Search

The EJP RD stewards' team was established in July 2020 to support European Reference Networks (ERNs) in creating Findable, Accessible, Interoperable and Reusable (FAIR) patient registries. To track the evolving registry FAIR implementations, the steward's team collected the ERNs' reported updates in an implementation status inventory. We here analyze the progress registered to understand the impact of the team's approach on supporting ERNs. The team conducted two rounds of online interviews in 2020 and 2021 to collect the ERNs implementation status. Prior to the interview, they sent ERNs an inventory structured into five phases: a) data modelling; b) data collection; c) registering the registry; d) data accessibility, e) Data querying; and instructed ERNs to fill it using the terminology "Implemented", "Implementing assisted by expert", "Plans to Implement", "Needs Expert help" and "Not applicable". From a total of 24 ERNs, 20 completed the inventory. The results indicate an increase in the number of FAIR tools implemented from 2020 (32.6%) to 2021 (56.0%). Additionally, the percentage of tools that shifted from "Plans to implement" or "Needs expert help" to "Implemented" in 2021 were, respectively, 35.8% and 18.8%. In both years, the most significant increases in implemented tools are related to the modelling phase - 17.3% in 2020 and 29.2% in 2021 - and "making data accessible for querying" - 5.2% in 2020 to 13.2% in 2021. Out of all five stages, the tools more frequently marked as "Plans to Implement" in 2020 concerned the modelling phase (9.8%), which in 2021 decreased to 4.6%. Changes in staff and different people completing the inventory are limitations of this analysis, as they could result in different interpretations of categorizing the tools and, occasionally, the loss of reported evolution. Regardless, the joint approach between ERNs and Stewards appears to have resulted in more implemented tools, mainly in the modelling stage. It also allows prioritizing future efforts on accessibility and discoverability for querying, and data reuse. We expect the number of implemented tools to plateau and the tools planned for implementation or needing expert help to decrease as a result of the ERNs learning more about the options for FAIR data.&nbsp

The process of making data Findable, Accessible, Interoperable and Reusable (FAIR - FAIRification) varies across projects, depending on their objectives and domain needs. However, such variation can complicate identifying the best workflow for an efficient large-scale national implementation. It is therefore necessary to align the different FAIRification processes and workflows. With this aim in mind, we mapped a set of FAIRification workflows to understand the divergence among them and identify critical common FAIRification and domain-specific steps. This work was executed in the context of the DutchHealth-RIand theEuropean Joint Programme on Rare Diseases (EJP RD), which both rely on the FAIR principles to enable sustainable data reuse in the healthcare domain. First, we gathered a group of FAIR experts amongst the collaborators in the two initiatives. Next, we identified a set of relevant workflows used within the rare disease domain and the Dutch health research infrastructure context. A total of seven workflows were selected, including the EJP RDFAIRopoly, given its compilation of theDe NovoandGenericworkflows. Through several hands-on meetings, we discussed the interpretation and best mapping of the workflows’ steps. Finally, we considered any steps that were mapped in more than four workflows (>50%) as critical. Preliminary results show three critical steps: a) identification of FAIRification objectives and expertise, b) (meta)data assessment, definition and semantic modelling, and c) defining data licensing (or consent), access and use conditions. The first can impact the success of the FAIRification of projects, whereas (meta)data assessment and modelling are intrinsically related to the core of the FAIR principles: semantic interoperability and machine actionability of (meta)data. Finally, licensing (consenting), accessing and reusing conditions are equally crucial for healthcare data reuse as they protect sensitive information. Future work includes identifying and analysing domain-specific steps and detailing each critical step into recommended resources, checklists and needed expertise. Any training needs regarding the FAIR expertise capacity can then be identified from such descriptions. We expect that embedding critical steps in the national research infrastructure can improve FAIR data availability.

Background: Allele-specific expression (ASE) refers to divergent abundances of allelic copies of a transcript and is quantified by RNA sequencing. Multiple studies have shown that ASE likely plays a role in many diseases by modulating the pathogenesis or phenotypes. However, routine genome diagnostics is based on DNA sequencing and therefore neglects the regulation of gene expression such as ASE. To take advantage of ASE information in absence of RNA sequencing, ASE must be predicted using only DNA variation. Results: We have constructed ASE models from BIOS and GTEx that predict local ASE using DNA features. These models are highly stable as shown by performance metrics, cross-validation, and feature importance. Many different types of features are involved in these predictions, highlighting the complex regulation that underlies ASE. We applied the BIOS-trained model to population variants in three genes in which ASE plays a role in disease penetrance or modulation: BRCA2, RET, and NF1. This resulted in predicted ASE effects for 27 out of 8,957 variants. Ten of these were known pathogenic variants, while the rest are interesting candidates for further elucidation of disease etiology. Conclusions: We demonstrated that ASE can be stably and reliably predicted from DNA features using machine-learning models. Future efforts may further improve sensitivity and translate these models into a new class of genome diagnostic in silico prediction tools to prioritize candidate pathogenic variants or regulators thereof for follow-up validation by RNA sequencing.

Human biobanks, which are the core of BBMRI-ERIC medical research infrastructure, are repositories of biological material and data associated with the research participants (donors or patients willing to participate in the research). The associated data covers a broad range of data types: from data collected directly from the research participants and medical processes related to them, to data generated from the biological material. This document focuses on describing biobank data processing workflows that were selected for piloting in EGI-Engage the biobanks by the BBMRI.nl and BBMRI.cz (national nodes of BBMRI-ERIC) together with their associated biobanks. The main focus is on proteomics and genomics workflows, which cover both extremes of privacy-sensitive data processing spectrum: from relatively non-sensitive applications to very sensitive ones.