Search

Life Science (LS) communities must increase the sustainability of their data resources, software tools, and workflows, to allow their long-term re-use by the wider scientific community, especially in future applications involving machine-based analyses. Here, we describe key findings regarding sustainable LS resources, based on experiences from the EOSC-Life project. EOSC-Life involves 13 European LS Research Infrastructures. It lays the foundation for an open, digital and collaborative space for biological and medical research. We describe organisational, technical, financial and legal/ethical challenges that represent the main barriers to sustainability in the LS domain. Using lessons from 27 scientific projects selected via open calls, we demonstrate the efficiency of the EOSC-Life support model for sustainable FAIR data management and explore the complex sustainability needs for sensitive- and industry-related data resources. We describe cross-disciplinary sharing of best practices, and how this contributes to the sustainability of knowledge across communities, through building of training resources. We formulate a set of recommendations, focusing on working with experts, and communicating outcomes, to establish strong credibility and recognition of research projects. Furthermore, demonstrating and practising also permit a more sustainable adoption of reproducible research outputs. Recommended measures include sustainable training methodologies, as well as ensuring high quality metadata to drive the sustainable reusability of scientific objects. Responsibilities and roles to ensure sovereignty, sustainable services and tools must also be attributed. Finally, our analysis demonstrates how data harmonisation facilitates interoperability of tools, data, solutions and a better understanding of concepts, functionalities and semantics in the LSs., This research is mainly a product of the EOSC-Life European programme funding from the European Union's Horizon Europe research and innovation programme under grant agreement Nº824087. Complementary support was provided through EU funded project AgroServ (grant agreement Nº101058020), EU funded project BY-COVID (grant agreement Nº101046203), EU funded project DANUBIUS-IP (grant agreement Nº101079778), EU funded project EMPHASIS-GO (grant agreement Nº101079772), EU funded project FAIRplus (IMI grant agreement Nº802750), EU funded project FAIRsharing (Wellcome grant agreement Nº212930/Z/18/Z), EU funded project ISIDORe (grant agreement Nº101046133), EU funded project Precision Toxicology (grant agreement Nº965406), UKRI DASH (grant agreement NºMR/V038966/1). Special thanks to T. Biro and her radical collaboration team from Research Data Alliance who gave us great inspiration on how to lead this radical collaboration work.

The original scope of task 8.3 is to develop a framework to assess the impact on reproducibility of the availability of life-science open data and workflows in the cloud.Such a framework should include the understanding of the tools and services produced and made available, the gaps they are filling up and their expected impact.A great part of the activities within EOSC-Life is actually related to reproducibility and provides in several ways tools that will have an impact. We describe here such activities and explain why and how they will have an impact on reproducibility in life sciences. In addition, for each action, we provide an indication of how we could measure its impact in practical ways, using adequate proxies. For these reasons, we changed the title of the deliverable, from “framework to assess ...” to “framework to enhance reproducibility”.First of all, we reasoned on what can be the contribution of open science to improve the reproducibility of research. Publicly sharing data, protocols, tools and computational workflows makes it possible to compare or combine the data and outcomes from different studies within a discipline as well as integrate data across scientific domains. It allows conclusions to be validated and possibly corrected as well as being reinforced by meta-analyses. Replication data and test/training data can also be used in many applications to contribute to reproducible research. Moreover, new hypotheses, different from the original aims of the study, can be explored. Datasets can be re-used to develop and test new methods, to conduct scientific and technical benchmarking activities and to support training activities. Therefore, in addition to generating more value from research investments, data sharing has the potential to increase confidence in research outcomes and increase knowledge dissemination. These benefits of open sharing have long been recognized in some fields such as bioinformatics, which has a long history of publicly sharing data with, for example, public repositories for nucleotide sequences going back 30 years and the Protein Data Bank (PDB), a repository of information about the 3D structures of proteins, nucleic acids, and complex assemblies, that celebrates its 50th birthday this year.In this notion, any improvement in sharing of data, tools and workflows among scientists and across disciplines, which is the aim of EOSC-Life and the wider EOSC, will contribute to reproducible science. In addition to this general scope, several specific actions to frame transparency in the reporting of experimental protocols, data and analytical workflows warrant the reproducibility of every single object (experimental results, data, or workflows) that is made available on the cloud.We describe here the initiatives in EOSC-Life to implement existing tools for reproducibility as well as to develop new tools for its enhancement. As the final goal of EOSC-Life is to make data resources available to the wider community of life scientists, although necessarily technical in several points, this document aims at a general readership, including experimental in addition to data scientists.

The original scope of task 8.3 is to develop a framework to assess the impact on reproducibility of the availability of life-science open data and workflows in the cloud. Such a framework should include the understanding of the tools and services produced and made available, the gaps they are filling up and their expected impact. A great part of the activities within EOSC-Life is actually related to reproducibility and provides in several ways tools that will have an impact. We describe here such activities and explain why and how they will have an impact on reproducibility in life sciences. In addition, for each action, we provide an indication of how we could measure its impact in practical ways, using adequate proxies. For these reasons, we changed the title of the deliverable, from “framework to assess ...” to “framework to enhance reproducibility”. First of all, we reasoned on what can be the contribution of open science to improve the reproducibility of research. Publicly sharing data, protocols, tools and computational workflows makes it possible to compare or combine the data and outcomes from different studies within a discipline as well as integrate data across scientific domains. It allows conclusions to be validated and possibly corrected as well as being reinforced by meta-analyses. Replication data and test/training data can also be used in many applications to contribute to reproducible research. Moreover, new hypotheses, different from the original aims of the study, can be explored. Datasets can be re-used to develop and test new methods, to conduct scientific and technical benchmarking activities and to support training activities. Therefore, in addition to generating more value from research investments, data sharing has the potential to increase confidence in research outcomes and increase knowledge dissemination. These benefits of open sharing have long been recognized in some fields such as bioinformatics, which has a long history of publicly sharing data with, for example, public repositories for nucleotide sequences going back 30 years and the Protein Data Bank (PDB), a repository of information about the 3D structures of proteins, nucleic acids, and complex assemblies, that celebrates its 50th birthday this year. In this notion, any improvement in sharing of data, tools and workflows among scientists and across disciplines, which is the aim of EOSC-Life and the wider EOSC, will contribute to reproducible science. In addition to this general scope, several specific actions to frame transparency in the reporting of experimental protocols, data and analytical workflows warrant the reproducibility of every single object (experimental results, data, or workflows) that is made available on the cloud. We describe here the initiatives in EOSC-Life to implement existing tools for reproducibility as well as to develop new tools for its enhancement. As the final goal of EOSC-Life is to make data resources available to the wider community of life scientists, although necessarily technical in several points, this document aims at a general readership, including experimental in addition to data scientists.

This deliverable 3.2 is a report on the demonstrator projects, the eight scientific and technical pilot projects that were selected to provide concrete scientific use-cases and guide and structure the work done in EOSC-Life to build an open digital and collaborative space for biological and medical research. We report in this deliverable the process of integration of the demonstrators within EOSC-Life, the achievement of the demonstrators who developed and made available to the scientific community several valuable resources (databases, workflows, web platform...), the actions undertaken within EOSC-Life to disseminate the demonstrator achievement and finally the results of the demonstrator survey to learn from the demonstrator experience and improve the integration of the new pilot project within EOSC-Life.

The original scope of task 8.3 is to develop metrics to assess the impact on reproducibility of the availability of life-science open data and workflows in the cloud. A great part of the activities within EOSC-Life is actually related to reproducibility and provides in several ways tools that will have an impact. Therefore, we decided that it would be more informative to describe such activities and to explain why and how they will have an impact on reproducibility in life sciences, instead of providing abstract metrics for such an impact. For these reasons, we changed the title of the deliverable, from “framework to assess ...” to “framework to enhance reproducibility”. First of all, we reasoned on what can be the contribution of open science to improve the reproducibility of research. Publicly sharing data, protocols, tools and computational workflows makes it possible to compare or combine the data and outcomes from different studies within a discipline as well as integrate data across scientific domains. It allows conclusions to be validated and possibly corrected as well as being reinforced by meta-analyses. Replication data and test/training data can also be used in many applications to contribute to reproducible research. Moreover, new hypotheses, different from the original aims of the study, can be explored. Data sets can be re-used to develop and test new methods, to conduct scientific and technical benchmarking activities and to support training activities. Therefore, in addition to generating more value from research investments, data sharing has the potential to increase confidence in research outcomes and increase knowledge dissemination. These benefits of open sharing have long been recognized in some fields such as bioinformatics, which has a long history of publicly sharing data with, for example, public repositories for nucleotide sequences going back 30 years and the Protein Data Bank (PDB), a repository of information about the 3D structures of proteins, nucleic acids and complex assemblies, that celebrates its 50th birthday this year. In this notion, any improvement in sharing of data, tools and workflows among scientists and across disciplines, that is the aim of EOSC-Life and the wider EOSC, will contribute to reproducible science. In addition to this general scope, several specific actions to frame transparency in the reporting of experimental protocols, data and analytical workflows warrant the reproducibility of each single object (experimental results, data or workflows) that is made available on the cloud. We will describe here the initiatives in EOSC-Life to implement existing tools for reproducibility as well as to develop new tools for its enhancement. As the final goal of EOSC-Life is to make data resources available to the wider community of life scientists, although necessarily technical in several points, this document aims at a general readership, including experimental in addition to data scientists.

This deliverable reports on the outcome of Task 9.1 ”Analysis of the training needs of different EOSC-Life users”. An online survey was developed to get structured feedback on the training landscape from the Biological and Medical ESFRI research infrastructures (BMS RIs). One of the aims was to identify gaps in existing training offers and developing a plan on how to address these gaps. 35 responses were received, from 13 research infrastructures (RIs). All 13 RIs provide training as part of their activities, which is a great starting point for the EOSC-Life training strategy. The training needs with respect to EOSC-Life topics were analysed, which RI(s) expressed interest in the training topic, which RI(s) already provide training and which RI(s) are interested in providing training in the future. The survey results underline that the 12 pre-selected topics in the grant proposal cover the EOSC-Life training needs extremely well. However, the survey also highlighted that there is a need for outreach and awareness raising about EOSC in general, and EOSC-Life specifically, to RI staff within and outside of the project. In addition training needs were collected from the other EOSC-Life WPs and discussed during the project retreat in October. As a next step, we will facilitate that training in high demand areas will be expanded, e.g. by facilitating exchange of experience between RIs that expressed interest in providing training in a given topic with those already providing this training. In addition, as response to a high demand of introductory training on EOSC-Life and the clear need for EOSC-related outreach, a Translator group was initiated.

The imaging use-case addresses interoperability of large-scale image data sets, which is required for reusing and comparatively analyzing data sets generated by different sources. Cellular phenotypes are good reporters of underlying cellular processes. This observation motivates the use of experimental systems such as RNAi screens or gene knock-out mice from which the molecular underpinning of cellular phenotypes can be inferred. The same information is however difficult to get for human diseases, usually requiring costly genome-wide association studies that often end up being inconclusive. Comparing cellular phenotypes of human disease samples to those of experimental systems could be a faster and cheaper alternative by allowing for transfer of information suggesting likely molecular mechanisms or biological processes affected in the human samples. This information could then be used to predict biomarkers for the corresponding diseases. However, because annotations of cellular phenotypes are made of free text descriptions that are often ambiguous, comparison of cellular phenotypes within and across domains is currently not possible. To harmonize the annotation of cellular phenotypes, WP6 partners have developed the Cellular Microscopy Phenotype Ontology (CMPO), a species neutral ontology for describing phenotypic observations at the cellular level. CMPO was used to annotate images from gene knock-out mice, human cancer samples and RNAi screens to try and recover preselected candidate biomarkers. Analysis of the resulting annotation data uncovered different annotation practices across RNAi screens, mouse and human samples that hamper comparisons across these imaging domains. Our observations suggest a possible reorganization of CMPO and indicate that, in addition to annotations standards, we also need standardized annotation practices.