MeDaX - bioMedical Data eXploration
ABOUT THE PROJECT
Evidence-based medicine uses high-quality, curated, and reliable data from health research, combines them with the respective individual expertise of the treating physicians and aims to make the best possible treatment decisions for individual patients. In order to generate the necessary knowledge from the accessible data sources, heterogeneous data must be combined and be made comparable – a complex task from the perspective of data sciences. For a reliable data basis of high quality, good documentation of the research process and the resulting data is crucial. This also includes the process of integrating heterogeneous data from different sources, as the quality of the database depends to a large extent on the quality of the data sources, but no less on the quality of the data integration.
MeDaX - bioMedical Data eXploration
GOALS
With a focus on the generation of FAIR data and the integration of standardized data formats, the junior research group MeDaX is developing an information and research platform for (bio)Medical Data eXploration. For this purpose,
- heterogeneous (bio)medical data from various sources are merged into a knowledge graph (graph database) via so-called ETL processes (ETL: Extract-Transform-Load),
- new or improved methods for computer-aided semantic enrichment, quality control, provenance and comparability are developed and
- algorithms and tools for further processing of the data are integrated.
Various data sources, such as the Medical Informatics Initiative core data set, local population studies, biomedical ontologies, and publicly available information portals, will serve as starting points. In addition, the needs of patients and data-collecting physicians and scientists must be elicited during the set-up phase to ensure the greatest possible benefit from the MeDaX platform.
The protection of personal data and the transparent documentation and communication of scientific work as the basis for sustainable science are particularly important to the group.
NEWS
- MILA Seminar - PD Dr. Susanne Neufang on gender bias in AI models
- MILA-Seminar - Daniele Liprandi on EvoNEST: a modular open-source software for integrated organismal data management and analysis
- MILA-Seminar - Heiko Dunkel on Implicit Egotism among German Academics
- MILA Seminar - Anne-Christin Hauschild on Addressing the Hurdles in Developing Predictive Models in Healthcare
- MILA Seminar - Lena Baum on HERALD, a domain-specific query language for longitudinal health