community and a number of national and international projects and initiatives. Successful EPPN activities have led to the creation of the EMPHASIS project which was initiated by EPPN core members and has been listed in the ESFRI roadmap. EMPHASIS will facilitate structured development and use of plant phenotyping infrastructure in Europe based on the foundation of EPPN. Additionally, EPPN members represent the group of the International Plant Phenotyping Network (IPPN) which has been initiated as an association and an important hub for networking activities to successfully continue the integration of the plant phenotyping community on a global scale.

Project Context and Objectives:
2. Summary description of project context and objectives.
2. 1 Content and overall objectives of the project
Plant productivity is at the centre of major economic, ecological and societal challenges. Plants are the key resource for food and feed that needs to be increased by quantity and quality to meet the increasing nutritional demand of the growing human population (FAO, Rome 2009). At the same time, plants shall be increasingly utilized as renewable energy source, and as raw material for chemicals and other industrial products. Climate change and scarcity of arable land constitute significant challenges for sustainable agricultural production. It is becoming increasingly necessary and urgent to strengthen the impact of plant sciences through practical breeding for varieties with improved performance in very diverse agricultural environments. Specifically, quantitative assessment of plant structural and functional properties, plant phenotyping (Fig. 1) has become a major limiting factor in genetic and physiological analyses, in plant sciences as well as in plant breeding. Molecular plant biology and molecular based breeding techniques have developed rapidly within the last decade, an increasing number species have been sequenced and the large collections of mutants, accessions and recombinant lines produced allow analysis of gene functions, high-definition genotyping is carried out on thousands of plants in an automated way, allowing association genetics and determination of multi-parental Quantitative Traits Loci and, for transcriptomic, proteomic and metabolomic analyses, large robotized platforms are available allowing detailed characterization of the biochemical status of plants at a reasonable cost. In contrast, the understanding of the link between genotype and phenotype has progressed more slowly. Faster progress is currently hampered by insufficient capacity (both technical and conceptual) of the plant science community to analyse the phenotypes of existing genetic resources for their interaction with the environment. Because growth time and sufficiently large experiments are simply costly and time consuming. Improvement in phenotyping is therefore a key factor for success in modern breeding, as well as for basic plant research.
Non-invasive quantitative analysis of plant structure and function is essential to effectively advance the understanding of plant-environment interactions which is needed for research and application to sustainable and resource-efficient crop production in the context of climate change and varying agricultural production conditions. Novel genetic methods have mostly been tested on relatively simple phenotypic parameters, such as flowering time, which do not require sophisticated phenotypic measurements. The use of the same genetic methods in analyses of complex traits related to plant productivity or tolerance to biotic and abiotic stresses require the analysis of defined genetic material in multiple, well characterized environmental scenarios. Key plant functional traits such as yield, photosynthesis or plant water relations as well as the role of plant architecture for resource use efficiency above- and belowground need to be quantified during the dynamic plant responses to the environmental conditions throughout their development. Innovative technologies and novel approaches are required for the European plant science and plant industry. The capacities of existing platforms need to be combined in order to efficiently obtain data relevant for the performance of plants in agriculture and natural vegetation systems. Europe has reached a leading role in plant phenotyping and several institutions participating in EPPN are globally recognized as the hotspots for the development and implementation of plant phenotyping approaches such as development and implementation of novel technologies as well as data analysis tools. The European research area has a high potential to maintain and expand its leading role in this key technology by coherently integrating the expertise and capacities while, at the same time, reducing the fragmented landscape within this research area in Europe.
The overall aim of the EPPN project was to create structural and functional synergies between the leading plant phenotyping institutions in Europe as the core for the developing European Plant Phenotyping Network by linking phenotyping experts, user communities and technology development. This effort provided the basis for novel scientific approaches in the utilisation of the existing facilities through Transnational Access to leading European phenotyping facilities embedded in a strong scientific background. The project fostered their development into an effective European infrastructure including the required human resources and expertise to support Transnational Access to the diverse user communities. Joint research activities developed and adapted novel sensors and methods/assays for application in plant phenotyping. Additionally, novel approaches to sensor development and environment simulation and monitoring were made available to address the specific needs of the plant phenotyping user community. Dedicated networking activities provided a link between different stakeholders within Europe and beyond. EPPN followed the vision that leading phenotyping infrastructures – often with a national character in the member states – form the network and the nucleus that provides an efficient development of a plant phenotyping community in Europe and beyond by integrating centres outside of Europe such as the Australian Plant Phenotyping Facility (APPF) and fostering interaction in this research field on a global scale.
2. 2 Specific objectives of EPPN during the project
2.2.1 Management
The management activities within the project included the establishment of administrative, financial, legal, and technical issues to enable efficient management, cooperation within the network and, dissemination of the results. In particular, effective management needed to be established to guarantee the efficient flow of information between the three pillars of the project: networking, joint research activities and transnational access. Specifically, the Transnational Access played an important role with the implementation of the access procedures at the beginning of the project and continuous evaluation during the project to enable demand driven access of users to the phenotyping facilities. Networking activities were linked to the Transnational Access by events related to the information of diverse user community about the access opportunities, while at the end of the projects the focus of the networking activities shifted towards dissemination of the results obtained during the project.
2.2.2 Networking
Plant phenotyping science is an emerging discipline which has been widely recognized and has led to the development of a number of networks, projects or initiatives. EPPN was one of the major plant phenotyping projects in Europe with the goal to integrate the plant phenotyping community. EPPN generated an efficient interface for the users requiring phenotyping platforms with scientists working at the forefront of developing phenotyping approaches. This allowed a faster and targeted transfer of know-how from method development to the phenotyping platforms and application, as well as the efficient identification of questions in phenotyping applications to further stimulate work on technological challenges, assay development and the establishment of data management solutions.
Different networking levels within EPPN fostered pan-European interaction, and attracted also researchers from outside of Europe. Through the close involvement of research institutes, universities, industry and SMEs related to EPPN but also development of new sensors, approaches and databases, novel interdisciplinary interfaces were generated to develop a culture of interaction. To achieve this goal, communication, networking and education needed to unfold at four different levels:
i) between existing and newly developing phenotyping platforms: At this platform level the exchange of knowledge on technologies, protocols and data acquisition was essential. The exchange of information happened between experts and was supported by specific workshops, as well as dedicated visits.
ii) between phenotyping platforms and users from academy and industry: The communication between platform scientists and users covered the information of opportunities of phenotyping platforms to new users through announcements and sessions in user community conferences as well as specific workshops in new member states. The on-site information and education of potential users to illustrate potential experiments included on-site in workshops and training courses with information on basic knowledge about phenotyping methods, technologies and procedures.
iii) between platforms scientists, developers and users: Networking between platform scientists, developers and users was essential. To address this challenge, bilateral meetings of platform scientists with developers and users needed to be established as community building exercises before joint strategic discussions can be implemented fruitfully.
iv) with other leading international phenotyping centres: International collaboration was key to exchange ideas and for standardization/benchmarking. EPPN organized international phenotyping meetings as well as workshops in the context of the International Plant Phenotyping Network (IPPN). The goal of these workshops was to discuss Good Phenotyping Practice developed by EPPN.
2.2.3 Joint research activities
The Joint Research Activities addressed important steps for further development of plant phenotyping, namely the development of improved technologies to address diverse needs defined by users of the platforms, the definition and establishment of accepted standards of good phenotyping practice and the management of the large datasets of phenotyping platforms and their integration with genotypic databases. Joint Research Activities also further improved the quality and quantity of access and stimulated the use of a range of phenotyping approaches within the EPPN, and the plant science community at large with tree major pillars.
Novel Instrumentation for Plant Phenotyping
The overall goal of this work package was to provide the foundation for a novel generation of instruments to allow the quantification of phenotypic traits related to plant performance. Within this work package selected instruments as well as methodologies were made available within EPPN to allow benchmarking of the novel solutions with reference to existing methodologies and, experimental proof-of-concept towards the development of dedicated screening procedures for relevant plant performance traits. Specific technical goals of this work package were categorized into two instrument and technology development tracks: i) non-invasive imaging; ii) automation and robotics for plant cultivation.
Good Practice in Phenotyping
The main objective of this work package was to define the key elements of phenotyping procedures to be able to optimise protocols for phenotyping experiments from experimental design to data interpretation and, in particular to establish methods for the evaluation of the responses of different genotypes to environmental conditions. The work package developed guidelines of Good Phenotyping Practice based on experience from integrated phenotyping platforms.
Another objective of this work package was to test the reproducibility of data in phenotyping platforms and to test and elaborate methods to evaluate the responses of different genotypes to environmental conditions, by exploiting the contrasts of environmental conditions within the network of platforms. This was strongly depending on the accuracy of environmental characterisation in platforms, on the possibility to associate individual traits with plant performance and on the type of traits which were measured. We addressed this point with a set of experiments comparing different platforms at different stringency of environmental conditions with respect to the ranking of genotypes on key traits. This approach included not only the variability between platforms but also within a platform and analysed the robustness of results in repeated phenotyping experiments performed in similar environmental scenarios. This effort aimed at identifying the sensitive parameters that need to be controlled for robust and reliable results.
IT for high throughput phenotyping
Currently, many plant phenotyping databases and solutions exist and several state of the art analysis tools have been developed. However, a major problem is the lack of commonly accepted standards, ontologies or APIs which need to be defined to allow cross-site phenotyping approaches and most efficient use of resources. Novel challenges arise in the area of whole genome integration and novel image extraction algorithms. The primary objectives of this work package were thus i) to develop and define common structures and variables, ii) to establish a framework to access data management infrastructure, (iii) to derive numerical or categorical data from raw (image) data, (iv) to provide bioinformatics tools for data analysis and interpretation and (v) to provide user support and community tools to build a vivid plant phenotyping community.
2.2.4 Transnational Access
The EPPN phenotyping platforms provided the full range of state-of-the-art phenotyping systems and approaches for Transnational Access, offering users the full potential of modern plant phenotyping. Twenty three different installations within EPPN were available for Transnational Access to address the diverse needs of the users from simple scalar analysis of e.g. biomass of a plant, to complex 3D-structures, from physiological key functions like growth and photosynthesis, to chemical composition. Trait measurements under well controlled environmental conditions control and experiments included the ability to simulate water stress, complex light spectra, dynamics and intensities, nutrient relations, and biotic stresses which can be addressed under relevant throughput.
The modality of access required an application by the users which are evaluated in an independent peer reviewing process by scientists outside of EPPN. Based on the recommendation of the reviewers and the feasibility of the proposed projects which were assessed by the respective infrastructure scientists, proposed projects were ranked and assigned a priority. The Transnational Access opportunity directly benefited from the networking activities with an effective information about the opportunities of access to the EPPN installations. The Transnational Access was also evaluated throughout the duration of the project to enable a demand driven access to available installations.

Project Results:
3. Description of the main S&T results/foregrounds
3.1 Networking
The EPPN project aimed to develop a pan-European community which was focused on plant phenotyping, and organized around the EPPN phenotyping facilities. To achieve this goal, communication, networking and education needed to unfold at four interdependent levels described in the tasks:
• between existing and newly developing phenotyping platforms,
• between phenotyping platforms and users from academy and industry,
• between platforms, developers and users,
• with other leading international phenotyping centres.
Task 1 Virtual EPPN network for sharing information
A number of networking tools was established to enable communication within EPPN and the external user community. The EPPN website was set up at the beginning of the project and has proven to be a highly successful tool for communication with over 650.000 visits, in which attention was paid to at least two webpages per visit. The website included the general information about EPPN such as project structure, governance as well as announcements about EPPN activities, EPPN supported publications and standards as well as all details about Transnational Access procedure. We put a special attention on the Transnational Access webpages which included the overview and description of all installations as well as tools to select the installations which suites best the needs of the user. For internal communication within the consortium an intranet was set up with all relevant information for the consortium members.
Communication with the internal and external community was also addressed by the establishment of an EPPN working group in LinkedIn, which is a networking platform for scientists and industry to discuss different topics related to plant phenotyping but also to advertise upcoming events. Additionally, users could subscribe for the EPPN newsletter. However, based on a survey within the community with nearly 200 participants (see http://www.plant-phenotyping.org/IPPN-SURVEY) which was co-organized by EPPN, most survey participants prefer to be informed about upcoming activities or other news information by emails. Therefore the over 230 subscribers of the EPPN newsletter were regularly informed about different activities including upcoming events, new publication or achievements by email.
Task 2 Communicating the Transnational Access projects to users through presentations and dedicated user workshops
The existence of the facilities, scientific benefits, and procedures for the Transnational Access to the EPPN installations were widely discussed and advertised in user- specific journals, at conferences in oral and poster presentations as well as by circulating flyers at different events and emails in different emailing lists in different disciplines of plant biology. A publically accessible EPPN-webpage (See Task 1 Virtual EPPN network) specifically addressed Transnational Access as well as networking and education activities which were also closely related to Transnational Access opportunities.
A number of workshops and meeting were organized in different member states specifically 4 workshop addressed the information of potential users about the Transnational Access including the facilities, benefits, and limitations of EPPN phenotyping platforms.
• Warsaw, Poland, May 2012 with over 120 participants,
• Porto Heli, Greece, September 2013 as a satellite meeting of the EPSO conference with 60 participants,
• Copenhagen, Denmark, September 2014 with over 60 participants,
• Barcelona, Spain, November 2015 final EPPN symposium with over 160 participants.
The workshops elaborated on general aspects of plant phenotyping and the opportunities offered by the EPPN platforms. But we also openly discussed with different users about the needs and requirements of plant phenotyping to specifically address the needs of the plant phenotyping community. Additionally, EPPN members actively advertised the Transnational Access opportunities in a range of dissemination activities and by organizing dedicated sessions at different events (see Task 8 Outreach).
Task 3 Workshops with technology developers
With these workshops we specifically aimed to address technology development which could be used and implemented into plant phenotyping platforms. The goal of these workshops was the establishment of an interaction between technology developers and platform operators to facilitate and define potential for novel sensors and approaches for phenotyping to optimize and extend the pool of measurable plant traits as well as the adaptation and standardization of the novel plant trait. Two dedicated workshops were organized:
• A developer workshop was organized in Wageningen in September 2012. The scope of this workshop was to identify technologies that are not yet utilized in the phenotyping platforms but have the potential to be implemented in phenotyping platforms including groups outside EPPN. The workshop attracted about 70 participants and enabled close interaction between technology developers from very diverse fields from the industry and academia with platform operators and users.
• While the fist developer workshop was directed to the identification of different technologies which can be implemented into plant phenotyping platforms, the second workshop, in Prague in July 2014 focused on technology and standards related to the assessment of plant water relations. A number of technologies are currently available and implementation as well as relevant use of these technologies is rather important. Therefore we decided to focus in this workshop on technologies as well as standards needed for a relevant assessment of plant water relations. The workshop was limited to approximately 20 participants and included invited experts in the field of plant water relations to effectively address the topic.
Task 4 EPPN Round table meetings for phenotyping expert
Two round table meetings were organized to discuss, set and distribute best phenotyping practices, standards and protocols:
• The first Round Table meeting was held in Wageningen in July 2013 with the goal to discuss how to distribute phenotyping protocols within and beyond EPPN.
• The second Round Table meeting was held in Nottingham in October 2014 specifically focusing on the extension of available protocols and their distribution as well as interaction with other projects.
Both Round Table meetings specifically addressed the “reference experiment” (see WP4 Task 1 and 2) to analyse and evaluate the responses of known genotypes to different environmental conditions at different platforms with contrasting environmental conditions in greenhouses across Europe and Australia. The essential step in this experiment was the definition of the standards for both trait assessment and environmental monitoring and data management as the key to elaborate and test methods to evaluate the responses of different genotypes to environmental conditions. The reference experiment provided thus an important step towards standardization of phenotyping experiments. The collection of protocols also goes beyond the interests of the EPPN participants and addresses plant phenotyping community at large. Additionally plant phenotyping represents very specific requirements for best practice in experimental procedures with respect to a large number of plants which need to be screen in an automated mode with non-invasive technology. Therefore, we established a dedicated website for plant phenotyping approaches with a collection of protocol, guidelines and literature specifically dedicated to phenotyping experiments which were continuously extended based on the growing expertise of the EPPN consortium (http://www.plant-phenotyping-standards.net/).
Task 5 EPPN Summer schools
Training school type courses were held to provide practical training at phenotyping platforms for small groups of potential users with focus on early career scientists. The training schools covered a wide range of topics highly relevant for plant phenotyping from molecular breeding, trait assessment with non-invasive methods as well as data analysis and modelling in both dedicated lectures and practical exercises. Lectures were held by renowned phenotyping experts and practical exercises were supported by local stuff working in phenotyping platforms on a daily basis and allowed the participants the application of important experimental tools in realistic phenotyping scenarios. Additionally, excursions to different facilities provided an inside into phenotyping experiments both under controlled lab and greenhouse as well as under field conditions (for more details see also: http://www.plant-phenotyping-network.eu/eppn/summer_school). The one week training schools were held:
• at the Biological Research Centre in Szeged, Hungary in July 2013,
• at the Aberystwyth University, UK in March 2015,
• additionally, EPPN supported the DPPN/EURoot Winter School on Root Phenotyping at the Forschungszentrum Jülich in November 2015.
Task 6 International Phenotyping Conferences
Plant phenotyping is becoming a central field of research and application in academia and industry resulting in the development of new phenotyping platforms and methods usually driven by the strong demand by users. There is a need interaction between different stakeholders and an exchange of experience and information related to plant phenotyping technologies, use and application of these technologies in dedicated experiments as well as data analysis and management approaches. To effectively address these goals plant phenotyping requires close interaction on the international level beyond Europe. EPPN was thus closely involved in supporting two international symposia:
• EPPN organized the 3rd International Plant Phenotyping Symposium in Chennai, India in February 2014. The symposium was organized with the International Plant Phenotyping Network (IPPN) and the M. S. Swaminathan Research Foundation. The symposium attracted approximately 200 participants from industry and academia and provided an excellent networking opportunities between users, platform operators or yet-to-become platform operators, as well as experts in technology development and integration. For more details see: http://www.plant-phenotyping.org/chennai_2014.
• EPPN organized an international symposium in Barcelona in November 2015 whit the major goal to disseminate EPPN activities and give the users of phenotyping platforms within EPPN Transnational Access scheme the opportunity to present their work as a basis for discussions between users, scientists operating phenotyping platforms as well as technology developers about the needs and requirements of plant phenotyping. The symposium attracted over 160 participants and covered different topics relevant for plant phenotyping For more details see: http://www.plant-phenotyping.org/Home_EPPN_Symposium
Task 7 Modality of access
This task refers to the Transnational Access to EPPN platforms. Potential users submitted their application to the EPPN office after discussing with the scientists responsible for a specific installation about the feasibility of their experiment. All proposals were subject to a peer reviewing process by independent experts, outside the EPPN consortium. The reviewers were selected by providing expertise on the scientific field of plant biology concerned with the proposed project. The reviewers were requested to submit confidentially written assessments to the EPPN office and are part of the documentation of the selection process. Based on the recommendation of the reviewers and the feasibility of the proposed projects these projects were ranked and assigned a priority. Accordingly, the management committee decides with the scientist of the infrastructure offering transnational access if the proposed project can be executed. All selected and finalized projects are listed on the EPPN webpage with a short project summary (http://www.plant-phenotyping-network.eu/eppn/selected_projects).
Task 8: Outreach to new users
A special effort was made to inform existing and attract new users in a range of different activities. The EPPN consortium was involved in a number of activities dedicated to discuss the needs and requirements of the plant phenotyping and to identifying opportunities, future challenges with the diverse community at large. For example, the EPPN participants were involved in over 260 dissemination activities across Europe throughout the duration of the project. EPPN became well known beyond Europe and for example the EPPN website attracted over 650.000 visitors during the project period.
Specifically, in addition to organizing the EPPN meetings such as the workshops to inform users about the Transnational Access opportunity (Task 2), workshops with technology developers (Task 3) round table meetings (Task 4) and training schools (Task 5) EPPN was represented at a number of dedicated events to address different communities and to discuss the opportunities of plant phenotyping with the diverse community at large. A summary of the events organized by EPPN is on the website: http://www.plant-phenotyping-network.eu/eppn/previous_events. In short EPPN organized:
• Imaging workshop in Nottingham in September 2012: EPPN supported a session at the “International Workshop on Image Analysis Methods for Plant Science” with the goal to specifically address plant imaging experts and discuss imaging approaches dedicated to plant phenotyping.
• Dedicated session at the 57th SIGA Annual Congress, in September 2013. The congress of the Italian Society of Agricultural Geneticists attracted international audience with focus on genetic approaches in breeding. The goal of the plant phenotyping session at this congress organized by EPPN was to inform breeders and geneticists about the available plant phenotyping platforms across Europe and to discuss the needs and requirements of plant phenotyping.
• Dedicated session at the FESPB/EPSO Congress, June 2014. EPPN organized a session at the international FESPB/EPSO congress to reach and inform a very wide and diverse user community attending this congress about plant phenotyping. FESPB/EPSO congresses are usually attended by over 1000 participants.
• Workshop co-organized with the EU funded project WATBIO in Abersytwyth in September 2014. The workshop aimed at bringing together researchers and students for a series of talks, discussions and practical exercises on different aspects of plant phenomics and metabolomics.
• CropSense Symposium in Bonn, Germany, in September 2014. The goal of this co-organized symposium was to address topics which have gained increasing importance such as field phenotyping and transition from phenotyping under controlled greenhouse to the heterogeneous field.
• Technical session: Plant Phenotyping for the responses to environmental conditions in Montpellier June 2015. This session specifically addressed discussions about practical solutions of plant phenotyping applications.
• Wheat Initiative, Face to Face meeting of the Expert Working Group Plant Phenotyping in Barcelona in November 2015. EPPN co-organized the meeting to closely interact with the plant phenotyping community beyond Europe.
Further outreach activities include:
• Organizing a Special Issue on Plant Phenotyping in the Journal of Experimental Botany which resulted in numerous EPPN contributions in this issue. For more details see the editorial of the Special Issue: http://jxb.oxfordjournals.org/content/66/18/5385.full
• A survey about the needs, limitations, challenges of plant phenotyping which was performed together with the International Plant Phenotyping Network (IPPN, http://www.plant-phenotyping.org/) the COST Action (http://costfa1306.eu/) The German Plant Phenotyping Network (DPPN, http://www.dppn.de) and the Wheat Initiative (http://www.wheatinitiative.org/).
• EPPN is included in the PAERIP data base which combines infrastructure projects indicating potential partnerships between EU and African projects and research facilities
http://inventory.paerip.org/paerip/
• EPPN is included in a European research infrastructure database MERIL http://www.esf.org/activities/science-policy/research-infrastructures/meril-mapping-of-the-european-research-infrastructure-landscape.html
• There was a close interaction with the EU funded project called transPLANT (http://www.transplantdb.eu/) with respect to recommendation of data standards (for more details see WP4 Task 1).
Throughout the duration of the EPPN project we also closely interacted with national platforms in Europe such as the German Plant Phenotyping Network (DPPN, http://www.dppn.de) French Plant Phenotyping Network (https://www.phenome-fppn.fr/) UK-Plant Phenotyping Network (http://www.ukppn.org.uk/) as well as the Australian Plant Phenotyping Facility (APPF, http://www.plantphenomics.org.au/). The institutions leading these national platforms are also members of the EPPN project. Most members of the EPPN consortium have also joined the Phenomen-ALL COST action. The European national platforms which were members of EPPN represent also the core institutions in the EMPHASIS project which has been listed on the ESFRI road map.
Finally, we established a close cooperation with the International Plant Phenotyping Network (IPPN, http://www.plant-phenotyping.org/) e.g. by co-organizing the International Plant Phenotyping Symposium (see Networking Task 6). This cooperation enabled a continuation of successful activities of EPPN beyond the duration of the EPPN project and, to address plant phenotyping community beyond Europe. For example IPPN will host and extend the standards established and collected by EPPN after the end of the EPPN project, EPPN networking activities will be continued and implemented into IPPN activities to successfully continue the integration of the plant phenotyping community.
Task 9: Training of users and support in data analysis
This task addressed specifically the support of uses that obtained Transnational Access to train and support them with respect to plant phenotyping approaches at each institution which included experimental design, performance of the experiment and data analysis. Additionally, logistic support was provided to all users.
A local staff member, a mentor, was assigned to each user gaining access through Transnational Access projects. This staff member assisted the user in the experimental work by providing assistance with operating the phenotyping platforms, training, data analysis and interpretation. The aim of the mentoring program was to train the user to become an independent researcher with respect to phenotyping approaches. At each institution selected users were integrated in the local working groups to ensure that users receive sufficient feedback from the local working groups. Additional expertise of the staff members included support in plant cultivation, samples preparation or experimental design. Logistic support was offered to all users for travel and accommodation during the Transnational Access experiments, as well as for preparatory visits before the experiment and visits after finalizing the experiment to discuss the results or prepare manuscripts.
3.2 Novel Instrumentation for Plant Phenotyping
Task 1: Development of non-invasive imaging technology
The present generation of operating instruments for automated or semi-automated measurement of plant phenotypic traits use 2D imaging usually in the visible and near infrared portion of the light spectrum. A second generation of instruments has to address the need of providing additional layers of relevant information by introducing one additional physical dimension (2.5D-projections or full 3D), by the overlay of image features acquired by cameras capturing specific wavelengths (e.g. characteristic leaf pigments, chlorophyll fluorescence), or by hyper-spectral instruments that use a large portion of the electromagnetic spectrum to reconstruct specific spectral signatures that need to be correlated to plant metabolic traits.
The task dealt with the generation of operational non-invasive instruments for automated or semi-automated measurement of plant phenotypic traits and was divided into 5 sub-tasks related to:
• 3D reconstruction from imaging in the visible range (Task 1.1)
• overlay od a 3D reconstruction with features acquired by cameras capturing wave lengths in the near infra-red range (Task 1.2)
• overlay with bands related to chlorophyll fluorescence (Task 1.3)
• 3D information extracted from laser scanning of small plants (Task 1.4)
• assessment of the entire spectral range (Task 1.5).
• comparison of different 3D approaches (Task 1.6)
Task 1.1: 2.5D shape and colour sensing system (Partner 8, DLO)
Objectives:
An important part of the phenotype of plants is expressed in their 2.5-dimensional shape (2.5D). This task aimed to develop a 3D sensor that allows accurate and fast measurement of the shape of a plant. Moreover, the colour of the plant should be projected on the shape model. This way, the model of the plant is as realistic as the actual plant that may be viewed by human eyes. With this sensor, the changes of the phenotype over time can be measured automatically and compared to other plants from the same batch. This allows an objective measurement of the shape descriptors of a plant.
Achievements:
Two measurement systems have been assembled for the measurement of the 3D plant structure and both systems are fully operational. These systems can be used to measure plants of up to a height of 70 cm for an accurate reconstruction of 3D plant models. The systems use multiple cameras which observe plants from a number of different viewing angles by utilizing a fast 3D reconstruction method which results in a 3D model within a fraction of a second (± 45 ms). The system is very flexible and can use a variable number of cameras. The quality of the 3D reconstructions generally improved when a larger number of cameras were used. The silhouettes of the plant in the acquired images by each camera are used to reconstruct the plants in 3D by using the so called shape-from-silhouette method. The 3D reconstructions of the whole plants can then be segmented in stem and leaves and, the quality of different traits such as plant volume, leaf area or stem length can be calculated. The system was extensively validated by ground truthing with manually measured traits with the traits based 3D model reconstruction. A very good correlation between the traits based on the 3D reconstruction and the manually assessed traits was obtained. A detailed description of the system which includes the camera calibration procedure, the analysis software and the validation experiments is published in a Special Issue of the journal Machine Vision and Applications (Golbach F et al 2015 Machine Vision and Applications, DOI: 10.1007/s00138-015-0727-5).
Task 1.2: Near Infrared sensing system (Partner 8, DLO)
Objectives:
Many quality characteristics of the plant can be measured even before they are expressed in shape or visible light. By using near infrared (NIR) sensors, we can overlay the 3D model with a near infrared layer. Such a layer may allow us to detect characteristics such as water content (related to water status, maturity or ripeness) in specific plant parts by looking at the active components in a plant. The use of non-visible light complements the 3D colour model (WP3, Task 1.1) and allows for a fuller description of the plant phenotype.
Achievements:
A prototype sensor probe has been constructed with the aim to measure the water content of an individual leaf. The probe consisted of a clip placed around the leaf in which a light source and light fibre attached to a Near Infra-Red (NIR) spectrometer with a sensitive range of 400-1700nm were integrated. The probe was mounted on a robotic arm to bring the sensor to a specific plant part / location in a single plant. The position in the plant where the robot should place the sensor was calculated by the measurement system developed in WP3 task 1.1. The shape of the plant was captured in a 3D model and individual leaves were segmented. Based on the architecture of the plant a single leaf was selected and a path was generated to guide the sensor on the robotic arm to that particular leaf. To measure the water content in a single leaf the light emitted by an optical fibre was shone onto the leaf and a light receiving optical fibre captured the transmitted light through the leaf towards the Near Infra-Red spectrometer to capture the light spectrum. A prediction model was developed which linked the captured spectra of maize leaves with reference data, which were leaves with known leaf water content levels. The resulting model in this feasibility study was able to predict the water content in a maize leaf with a convincing and promising correlation. This sensor / actor approach has several advantages: the potential to integrate the robot-arm/ probe combination into a high throughput phenotyping system and most importantly, it combines the accuracy of a local and reliable transmission measurement with automated repeatability using robotics.
Task 1.3 - 3D photosynthesis sensing system (Partner 11, CzechGlobe)
Objectives:
Chlorophyll fluorescence emission is one of the most powerful optical signals that non-invasively reports on the processes of light capture and its photosynthetic use. The currently available chlorophyll fluorescence imagers capture the signal flattened to only two spatial dimensions and time. In this task we construct a system that performs a highly accurate 3D reconstruction of a plant from multiple real colour images and projects the fluorescence transients on such a reconstruction providing 3D information about photosynthetic efficiency.
Achievements:
Chlorophyll fluorescence is a powerful tool yielding information about the light use efficiency and plant response to biotic and abiotic stress. In imaging applications, fluorescence transient is usually recorded on single leaves. Here, we have developed a method to co-register the 3D plant structure with the fluorescence signal. The 3D photosynthesis system was developed and demonstrated at a workshop at the Forschungszentrum Jülich. The system consists of a 3D model acquisition, where an object is scanned, data transferred to the PC and basic model processing is performed. The model of a plant is constructed with use of commercial scanner Artec MHT™ using the structured light approach. The procedure of co-registration of chlorophyll fluorescence image to 3D model is based on a binary image of a model and corresponding chlorophyll fluorescence image. After registration the 3D model of chlorophyll fluorescence is achieved. This approach may be able to bring a number of additional information for plant phenotyping, such as vertical distribution of light use efficiency, progress of senescence from older (lower) to younger (upper) leaves in relation to the effects of stress conditions, the vertical distribution of nutrients (e.g. nitrogen) as a sign of nutrient use efficiency etc.
Task 1.4: Laser scanning for phenotyping of small plants. (Partner 12, ABER)
Objectives:
In phenotyping of small plants there is a significant problem in assessing canopy structure in 3D in a high-throughput manner. This task aimed to investigate the value added in determination of canopy physical structural and components specifically for pure swards and mixtures by the addition of laser scanning compared with the more usual 2D visual determinations.
Achievements:
Recently devices to measure 3D plant structure have become widely available and offer the possibility to directly obtain 3D models of single plants or canopies. These devices are based on different measurement principles and can be applied in different scenarios. A comparison of three different approaches is described in Task 1.6. Here we analysed specifically laser scanner to capture the 3D surface of a number of plants, including brassicas, wheat, oats, Brachypodium and forage grass swards. Detailed point clouds can be captured and visualised. Parametrization to extract meaningful information in particular in forage grasses on sward structure proves rather difficult because of occlusion where plant structures overlay each other and lack of colour, both of which leads to loss of information. To overcome this, a supplementary imaging method was developed which is based on multiple-view 2-D imaging followed by reconstruction of 3-D models based on structure-from-motion computer vision principles. Parameters extracted include colour, area of organs and relative angles. Time series also provide growth and movement information and have been published in book chapters acknowledging the support of EPPN (Lou L 2014 In “Image Analysis and Recognition Lecture Notes in Computer Science", pp 349-356; Lou L 2014 In “Advances in Autonomous Robotics Systems Lecture Notes in Computer Science" Volume 8717, pp 221-230).
Task 1.5 - Hyper-spectral imaging (Partner 1 – Juelich)
Objectives:
Hyperspectral analysis allows quantifying spectral changes at many wavelengths simultaneously. It thus provides an unbiased approach that does not require previous knowledge on the spectral range, in which a plant process is affected during the phenotyping experiment. The task was build on existing hyperspectral equipment at Juelich and measure spectral properties of plants exposed to different stress levels. Technologies to deconvolute the signals and to identify characteristic spectra for the stress responses has been determined.
Achievements:
We integrated a hyperspectral imaging line scanning spectrometer in a ‘sensor to plant’ system with an automated routine for positioning which allows in situ reflectance measurements at the level of the whole shoot. Four species (rapeseed, tomato, maize, barley) with contrasting shoot architectures were grown at varying nitrogen and phosphorus levels. Plant pigment content (chlorophyll, carotenoids, and anthocyanin) was measured manually and canopy spectral reflectance (400–1000 nm) were assessed. The data were analysed and we observed that nitrogen deficiency suppressed growth and decreased plant chlorophyll content. The decrease in plant chlorophyll content, as a result of reduced nitrogen fertilization, was more prominent in the two monocotyledonous species. The spectral reflectance signature of plants grown under three nitrogen levels was considerably different. Linear relations between vegetation indices (VIs), based on canopy reflectance, and plant leaf area or plant chlorophyll content were observed. However, the correlation coefficients of these relationships were strongly species dependent. The work performed here forms the basis for a manuscript that however will need to be integrated with additional data before submission for a publication. We aim at completing this work contributing additional experiments in 2016, which are not directly funded by the EPPN project.
Task 1.6 (NEW Task): Comparison of 3D methods (Partner 8, DLO; Partner 11, CzechGlobe; Partner 12, ABER)
Objectives:
Within WP3, three different 3D sensors have been developed, based on a) the silhouette-from-shape method (Task 1.1) b) structured light (Task 1.3) and c) laser scanning (Task 1.4) each with its own merits and demerits. In this newly defined task, we aim to compare these3D sensors for the specific purpose of 3D reconstruction and analysis of plants.
Achievements:
Quantitative comparison of 3D approaches may be rather challenging because of the difficulty of i) defining a set of measures that can define the benchmark and ii) establishing a suitable ground truth data set and neither benchmarking nor comparing the different methods was possible. Therefore we decided to describe and test the 3D setups instead of quantitative comparison. Three different approaches were extensively described and evaluated based on i) the silhouette-from-shape method (Task 1.1) ii) structured light (Task 1.3) and iii) laser scanning (Task 1.4) each with its own merits and demerits. Comparing these 3D approaches has shown that depending on the goal of the experiment, the shape of the plant/canopy, the required speed/capacity of the experiment, the costs of 3D system, the amount of necessary details of small plant parts etc. different choices have to be made which result in one of even more 3D sensor/ software options to address the needs of a specific experiment.
Task 2: Automation and robotics
This task addresses two different levels related to automation. The first Task relates to basic phenotypic implementations involving plant movement or sensor movement in phenotyping platforms and is linked to the evaluation of potential impact of these two operational modi on workflows and experimental designs (Task 2.1). The second Task addresses practical solutions for the establishment of a field-scale operable system (Task 2.2).
Task 2.1 Concept study to compare the “plant-to-sensor” vs.” sensor-to-plant” approaches (Partner 1, Juelich)
Objectives:
Most commercial phenotyping platforms in growth chambers and greenhouses are built based on the concept to move plants (individually or in groups) to the sensing system. One argument in favour of these plant-to-sensor systems are the constant conditions of imaging at the location of the measurement. On the other hand, sensor-to-plant systems could provide significant advantages, when assays for phenotyping require increased complexity of environmental conditions during plant growth. Systematic approaches to address benefits and drawbacks of both alternatives have been addressed in this task.
Achievements:
During the lifetime of the EPPN project we were able to provide a first assessment of potential effects of the automation system SCREEN-House at FZJ-IBG2 - a plant to sensor system - on the development of typical monocot (barley) and dicot (rapeseed) species of interest for this platform. We were able to develop experimental setups and statistical analyses to assess quantitatively whether the movement of the plants to the imaging station (realized via a crane equipped with a pot grabber in our automated system) has any undesired effect on plant growth and leaf area establishment that can potentially introduce undesired variability in screening experiments. Typically, routine experiments in this and other similar phenotyping platforms are designed to compare genotypic responses under non-limiting and limiting growth conditions, such as reduced soil water or nutrient availability. Based on the experiments performed we conclude that in our shoot phenotyping system SCREEN-House that was also used for EPPN Transnational Access experiments and for the EPPN reference experiment (WP2 Task 2), there were no significant effects of transporting plants via a crane system to the imaging station by using typical transport frequencies that would be required in a typical screening experiment (for instance, once or twice each week during the experiment). Additionally, we were able to highlight that very detailed randomization protocols need to be designed for our (and also for similar) phenotyping systems to avoid likely positional effects due to micro-environmental conditions with particular reference to light heterogeneity, direct temperature radiation effects and edge or border effects. These findings additionally support the conclusions from the reference experiment (WP4 Task 2) that the microenvironment in the greenhouse has a substantial influence on pant properties and may make any comparison between experiments from different greenhouse very difficult without careful characterisation of the environmental conditions.
Task 2.2 Design of field-operable automated phenotyping system (Partner 1, Juelich)
Objectives:
Phenotyping in the field could benefit significantly from automated systems which quantify structural and functional traits of plant at the plot scale with focus on stand properties. The aim of this task was to evaluate different field screening system specifically systems that have already been developed at Juelich.
Achievements:
This task involved the development of a sensor positioning systems for experiments related to canopy phenotyping to be tested and used in plant breeding plots. Several different options for positioning systems were thoroughly assessed if they are suitable for field phenotyping, based on the following five criteria: i) implementation, ii) transferability, iii) operation, iv) sustainability, and v) field usage. We provided an in depth assessment weighing the advantages and disadvantages of possible technical solutions for the realization of semi-automated in-field phenotyping by proximal or remote sensing approaches. Some prototypes for field phenotyping were developed, yet these technical developments were not financed by EPPN but by other national and international projects in Jülich. Finally, a number of these approaches were benchmarked for their applicability to assess relevant traits under conditions with the goal to establish phenotyping concepts that allow the measurement of structural and functional traits under dynamic field conditions. The approaches and some case studies are summarized and reviewed in a book chapter with the title “Field phenotyping – Concepts and examples to quantify dynamic plant traits across scales in the field” which is was submitted to a review process and will be published approximately in summer 2016 in the book: “Terrestrial Ecosystem Research Infrastructures: Challenges, New developments and Perspectives” edited by Abad Chabbi and Jacques Roy.
3.3 Good Practice in Phenotyping
Task 1: Collection of protocols and standardized formats for phenotypic traits (Partner 5, INRA, all platforms)
Task 1.1 Environmental data / Task 1.2 Trait measurements
Objectives:
The description of the Task 1.1 “Environmental data” and the Task 1.2 “Trait measurements” is combined as both tasks are closely linked. Plant phenotyping deals with the assessment of phenotypic traits under well-known environmental conditions, both needs to be well defined to enable comparability between labs but also use of the data in meta-analysis approaches. Within these tasks we collected standards of phenotyping experiments required for the characterization of the environmental conditions (Task 1.1) and of the respective trait measurements (Task 1.2).
Achievements:
Currently many plant labs use different methods and protocols for very specialized as well as very common measurements. Few recent publications have aimed to bring together the knowledge that is necessary for a plant biologist to set up experiments and apply the environmental conditions that are appropriate to answer questions of interest (Poorter et al. 2012 Functional Plant Biology 39: 821-838), very specific experimental issues such as pot size (Poorter et al. 2012 Functional Plant Biology 39: 839-850) or methods to analyse plant response to the environment (Tardieu 2013 Frontiers in Physiology 4, 17). Plant phenotyping adds another aspect to standardization of experimental procedures since such experiments require screening of large number of plants using non-invasive methodology often in automated mode and, additionally a detailed monitoring of the dynamic environmental conditions.
In this task the EPPN consortium initiated a collection of protocols which at the beginning were directly linked and also practically tested in the reference experiment (WP4, Task 2.1 2.2). The standards have been continuously extended throughout the project and include: i) minimum requirements for environmental monitoring of light conditions, temperature, water relations and soil characteristics; ii) traits assessment related to phonological stages, plant imaging, destructive sampling, photosynthesis and a collection of optical methods for stress detection. The standards are published on the website: http://www.plant-phenotyping-standards.net/.
Furthermore to address a wider plant phenotyping community we established a cooperation with the EU funded project transPLANT (http://www.transplantdb.eu/) resulting in a tangible outcome in a manuscript that was published in 2015 in a Special Issue of the Journal of Experimental Botany dedicated to plant phenotyping: “Recommendations for metadata and data handling in plant phenotyping experiments” (Krajewski. et al. 2015 Journal of Experimental Botany, 18: 5417-5427). The manuscript includes recommendations addressing three different aspects required for standardization of plant phenotyping experiments: i) define a detailed checklist of attributes summarizing the data content following a Minimum Information principle as pioneered by other communities in bio-medical sciences. The MIAPPE (Minimum Information About Plant Phenotyping Experiment) checklist of attributes is registered at Biosharing repository and can be found at: http://cropnet.pl/phenotypes; ii) annotate attributes and phenotypes as much as possible using publicly available vocabularies and ontologies; iii) use of a ISA-TAB format specifically designed according to MIAPPE definitions http://cropnet.pl/phenotypes. A second manuscript that will acknowledge EPPN funding is in preparation, in collaboration with colleagues of the TRANSPLANT project and with bio-informaticians and biologists working on plant phenotyping national initiatives (mainly German Plant Phenotyping Network; DPPN and French Phenome project). In this manuscript we further expand the details of MIAPPE and provide ISA-TAB formats to implement it. The draft "Minimum Information about Plant Phenotyping Experiment" (MIAPPE) comprises a list of attributes that, in our opinion, are necessary for a useful description of plant phenotypic data. The MIAPPE checklist contains the following sections: General metadata; Timing and location; Environment; Biosource; Treatments; Experimental design; Sample collection, processing, management; Variables; Observations.
Finally, EPPN has continued close cooperation with the International Plant Phenotyping Network (IPPN, http://www.plant-phenotyping.org/) which was established end of 2015 as an international association to further spread these standards established by EPPN to the global community. Also, the Expert Working Group on Plant Phenotyping of the Wheat Initiative (http://www.wheatinitiative.org/) is actively cooperating with EPPN and IPPN on the establishment on standards for wheat phenotyping and distribute them in the near future at the EPPN and IPPN websites.
Task 2: Construction and analysis of data matrices allowing to compare outputs of platforms
Task 2.1 Reference experiment performed at different platforms (Partner 1 Juelich, Partner 2, IPK; Partner 3 HGMU, Partner 5 INRA; Partner 6 AA, Partner 8 UoN, Partner 9 HAS, Partner 12 ABER, Partner 10 APPC1)
Objectives:
This task includes data captured and processed by different phenotyping platforms and approaches to analyse and evaluate the responses of known genotypes to different environmental conditions at different platforms with contrasting environmental conditions in greenhouses across Europe and Australia. The performances and ranking of tested genotypes differs between locations because of the genotype-environment interaction. Therefore it is important to use highly standardized measurement of traits as well as environmental monitoring to be able to evaluate the sensitivity of these genotypes to environmental variables.
Achievements:
The reference experiment pursues two major goals: i) establishment of standardized experimental protocol at each platform, which includes the use of the same ontologies, standardized environmental monitoring, standardized instrument calibration (in particular cameras) and procedures to assess a specific plant trait; ii) elaborate and test methods to evaluate the responses of different genotypes to environmental conditions, by exploiting the contrasts of environmental conditions within the network of platforms. The performances of tested genotypes is supposed to differ between different locations as well as their ranking is also expected to differ because of the genotype - environment interaction. We therefore aimed at measuring common variables in all sites (e.g. maximum radiation use efficiency) and at elaborating methods that use the network to evaluate the sensitivity of genotypes to environmental variables such as evaporative demand, temperature and light.
A set of representative genotypes of two species (maize and canola) was selected with a wide phenotypic plasticity to cover a wide range phenotypic traits of interest and the seeds were circulated to all involved labs. Prior to the performance of the experiment we defined the establishment of basic standards for environmental characterization of an experiment as well as experimental protocols. Specifically we established minimum requirements for environmental monitoring and elaborated standards for basic experimental procedures with special attention to the calibration and comparability of the imaging systems at different locations by comparing measurements of artificial plants made of metal and plastic which were circulated to all involved platforms. The requirements for environmental monitoring and experimental standards are published on the EPPN website and are available for the community (see also WP4 Task1, http://www.plant-phenotyping-standards.net). Finally, the experiment with the real plants was performed and completed at all locations with focus on traits related the phenological development and growth. The results are summarized in more detail in WP4 Task 2.2.
Task 2.2 Analysis and classification tools (Partner 5, INRA and Partner 12, ABER)
Objectives:
The data obtained in the reference experiment has been used to determine the best methods to normalise and integrate data captured in different platforms (Task 2.1). The data has been used to develop range of analysis tools and optimise models discriminating between genotypes.
Achievements:
Phenotyping platforms in controlled conditions are most often placed in a greenhouse, i.e. in often fluctuating conditions which change within very different time frames. For example light conditions within different locations in a greenhouse can vary between seconds and days as clouds and sun is moving. There is often also substantial spatial variability within a greenhouse due to shading by parts of the greenhouse constriction, the position of cooling/heating devices and temperature if plants are placed close to the doors, inner or outer walls. Additionally, multi-site experiments involve different environmental conditions in terms of light, temperature and the resulting evaporative demand at each location. It is therefore essential that the environmental data are carefully collected and analysed. The reference experiment was finalized and the data are being analysed as described in WP4 Task 2.1. The major goal of the reference experiment was to elaborate and test methods to evaluate the responses of different genotypes of maize and canola to environmental conditions by exploiting the contrasts of environmental conditions.
While using thermal time, which represents equivalent days of growth at 20°C or in degree days has proven as a robust analysis tool specifically for maize. We observed leaf appearance that closely correlated with thermal time across all participating platforms. Normalisation of biomass, leaf area and radiation use efficiency proved more difficult because of the heterogeneous conditions within the greenhouses. Plant growth and biomass was measured over time in all platforms. Plant biomass and intercepted light are closely correlated and this relationship is definition by the radiation use efficiency. Analysing radiation use efficiency across experiments may thus allow to compare biomass corresponding to different genotypes by estimating the balance of photosynthesis and respiration of these genotypes. However, the analysis did not prove successful because of i) an insufficiently precise measurement of light in the network of greenhouses and ii) insufficiently precise calibration for biomass and leaf area calibration across platforms. The participating platforms have decided to improve both aspects in the future.
The experiment clearly indicate that normalization of environmental and phenotypic data is of substantial importance for the future comparison of phenotypic data obtained at different platforms. In particular, methods for the estimation estimating of spatial variability of environmental conditions needs to be clearly defined and implemented in different platforms to be able to facilitate joint analyses of experiments.
Task 3 Development of exemplary assays for structure and function analysis
Objectives:
The main objective of this task is to develop methods for high throughput detection and diagnostics of plant responses to abiotic and biotic stress factors using optical reporter signals. The assays are based on changes in chlorophyll fluorescence emission transients, changes in excitation/emission spectra of chlorophyll (red/far red) as well as spectral properties of leaves/plants (reflectance) as the important reporter signals for plant phenotyping under various stress conditions. Feature selection procedures for individual stress/optical signal combination are trained and validated on representative datasets and the developed optical reporter parameters and assay methods are available to the consortium for use within different phenotyping platforms and for validation.
Achievements:
Within this WP we developed, validated and compared a number of different non-invasive methods which have the potential to be implemented as high-throughput optical methods for the evaluation of plant response to the effects of environmental biotic and abiotic stress conditions. A literature overview with nearly 100 publications presents a matrix with measurement parameters obtained by different optical methods and the environmental stress they potentially can be used for. The resulting protocols and literature references are published on the website: http://www.plant-phenotyping-standards.net/.
Additionally within this WP we also specifically tested some of these optical methods such as different parameters related to chlorophyll fluorescence imaging, spectral and thermal imaging with different plant species exposed to different environmental stress conditions. The experiments defining different treatments are summarized below:
• Drought stress
Different Arabidopsis thaliana accessions were used to evaluate the potential of thermal imaging, spectral reflectance and chlorophyll fluorescence imaging for the non-destructive detection of responses to drought stress and compared with physiological measurements of growth, gas exchange and relative water content. Thermal imaging and spectral reflectance measurement proved reliable tools for tracking drought-induced changes and compared well with physiological measurements across all accessions. Chlorophyll fluorescence approaches, however, did not result in similar response for the used accessions. Therefore combinatorial imaging based on advanced statistics classifiers and feature selection methods on the time series of images was used for the analysis of different chlorophyll fluorescence parameters. Different algorithms were compared and some classifiers proved very useful to give clear distinction between well-watered and stressed sets of plants among chosen A. thaliana accessions (Klem K et al. Journal of Experimental Botany, submitted; Klem. K et al. New Phytologist, submitted).
• Low temperature stress
Cold sensitive and tolerant Arabidopsis thaliana accessions were used to analyse the changes of chlorophyll fluorescence transients between the non-acclimated, cold acclimated, and sub-zero temperature treated plants. A number of different chlorophyll fluorescence parameters can be used to discriminate between the cold sensitive and tolerant plants. Statistical classifiers with the sequence of captured chlorophyll fluorescence images can further improve the discrimination of the sensible and tolerant Arabidopsis accessions (Mishra A et al 2014 Plant Methods, 10: 38).
• UV radiation
Barley varieties differing in their sensitivity to ultraviolet (UV-B) radiation were exposed to different UV-B at different levels of photosynthetically active radiation (PAR). Different photoprotective mechanisms in barley leaves resulted and different levels of constitutive flavonoids and xanthophyll-cycle pigments which were induced by different levels of high radiation stress. Acclimation to UV and PAR was able to ameliorate the negative consequences of high radiation levels on photosynthesis. Both total contents of epidermal flavonoids and the total pool of xanthophyll-cycle pigments were closely correlated with gas exchange and chlorophyll fluorescence measurements (Klem K et al 2015 Plant Physiology and Biochemistry, 93, 74-83).
• Biotic stress: powdery mildew
Different spring barley genotypes with different resistance genes to powdery mildew were monitored with optical methods after infection. In particularly thermal imaging proved very useful in detecting the infection before the appearance of visual symptoms. Some chlorophyll fluorescence parameters also showed that they could be used as early detection tool for infection but the discrimination on the basis of virulence and reaction type was not conclusive (Klem et al, Functional plant Biology, re-submitted).
3.4 IT for high throughput phenotyping
Task 1: Collection of details of existing databases, ontologies and IT infrastructure including capabilities and implementation details (HAS, IPK, AB, MPI, KeyGene, Phenome, INRA, FZJ, HMGU, APPF)
Objectives:
A set of different IT solutions are available to the phenotyping community. These comprise phenotyping platforms with integrated databases, used ontologies, databases storing environmental characterisations and/or systems providing visualization and evaluation capabilities which are necessarily heterogeneous. The different implementations present in EPPN were to be collected and compared and based on this compendium a set of recommendations and conventions for phenotypic databases were to be assembled which will serve the community when developing novel phenotyping solutions.
Achievements:
The task started out from a large set of partners with diverse IT solutions, ontologies, controlled vocabularies etc. at the different partner sites. At the time of the beginning of the project, there was no clear data exchange format nor was there a consensus on how to best facilitate this and which formats could be suitable. Therefore a first task was the mapping of the different datasets and databases and their comparison in terms of variables that were held and stored in the databases. This was also coordinated together with the Australian partners (especially the PODD database) and with the TRANSPLANT EU project. It turned out that this collaboration was very fruitful, as this led to a first joint recommendation of EPPN and TRANSPLANT on how to deal with phenotyping data and that certain ontologies should be used. Also as a conclusion it was agreed that the simple but extendible ISA-TAB format could and should be used for a data exchange and potentially archival. This is because this format will be further maintained and because it is used in a variety of biomedical investigation fields. Furthermore it reflects the typical way in which plant biologists and physiologists think about their data, as data is organized in sheets not to dissimilar from an Excel Worksheet.
However it also became clear that it was absolutely necessary that the contents of the data contained needed to be harmonized and that there is really something like a minimal necessary information that is needed. Therefore the “Minimal information about a plant phenotyping experiment” MIAPPE was developed giving clear recommendations what kind of data is necessary and if voluntary information is provided how this is best described. To a great extent however some existing Minimal information standards could be reused. This recommendation and its description has been published as a recommendation paper in the Journal of Experimental Botany (Krajewski et al 2015).
In addition, EPPN and transPLANT have already drafted a second manuscript detailing implementations and further updates.

Task 2: Adaptation of databases and development of an adapter and translation framework (HAS, IPK, AB, KeyGene, Phenome Networks, INRA, MPI, FZJ, HMGU, APPF)
Objectives:
The plethora of existing phenotyping IT solutions is only partly standardized. However, databases need to be able to communicate with each other by sharing data and analysis capabilities to enable efficient use of resources. To communicate between the different databases, it is not only important to develop common standards (Task 1 and WP Ontologies) but to also have efficient standardized interfaces so that different modules can communicate with each other and to be able to re-use tools (see Task 3)
Achievements:
Based on the achievements in Task 1, it became possible to focus on the mainstay of the data and develop tools dedicated to the analysis and quantification of data. As ISA-TAB was chosen as a standard data format the IAP developed by IPK Gatersleben was adapted and developed to use this exchange format and to build on this data standard. The IAP platform now allows the integration of different data sets, analysis of data and features a very user friendly interface to accommodate the wide phenotyping variety. As such it is also compatible with other EPPN developments. The IAP tool was published in the Plant Cell and it now integrated at several platforms (Chen et al 2014).
Task 3: Framework improvement for efficient data pre-processing (KeyGene, AB)
Objectives:
Plant phenotyping data exchange between sites, data analysis using, and data visualization require data to be standard compliant. But often the data also needs to be represented in numerical or categorical form. Thus, novel pipelines complementing existing approaches have been developed to transform image data to a set of image derived digital phenotypes (sizes, colours, shapes, orientations, skeletons) for each image.
Achievements:
One of main problems in developing methods for plant phenotyping is the “looseness” of plant phenotyping when it comes to imaging. This is because there is no “standardized” plant. Plants come in different forms, heights, and feature many different structures that can be captured through different setups. However in all cases it is absolutely necessary to translate this data back into the structural elements of a wide range of plants by keeping things abstract and defined in the individual plant species. This is because raw image data cannot be computed on for e.g. studies on yield in different genetic plants for GWAS studies. Therefore, only capturing the different plant structures allows a quantitative analysis of effects of genetic and environmental variation on plant structure to be assessed.
As almost all plant based phenotyping experiments are image based in the wider sense however, methods that allow extraction of 3D information about plant structures from these data sets would be an important pre-processing step and likely a bottleneck in many workflows. This is because this pre-processed data set could be further processed to be divided into structures such as leaves, stems and reproductive structures depending on the plant species. Thus a method was developed to partition multiple images of plants using structures from motion principles which has successfully been developed for 3D point clouds of the plant structure. In order to allow this approach not to be tied into a specific plant species, this was developed for the contrasting species on Arabidopsis (a small model dicot), Maize (a large monocot crop species) and others. The method comprises image acquisition using a digital camera and turntable; extraction of local invariant features and matching these from overlapping image pairs; estimation of camera parameters and pose based on Structure from Motion; and the employment of a patch based multi-view stereo technique to develop a dense 3D point cloud. As this initial prototype was judged to be successful it was then extended to e.g. wheat and Brachypodium (as two monocot species). In addition to speed up processing, Graphics Processing Unit computing was tested. Furthermore, as it is important to correctly pair the data between images, improved image pair matching was achieved by algorithmic improvements. In total this resulted in an error for less than 5% when looking at the model plant Arabidopsis.
The method has been published including updates as proceedings as is usual in the computational field: (Lou 2014a and Lou 2014b).
Task 4 Adaptation of existing tools for re-use and development of novel tools for analysing phenotypes and associating them to genetic and genomic information (PhenomNetworks, Keygene, HAS, IPK, INRA, MPI)
Task 4.1 Adaptation of existing tools and development of novel data visualization capabilities (KeyGene, PhenomeNetworks, MPI, HAS, IPK)
Objectives:
Already existing tools need to be adapted to be able to use the standardized API established in Task 2, based on the experience of PhenomeNetworks with plug-in structures and existing frameworks to integrate R. This can permit to re-use the same tools at multiple different sites. In some cases this might require training of potential developers to develop their own analyses as plug-ins (e.g. PhenomeNetworks). Furthermore, keeping the audience of the phenotyping data in mind, data mining and visualization aspects needs to be integrated using novel easily comprehensible visualization techniques e.g. a Phenome Browser in analogy to a Genome Browser.
Achievements:
Data analysis in digital phenotyping projects is cumbersome. This is caused by the large number of images, image angles, time points as well as a large number of digital traits per image. Reduction of complexity to answer very specific questions will thus allow many more researchers to adopt and take advantage of the power of digital phenotyping. Therefore on the one hand side the IAP framework was developed (see also previous Tasks). The IAP framework at the IPK comprises data visualization modules such as growth curves, means etc. In addition, it allows to conduct reproducible analyses, hierarchical clustering, growth models etc. The data can be compiled into a “phenoreport” given the most important analyses results. Also this has led to the improvement of web based tools and suites such as the one from PhenomeNetworks. Indeed, this has led to e.g. the fully integrated website PhenomeNetworks and additional ideas have flown into the IAP framework, this the partners have truly cross fertilized developments. Online it is now possible to browse phenotypic values, to visualize these values and to provide simple plots for phenotypic values. This comprises amongst others histograms, scatter plots, PCA biplots etc. Furthermore similar complementing plotting capabilities have been developed by the partner RWTH. These also comprise the capability to map phenotypic data onto world maps in cases where plant populations with passport data are available to extend beyond phenotypic data and bridge to data interpretation.

3.5 Transnational Access
Plant phenotyping is an emerging field of research based on a strong demand by crop breeders, agricultural industry, and academia. EPPN brings together the leading European phenotyping platforms and enabled access to 23 different plant phenotyping installations. The requirements for phenotyping are very diverse as demonstrated in the Transnational Access experiments with respect to i) plant species ranging from model species Arabidopsis to trees, ii) experimental design which needs to address various and often combination of different stress levels or a very detailed and specific scientific questions, iii) traits that need to be assessed above- and/or belowground with a different levels of detail. In total EPPN has made 66 Transnational Access experiments possible.
Overview of the EPPN installations available for Transnational Access:
Jülich
• SCREEN Chamber shoot analysis system in growth chambers
• SCREEN House shoot analysis system in growth chambers
• SCREEN Root LP analysis of root and shoot of plants in rhizotrons in a greenhouse
• SCREEN Root SP analysis of roots growing in perti dishes in growth chambers
IPK
• APPP shoot analysis of small, intermediate, and large plants
• MP destructive metabolite analyses
HMGU
• expoSCREEN four walk-in size chambers with controlled environmental conditions
• sunSCREEN chambers with a small sun simulator for controlled environment
INRA
• DIAPHEN high throughput facility in the field
• Phenody/Phenoarch shoot analysis in greenhouse facility
• Phenopsis shoot analysis of Arabidopsis thaliana plants
• PPHD root analysis
HAS
• RSDS root analysis in a greenhouse
• SSDS shoot analysis in a greenhouse
UNOTT
• MicroCT 4D root analysis
• Root Trace analysing root structure and growth in agar plates.
• SCREEN Field root assessments on a plot-scale in field experiments from soil cores.
• SCREEN Glasshouse large plant root analysis system in glasshouses
• Vertical Confocal confocal microscopy for non-invasive analysis of root cells
ABER
• FTIR/NMR Fourier transform infrared spectrophotometer
• IPC shoot analysis in a greenhouse
• Micro Raman images of plant samples for chemical composition.
• TGA-py GC/MS Thermogravimetric analysis on plant cell wall composition
Selection and access procedure to EPPN installations
The access and selection procedure for Transnational Access to the EPPN installations was the same for all institutions offering Transnational Access within the framework of the EPPN project. The installations within the Transnational Access scheme were very diverse with respect to the scientific questions they can address and the duration of the potential Transnational Access project ranging from sample analysis which usually can take few days to the analysis of physiological and structural plant properties which may take several weeks or even months. Therefore the number of experiments per installation as well as the preparatory phase of these experiment was very different between installations. To account for these differences we decided to have a permanently open call for applications for access with dedicated deadlines for each installation which were summarized in an access calendar on the EPPN website.
The whole Transnational Access process was defined by three major steps:
Application
The users could select the installation that may fit best their requirements based on the description on the website which included filter tools designed to select the traits of interest and the corresponding installation as well as to identify the application deadline for that specific installation. Prior to the submission of an application every user was specifically asked to contact the scientists responsible for the installation to discuss the feasibility of the experiments envisioned by the users in detail. Based on this discussion the users could adapted the description of the proposed project which should include only feasible experiments as confirmed by the responsible scientist. Only feasible experiments were then introduced into the peer review process after the proposals were submitted to the EPPN office.
Review and selection
Submitted proposals were checked at the EPPN office for eligibility and were then subject to a peer review process by independent experts outside the EPPN consortium. The reviewers were selected by providing expertise in their scientific field of plant biology related to the proposed user projects. The reviewers were requested to submit confidentially written assessments to the EPPN office. The evaluation criteria included: i) technical feasibility of the project, ii) scientific excellence, iii) scientific progress leading to a publication was expected, iv) special support was given to young scientists. Finally, priority was given to first time users from countries lacking phenotyping facilities.
Based on the recommendation of the reviewers and the feasibility of the proposed projects these projects were ranked and assigned a priority. Accordingly, the management committee decided with the responsible scientist operating the infrastructure if the proposed project can be executed. All finalized projects are listed on the EPPN webpage with a short project summary (http://www.plant-phenotyping-network.eu/eppn/selected_projects.
Access
Users selected for access obtained all the logistic, scientific and technological support needed to successfully complete the proposed experiments. User group members were very closely involved in the experiment and were usually integrated into the local working groups during the experiment. A local stuff member, a mentor, supported the user throughout the experimental period, by i) discussing and designing in detail the experiments, ii) training the users in standard operating procedures applicable to the installations, iii) supporting the users during the experimental phase, iv) enabling data processing and analysis, v) assisting in data interpretation including the preparation of manuscripts and reports as appropriate.
All users were accommodated close to the facilities in good quality hotels or guest houses. In addition the institutions assist in renting economic rooms for long-term users.
The results of the Transnational Access scheme
Transnational Access played an essential role in integrating the European plant phenotyping community. EPPN organized a number of workshops, meetings and other activities (see chapter Networking) to specifically inform about the opportunity of Transnational Access and to discuss the needs and requirements of plant phenotyping with the diverse user community. The Transnational Access scheme was very well received by the plant science community and also demonstrated the high demand for plant phenotyping. We received a large number of requests for access including users from outside of Europe and a very close discussion between the phenotyping centres and users with very diverse requirements was initiated. This discussion is essential for the future development of the plant phenotyping science to meet the requirements of breeders, and the plant science community at large. The practical outcome of the EPPN Transnational Access resulted in 88 applications from which 66 were selected for Transnational Access. All experiments were finalized and data analysis has continued beyond the duration of the EPPN project. An overview of the users who obtained access and the experiments performed in the Transnational Access experiments is summarized below.
Users
EPPN addressed a very wide user community. We received applications from 21 countries representing the institutions of the user group leader. The members of the user group who were directly involved in the Transnational Access experiments represent 174 people from 32 different countries. From these users 47% were female and we addressed both young and experienced scientists. The experienced scientists represent very often the supervisors of early career scientists who were then directly involved in executing the experiments at the phenotyping platform. The age of all the users in a user group:
- 34% younger than 35,
- 33% ranged between 36-45,
- 34% older than 46.
Most of the users, approximately 83%, we attracted in EPPN were new users and no previous collaboration was established. Thus, EPPN successfully managed to extend the existing networks and to address very different user groups from diverse research fields who could benefit from the access to the EPPN installations. This large number indicated that there is growing demand for use of phenotyping infrastructure for a growing community.
Experiments
In all approved Transnational Access experiments 20 different plant species were addressed from algae to trees (poplar) indicating the diversity of the platforms and the addressed experiments. Most experiments were performed with the model species Arabidopsis followed by common crops such as wheat, maize and barley. The diverse needs of the user community are also indicated by the range of treatments which were performed at different phenotyping platforms. In general, most experiments addressed abiotic stress specifically plant performance under limited water or nutrient availability. Additionally, complex experiments which focus on multiple stress factors such as combination of drought and nutrient stress, drought and temperature, drought and salinity were performed at a number of platforms as well. There was a large demand specifically for phenotyping of root traits with different level of detail.
Scientific outcome
Currently 34 publications were completed within the EPPN project and about 10 have resulted so far from the Transnational Access activities. We expect that the number will substantially increase because a large number of data from Transnational Access experiments are still under evaluation and the lag time between the performance of the experiment and publication of the data is on average in the range of 2-3 years. Based on the EPPN organized symposium on plant phenotyping in Barcelona on 11-12th of November, 2015 which specifically addressed work obtained by users within the Transnational Access to EPPN installations (http://www.plant-phenotyping.org/Home_EPPN_Symposium) we can expect quite a substantial number of publications resulting from the Transnational Access activities.
In summary, Transnational Access provides a very powerful approach to address important questions of and by the users who usually would not have access to the infrastructure enabling them to do this research addressed in EPPN. The close interaction of the users with the responsible scientists at the phenotyping platform adds another important aspect which is a very close interaction needed to integrate the plant phenotyping community. In particular the large number of new and young users may provide a basis for future and long term interactions. In a number of cases this interaction resulted already in follow-up activities such as joint research proposals.

Potential Impact:
4. The potential impact of EPPN
Why do we need plant phenotyping?
Food security and increasing plant biomass in a sustainable manner for human nutrition and bio-industries is the key challenge for the coming decades. Food and biomass production will have to be doubled by 2050 to match the increasing demand (FAO) which is even more challenging in times of climate change (IPPC). In the past, agriculture expanded the production by increasing inputs of nutrients and selective breeding as well as by occupying larger agricultural areas. However, the area of productive land is actually decreasing in many European countries and further increase of nutrients will hardly increase plant productivity but lead to leaching and runoff from cropland causing environmental pollution. This combination of factors requires integrated solutions and new technologies to improve plant production, based on knowledge-driven innovation to the farming sector as well as agricultural and seed industries.
Today´s breeding approaches produce typically about 1% yield increase per year for the key crops such as wheat or maize. This rate is far below the required 1.5 - 2.2% improvement needed to meet the increased global demand in particular in times of climate change which additionally negatively affects plant productivity. In the past decades, genetic improvement has been the major source of yield increase. While genotyping is now very rapid and costs are decreasing due to industrialized sequencing pipelines, phenotyping which is the analysis of crop performance with respect to structure, function, quality and interaction with the environment, remains the major bottleneck in plant sciences as well as for the exploitation of crop genetic diversity. Because of the rapid developments in plant molecular biology and in molecular-based breeding techniques, (i) an increasing number of species have been sequenced and large collections of mutants, accessions and recombinant lines allow now analysis of the gene functions. (ii) High-definition genotyping can be carried out on thousands of plants in a highly robotized way allowing association genetics or multi parental Quantitative Traits Loci (QTLs). (iii) For transcriptomic, proteomic and metabolomics analyses, large robotized platforms are increasingly available so the biochemical status of plants can be investigated in detail at reasonable costs. In contrast, the understanding of the link between the genotype and phenotype has progressed more slowly, despite of recent advances and establishment of new plant phenotyping platforms. Additionally, the transition from lab to field, i.e. from controlled conditions with focus on single plants to stands in a heterogeneous field represents a challenge. Faster progress requires new technical and conceptual approaches to quantitatively analyse the existing genetic resources for their interaction with the environment. Advances in phenotyping are therefore a key factor for success in modern breeding as well as for basic plant research.
EPPN integrates the plant phenotyping community
EPPN specifically addressed the phenotyping bottleneck in particular by integrating the European plant phenotyping community and by advancing the further development and implementation of methodology, infrastructure and the emerging science field to maintain Europe´s leadership in plant phenotyping, specifically by addressing a number of important points:
i) by providing access to the relevant European plant phenotyping infrastructure to the scientific community and allowing a systematic application of quantitative phenotyping methodologies to address the gene-environment interaction related to the diverse needs of the user community;
ii) by delivering and testing new in particular non-invasive technologies for the utilisation within the EPPN phenotyping platforms, and the plant science community at large;
iii) by developing of standards of good phenotyping practice (GPP) as the basis for experimental protocols including minimum requirements for environmental monitoring and trait assessment which allow generalized analysis with flexibility to include new procedures to efficiently utilize synergies between phenotyping facilities and set benchmarking references for future phenotyping efforts and developments;
iv) by developing efficient processes to integrate molecular and phenotypic data in a robust and reliable way to identify molecular control of phenotypic responses to relevant environments;
v) by fostering pan-European interaction which also attracted researchers from outside of Europe through close discussion between plant phenotyping scientists from phenotyping centres, users from academia and industry as well as technology developers and decision makers. Thus, EPPN has played an important role in advancing the infrastructure and the emerging science field to maintain Europe´s leadership in plant phenotyping.
In summary, EPPN very effectively addressed and integrated the needs of plant phenotypic community in Europe. In particular, the demand of the community and the broad interest of using phenotyping infrastructures and methodology in many countries in Europe has been frequently expressed by both representatives from academia and industry. The core members of EPPN responded to this demand by an application to the ESFRI forum with a project called EMPHASIS. The proposal was very positively evaluated and EMPHASIS is now listed on the ESFRI roadmap which will facilitate as structured development and use of plant phenotyping infrastructure. Additionally, the plant phenotyping bottleneck is of course well known beyond Europe. The core members took here the initiative as being the leaders in plant phenotyping to initiate an International Plant Phenotyping Network (IPPN) as an association and a global networking hub which specifically addresses plant phenotyping. Thus, EPPN played an essential role in combining and linking this field to effectively address future grand challenges.
Three pillars of EPPN
The EPPN project was defined by three closely related pillars which were supposed to address different aspects of plant phenotyping and the community at large and include: i) networking and close interaction with users and technology developers from academia and industry, ii) enabling access to the leading European plant phenotyping infrastructure and iii) developing technology and standards which can be implemented into the phenotyping platforms. These three pillars are interlinked and their socio-economic impact and the societal implications addressed by EPPN during the duration of the project are described in more detail below.
Networking
The plant phenotyping community represents a very diverse needs and expertise from plant biologists, geneticist, breeders, imaging experts, automation and software engineers, IT experts etc. Therefore the goal of the networking activities was to focus the networking activities around the EPPN phenotyping facilities as a common integrating objective. Networking, communication and education was created at different levels: i) between existing and newly developing phenotyping platforms, ii) between phenotyping platforms and users from academy and industry, iii) between platforms, developers and users and iv) with other leading international phenotyping centres.
At the level of phenotyping platforms we fostered the exchange of knowledge on technologies, protocols and data acquisition. Specifically, the exchange of information happened between experts which was supported by dedicated workshops. Additionally, there was a close interaction within the joint research activities between the platforms with respect to the development of experimental standards and data management. In both cases providing a tangible outcome available for the wider community. Experimental standards which include environmental monitoring and trait assessment were published on the website, while recommendation for data management were published in a joint manuscript with the EU funded project transPLANT (for more details see WP 4 Task 1). Establishment of common standards represents a very essential step in integrating and advancing the plant phenotyping community.
The communication between platform operators and users covered the information of opportunities of phenotyping platforms to users of the phenotyping infrastructure from academia and industry. This networking activity is very closely linked to Transnational Access and the information of the potential users about the opportunity for access to EPPN infrastructure. The interaction with the user community was promoted on different levels. We started a dedicated workshops to inform about Transnational Access and to discuss the need and requirements of plant phenotyping in countries without members in the EPPN project. Additionally, we specifically addressed the very divers users by organizing sessions or actively participate at many different conferences focusing on breeding, plant genetics as well as imaging or precision agriculture (for more details see WP2 Task 8). Additionally, we also focused on the early career scientists by organising dedicated training schools at phenotyping platforms. These activities were very successful, for example the visits on EPPN website continuously increased up to 25.000 visitors per month at the end of the project totalling in over 650.000 visitors.
Networking between platform operators, developers and users is essential, but difficult to achieve in a single step. To address this challenge, in a first step bilateral meetings of phenotyping operators with developers and phenotyping operators with users were established to build an interactive community. In the second step strategic discussions were implemented such as dedicated developer workshops which aimed at identification of novel technology which may be valuable to address specific traits and, identification of methods and benchmarking of these methods to specifically address defined topics such as plant water relations (for more details see WP2 Task 3).
International collaboration is key to exchange ideas and for standardization and benchmarking of different approaches beyond Europe. EPPN platforms organized international meetings related to different aspects of plant phenotyping with different experts. Specifically, EPPN organized these meeting in the context of the International Plant Phenotyping Network (IPPN) which is an association with the goal to become a global networking hub in plant phenotyping. Furthermore, EPPN established a close cooperation with other national and international projects and initiatives by co-organizing workshops and training schools for example with the national initiatives such as the German Plant Phenotyping Network (DPPN), the French Phenotyping Network (FPPN) and the UK-Plant Phenotyping Network (UP-PPN). These national initiatives form also the core of the EMPHASIS project listed in the ESFRI roadmap. Additional activities include close cooperation with EU funded projects, EPPN was also involved in a close discussion with the ESFRI project ELIXIR about the integration of plant phenotypic and genomic data. EPPN organized also workshops with the EU funded WATBIO project and published important recommendations for data management with the EU funded transPLANT project. Close cooperation was established with the COST Action Phenomen-ALL with members from 26 Member States (http://costfa1306.eu/). Finally, EPPN imitated together with the IPPN, the DPPN, the COST Action and the Wheat Initiative (http://www.wheatinitiative.org/) a global survey about the status, needs and development of plant phenotyping, which represents the first large scale mapping of the status of the plant phenotyping community (http://www.plant-phenotyping.org/IPPN-SURVEY).
In summary, the networking activities addressed a wide range of users, developers and new emerging platforms within Europe and beyond. These activities clearly indicated the need for plant phenotyping infrastructure which will be addressed in the ESFRI project EMPHASIS as a potential long term development. Additionally the EPPN consortium represents a core group for an application for a follow-up project of EPPN which is called EPPN2020 (INFRAIA-01-2016-2017: Integrating Activities for Advanced Communities). Nearly 70 installations were identified across Europe which could be included into the proposal representing a very active and vivid community. At the moment of writing his report the EPPN2020 committee is evaluating the priority of installations which correspond best to the needs of the user community in Europe.
Transnational Access
The major aim of EPPN was to create synergies between the leading plant phenotyping institutions in Europe as a nucleus for the development of a strong network which fosters the development of an effective European infrastructure including human resources, expertise and communication needed to support Transnational Access to user communities. The Transnational Access to the leading plant phenotyping facilities represented here the key activity of the project. Transnational Access opportunity was very extensively and widely discussed and advertised within the community at large as describe in the networking paragraph and Transnational Access benefited largely from this activities. We addressed potential users all over Europe resulting in a large number of requests, applications and finally experiments which were performed at the phenotyping facilities.
The access procedure included a review process of applications which were evaluated as feasible by the responsible scientist at the phenotyping platform based on a close interaction with the applicant prior to the submission. The reviewers were independent scientists and experts in specific experiments for access to a specific installations. Thus in total over 60 different independent reviewers were involved in evaluating the Transnational Access which also became an additional effective instrument to inform the plant science community about the access opportunities. The demand for access was very high and resulted in 88 applications from which 66 were selected for Transnational Access. However, there was a substantially higher amount of requests from users whose demands would require establishment or implementation of methods which according to the technical limitations cannot be routinely used at the required throughput. For instance, phenotyping of root properties under realistic scenarios in soil was highly demanded. Within EPPN only one installation could address this demand with a throughput which was often below the requirements of the users to measure e.g. mapping populations with often 200 and more genotypes. Thus, there is still a substantial demand for the development of technological and conceptual approaches which can be implemented into routine applications to address the needs of diverse users in particular for high throughput plant phenotyping. Additionally we received a number of requests from non-eligible users for instance from outside of Europe.
The user groups who obtained access included both early career and experienced scientist who were involved in discussing and planning the selected experiments. About 50% of the user group members were female scientists. This interaction addressed in many cases new users (83%) fostering the interaction and establishment of new partnerships which also resulted in follow up projects beyond EPPN. During the Transnational Access experiments, the members of the user group were directly involved in the performance of the experiments at the phenotyping installations and spent often weeks and months at the respected institution providing access. Here we specifically addressed early career scientists, who being directly involved in the practical execution of the experiments obtained training to become familiar with phenotyping approaches while achieving the set scientific goals for each project. Users were also closely integrated into the local often very multidisciplinary groups, they participated at all group or lab meeting, presented and discussed the obtained results in these meeting, which formed an important basis for the establishing of a network, particularly important for young scientists.
Beyond fostering close cooperation and networking between scientists, Transnational Access addressed of course important scientific questions of users who otherwise would not be able to address these questions. A number of publications resulted directly from the activities within the Transnational Access and we expect at least 15 or more publications because the lag time between the performance of the experiment and the publication of the data is often longer than two or three years and most experiments were performed in 2014. The Transnational Access obtained also opportunities to disseminate the results e.g. in an international Symposium in Barcelona organized by EPPN and dedicated to the dissemination of Transnational Access results.
In summary, the Transnational Access scheme represented an important approach for the users to obtain access to installations otherwise they would not be able to use, to train early career scientists in phenotyping approaches and to establish important contacts and networks within the community. Additionally, Transnational Access represented an important test case for the access providers and projects related to phenotyping infrastructure underlining once more that there is a phenotyping bottleneck which has to be addressed. Finally the need to further develop plant phenotyping as a scientific discipline to quantitatively assess required plant traits and plant phenotyping to become a tool for breeders or the agro-industry at large is essential. A close interaction between phenotyping platforms and diverse users is essential to effectively advance plant phenotyping.
Joint Research Activities
Joint research activities resulted in an adaptation and development of novel sensors, methods and standards for application in plant phenotyping. Innovative phenotyping concepts which integrate mechanistic, medium- and high throughput phenotyping were made available to the community. The joint research activities were divided into three distinct parts summarized below.
Novel instrumentation for plant phenotyping
The overall goal of this WP was to provide the foundation for a novel generation of instruments allowing accurate quantification of phenotypic traits related to plant performance. During the course of the project selected instruments were used for benchmarking and experimental proof of concept. The major focus was the use of imaging approaches for trait assessment in particular different approaches to obtain 3-dimensioanl plant models were developed (Golbach et al 2015 Machine Vision and Applications; Kjaer 2015 Sensors; Lou 2014 Advances in Autonomous Robotics Systems; Lou et al 2014 Image Analysis and Recognition) and spectral imaging (Bergsträsser 2015 Plant Methods) was evaluated. Imaging is commonly used in most phenotyping platforms and evaluation of different modes as well benchmarking of different approaches e.g. for tiller counting (Boyle et al 2015 Machine Vision and Applications) is very valuable for the plant phenotyping community. Additionally the sensors for imaging can be applied in different modes either the sensors are transported to the plants (Fanourakis et al 2014 Plant Methods) or the plants to the sensor. Both approaches have obviously advantages and disadvantages and their evaluation is crucial for the design and establishment of new platforms.
Good Practice in Phenotyping
The main objective of this work package was to define essential elements of phenotyping procedures to optimise protocols aimed at answering biological questions, from experiment design to data interpretation, and to compare results from different platforms and experiments. Specifically we compared the same genotypes grown at different locations under different environmental conditions to evaluate the response of these genotypes to the environment and to test the reproducibility of data in phenotyping platforms. An essential step here was the establishment of agreed on standards which were also published on the EPPN website and are freely available to the community. During the project the set of standards was continuously updated and extended, for example with approaches for the use of optical method for stress detections which were partly also based on studies performed within EPPN (Šebela et al 2015 Field Crops Research, Klem et al 2015 Plant Physiology and Biochemistry, Mishra 2014 Plant Methods). Standards in plant phenotyping represent an essential step to enable for comparison of phenotypic data from different platforms. EPPN initiated an important step for the establishment of good phenotyping practice and implementation of these standards into platforms which will have to be further continued beyond EPPN and has been recognized as an essential aspect in EPMPHASIS, IPPN and EPPN2020 provided the project is funded.
IT for high throughput phenotyping
Many plant phenotyping platforms have been established within the last years having often their own databases and solutions. Scaling up experimentation increasingly requires appropriate data management schemas to make use of quantitative analyses. In particular, the lack of commonly accepted standards, ontologies or APIs represents a challenge which we addressed in EPPN. Usually common set of parameters is present in all databases including proprietary system. In phenotyping the extraction of a phenotypic traits from non-destructive plant imaging plays an essential role and were defined in a case study with barley (Chen et al 2014). Further development included an establishment of a working demonstration pipeline for data management, image analysis, and result visualization of large-scale phenotypic data sets (Klukas et al 2014). This activities have led to a cooperation with the EU funded project transplant which resulted in a publication with recommendations addressing wider plant phenotyping community (Krajewski et al 2015). Three different aspects for standardization of plant phenotyping experiments were specified: i) define a detailed checklist of attributes summarizing the data content following a Minimum Information principle as pioneered by other communities in bio-medical sciences; ii) annotate attributes and phenotypes as much as possible using publicly available vocabularies and ontologies; iii) use of formats specifically designed according to MIAPPE definitions http://cropnet.pl/phenotypes. The interaction with transPLANT project has been currently extended beyond the EPPN project by establishing a close discussion within the plant phenotyping community which will be led by ESFRI projects, EMPHASIS and include close interaction with ELIXIR.
In summary, the joint research activities addressed important aspects of plant phenotyping centred on developing and benchmarking of sensors and methods highly relevant for plant phenotyping as well as standards essential for experimental design and data management. The activities have been widely disseminated in scientific publications in different meetings and address a wide community including scientists as well as the industry sector and technology developers. Future continuation of these activities will be ensured by a number of projects and networks initiated by EPPN.
Dissemination activities and exploitation of results beyond the duration of the project
Plant phenotyping addresses important solutions to grand societal challenges which are important beyond Europe. Integrating the plant phenotyping community was the overarching goal of the EPPN project and to strengthen the competitiveness of European research and industry. EPPN became a nucleus to enable discussion and interaction between different stakeholders from academia and industry and played an important role for the initiation of future projects and activities.
One major aim of the project was to provide a platform for knowledge and information exchange to the scientific groups working in the field of plant phenotyping such as agricultural and plant science, breeders, agroindustry and technology developers. We focused on integrating the interdisciplinary endeavour to strengthen the development of plant phenotyping infrastructure by a systematic exploration of the area, and the identification of common problems and solutions to advance plant phenotyping at large. We pursued the goal to advance plant phenotyping in basic research as well as in application for breeding and agricultural management across Europe. To reach these gaols we effectively deliver information to the communities by identify leading groups, including EU projects that were installed during the operating time of EPPN and, that have the capacity to radiate and implement relevant expertise and information. Specifically, synergy building was fostered rather than doubling efforts in establishing the same approaches or equipment within existing and developing plant phenotyping platforms. Therefore EPPN aimed to provide help to installations at other places to become more cost- and labour-effective. This includes support for integrated platforms as well as for groups that build expertise in a specialized setup and field. Additionally the long term development of this field requires the advancement of dedicated skills and expertise in the next generation of scientist. EPPN integrated the different disciplines and expertise in plant phenotyping by interaction with young scientist who were directly addressed in training activities and in mobilisation human resources for the European and global research community in Transnational Access where mostly young scientists were involved in close intense interaction with phenotyping experts at European in phenotyping centres.
Addressing wider plant phenotyping community
Dissemination activities were effectively addressed by a number of networking activities described earlier. Beyond the aims described in Networking we fostered active interaction and sharing of expertise between research organizations to their mutual benefit. EPPN partners illustrated the activities of the project as “Best Practice Examples” on plant phenotyping and present them to the respective communities. EPPN experts were available for questions on plant phenotyping applications and installations in Europe with respect to purchase, implementation and experimental setups. For example EPPN consulted the institutions which were interested in the establishment of national platforms in The Netherlands, Poland, Romania and a several other countries.
EPPN also fostered a close cooperation with other national and international networks and projects. A close cooperation was established with the COST Action with 26 Member states as members. Dissemination of the EPPN activities during COST activities addressed a wide community initiating a number of follow up activities such as common projects. EPPN was also presented as a “Best Practice Examples” to support the activities of the Nordic Plant Phenotyping Network, the Expert Working Group on Plant Phenotyping of the Wheat Initiative. Additionally, EPPN core members initiated the International Plant Phenotyping Network, an association dedicated to networking activities on a global scale. This will extend the EPPN network of European and Australian partners and include important phenotyping centres across the globe into the discussion about future developments of plant phenotyping. Currently, 23 different institutions across the globe are members of IPPN. Within Europe EPPN became the nucleus for the organisation of the project EMPHASIS. Thus, expertise and experience gathered during the EPPN project represents an important test case for future implementation of European plant phenotyping infrastructure within the framework of ESFRI.
EPPN also strongly fostered the development of common standards in plant phenotyping related to both experimental standards including trait assessment and environmental monitoring as well as data standards. In particular addressing the wider plant phenotyping community with recommendations for standardization for instance with common publication with other projects such as the EU funded project transPLANT represent an important step in implementing this important aspect of plant phenotyping. Plant phenotyping data accumulate very quickly giving us a great potential for expediting both discovery and application if these data are made publicly available for analysis and are sufficiently managed by standards that would ensure interoperability among data providers. This important activity will be continued by EMPHASIS.
Education and Training of scientists
The expertise combined in EPPN covered a broad range comprising scientists and technologists, engineers, sensor and IT specialists. In other to exploit such multidisciplinary nature, the consortium integrated the different disciplines and expertise. We organised and facilitated workshops where these cross discipline approaches and integrating skills were experienced. The EPPN project contributed to training researchers (and staff) across traditional scientific fields and barriers and pragmatically implemented the policy on lifelong learning. As a consequence, the staff to be employed for and trained by completing the project added to the European human capital, reinforcing the European resource base fundamental for the establishment of Europe as the most competitive knowledge-based society.
EPPN provided also dedicated training activities related to relevant aspects of plant phenotyping (see Networking) and practical training at the plant phenotyping centres during Transnational Access experiments (see Transnational Access). In both cases over 100 early career scientists were addressed representing an important knowledge transfer and the most important basis for future development of the next generation of plant phenotyping scientists.
General Public
The general public was made aware of the science and technology aspects within the EPPN project. Plant phenotyping is urgently required to improve crop varieties, while simultaneously providing huge potential to raise the interest of the public for plant sciences in general. It often provides images or image sequences that illustrate usually not realized processes in plants, soils and food stuff. EPPN partners presented regularly the activities of the network beyond the usual scientific audience.
The overarching dissemination goal was to increase the acceptance and awareness of scientific achievements of the kind obtained in the EPPN project with respect to the bioeconomy goal of the European Union. The EPPN partners used examples with high impact potential for the media by: i) illustrating the results not only in peer-reviewed science journals, but also by dissemination through public media; ii) setup of the EPPN website including a newsletter and the establishment of a mailing list of interested persons. This was not only open for scientists, but also for interested public; iii) the consortium participated and present its work also within more general conferences e.g. scientific workshops but also more strategic meetings such as the biannual meeting of the European Science Organization (EPSO) including science policy issues.
References
Bergsträsser S et al (2015) HyperART: non-invasive quantification of leaf traits using hyperspectral absorption-reflectance-transmittance imaging Plant Methods 11: 1
Boyle RD et al (2015) Automated estimation of tiller number in wheat by ribbon detection. Machine Vision and Applications, in press.
Chen D et al (2014) Dissecting the phenotypic components of crop plant growth and drought responses based on high-throughput image analysis. The Plant Cell 26, 4636-4655.
FAO, Rome (2009) The state of food and agriculture. FOOD AND AGRICULTURE ORGANIZATION OF THE UNITED NATIONS.
Fanourakis D et al. (2014) Rapid determination of leaf area and plant height by using light curtain arrays in four species with contrasting shoot architecture. Plant Methods 10: 9
Golbach F et al (2015) Validation of plant part measurements using a 3D reconstruction method suitable for high-throughput seedling phenotyping Machine Vision and Applications. In press.
Kjaer KA & Ottoson CO (2015) 3D Laser Triangulation for Plant Phenotyping in Challenging Environments. Sensors, 15: 13533-13547
Klem K et al (2015) Ultraviolet and photosynthetically active radiation can both induce photoprotective capacity allowing barley to overcome high radiation stress. Plant Physiology and Biochemistry 93, 74-83
Klukas C et al. (2014) Integrated Analysis Platform: An Open-Source Information System for High-Throughput Plant Phenotyping. Plant Physiology 165: 506-518
Krajewski P et al. (2015) Towards recommendations for metadata and data handling in plant phenotyping. Journal of Experimental Botany, 18: 5417-5427
Lou L et al. (2014a) A Cost-Effective Automatic 3D Reconstruction Pipeline for Plants Using Multi-view Images. In Advances in Autonomous Robotics Systems Edited by Michael Mistry, Aleš Leonardis, Mark Witkowski, Chris Melhuish. Pages 221-230
Lou L et al. (2014b) Accurate Multi-View Stereo 3D Reconstructionfor Cost-Effective Plant Phenotyping. In Image Analysis and Recognition Edited by Aurélio Campilho, Mohamed Kamel. Pages 221-230.
Mishra A et al (2014) Chlorophyll fluorescence emission can screen cold tolerance of cold acclimated Arabidopsis thaliana accessions. Plant Methods 10: 38
Poorter et al. 2012 The art of growing plants for experimental purposes: a practical guide for the plant biologist. Functional Plant Biology 39: 821-838
Poorter et al. 2012 Pot size matters: a meta-analysis of the effects of rooting volume on plant growth. Functional Plant Biology 39: 839-850
Šebela D et al (2015) Temporal chlorophyll fluorescence signals to track changes in optical properties of maturing rice panicles exposed to high night temperature. Field Crops Research 177: 75-85
Tardieu F (2013) Plant response to environmental conditions: assessing potential production, water demand, and negative effects of water deficit. Frontiers in Physiology 4, 17

List of Websites:
Website: http://www.plant-phenotyping-network.eu/
Contact details:
Coordination: Prof Ulrich Schurr (u.schurr@fz-juelich.de)
Project management: Dr. Roland Pieruschka ((r.pieruschka@fz-juelich.de)
Administration: Petra Insberg (p.insbertg@fz-juelich.de)

final1-logo-eppn.jpg

final1-fig-1-gxe-triangle.jpg

Final Report Summary - EPPN (European Plant Phenotyping Network)

Share this page

Download