Development of a framework for data management, harmonisation, visualisation, and analytics of multi-omics clinical trials data in the context of autoimmune diseases
[eng] (Multi-)omics technologies generate vast amounts of data that require proper organisation and storage for efficient analyses and interoperability. Historically research organisations managed data locally, but in today’s large-scale international collaborative projects, this hinders efficient d...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis doctoral |
| Estado: | Versión publicada |
| Fecha de publicación: | 2025 |
| País: | España |
| Institución: | Universidad de Barcelona |
| Repositorio: | Dipòsit Digital de la UB |
| OAI Identifier: | oai:diposit.ub.edu:2445/225791 |
| Acceso en línea: | https://hdl.handle.net/2445/225791 http://hdl.handle.net/10803/696360 |
| Access Level: | acceso abierto |
| Palabra clave: | Genòmica Bioinformàtica Bases de dades Immunologia Genomics Bioinformatics Databases Immunology |
| Sumario: | [eng] (Multi-)omics technologies generate vast amounts of data that require proper organisation and storage for efficient analyses and interoperability. Historically research organisations managed data locally, but in today’s large-scale international collaborative projects, this hinders efficient data exchange. A case in point, is the use of SFTP file server for data sharing purposes, which requires familiarity with specific software or command-line tools. Commercial cloud storage (e.g., Dropbox or Google Cloud) offers an alternative but demands trust in black-box solutions and their providers. Open-source alternatives exist, each with specific strengths and limitations. In this thesis, I evaluated the applicability of these existing open-source data warehouse solutions for the 3TR project (3tr-imi.eu). It is a public-private partnership tackling seven immune-mediated diseases across 15 European countries involving 69 partners. As none of the tested software solutions met the needs of the 3TR project consortium I developed OmicsDM, a web-based platform that facilitates data storage, sharing, visualisation, and analysis. In 3TR, diseases are analysed both independently and jointly. Regarding the latter, the heterogeneity of the pheno-clinical data across studies poses a major challenge. To address this, I participated in the development of Convert-Pheno, a software toolkit for the interconversion of common data models for phenotypic data. Thus, laying the foundation for the envisioned cross-disease analysis. Next, the consortium needed a method to group patients by phenotypic resemblance rather than diagnostic labels alone. Thus, enabling researchers to construct phenotype-based cohorts whose analysis can reveal shared mechanisms or treatment responses across diseases To aid in the grouping we built Pheno-Ranker. It leverages semantic similarity to compare phenotypic data stored in GA4GH standards and beyond. For both Convert-Pheno and Pheno-Ranker I built intuitive web user interfaces, each with its own public playground to let users explore the capabilities of each of the tools. To sum it up, the developments in this thesis are an important contribution to the improvement of the biomedical data management, harmonisation, visualisation and analytics infrastructure in Europe and beyond. |
|---|