Periodic Reporting for period 1 - GenoMed4ALL (Genomics and Personalized Medicine for all though Artificial Intelligence in Haematological Diseases)
Periodo di rendicontazione: 2021-01-01 al 2022-06-30
GenoMed4All is a European initiative to transform the response to haematological diseases by seizing the power of Artificial Intelligence (AI) and standardized interoperable sharing of cross-border data. For doing so, the project will support the pooling of genomic, clinical data and other “-omics” health data through a secure and privacy respectful data sharing platform based on the novel Federated Learning (FL) scheme, to advance research in personalised medicine.
GenoMed4All will make use of the existing infrastructures and initiatives, including powerful High-Performance Computing (HPC) facilities, hospital registries, data processing tools, and pre-existing repositories towards facilitating personalised medicine in common, rare and ultrarare haematological diseases to demonstrate the versatility and utility of the solutions, as can be seen in Figure 1. The strategy is defined to start to integrate the 10 clinical partners to later enlarge it to the overall 66 relevant clinical repositories in 15 Member States that ERN-EuroBloodNet members involve, besides 20 repositories not present in EuroBloodNet.
GenoMed4All will demonstrate the potential and benefits of trustable and explainable AI technologies, with a novel approach to AI models and algorithms (deep learning, variational autoencoders, generative models, besides combining with advanced statistical and Machine learning) to exploit the powerful set of “-omics” data which will be at researchers’ disposal leading to more reliable and meaningful outcomes for advancing research and personalised medicine, with 3 use cases covering oncological and non-oncological Haematological Diseases (HD). It will allow for identifying new knowledge, to support clinical research and decision making by linking Europe's relevant genomic repositories in haematological diseases, while ensuring full compliance with data protection legislation and ethical principles and increasing the AI trust for personalized medicine and impact.
GenoMed4All will make use of the existing infrastructures and initiatives, including powerful High-Performance Computing (HPC) facilities, hospital registries, data processing tools, and pre-existing repositories towards facilitating personalised medicine in common, rare and ultrarare haematological diseases to demonstrate the versatility and utility of the solutions, as can be seen in Figure 1. The strategy is defined to start to integrate the 10 clinical partners to later enlarge it to the overall 66 relevant clinical repositories in 15 Member States that ERN-EuroBloodNet members involve, besides 20 repositories not present in EuroBloodNet.
GenoMed4All will demonstrate the potential and benefits of trustable and explainable AI technologies, with a novel approach to AI models and algorithms (deep learning, variational autoencoders, generative models, besides combining with advanced statistical and Machine learning) to exploit the powerful set of “-omics” data which will be at researchers’ disposal leading to more reliable and meaningful outcomes for advancing research and personalised medicine, with 3 use cases covering oncological and non-oncological Haematological Diseases (HD). It will allow for identifying new knowledge, to support clinical research and decision making by linking Europe's relevant genomic repositories in haematological diseases, while ensuring full compliance with data protection legislation and ethical principles and increasing the AI trust for personalized medicine and impact.
GenoMed4All is composed by 9 different work packages that have been organised to reach the balance between the technologies to be developed and the medical and impact factors to guarantee sustainability and proper fitting to the HD AI genomics needs. The main results achived in this period are:
- Setting up the data protection and legal framework and fulfilment of ethical requirements
- Completion of the user and functional Project requirements
- First release of the federated data platform incliding the development of a demo
- Pre-processing pipeline defined and first implementations (Covering genomic, clinical, radiogenomic, etc. kinds of data)
- AI set of algorithms defined and agreed between AI and clinical partners
+First experiments took place with data provided by clinical partners
+Preliminar results with AI are showing a great promise with good advances over the SoA
+Radiogenomics AI and multimodalities are advancing at good pace
- Validation protocol in place for the whole project activities WP4-5-6-7: first approach was developed
- Preparation of the first pilots in the use cases and data organization:
+Work which happened in all 3 use cases
- Engagement strategy to extend to more nodes in place and providing new members of the federation
+Both technical, organizational and legal dimensions
- Standardization of genomic information
- Initiation of the GenoMED4ALL value proposal
- Discussion and contact with other initiatives (B1Mgenomes, other projects funded under the same cluster…)
- Setting up the data protection and legal framework and fulfilment of ethical requirements
- Completion of the user and functional Project requirements
- First release of the federated data platform incliding the development of a demo
- Pre-processing pipeline defined and first implementations (Covering genomic, clinical, radiogenomic, etc. kinds of data)
- AI set of algorithms defined and agreed between AI and clinical partners
+First experiments took place with data provided by clinical partners
+Preliminar results with AI are showing a great promise with good advances over the SoA
+Radiogenomics AI and multimodalities are advancing at good pace
- Validation protocol in place for the whole project activities WP4-5-6-7: first approach was developed
- Preparation of the first pilots in the use cases and data organization:
+Work which happened in all 3 use cases
- Engagement strategy to extend to more nodes in place and providing new members of the federation
+Both technical, organizational and legal dimensions
- Standardization of genomic information
- Initiation of the GenoMED4ALL value proposal
- Discussion and contact with other initiatives (B1Mgenomes, other projects funded under the same cluster…)
GenoMED4All progressed beyond the SoA in different areas: novel insights in haematological diseases thanks to the use of AI, federated learning provision adapted to this area needs, new preprocessing of multiomics information and improved HD pipelines.
As reported in the KPIs evaluation, during this initial period a special effort has been made to first define how and where GenoMed4All should impact in order to disseminate and increase the impact of the project, and then to start performing the different actions already planned for doing so. In this regard, a set of more than 430 posts have been created in social media, 24 blog entries, 10 Zenodo publications (and 10 publications in the pipeline).
Moreover, a total of 6 events have been featured, 1 workshop hosted and 3 webinars have been attended. With regards to the online community, we have more than 1.3k followers in Twitter and LinkedIn and finally more than 250 hard copies (tri-fold brochure and stickers) have been distributed.
The enlargement of the community and federated started thanks to the engagement strategy which will allow to increase the number of clinical partners and sites in the haematological domain.
The collaboration with initiatives such as B1Mgenomes and other projects will enhance the project impact.
More details are offered in D9.1 that includes a complete definition of the impact master plan for the entire project together with D9.3 due to M23, that will provide an initial assessment of the impact, exploitation and sustainability of the project during those first months.
As reported in the KPIs evaluation, during this initial period a special effort has been made to first define how and where GenoMed4All should impact in order to disseminate and increase the impact of the project, and then to start performing the different actions already planned for doing so. In this regard, a set of more than 430 posts have been created in social media, 24 blog entries, 10 Zenodo publications (and 10 publications in the pipeline).
Moreover, a total of 6 events have been featured, 1 workshop hosted and 3 webinars have been attended. With regards to the online community, we have more than 1.3k followers in Twitter and LinkedIn and finally more than 250 hard copies (tri-fold brochure and stickers) have been distributed.
The enlargement of the community and federated started thanks to the engagement strategy which will allow to increase the number of clinical partners and sites in the haematological domain.
The collaboration with initiatives such as B1Mgenomes and other projects will enhance the project impact.
More details are offered in D9.1 that includes a complete definition of the impact master plan for the entire project together with D9.3 due to M23, that will provide an initial assessment of the impact, exploitation and sustainability of the project during those first months.