Optimized High-Performance Compute Infrastructure for Precision Medicine Workflows at Hospitals

Sukumar, Sreenivas ;   Lim, Seung-Hwan ;   Tourassi, Georgia

We have designed and prototyped a futuristic and unified health information technology system for healthcare workflows for the precision medicine era ? medical workflows that involve genomics, sensors, past history of symptoms, diagnosis and procedures, along with clinical images. Our approach successfully addresses the following requirements emerging for precision medicine workflows: (i) store, process and retrieve massive multi-modality biological databases (genomics, claims, biological pathways, etc.); (ii) generate high-resolution molecular and clinical data from each patient; and (iii) gain the insight through scalable data analysis/knowledge discovery algorithms to formulate effective treatments for each individual's illness. Furthermore, our system offers scale-up (with the ability to store data collected from millions of patients in one system) and speed-up (what used to take several hours on state-of-the-art health IT systems takes only a few minutes), leverages algorithms that could not be unleashed on the data because of limited computing capacity and extends to ensemble research studies in population health.