AI Moves From Reading Genomes Toward Designing Biology

Why in the News ?

A Stanford University–Arc Institute experiment demonstrated that AI models can generate complete bacteriophage genomes that scientists can synthesise and test. The development highlights opportunities for phage therapy and biomedical research, while raising significant biosecurity and AI-governance concerns.

From Genome Reading to Biological Design

●     Scientists have progressed from sequencing viral genomes to synthesising and modifying them; AI is now entering the design stage.

●     Researchers used Evo 1 and Evo 2, genome-language models trained on DNA patterns, to generate novel genomes related to the bacteriophage ΦX174.

●     Unlike conventional language models that learn words, these systems learn patterns in the genetic alphabet — A, C, G and T.

●     Out of 285 AI-generated designs that were physically synthesised and tested, 16 produced functioning bacteriophages.

●     Some AI-designed phages successfully overcame bacterial resistance that had defeated the original virus.

●     Importantly, AI did not independently manufacture a virus; scientists selected the designs, synthesised the DNA and conducted laboratory experiments.

●     The significance lies in AI’s ability to explore complex combinations of genetic changes that humans may find difficult to test individually.

Medical Potential and Biosecurity Risks

●     The most immediate medical opportunity is phage therapy, where bacteriophages are used to target disease-causing bacteria.

●     This could help address Antimicrobial Resistance (AMR), as conventional antibiotics are becoming less effective.

●     AI-designed phages could potentially be customised against particular bacterial strains and resistance patterns.

●     Similar AI capabilities could support the development of vaccine antigens, antibodies, therapeutic proteins and viral vectors.

●     However, generative biology could also accelerate the design of harmful biological systems.

●     The concern is therefore capability amplification — AI could make knowledgeable and well-equipped laboratories substantially more effective at biological engineering.

●     Traditional DNA screening largely checks whether sequences resemble known pathogens or toxins; future safeguards may also need to assess the potential biological function of designed sequences.

●     Excessive restrictions could obstruct legitimate research, while inadequate safeguards could create biosecurity risks.

 About AI, Biotechnology and Biosecurity:

●     Bacteriophages are viruses that infect bacteria and are being explored as alternatives or complements to antibiotics.

●     Generative biology applies AI to predict, design or optimise biological molecules and systems.

●     Antimicrobial Resistance (AMR) occurs when microorganisms become resistant to medicines used against them.

●     AI can accelerate drug discovery, vaccine development, protein engineering and personalised therapeutics.

●     Effective governance requires safeguards across the entire chain: AI models, DNA-synthesis companies, laboratories and institutional biosafety mechanisms.

●     India needs greater investment in scientific AI, secure computing, high-quality biological datasets and trusted research-access frameworks.

●     The key policy challenge is achieving “controlled acceleration”—promoting beneficial scientific innovation while ensuring that safeguards become stronger as technological capability and associated risks increase.