topics|Code of life

Synthetic Biology

The field of writing DNA from scratch rather than merely reading or editing it.

Synthetic genomes

The first synthetic gene was created in the 1970s. Subsequent milestones in whole-genome synthesis:

2008 — Mycoplasma genitalium, the first synthetic bacterial genome
2010 — Mycoplasma mycoides
2019 — Escherichia coli
2025 — Baker's yeast (Saccharomyces cerevisiae), the first synthetic eukaryotic genome. Work began in 2006; the last of its 16 chromosomes was synthesised and announced in a paper published in January 2025. At 12m base pairs, it is the largest completed synthetic genome to date.

The Synthetic Human Genome (SynHG)

Launched on June 26th 2025, SynHG aims to develop the tools needed to create a human chromosome from scratch. The project is funded partly by Wellcome, a charity, and includes members from a handful of British universities. Its lead is Jason Chin, based at the University of Oxford. Dr Chin estimates the cost of synthesising a human chromosome at around $650,000. George Church, a biologist at Harvard University, puts the figure at more than $20m.

The smallest human chromosome—number 21—measures 45m base pairs, nearly four times larger than the yeast genome.

The project includes a programme called Care-full Synthesis to investigate the ethical dimensions of the work. Hank Greely, a lawyer and bioethics expert at Stanford University, has argued that testing whether a synthetic human chromosome functions normally would require putting it into babies, which would be illegal in most countries including Britain.

AI and genomic design

AI models are beginning to guide DNA design. In February 2025 the Arc Institute in Palo Alto released Evo 2, a generative AI model capable of devising new genomes based on short DNA sequences given as prompts. On June 25th 2025 Google DeepMind launched AlphaGenome, a deep-learning model that can predict how small genetic changes will affect cell function. Hani Goodarzi, one of the scientists at Arc who developed Evo 2, has said the two models could, when combined, allow scientists to produce new designs for human DNA that enable specific cellular functions.

Challenges

The principal difficulty is scale. DNA is built one nucleotide at a time, and longer stretches—made by combining short ones—are more difficult and costly to produce accurately. Gene editing via CRISPR may prove a cheaper and more reliable alternative for many applications; CRISPR techniques are now capable of making several simultaneous edits.