The DetectEverything Project

DetectEverything is a new pilot project using applied metagenomics in environmental contexts as an early warning system for infectious disease outbreaks. DetectEverything finds not only all biological threats from high-risk sources––including medical/lab equipment, live animal markets, farms, wastewater, and air––but the system uses AI to categorize massive amounts of genomic data into signals that tell us when changes in microbial communities signals an emerging health threat.

Problem & Concept

Intensive livestock farming, driven by urbanization and global population growth, has altered the human-animal interface, creating opportunities for the spillover and amplification of emerging zoonotic diseases. 

Spillovers are a critical barrier to poverty alleviation around the world, since the majority of livestock owners in LMICs practice backyard farming in close proximity to wild animals and people, with minimal biosecurity, thus increasing the risk of spillover in these regions. The rate of outbreaks of emerging, re-emerging, and novel infectious human diseases are accelerating globally, while research shows rural agricultural settings in Southeast Asia to be among the most likely places for future spillover—in particular the porous border region known as the “Golden Triangle”, covering southern China, Laos PDR, Myanmar, and Thailand. 

The human-livestock interface is both the most practical target for zoonotic disease surveillance and the most likely place where zoonotic transmission will occur next. Yet it is one of the least targeted approaches in our current pandemic prevention arsenal, while the majority of surveillance efforts are focused in countries from where the next pandemic is least likely to originate. And even when this critical interface is targeted, efforts are largely event-based, or limited in focus to economically significant animal pathogens, often missing unreported or asymptomatic events, or rely on routine testing for common pathogens, missing emerging, rare, or difficult-to-diagnose cases. 

Early detection of risks before the emergence of zoonotic disease in humans provides a window of opportunity to disrupt transmission before outbreaks become epidemics or pandemics. 

Solution

The DetectEverything solution deploys metagenomic sequencing in rural agricultural settings in Southeast Asia. The goals of the system are to:

1. Decrease time-to-detection for outbreaks
2. Increase sensitivity of government surveillance systems
3. Reduce impacts of animal disease on livelihoods of local communities.

Our approach combines anomaly detection AI algorithms with the burgeoning field of Precision Epidemiology, enabled by metagenomic sequencing technology, in order to create a near real-time pathogen detection system. The resulting system detects aberrations in the microbial community of samples “pre-spillover”, before infected animals have the opportunity to transmit to humans. 

DetectEverything provides a high-impact, low-cost approach that can be used as a blueprint for LMICs to implement effective early warning systems that can prevent outbreaks before they cause widespread economic damage. Governments can use the research findings to justify budget requests for scaling genomic surveillance, helping build sequencing capacity for regions most severely impacted by infectious disease outbreaks.

——–

Latest Research