Ginkgo is Developing Digital Biothreat Intelligence for the Bioradar of the Future

By Chris Pardee

Data Science Manager, Ginkgo Biosecurity

Biological risks are not uniformly distributed in the world, and yet they also do not respect political borders. The movement of potentially dangerous pathogens through complex human networks necessitates a platform that revamps traditional detection models to focus on earlier detection of developing threats in the interest of strong global biodefense and biosecurity. We are developing a “bioradar” system, which pairs traveler testing at airports, points of entry, and farms with pervasive digital biothreat monitoring, then feeding these datastreams into modeling tools to maximize the value of insights gained from these sources.  

Digital Biothreat Intelligence

One pillar of our bioradar is the collection of digital biothreat intelligence from diverse publicly-available sources. By monitoring thousands of official and unofficial sources of public health information, Ginkgo’s epidemiologists and analysts quickly identify developing infectious disease events and assess them, with assistance from AI and other tools, for their potential to cause significant economic, political, or societal disruption. In the process of doing so, Ginkgo fully structures unorganized public information into standardized datasets that provide the best available geographic and temporal information from multiple reporting sources. 

We have two tools that serve digital biothreat intelligence, which can help public health officials and national security experts make informed decisions. 

Epidemic Threat Monitor

Ginkgo’s Epidemic Threat Monitor presents a top-level view of all currently-developing infectious disease events being monitored by Ginkgo. Identified outbreaks are assessed daily by a team of analysts and proprietary algorithms to summarize the available information and prioritize each event based on its epidemiological characteristics. This information is then displayed in an interactive dashboard allowing users to search and filter for events of interest.

Human Epidemic Database

The Human Epidemic Database (HED) is one of the world’s largest repositories of high-quality, structured epidemiological data on historic and potentially emerging epidemics. This database incorporates temporal and geographic data collected from over 500 distinct reporting sources, providing information on more than 170 unique pathogens and over 3,200 distinct outbreaks in 245 countries. Approximately half of the HED are contemporary outbreaks identified and structured as they developed from roughly 2015-present. The rest of the HED are high-mortality viral outbreaks identified retrospectively from 1918-2014, but especially from roughly 1960-2014.

How we do it

Ginkgo’s digital biosurveillance intelligence starts with the continuous monitoring of thousands of official and unofficial, open-source intelligence feeds. These include social media, traditional media, international non-governmental organizations, and both public health and non-public health authorities at the local, state, national, and international level. Our team continuously structures summary-level data on developing infectious disease events, and meets daily to review these datasets and confirm each event’s algorithmically-generated prioritization score. These scores incorporate the structured information on each event’s pathogen, geographic scale, and epidemiological severity. Summary data for currently-monitored infectious disease events are piped automatically to the Epidemic Threat Monitor, where additional automated processing conducts further trend analyses and produces each event’s threat score.

A subset of monitored outbreaks that reach a certain prioritization score are also escalated for structuring into the HED. For these events, Ginkgo’s analysts review all available sources and select as many as necessary to provide the fullest possible view across geographic and temporal granularities. Every selected source is then structured according to comprehensive data interpretation methodologies, starting from the earliest relevant report and continuing until the event is declared over or the source discontinues its coverage. Structured datasets go through multiple layers of quality assurance and expert review before being incorporated into the HED. This process is repeated every time a new source report is published. Whenever necessary, our team will also combine multiple reporting sources into a fusion dataset providing the single best representation of all available data for a given outbreak. 

Giving us an early warning

Ginkgo’s digital biothreat intelligence activities allow for the early detection of developing infectious disease events. Our team’s continuous monitoring of diverse intelligence feeds enables us to rapidly identify developing infectious disease events shortly after reporting sources become publicly available. The system’s daily refresh further helps ensure that any newly identified or recently updated events are quickly available for users in the Epidemic Threat Monitor.   

In 2022, our digital biothreat intelligence system first detected the ongoing international monkeypox (mpox) outbreak on May 16, 2022, as a small cluster of cases being investigated in the United Kingdom. Our team began actively structuring the event on May 18, 2022 when similar cases were reported in Portugal, and the WHO issued its first report on the UK cluster. The score for this event escalated quickly from moderate to high, because cases were appearing in close succession in multiple countries where mpox is not usually seen. The system had similarly detected the cluster of pneumonia cases that would eventually be identified as SARS-CoV-2 on December 30, 2019. Our team began to actively monitor that event on January 2, 2020, and began structuring the event on January 5, 2020 when the WHO began reporting on the outbreak. 

The uneven distribution of biological threats and the highly networked nature of the modern world mean that all countries are constantly at risk for the spread of infectious diseases. Ginkgo’s bioradar—driven by biothreat intelligence systems—empowers users to quickly identify threats to their people, operations, or supply chains, and to mobilize responses as early as possible to mitigate potential disruptions.