Biotechnology and Policy · global
As AI Designs Complete Bacteriophage Genomes, the U.S. Makes Nucleic Acid Screening a Requirement for Research Procurement
Sixteen AI-designed bacteriophages viable in the laboratory offer another avenue for antibacterial research. A U.S. nucleic acid synthesis policy announced in November seeks to establish a verifiable safety checkpoint before digital sequences become physical materials.
When genetic sequences designed by artificial intelligence can become functional bacteriophages in the laboratory, the questions facing biotechnology extend beyond what models can write to who can manufacture it. On November 5, the U.S. White House announced a new nucleic acid synthesis screening policy linking suppliers’ safety reviews to federal research funding, making the procurement of synthetic nucleic acids a more binding checkpoint.
In its annual biotechnology analysis on November 17, OODA Loop placed AI-designed genomes alongside this policy, illustrating how technological capabilities and safety governance are advancing in tandem. One specific study was a bacteriophage design study published in Science in August. According to the paper’s abstract hosted by the Broad Institute, the research team used the ΦX174 bacteriophage as a template and a genome language model to generate complete genomes. Experimental testing yielded 16 viable bacteriophages that showed different fitness profiles under laboratory conditions.
Bacteriophages are viruses that infect bacteria and tools for exploring antibacterial therapies. The study further found that mixtures of the generated bacteriophages could overcome bacterial strains resistant to ΦX174 in experiments. The significance is that AI design has moved beyond sequences that merely appear plausible to measurable biological function, providing research leads for addressing bacterial resistance.
However, antibacterial performance in the laboratory remains a long way from treatment in humans. These results do not yet demonstrate safety and efficacy in patients, nor can they directly support the inference that the same method applies to more complex viruses or other biological systems. The study used an existing bacteriophage as its design template. That premise must also be retained when interpreting the findings, to avoid extrapolating success in a specific system into a general ability to design genomes.
The White House’s NSTM-6 policy focuses on the point at which sequences enter the manufacturing process. Covered recipients and subrecipients of federal funding must procure synthetic nucleic acids and benchtop nucleic acid synthesis equipment from suppliers that meet the policy’s requirements. For applicable orders involving sequences at least 50 nucleotides long, suppliers must use screening systems validated by the U.S. National Institute of Standards and Technology (NIST), verify customer identities, report flagged orders, and retain relevant records for at least eight years.
These requirements will be implemented in phases, rather than taking full effect on the day of the announcement. The policy calls for NIST to issue standards and testing and evaluation specifications within 90 days, and requires covered customers to switch to compliant suppliers for synthetic nucleic acid procurement within 250 days; the procurement deadline for benchtop synthesis equipment is 340 days. For research institutions, this means supplier selection, ordering procedures, and equipment procurement must all be adjusted to meet the new funding conditions.
The practical effect of the new system will still depend on screening standards, validation capacity, and subsequent enforcement. It constrains procurement through conditions on federal funding, which does not mean that all private transactions are subject to the same rules; completing screening also does not mean biological risks have been eliminated. AI-designed bacteriophages demonstrate a research capability that can be put to use, while the new policy addresses access to materials and accountability. Their intersection concerns how to bring valuable biological designs into validation while ensuring safety reviews keep pace with manufacturing capabilities.