UC Davis leads global AI collaboration to flag high-risk viruses before human spillover

New global collaboration uses UC Davis experts and AI to spot the next pandemic

UC Davis and UC Davis Health are joining the Coalition for Epidemic Preparedness Innovations (CEPI) and Boston University's BEACON project to flag viruses most likely to spill over into humans and cause severe disease. The effort integrates BEACON's open-source surveillance with UC Davis's Virus Intelligence & Strategic Threat Assessment (VISTA), delivering near real-time risk rankings with AI-assisted analysis and expert oversight.

"By identifying high-risk viruses and providing assessments of their pandemic risk potential, we're not just reacting to pandemics - we're working to prevent them." - Angel Desai, UC Davis Health

Ranking viral risk for early pandemic warning

UC Davis researchers built the SpillOver platform (now VISTA) from over half a million animal samples across 28 countries and public records. That dataset helped rank the spillover potential of nearly 900 wildlife viruses. With support from CEPI, the work expanded beyond wildlife to include viruses from domestic animals and other vectors, combining AI methods with expert input to assess pandemic potential.

What VISTA + BEACON adds

VISTA now connects with BEACON's large language models and global expert network to collect, analyze, and share intelligence on emerging infectious diseases across humans, animals, and the environment. BEACON is an open-access system that pulls signals from disease-tracking websites and systems, applies AI to score potential threat, and produces a human-verified summary report. CEPI is providing up to $1 million to support data integration between the two platforms.

Why this matters for scientists and public health teams

• Prioritize targets earlier: Use risk scores to focus surveillance, sampling, and assay development on viruses most likely to spill over and spread.
• Faster situational awareness: BEACON's human-verified summaries can streamline literature triage and horizon scanning for program leads.
• Better study design: Align fieldwork and sequencing with host, vector, and geography signals that correlate with higher risk.
• Resource allocation: Direct limited lab, sequencing, and biosafety capacity to the highest-impact viruses.

How to plug this into your workflow

• Adopt a weekly review: Track VISTA's risk rankings and BEACON summaries to update watchlists and sampling plans.
• Build a triage pipeline: Pair LLM-based literature screening with human verification to mirror BEACON's model for internal reports.
• Coordinate across teams: Connect epidemiology, genomics, and field units on shared shortlists of high-scoring viruses.
• Document assumptions: Keep versioned notes on risk criteria and thresholds to support reproducibility and audits.

What to watch next

• New high-risk candidates as integration matures and data coverage broadens.
• Signals that shift priority hosts, vectors, or regions for targeted surveillance.
• Human-verified reports that change confidence in early AI flags.

Learn more about CEPI's role in pandemic preparedness at cepi.net.

If your team is formalizing LLM triage and reporting, you can explore practical prompt engineering workflows here: Complete AI Training - Prompt Engineering.