What is IMIM’s main focus?
Our research group Integrative Biomedical Informatics, which belongs to IMIM (Institut Hospital del Mar d’Investigacions Mèdiques) and UPF (University of Pompeu Fabra) located in Barcelona, Spain, is focused on developing bioinformatic tools to gain insight from biomedical data for improving human health. We use data mining and integration and natural language processing techniques to identify relevant information from textual resources (e.g. scientific publications, electronic health records and clinical trials) and standardise it to make it amenable for searching, analysis and integration with other resources. Our knowledge of network biology enables us to combine text-mined data with omics datasets to build network models of diseases and drug mode of action. We are particularly interested in assessing the impact of human genomic variation on disease and drug effect and we use these systems models for this purpose. Last but not least, our group provides software and resources for the community, among them the DisGeNET platform on human disease and their genes.
How has participating in TransQST impacted your organization?
To be able to participate in a consortium like TransQST gave us the opportunity to learn what are the current needs in the pharmaceutical industry regarding computational modelling for drug toxicology. This prompted us to develop new bioinformatic solutions to address their needs.
One example is DisGeNET. DisGeNET is a knowledge platform that integrates information from different sources including the scientific literature on the genomic variants and genes associated with human diseases. A major bottleneck in drug R&D programs is prioritizing information on targets associated with diseases and identification of their toxicity profiles. Thanks to TransQST, we expanded DisGeNET functionalities to support these specific needs. In addition, we implemented advanced visualization capabilities based on network biology to support the analysis of complex and heterogeneous datasets in combination with DisGeNET. This has been implemented as an open-source software that is part of the Cytoscape ecosystem of tools, one of the most widely used network analysis platforms. Finally, being part of the consortium allowed us to explore synergies with other partners that provide open source tools and resources, which resulted in the integration of DisGeNET data in other resources such as Reactome and OmniPath.
What challenges has IMIM faced across the project? What are your main achievements in TransQST?
The first challenge for us was to understand the needs of the partners from the pharmaceutical industry and tailor our approaches and methods to address those needs. Another challenge we faced throughout the project was to communicate and make other partners understand the bioinformatic approaches we develop and apply to address project’s needs.
We had the opportunity to contribute in different ways to the project. We coordinated the activities of WP4, devoted to the development of the bioinformatic and systems modelling tools that have to be applied for the development of QST models across the different organs. In addition, we also contributed with the development of different bioinformatic solutions to support data collection and harmonization, omics data analysis and the development of network models of drug toxicity. In addition to DisGeNET, we developed the software iPATH, for the discovery of protein pathways underlying drug mode of action and toxicity. Finally, we also developed a bioinformatic pipeline for the data gathering, analysis, and integration of human genomic variability in drug toxicity assessment.
Which TransQST outcome/s do you think have been the most significant? Why?
We would like to highlight the portfolio of bioinformatic tools and models that are the legacy of the project. Ten tools were created and/or further developed thanks to TransQST. Most of them are open-source/public tools and are used within TransQST and beyond, including other IMI and H2020 projects. These tools have addressed some of the goals of the project such as integrating a wide array of data types and providing mechanistic insight on drug toxicity, along with providing advanced visualization and analytics. Many of the tools follow the FAIR data principles and therefore are aligned with initiatives related to data discoverability, standardization and reproducibility.
What do you believe is the main contribution of TransQST in the field?
The consortium was able to implement several QST multiscale models for 4 organ systems (liver, kidney, heart and gastro-intestinal (GI) system). Some models have to be developed from scratch (e.g. GI system) while others we built upon previous developments already available (e.g. cardiac system). From the point of view of model development, it was a very interesting experience to see the development of the models from these very different starting points, and also with different levels of data available to build and evaluate the models. Finally, these models have been developed through collaboration among scientists from academia and the industry, which is the path to go if you want to implement innovative approaches that are fit for purpose for real problems in the industry.
Impact of TransQST on further development of DisGeNET and on strengthening the european industry base of small and medium-sized enterprises.
TransQST has actively participated in the further development of the DisGeNET knowledge management platform. DisGeNET is an Elixir Recommended Interoperability Resource, and though its existence is prior to the project, the platform has been updated and improved by the TransQST consortium and is publicly accessible at https://www.disgenet.org/. To take the platform one step further, a commercial solution, DISGENET plus, has been set up and is being commercialised by the SME MedBioInformatics Solutions (MBIS), founded by TransQST researchers during the project lifetime. The collaboration of the TransQST consortium in the development of the initial open-source platform has helped to provide visibility to DisGeNET both as an open-source tool to enhance biomedical research, as well as a commercial solution for advanced use by the pharma sector in support of drug discovery pipelines.