Big Data and Personalized Medicine (BDPM) is one of the four departments of CSS, established to perform data driven Research and Development (RnD). BDPM is built as a translational research structure gathering the necessary complementary competencies to achieve the Personalized Medicine objectives. The purpose of BDPM is to explore the globally available solutions for harvesting Big Data using Artificial Intelligence, select the most efficient ones for the Danish data scene and develop prototypes that will accelerate research and the translation of research results into improved diagnoses, prognoses, and treatments.
Research domains
- Data modeling
Unify the different codding and classification of biomedical data with the use of international standards while achieving the highest level of quality and the lowest reduction in granularity. Describe data in a clinical-friendly way.
- Data integration
Normalize and correlate the different data types and dimensions.
- Data exploration
Enable healthcare actors to navigate within the different types of biomedical data (phenotypic, OMICs, demographic, socioeconomic, etc.).
- Data analysis and cross-analysis
Define cohorts using clinical/medical terms in order to extract meaningful information from specific datasets and combine information from different analyses.
- Data visualization and reporting
Presenting research results, especially multidimensional data without standardized visualization forms, in a medically/clinically meaningful way.
- Artificial intelligence, machine learning and deep learning
Develop, train and validate methods to extract useful information from the data universe (e.g. predictions and estimations).
- Patient-Like-Me (PLM) tools
Translate Personalized Medicine research results from all the points above into decision support solutions for healthcare actors.
BDPM projects are following the principles of open-science, FAIR data and the GCP aiming to be easily adapted to the needs of other medical domains.
From real world data to real world evidence for improved personalized healthcare services
Patient-Like-Me tools
PLM is a suite of clinician-friendly tools aiming at bringing validated results from Personalized Medicine research to the bedside’s daily work. PLM is strongly dependent on various domains of expertise dealing with medicine, biomedical data integration, modelling and analysis. Actively developing one of the world’s most efficient PLM solution requires important effort in domain-specific data curation and validation that can only be performed by a medical doctor’s level of knowledge. The quality of the data, the deep understanding of the data relations and the link between data values and biological properties is the angular stone of the PLM, upon which the other specialists can produce powerful and reliable results. BDPM is designed to combine the expertise of biomedical researchers, data scientists, machine learning experts and scientific programmers in a structure providing all the necessary latitude for highest quality for research development.
Collaborators
-
Region Zealand
Data and Development Support
-
NGC
Danish National Genome Center
-
OHDSI
Observational Health Data Sciences and Informatics
-
EHDEN
European Health Data & Evidence Network
-
Erasmus University Medical Center
Health Data Science Group
-
Columbia University Medical Center
Department of Biomedical Informatics