The problem. The genetic code is redundant — 64 codons for 20 amino acids plus stop — so in principle you don’t need all 64. But “in principle” is cheap. Could you actually rewrite an entire genome to use fewer codons, across every gene at once, and have the organism survive?

The idea. Chin’s group designed and chemically synthesized a 4-megabase E. coli genome, Syn61, in which every occurrence of three codons was swapped for a synonym: two serine codons (TCG→AGC, TCA→AGT) and the TAG stop codon (TAG→TAA). That’s ~18,000 edits genome-wide, designed in silico, synthesized in segments, and stitched into the chromosome by stepwise replacement (REXER/GENESIS), continuously checking viability. The recoded strain grows — a little slower, but it lives.

Why it matters. This is the vivid, tangible answer to what genome-scale Build looks like, and it’s really a computation story wearing a lab coat: the genome is designed as text, validated by software, and only then fabricated. The payoff is downstream — freeing codons means you can reassign them, e.g. to encode non-canonical amino acids or to engineer genetic isolation and phage resistance (the follow-up Syn61Δ3 work). It’s the strongest single illustration that genomes are editable at the whole-organism level, and that the hard part is increasingly design and verification, not chemistry.

Verdict. A landmark, and an honest one — the reduced growth rate and the sheer cost/effort are stated plainly, and “removed three codons” is a long way from “designed a genome from function.” But as a demonstration of scale it’s unmatched, and it’s the cleanest on-ramp to the DBTL framing. My favorite “genome as editable file” post.