nature.com/lw767/magazine-assets/d41586-02970 w/d41586-02970 w_2769007070707070.JPG? https://media.nature.com/lw319/magazine-assets/d41586-02970-w/d41586-02970-w_27690070.jpg?as=webp 319W "SISES =" (Max-Width) 319PX, 100VW, 767PX ">

Artificial intelligence (AI) helps to re -draw the family tree of viruses. Predicted protein structures, with Alphafold and Chatbot-Inspired "Protein language models" have covered surprising connections in a family of viruses that infect and arise threats.

A large part of the scientists' understanding of virale Evolution is based on the comparison of genomes. However, the lightning-fast evolution of viruses, in particular from those with RNA genemen, and their tendency to adopt genetic material from other organisms that genetic sequences can hide deeper and distant relationships between viruses that are to vary depending on the examined.

In contrast, the shapes or structures of the proteins encoded by viral genes tend to change, which enables these hidden evolutionary connections to be recognized. However, it was not possible to compare protein structures across an entire virus family until tools such as Alphafold, which can predict protein structures on a large scale, says Joe Grove, a molecular virologist at the University of Glasgow, Great Britain.

In an article published this week in Nature 1 show the power of a structured-based Approaches among the flaviviruses-a group that includes the hepatitis-C, dengue and zika virus as well as some important animal pathogenic and species that could represent emerging threats to human health.

How to penetrate viruses

Understanding the researchers about the evolution of the flaviviruses is mainly based on sequences of slowly evolving enzymes that copy their genetic material. However, it is remarkably little known about the origins of the "Viral Entry" proteins that use flavivirus to penetrate in cells and that determine the host that you can infect. Grove argues that this knowledge gap is the development of an effective vaccine against Hepatitis C , which killed hundreds of thousands of people every year.

"At the sequence level, things are so diverged that we cannot say whether they are related or not," he says. "The breakthrough in the prediction of protein structures opens the entire question and we can see things quite clearly."

The researchers used Deepminds Alphafold2 model and Esmfold, a Structural- Predictions tool, which was developed by the technologie Meta , for more than 33,000 predicted structures of 458 Flavivirus-Spezies to generate. Esmfold is based on a voice model that was trained with tens of millions of protein sequences. In contrast to Alphafold, it only takes one input sequence instead of racing on several sequences of similar proteins, which could make it particularly useful for examining the mysterious viruses.

hepatitis c virus e1 glycoprotein predicted using colabfold-alpahfold2.

The predicted structures made it possible for the authors to identify viral entrry proteins, the sequences of which differ greatly from those of well-known flavivirus. They found some unexpected connections. The virus group, which comprises hepatitis C, uses a system to infect cells that are similar to the pestiviruses - a group that includes the classic swine flu virus, causes hemorrhagic fever in pigs, and other animal pathogenic.

The AI-based comparisons showed that this input system differs from the many other flaviviruses. "We don't know where your input system comes from for hepatitis C and his relatives. It could have been invented," says Grove.

stolen by bacteria

The predicted structures also showed that the well-examined entrance proteins of Zika and Dengue viruses have the same origins as those of the "strange and wonderful" flaviviruses with huge genomes, including the Haseki tick virus, which can trigger fever in humans. Another big surprise was the discovery that some flaviviruses have an enzyme that was apparently stolen by bacteria.

"That would be unprecedented," says virologist Mary Petrone from the University of Sydney, Australia, it would not be for the discovery of her team this year of a similar theft with a particularly "strange and wonderful" flavivirus art 2 . "Genetic piracy could have played a greater role in the evolution of the flavivirus than previously thought," she adds.

David Moi, a computer-aided biologist at the University of Lausanne, Switzerland, says that the Flavivirus study is only the top of the iceberg and that the evolutionary stories of other viruses and even some cellular organisms are probably told with AI. "Now that we can take a look, all of these things have to get a small update," he says.