curaited.dev
  1. Home
  2. » 2025-09-14
  3. » Optimization

AI Breakthrough Accelerates DNA and RNA Therapeutic Design

Google Research and Move37 Labs have introduced NucleoBench, the first large-scale benchmark for nucleic acid design, alongside AdaBeam, a revolutionary AI algorithm significantly advancing the creation of novel DNA and RNA sequences for therapeutic properties. This original work demonstrates that AdaBeam, a hybrid adaptive beam search, outperforms existing methods on 11 of 16 biological tasks evaluated by NucleoBench. AdaBeam achieves superior scalability and memory efficiency through fixed-compute probabilistic sampling and 'gradient concatenation', promising to accelerate AI-driven drug discovery and enhance next-generation mRNA vaccines and gene therapies.

calendar_today 2025-09-11 attribution research.google/blog/

Smarter nucleic acid design with NucleoBench and AdaBeam

Designing novel DNA and RNA sequences with specific therapeutic properties is a formidable challenge, crucial for next-generation treatments. Google Research and Move37 Labs have introduced NucleoBench, the first large-scale benchmark for nucleic acid design, alongside AdaBeam, a revolutionary AI algorithm significantly advancing this critical biomedical frontier. NucleoBench rigorously evaluates 9 algorithms across 16 biological tasks, revealing that AdaBeam, a hybrid adaptive beam search, outperforms existing methods on 11 tasks. AdaBeam demonstrates superior scalability for long sequences and large models by utilizing fixed-compute probabilistic sampling and 'gradient concatenation' for memory efficiency. This open-source tool promises to accelerate AI-driven drug discovery responsibly, enhancing mRNA vaccines and gene therapies.
Good summary?
⊂