WP4 will establish a data-sharing platform that will effectively exploit the data that has been pulled from multiple sources. This platform will provide descriptive and predictive information on seven types of HMs, thus improving decision-making and the treatment of patients. The creation of statistical models will identify the best practices that will enable the best evidence to be generated. WP4 will use the most innovative methods of data analysis and the most advanced algorithms in order to achieve the optimum management and mining of the data that has been collected.
- Create a unified platform to exploit data pulled from multiple sources
- Deploy automatic learning techniques and statistical models to generate descriptive, comparative, and predictive information
- Establish best practices to generate relevant evidence and information
- Transform the data gathered on the data platform into specific findings addressing the unmet needs of patient management and overcome challenges associated with the development of new products and diagnostic methods
- Develop the pilot HARMONY platform, visualize how it will be exploited, and monitor the results
Bayer, Celgene, EBMT, ELN, EORTC, GMV, HULAFE, IBSAL, Janssen, LeukaNET, MediUni Wien, Menarini, Novartis, Takeda, Ulm University, UNIBO, University of York.
Considerable progress has been made in terms of establishing the Big Data platform and the methodology for analyzing data:
- HARMONY's Big Data platform has been established;
- A common data model that adheres to the FAIR (findable, accessible, interoperable, and reusable) data-sharing principles has been created;
- HARMONY has begun developing and testing models based on available data on AML (TCGA public data, UNIBO internal data and Sanger Institute data);
- HARMONY has started analyzing the description of datasets;
- Methods working on methods for gene network analyses are beginning to be developed.
Workpackage WP4 will focus in 2018 on the establishment of a data governance process, the definition of the minimum data sets and the implementation of the first research question, focusing on AML.
- Data governance process and tool implementation;
- Definition of minimal data sets for selected HMs;
- Execute upon gender differences in diagnosis and outcomes for the first set of HMs;
- Rule definition and implementation for the first research question (AML);
- Definition and operationalization of user access.