Abstract

The AIMed Cloud to Empower Precision Medicine and Drug Discovery

Author: Spyro Mousses

There is an urgent unmet need for shared data analysis resources capable of supporting both better care and faster cures for pediatric diseases. Traditional approaches to data analytics have two major limitations. The first is that the clinical datasets are in healthcare IT silos, making it difficult to conduct collaborative. The second limitation has to do with conventional approaches to analytics, which largely rely on statistical modeling. The vast majority of data analyses are therefore conducted at individual institutions and often do not truly map data to ‘knowledge networks of disease’. Here we provide a proof of concept system, called AIMed Cloud, to address these needs, and provide an innovative solution to the computational challenges of modeling complex multiscalar biomedical knowledge. Key to this system is the use of artificial intelligence approaches, like Deep Learning, and cognitive computing platforms for advanced knowledge representation. This unique cloud based computational solution represents a first of it’s kind platform, open to the entire pediatrics community, and represents a powerful new resource capable of empowering both precision medicine for better care, and sophisticated translation research for faster cures. Towards that vision, Innovation Institute and it’s partners, including Dell, Intel, and Systems Imagination, have aligned to create a shared cloud based AI resource to enable healthcare providers to intelligently integrate and mine patient clinical and genomic datasets that are emerging from the practice of precision medicine. Proof-of-concept projects are underway to demonstrate how cognitive computing and big-data could be applied in a shared cloud environment for predictive analytics and intelligent cohort-discovery. Specifically, one of the projects is aiming to integrate about one hundred genomes to discover genes that could be predisposing children to cardiomyopathy, including chemotherapy-induced cardiomyopathy. This analysis will involve both expert-based systems modeling as well as cognitive computing platforms to discover patterns of combined clinical and genomic features that could serve as multi-parametric predictors of heart disease, as well as novel targets for drug discovery and development.

Funding Acknowledgement (If Applicable): The Sharon Disney Lund Foundation The Innovation Institute