Martin Lablans1, Dennis Kadioglu2, Sebastian Mate3, Ines Leb3, Hans-Ulrich Prokosch3, Frank Ückert4. 1. Medizinische Informatik in der Translationalen Onkologie, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, 69120, Heidelberg, Deutschland. m.lablans@dkfz.de. 2. Institut für medizinische Biometrie, Epidemiologie und Informatik (IMBEI), Universitätsmedizin Mainz, 55101, Mainz, Deutschland. 3. Lehrstuhl für Medizinische Informatik, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Deutschland. 4. Medizinische Informatik in der Translationalen Onkologie, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, 69120, Heidelberg, Deutschland.
Abstract
BACKGROUND: Medical research projects often require more biological material than can be supplied by a single biobank. For this reason, a multitude of strategies support locating potential research partners with matching material without requiring centralization of sample storage. OBJECTIVES: Classification of different strategies for biobank networks, in particular for locating suitable samples. Description of an IT infrastructure combining these strategies. MATERIALS AND METHODS: Existing strategies can be classified according to three criteria: (a) granularity of sample data: coarse bank-level data (catalogue) vs. fine-granular sample-level data, (b) location of sample data: central (central search service) vs. decentral storage (federated search services), and (c) level of automation: automatic (query-based, federated search service) vs. semi-automatic (inquiry-based, decentral search). All mentioned search services require data integration. Metadata help to overcome semantic heterogeneity. RESULTS: The "Common Service IT" in BBMRI-ERIC (Biobanking and BioMolecular Resources Research Infrastructure) unites a catalogue, the decentral search and metadata in an integrated platform. As a result, researchers receive versatile tools to search suitable biomaterial, while biobanks retain a high degree of data sovereignty. CONCLUSIONS: Despite their differences, the presented strategies for biobank networks do not rule each other out but can complement and even benefit from each other.
BACKGROUND: Medical research projects often require more biological material than can be supplied by a single biobank. For this reason, a multitude of strategies support locating potential research partners with matching material without requiring centralization of sample storage. OBJECTIVES: Classification of different strategies for biobank networks, in particular for locating suitable samples. Description of an IT infrastructure combining these strategies. MATERIALS AND METHODS: Existing strategies can be classified according to three criteria: (a) granularity of sample data: coarse bank-level data (catalogue) vs. fine-granular sample-level data, (b) location of sample data: central (central search service) vs. decentral storage (federated search services), and (c) level of automation: automatic (query-based, federated search service) vs. semi-automatic (inquiry-based, decentral search). All mentioned search services require data integration. Metadata help to overcome semantic heterogeneity. RESULTS: The "Common Service IT" in BBMRI-ERIC (Biobanking and BioMolecular Resources Research Infrastructure) unites a catalogue, the decentral search and metadata in an integrated platform. As a result, researchers receive versatile tools to search suitable biomaterial, while biobanks retain a high degree of data sovereignty. CONCLUSIONS: Despite their differences, the presented strategies for biobank networks do not rule each other out but can complement and even benefit from each other.
Entities:
Keywords:
Data integration; Federation; Joint research; Registry; Research infrastructure
Authors: C Schüttler; N Buschhüter; C Döllinger; L Ebert; M Hummel; J Linde; H-U Prokosch; R Proynova; M Lablans Journal: Pathologe Date: 2018-07 Impact factor: 1.011
Authors: Sebastian Mate; Marvin Kampf; Wolfgang Rödle; Stefan Kraus; Rumyana Proynova; Kaisa Silander; Lars Ebert; Martin Lablans; Christina Schüttler; Christian Knell; Niina Eklund; Michael Hummel; Petr Holub; Hans-Ulrich Prokosch Journal: Appl Clin Inform Date: 2019-09-11 Impact factor: 2.342