In 2010, a team at the J. Craig Venter Institute built a bacterial genome from chemicals, put it into a cell, and got the cell to run on it. Reports put the cost at about $40 million and 200 person-years. The genome was 1.08 million bases long.

Fifteen years later, researchers published an E. coli whose 4 million base genome had been rewritten at more than 101,000 positions. The size of what we can write has grown. The more interesting change is what we write.

How big have synthetic genomes gotten?

Source: J. Craig Venter Institute and Wikipedia (2008 to 2016); Fredens et al., Nature, 2019; Science, 2025 (Syn57). Sizes are rounded. The 2010 size is the 1.08 million base genome reported by the JCVI team.

View data

milestone	kb
2008 M. genitalium	583
2010 JCVI-syn1.0	1080
2016 JCVI-syn3.0	531
2019 E. coli Syn61	4000
2025 E. coli Syn57	4000

Three of the five bars are JCVI genomes built in Mycoplasma, a bacterium with one of the smallest genomes of any free-living cell. The first, in 2008, copied a natural genome of 582,970 bases. The second, in 2010, was a 1.08 million base genome that ran a cell. The third, in 2016, was a design experiment: JCVI-syn3.0 cut the genome to 531,560 bases and 473 genes, and 149 of those genes had no known function.

The last two bars are E. coli, whose natural genome is bigger. Syn61 in 2019 and Syn57 in 2025 were both about 4 million bases. The size stopped growing. Something else was changing.

What changed instead?

The change was the genetic code. DNA is read three letters at a time. Each three-letter codon stands for an amino acid or a stop signal, and the natural code has 64 codons for about 20 amino acids. Many amino acids have several codons. That redundancy gives a synthetic biologist something to remove.

Source: Fredens et al., Nature, 2019; Science, 2025 (Syn57); MRC Laboratory of Molecular Biology. Syn61 uses 61 codons and Syn57 uses 57, of which 55 encode the 20 canonical amino acids.

View data

genome	removed
Syn57 (2025)	7
Syn61 (2019)	3
Natural E. coli	0

Syn61 removed three codons, down to 61. Syn57 went further. Its builders replaced every use of six sense codons and one stop codon with synonyms, which meant recoding more than 101,000 codons, and ended with a 57-codon code. Codons that no longer appear in the genome are free. A cell can be taught to read them as something new, such as an amino acid that does not occur in nature.

What about yeast?

Yeast is the next test. The Sc2.0 consortium has designed and built synthetic versions of all 16 chromosomes of Saccharomyces cerevisiae, plus a separate chromosome that gathers the transfer RNA genes. A May 2026 review in the journal Yeast notes that about one third of the haploid genome has been consolidated into a single strain so far. The natural yeast genome is about 12 million bases, three times the size of the E. coli genomes above.

Limits

A genome that runs a cell is not a product. These projects show what can be written, not what can be manufactured at a price. The 2010 cost figure comes from a Wikipedia summary of press reports, and the cost of the 2019 and 2025 projects is not stated in our sources, so we do not compare them. The sizes in the chart are for different organisms with different natural genome sizes.

Sources

1. Wikipedia, Mycoplasma laboratorium: genome sizes for M. genitalium and JCVI-syn3.0 (473 genes, 149 of unknown function), and the approximate cost and effort of JCVI-syn1.0.
2. Gibson et al., Creation of a bacterial cell controlled by a chemically synthesized genome, Science, 2010.
3. Fredens et al., Total synthesis of Escherichia coli with a recoded genome, Nature, 2019.
4. Escherichia coli with a 57-codon genetic code, Science, 2025.
5. MRC Laboratory of Molecular Biology, Syn57 represents a new chapter in the genetic code of life.
6. Sc2.0 review, One Yeast, Sixteen Synthetic Chromosomes, Infinite Possibilities, Yeast, May 2026.