D8.2 Documentation of the tools for the data manipulation and standard conversions in the rare-disease field

This document describes tools for the data manipulation and standard conversions in the rare-disease field. It describes a standard process to make rare disease data more FAIR compliant, services that support FAIRification, and a standardized process to make rare disease genome data comparable. The potential of following the principles of Findable¨ Accessible¨Interoperable¨and Reusable data for humans and computers (FAIR) is recognized in the rare disease community. Therefore, WP8 contributed to the cross project rare disease data linkage planto leverage documentation of tools for FAIR data manipulation and standard conversions from practical experiences with FAIRification of heterogeneous rare disease data resources. Seven steps are defined in the FAIRification process: (1) define driving user question(s) for scoping and testing; (2) analyse source data for planning the conversion towards FAIR data; (3) define the semantic model to represent data by ontologies and globally unique identifiers; (4) transform data records using the model of step 3; (5) define metadata to make the data resource more findable, accessible, and reusable; (6) deploy the FAIR data resource through an API that uses the metadata of step 5; (7) apply a query interface and test by executing the driving user questions. Because ontologies fulfill an critical role in step 3 of this workflow, we document the two ontologies that are currently IRDiRC recognized resources and the model that relates the two: the Orphanet Rare Disease Ontology (ORDO), the Human Phenotype Ontology (HPO) that defined phenotypic abnormalities, and the HPO-ORDO ontological Module (HOOM). Orphanet, the internationally recognized hub for rare disease information, provides HPO annotations of rare diseases defined in ORDO. HOOM qualifies the annotation between a clinical entity and phenotypic abnormalities according to frequency and by integrating the notion of a diagnostic criterion. HOOM was updated and expanded in the context of ELIXIR EXCELERATE WP8. Identifiers are central to interoperability, and this is reflected in the FAIR principles (Figure 2). Going to the level of data elements and data types, FAIR principles require provision of globally unique identifiers, for which this document refers to the URLs that are used in ontologies, the ‘10 simple rules for persistent identifiers2’, Identifiers.org, and an extension of Identifiers.org to add credit and attribution for data providers and databases. For steps 4 and 5 of the FAIRification workflow, Bioschemas isdescribed as an additional means to increase findability for resources that provide a web interface for data records. The RD-Connect platform ( https://platform.rd-connect.eu/ ) and Orphanet are also recognized resources to increase findability of rare disease resources. Finally, this document provides background for the standard pipeline that enables comparison of rare disease whole exome and whole genome sequencing data. The pipeline is implemented in the RD-Connect platform. Submission of rare disease genome data via this platform ensures that the data are comparable and stored in the EGA