A two-round Defense Advanced Research Projects Agency (DARPA) competition to develop next-generation computational tools for tracing the origins of engineered biological threats has completed its first phase and is now advancing to the harder problem of attribution. Round 1 of the Bio-Attribution Challenge — focused on detection and characterization of complex biological sequences — has concluded, with winners announced and invitations issued to qualifying teams for Round 2. With the competition open through June 15, 2026, and $180,000 in total prizes on offer, DARPA is actively seeking additional participants with expertise in bioinformatics, data science, high-performance computing, and machine learning.
Launched March 12, 2026, by DARPA’s Biological Technologies Office, the Bio-Attribution Challenge calls on innovators to develop computational tools capable of identifying, characterizing, and tracing the origins of modified biological sequences — proteins, genes, and genomes — within datasets far larger than current systems can handle. Round 2 targets the core attribution problem: determining the origins of engineered biological events.
Why Attribution Capacity Is a Biodefense Gap
Existing systems cannot analyze petabyte-scale datasets in near real-time — the speed at which attribution decisions may need to be made during a fast-moving biological event. DARPA program manager Abhishek Singharoy, Ph.D., framed the core problem as finding the “needle in a haystack” within data environments of unprecedented scale and complexity, and described rapid and accurate identification of a biological sequence’s source — whether natural or engineered — as a critical national security capability.
In a mass casualty event involving a suspected biological agent, early determination of whether the pathogen is naturally occurring or has been engineered affects treatment protocols, infection control decisions, public communication strategy, and the activation of law enforcement and national security response channels. Tools that can compress attribution timelines from days to hours — or hours to minutes — would materially change the clinical and public health response calculus.
Round 1 Results and Competition Structure
The challenge is purely computational: no actual pathogens or biological materials are involved. All data has been deliberately curated and developed by Lawrence Livermore National Laboratory to mimic realistic and complex scenarios without disclosing sensitive information. Participant software runs in a secure, government-controlled environment.
Team Crits-Christoph & Hakim took first place and $50,000; the Treangen Lab at Rice University placed second for $30,000; and Pastor21 AI Inc placed third for $10,000. Eight teams ranked in the scored tier, with several additional teams in an unscorable category. Round 2 invitations have been issued to qualifying participants.
Sources and Further Reading:
Translate your Bio-attribution Research into National Security Impact – DARPA