Science fiction books and films often portray artificial life forms, but this is not an entirely fanciful conceit. As early as the 1960s, Arthur Kornberg’s team in Stanford University synthesized the genome of the infective DNA phage phiX174 in vitro using DNA polymerase, ligase and a phage DNA template1. Some 30 years later, the same infectious phiX174 DNA was synthesized chemically by the J. Craig Venter Institute using synthesized oligonucleotides, validating that life is essentially chemistry2.

Thanks to further progress in DNA synthesis technology, larger genomes of mycoplasma, bacteria and yeast have subsequently been chemically synthesized, assembled and transplanted into host cells3,4,5, but genome synthesis in multicellular organisms remains unexplored territory. The report of the semi-syn18L moss genome in a recent issue in Nature Plants6 changed that and came as a surprise to much of the plant community, as genome synthesis was a topic most hadn’t thought about before.

This was a pilot experiment under the SynMoss project, which aims to de novo synthesize the whole genome of the terrestrial moss, Physcomitrium patens. According to the researchers, there were several reasons why P. patens was chosen. Sequence accuracy is the most important factor for genome synthesis because any sequencing errors can make the synthesized genomes unable to function in a cell and so fail to support life. Many high-quality genome sequences of P. patens exist, including a recently reported near telomere-to-telomere assembly7. Also, its short life cycle means that an engineered cell can regenerate quickly into a mature plant. Moreover, it has efficient homologous recombination and protoplast regeneration ability, both of which are essential for genome manipulations.

It seems quite reasonable that the first multicellular organism for genome synthesis happened to be a plant rather than an animal, given the higher totipotency of plant cells; plants with artificially synthesized genomes may also trigger fewer safety concerns in the general audience. From another perspective, plant cells contain both mitochondria and plastids, both of which originated from endosymbiosis8, implicating a potential to act as host cells for alien genomes, including synthesized ones.

For the semi-syn18L genome, a region of 155 kilobases (kb), which spans about one third of an arm of chromosome 18 was replaced with a redesigned, simplified, chemically synthesized fragment of roughly 68 kb. Semi-syn18L shows normal wild-type growth, produces spores and maintains an epigenetic landscape similar to the wild type, making it the first living multi-cellular organism to carry a partial artificial chromosome. The authors’ next steps will be to replace the whole chromosome 18 with chemically synthesized sequences, and then the whole genome, which is the ultimate goal of the SynMoss project6.

SynMoss was initiated by a group of Chinese scientists led by Dr. Junbiao Dai, a key member of both the synthetic yeast consortium (Sc2.0) and Genome Project-Write (GP-Write). In a Perspective in this issue of Nature Plants, Dr. Dai and his team highlight the design principles of the P. patens genome and call for international collaboration to work on this milestone project. They have also developed a bespoke online design platform called GenoDesigner, which presents a graphical interface that allows users to manipulate DNA sequences.

But why synthesize a genome? Surely not to show that humans can act as gods and create life? Nobel laureate Richard Feynman said: “What I cannot create, I do not understand?”, and synthetic biologists have a similar philosophy. For Feynman, a clear understanding of the mechanisms underlying a phenomenon can only be achieved with a bottom-up approach. This methodology has proved effective for many biological questions.

In recent years, we have seen the reconstitution of complex cellular processes and structures both in vitro and in vivo by using a minimal set of components or cell extracts. For example, the reconstitution of self-incompatibility response in the self-compatible Arabidopsis thaliana9 or last year’s demonstration by Kelly Dawe’s team at the University of Georgia that a protein-tethering approach can organize new functional centromeres in maize10. In that study, Centromeric Histone H3 (CENH3) protein was attached to synthetic repeat arrays in the maize genome and the ‘synthetic’ centromeres that resulted could both facilitate chromosome segregations and be inherited across generations.

Genome synthesis provides good opportunities for understanding the mechanism of life’s machineries and their evolution. The precise function of diverse genomic components will be revealed by this ‘build-to-understand’ methodology. The debugging process of the yeast Sc2.0 project allowed researchers to characterize the effects of tRNA genes and subtelomeres on gene transcription and silencing11. Similarly, in creating semi-syn18L, the presence of transposable elements (TE) was greatly reduced, but this had no obvious effects on the mosses’ phenotypes, indicating that what effects these TEs have are minimal or only become apparent under specific conditions.

Whole genome synthesis could have many practical applications. For instance, synthetic virus can direct rapid and large-scale production of vaccines; or synthetic bacteria could function as cellular factories producing various substances that satisfy our needs or can have applications in environmental protection. In a similar way to how gene editing has been used to redirect metabolic flux to accumulate provitamin D3 in tomatoes12, synthetic chromosomes could drive the synthesis of important natural products by clustering the necessary genes in a pathway together, offering a more direct and universal way to manipulate plant metabolism. New microbes or plants that may provide better food or medicine could emerge via this technology.

Modifying genomes through gene editing is already a technology. Genome synthesis is certainly more challenging than tweaking the genome using CRISPR, but it opens more spaces for exploring genome biology and more possibilities of engineering. The pioneering geneticist George Church has opined that gene editing will remain the choice for most applications of genetic engineering, but genome design will be useful for specialized applications, such as recoding an entire genome to incorporate new amino acids4,13.

Synthesizing genomes de novo involves more design and greater changes to the original genomes and the features of the organisms. Therefore, discussion is needed on how it should be regulated considering the potential risks of environmental contamination, species invasion, food insecurity and worse risks that science fiction has made us all too aware of. However, powerful technologies are always double-edge swords. With comprehensive, open and continuous discussions among scientists, policymakers and the public, a powerful new tool is being added to our biotechnological toolbox.