07/01/2026 | Press release | Distributed by Public on 07/01/2026 13:25
Identifying the exact source of a biological threat - whether it stems from a naturally occurring outbreak, an accidental containment failure, or an intentionally engineered pathogen - is a complex but vital national security challenge. Rapid and precise attribution requires investigators to sift through an overwhelming amount of genomic and environmental data to pinpoint microscopic anomalies.
The DARPA Bio-Attribution Challenge launched in March 2026 to accelerate the development of new computational tools to solve this problem. Today, DARPA announced the final winners of the competition, concluding a rigorous virtual event that awarded a total of $180,000 to top-performing teams.
Eight Round 2 teams successfully completed the conclusive "Determination - Attribution" phase of the challenge and are being formally recognized for their contributions to national security during a virtual awards ceremony. Processing hundreds of terabytes of data under pressure
DARPA designed the final phase of the competition to test the absolute limits of current computational capabilities. Participants processed massive datasets ranging from 600 to 800 terabytes within a strict 24-hour window. Competitors were evaluated on their ability to accurately identify indicators of a potential pathogen release or manipulation. To succeed, teams had to successfully combine deep sequence screening with precise geospatial and temporal metadata. Very minimal information was provided to the competitors about the samples to ensure the most realistic scenario possible.
"The ability to rapidly and accurately identify, characterize, and determine the source of a biological sequence, whether natural or engineered, within an enormously complex environmental sample background is a critical national security capability and U.S. government priority," said Cmdr. Cody Thornton, deputy program manager for the Bio-Attribution Challenge. "We called on creative researchers to help us catalyze a new generation of tools, and these top performers have successfully demonstrated methodologies that provide orders of magnitude improvements over current tools."
To ensure a realistic yet completely secure testing environment, the challenge was entirely computational. Lawrence Livermore National Laboratory meticulously curated all challenge data to mimic realistic and complex attribution scenarios. This approach allowed teams to test their algorithms against highly realistic data sets without the need to utilize actual pathogens or disclose sensitive information.
Following the announcement of the round-one winners in May 2026, the final evaluation distributes the remaining $90,000 of the total prize pool to round-two winners. In addition to the monetary awards, DARPA is also issuing non-monetary recognition to teams that achieved significant breakthroughs in speed, precision, and software efficiency.
| Team name | Award category | Prize / Recognition |
|
Team Crits-Christoph & Hakim |
First Place (Round 2) |
$50,000 |
| Team Kannadasan |
Second Place (Round 2) |
$30,000 |
| Aclid |
Third Place (Round 2) |
$10,000 |
|
Team Crits-Cristoph & Hakim |
Best in Show | Best overall performance across both rounds |
|
TwentyTwo Bio (from Round 1) |
Fastest Analysis | Fastest methods for executing data analysis |
|
Team Sylph (Carnegie Mellon University Yu Lab) |
Most Innovative Method | Most novel mathematical or computational method |
| Team Kannadesan | Most Data-Efficient | Smallest software package to achieve attribution |
| Pastor21 | Highest Precision | Closest measurements to actual true values |
| Treangen Lab | Highest Accuracy | Most consistency and reproducibility in measurements |
Media with inquiries regarding the Bio-Attribution Challenge should .