AI takes a step toward designing synthetic life
Researchers have used artificial intelligence to design entire genome sequences, marking a new step toward the creation of synthetic life, according to a study published in the journal Nature, reports a Qazinform News Agency correspondent.
The breakthrough relies on a DNA language model called Evo2, trained on trillions of genetic letters from organisms across the tree of life. Scientists used the model to generate several genome designs, including one inspired by the bacterium Mycoplasma genitalium.
The work builds on earlier milestones in synthetic biology. In 2008, researchers produced the first synthetic genome of a living organism by chemically synthesizing the genome of Mycoplasma genitalium. Later experiments “rebooted” similar genomes inside cells, which scientists described as the first example of synthetic life.
Despite the progress, experts say AI-generated genomes are still far from creating functioning organisms. “It’s cool, but it’s not there yet,” said Nico Claassens, a synthetic biologist at Wageningen University in the Netherlands.
Scientists note that designing genomes on a computer is only the first step. The sequences must still be synthesized and tested in real cells to confirm that they work.
Patrick Yizhi Cai, a genome engineer at the University of Manchester, described the potential impact of such technologies. “These AI models are the ‘ChatGPT moment’ for synthetic genomics,” he said. “You can start writing things that never existed in nature.”
Earlier research using Evo models successfully generated viral genomes capable of infecting bacteria, though those genomes were far simpler than those of bacteria or other living cells.
Experts caution that even small errors in a genome design could prevent it from functioning. “You cannot design life 70%,” Claassens said. “You can do that on a computer, but it will not be functional.”
Researchers say future advances may combine AI, robotics and automated laboratories to design and test new genomes more efficiently, potentially transforming biotechnology and genetic engineering.
Earlier, Qazinform News Agency reported that AI models used in a simulated nuclear war game escalated conflicts by threatening nuclear strikes in 95% of scenarios, according to new research conducted by King’s College London.
