AI for the First Time Used to Design New Viruses, 16 Bacteriophages Successfully Created

Artificial intelligence or AI is now entering further into the world of biology. Scientists in the United States have for the first time used generative AI to design new viruses that are not found in nature.

Anadolu Agency, quoted on Friday, August 7, reported that the study was conducted by scientists from Stanford University and the Arc Institute in California. The results were published in the journal Science.

The researchers used a generative AI model trained with millions of natural genomes to design bacteriophages in toto.

Bacteriophages are viruses that infect bacteria and can multiply in them. While the genome is the entire genetic material that carries the biological information of an organism.

In the study, scientists used two genomic language models called Evo 1 and Evo 2.

The way it works is similar to the large language model or LLM used in generative AI. The difference is that Evo 1 and Evo 2 do not predict a series of words, but rather a genetic sequence.

The training data comes from the genomes of viruses, bacteria, to more complex organisms such as plants and animals.

The researchers then chemically synthesized nearly 300 genomes and tested them in the laboratory.

The experiment produced 16 engineered bacteriophages with different sequences, structures, and survival and growth abilities.

When tested, the engineered bacteriophage was more effective in killing E. coli bacteria than the natural virus used as a comparator.

The study cited by Anadolu Agency said the approach could expand the capabilities of synthetic genomics, which is a field that designs or assembles genetic material artificially.

The researchers also looked at the possibility of using this method to develop phage therapy, which is the use of bacteriophages to fight disease-causing bacteria.

"Our approach expands what can be achieved by synthetic genomics, paving the way for adaptive and pathogen-resistant phage therapies that evolve rapidly, and building the foundation for larger and more complex genome designs," the researchers wrote.

They say this research is still a preliminary step. Similar methods in the future can be developed for sequencing and genome synthesis in larger biological systems as well as therapy development.