Federated cloud-based Artificial Intelligence-driven platform for liquid biopsy analyses

From 01-02-2021 to 31-01-2023

Description

“Liquid biopsy” represents an emerging minimally invasive screening tool that can potentially allow monitoring tissue dynamics in multiple human physiological and pathological conditions. In turn, the current extensive analysis of various liquid biopsy biomarkers associated with big data generation paves the way towards data-driven precision medicine by enabling patient-oriented proactive, predictive, preventive and participatory health management. We have recently developed an innovative massive parallel sequencing-based approach for genome-wide cell-free DNA (cfDNA) profiling and combined it with novel data mining approaches (patent pending), enabling sensitive detection of pre-symptomatic cancer, autoimmune disorders, pregnancy complications, and other pathologies. Multiple genetic centers and genetic testing sites worldwide reached out to us willing to improve their diagnostic potential by applying our analytical pipeline. However, to properly address these demands, an efficient and rapid end-user friendly platform including our analytical pipeline and processing data in a federated way is lacking. As such, the goal of this project is to tackle the main challenges for productization in order to create scalable federated liquid biopsy platform: (1) Creation of a scalable cloud environment enabling fast and efficient sample classification; (2) Establishment of a software suite for analysis, visualization of cfDNA profiles and reporting of clinically-relevant findings; (3) Expansion of the cloud-based platform to allow data analysis and reporting to distributed sites; (4) Aggregation of large datasets from distributed sites. The creation of a performant federated data analytics platform will pave the road for a successful spin-off creation.

Team

• Yves Moreau, Co-promoter

Financing

Funding: KU Leuven - Internal Funding KU Leuven

Program/Grant Type: C3 - KU Leuven Category 3 Funding