Search

RD-Connect is an unique global infrastructure project that links up databases, registries, biobanks and clinical bioinformatics data used in rare diseases research into a central resource for researchers worldwide. One of the aims of this project is to achieve the interoperability among patient registries. In order to allow data sharing, some strategies are being developing, like the implementation of ontologies for phenotypic description. Numerous ontologies already exist for the biomedical community. A review of existing biomedical ontologies and their characteristics is needed to identify gaps, avoid duplication of efforts and be able to make a recommendation. Our general objective was the definition of a strategy for the searching of ontologies useful to be implemented in RD Connect and other IRDiRC-funded projects. This searching strategy consisted of a multi-stage process: The two leading repositories of biomedical ontologies, the BioPortal and the OBO Foundry, were consulted for collecting a preliminary inventory of existing biomedical ontologies, which was composed by 377 ontologies.After the revision of the inventory, a list formed by the selected biomedical ontologies was obtained. This list contained 71 records.Afterwards, the selected biomedical ontologies were classified into the following groups: phenotype (n=3), clinic (n=7), classifications & nomenclatures (n=14), disorder/disease-specific (n=17), anatomy (n=7), patients-related data (n=7), pharmacy (n=6), epidemiology (n=3), health indicators (n=2), miscellaneous (n=1) and discarded (n=4).Because the degree of relationship among the selected ontologies is a very important aspect to consider for our objectives, a mappings study was performed. A special mention has to be aimed to Orphanet Ontology and Human Phenotype Ontology. Both are phenotype ontologies which fit exactly in RD-Connect aims. In addition to these ontologies, we propose the use of some ontologies focused on specific rare diseases, as well as some other ontologies as support tools.

The GRDR-GUID: a model for global sharing of patients de-identified data Yaffa R. Rubinstein1; Matthew McAuliffe2; Manuel Posada3; Domenica Taruscio4; Stephen C. Groft1 1Office of Rare Disease Research/NCATS/NIH, 2Center for Information Technology/NIH, 3 Institute of Rare Diseases Research, IIER-ISCIII, 4National Centre for Rare Diseases, Istituto Superiore di Sanità (Rome, Italy) The Global Rare Diseases Patient Registry Data Repository-GRDR aggregates de-identified patient data, using CDEs, from various rare disease registries utilizing a Global Unique Identifier (GUID)(1). The GUID is a unique random alpha-numeric set of characters assignedto each patient-data that GUID is NOT directly generated from personally identifiable information (PII). The GUID system allows the patient to be followed across studies and registries can be used also to link to biospecimens. GUID Process User executes the GUID tool client locallyPII is enteredPII is combined and one-way hash codes are generatedThe one-way hash codes are sent to the GUID serverIf the hash codes match the server's hash codes for an existing GUID, then that GUID is returnedIf the hash codes do not match, then a new random GUID is generated and returned To generate a GUID for the subject, the following PII is required (these elements are included in the ORDR/GRDR list of CDEs) (PII is never sent to the GRDR system): Complete legal given (first) name of subject at birthComplete Legal additional name of subject at birth (if the subject has a middle name)Complete legal family (last) name of the subject at birthDay of birth (1-31)Month of birth (1-12)Year of birth (####)Name of city/municipality in which subject was bornCountry of birthPhysical sex of subject at birth (M/F) The GUID is an ID that allows researchers to share data specific to a study participant without exposing PII and to track participants longitudinally, across multiple research sites and across multiple studies. RD-CONNECT (IRDiRC framework) is assessing possibilities to implement a GUID strategy in their work and the NIH-GUID is one of the possibilities under consideration. (1) Johnson SB, Whitney G, McAuliffe M, etal. Using Global Unique Identifiers to Link Autism Collections. J. Am. Med. Inform. Assoc. July 17, 2010. 10.1136/jamia.2009.002063.