The EU funded sysVASC project, a bench to bedside program with a systems biology approach to identify molecular targets for vascular disease treatment. Eagle Genomics leads the data analysis work package. Our task is to organise and harness the value of sysVASC scientific and patient data in the context of publicly available information on therapeutic gene targets. This allows rapid prioritisation of candidate targets for further experimental analysis.
The term CVD describes a multitude of clinical conditions and varies in degrees of severity. It is believed that the development of the disease is triggered by early functional and structural changes in the vessels that lead to multiple discomfort symptoms and even organ damage.
The early disease pathways (pathophysiological changes) of CVD are not very well understood. While earlier stages are, in part, reversible, their diagnosis is challenging due to the patients experiencing no symptoms. Screenings may need specialised equipment to characterise the underlying problem and often are not carried out due to the lack of symptoms. The vascular damage accumulates for years before patients are identified and offered therapeutic treatment. Advanced stages or critical events are not always the immediate consequence of earlier disease processes which adds further complexity to the disease and its diagnosis.
The sysVASC consortium compiled the required clinical resources, technological infrastructure, including advanced biotechnology tools, multi-disciplinary skills and expertise to tackle the challenge of defining a comprehensive, systems medicine-based characterisation of CVD to aid identification of novel therapeutic targets.
Eagle used our translational medicine platform to ingest, curate and annotate sysVASC and public patient data to build an e[catalog] of patient biomolecular data, clinical phenotypes and therapeutic targets. This restricted data repository provided sysVASC consortium members with federated access to all project resources. With the data prepared and integrated, biomolecular screens were performed using machine learning to predict associations between genes and clinical phenotypes.
We simultaneously captured categories required for data valuation (association of target with CVD, druggability of target, safety/criticality of target and novelty of target) and built a hierarchical AI valuation model.
The model was used to score all human genes according to their value as novel therapeutic targets for CVD. Candidate targets emerging from sysVASC biomolecular association screens were analysed in the context of the overall value distribution thus allowing us to prioritise those candidates to progress to the next stage of experimental validation.