Researchers try to cut the genetic code from 20 to 19 amino acids

2026-04-30 · Source: AI - Ars Technica · Field: Science & Research — Life Sciences & Biology, Research Methodology & Innovation · Depth: Advanced, medium

Summary

Researchers from Columbia and Harvard successfully engineered *E. coli* to function with a ribosome largely free of the amino acid isoleucine, testing hypotheses that early life forms may have utilized a reduced genetic code. The team employed AI-based protein design software and iterative testing to replace isoleucine with valine in essential genes and ribosomal proteins, focusing on the small ribosomal subunit. They managed to replace 20 of 21 small ribosomal subunit proteins, achieving a strain that grew at 70% the rate of unmodified cells, and eventually all 21, resulting in 60% growth, with no isoleucine restoration over 400 generations. This project highlights the power of "AI tools" in protein redesign, enabling changes that biologists might avoid, but also reveals current limitations in model explainability. While an astonishing achievement, its direct utility for understanding life before the universal common ancestor or creating a fully isoleucine-free genome remains a "maybe" category.

Key takeaway

Using AI-driven protein design, researchers successfully engineered *E. coli* to produce a functional small ribosomal subunit entirely free of isoleucine, a critical step towards reducing the universal genetic code. This involved replacing isoleucine with valine in 20 of 21 small subunit proteins, yielding a strain that grows at 60-70% the rate of unmodified cells and maintains stability over 400 generations. This advance demonstrates the profound capability of AI in synthetic biology, enabling radical alterations to fundamental cellular machinery for exploring reduced genetic codes and early life evolution.

Topics

• Genetic Code Evolution
• Amino Acid Reduction
• Ribosome Engineering
• AI-based Protein Design
• Isoleucine Substitution

Best for: Research Scientist, AI Scientist

Related on AIssential

• All life runs on 20 amino acids. These cells run key machinery on just 19 — Bacteria were reengineered to operate with 19 amino acids, leveraging AI for protein design.
• AI can write genomes — how long until it creates synthetic life? — The Evo2 genomic language model, an AI designed to read, interpret, and generate DNA, RNA, and protein sequences, marks a significant advancement in synthetic biology by designing whole genome sequenc…
• 🔬ESM: The Bitter Lesson is Coming for Proteins - Alex Rives, BioHub — Scaling protein language models with vast, diverse data reveals emergent biological capabilities and enables programmable design.
• GPT-5 lowers the cost of cell-free protein synthesis — GPT-5 connected to a robotic lab significantly reduced cell-free protein synthesis costs through autonomous closed-loop optimization.
• Programmable RNA translation through deep learning-driven IRES discovery and de novo generation — An AI framework unifies IRES identification, optimization, and de novo generation for programmable RNA translation.
• Leap in DNA Synthesis Slashes Time to Build New Genetic Sequences — Sidewinder enables rapid, high-fidelity DNA synthesis, resolving a key bottleneck for generative AI in biology.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI - Ars Technica.