- This event has passed.
Ethan Jackson (Microsoft Research)
May 25, 2016 @ 4:00 pm - 5:00 pm
Project PREMONITION: A Cyber-Physical System for Disease Surveillance
Emerging infectious diseases (e.g. SARS, MERS, Ebola, and Zika) pose significant human health, economic, and security risks. The unpredictability and global impacts of these diseases beg the question: “Could potential pathogens be detected early in the environment before they cause outbreaks?”. Because many of the diseases are maintained by and evolve in animal populations, such a detection system would need to provide insight into the microorganisms and viruses associated with animals. This is the goal of Project Premonition. Project Premonition begins with the idea of using a mosquito-as-a-device, which can locate an animal in the environment and collect a blood sample. There are over 3,600 known species of mosquitoes feeding on reptiles, amphibians, birds, and mammals in every continent except Antarctica; they are ubiquitous blood samplers. If mosquitoes could be autonomously collected at scale from the environment, they might provide broad insight into the epidemiology and evolution of potential pathogens. The major challenges of such a system are the high-throughput collection of wild caught mosquitoes and the computational analysis of their body contents. These are the technical challenges Project Premonition is attempting to solve. In this presentation, I will describe our on-going efforts to implement a Project Premonition system. Our overall system architecture is designed to: (1) autonomously obtain blood samples from animal populations via robotically-collected wild-caught mosquitoes, (2) apply high-throughput gene sequencing (NGS) to convert samples into metagenomic data, (3) use cloud-scale computational genomics to extract potential threats. The vision is a low-cost and high throughput system that exploits commodity robotics, drones, and NGS platforms to sample nucleic acids from many hosts and reduce pathogen detection to computational problems. This is joint work with Microsoft Research, Johns Hopkins University, University of Pittsburgh, Vanderbilt University, and the University of California at Riverside.