EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
Hacker News·Aug 15·5 min read
AI Can Now Design Functional Viruses. Should We Worry?
TL;DR
Researchers at Stanford University have created 16 functional bacteriophages using a genomic language model, successfully targeting antibiotic-resistant E. coli. This breakthrough could lead to personalized phage therapies for bacterial infections.
✦ Why It Matters
Explore opportunities to integrate AI-generated phage therapies into your research or clinical practices today.
Key Takeaways
- Explore AI tools for designing custom phage therapies.
- Stay informed about biosecurity regulations related to synthetic biology.
- Collaborate with interdisciplinary teams for advanced research.
- Evaluate the ethical implications of AI in biotechnology.
How It Works
The Evo 2 model was trained on a vast dataset of phage genomes, allowing it to generate new viral sequences. Researchers fine-tuned the model to focus on a specific phage, ΦX174, and applied computational constraints to enhance the likelihood of creating viable phages.
This process involved synthesizing candidate genomes and testing their ability to infect bacteria.
⚠ The Catch
Despite the success in creating functional phages, the AI-generated genomes were largely similar to existing phages, with an average of 97% identity to the ΦX174 template. This suggests that while the AI can rearrange genetic material, it may not yet produce fundamentally novel viral behaviors.