Mankind knows several thousand different viruses and regularly find new ones (including those dangerous to humans). However, artificial intelligence was able to find as many as 6,000 previously unknown viruses. And he found them there, where it would be extremely problematic for a person to seek.
Recently, scientists are increasingly looking for viruses through DNA sequencing. The fact is that many viruses “almost tightly” merged with cellular structures, and to reveal them in a “pure” form is extremely difficult. To identify viruses, scientists use genetic signatures of known viruses and bacteria. At the same time, researchers, for clarity, compare their method with the search function inside a text document, when we search for individual words and letters. But in this way, there is one problem: scientists cannot search for what they do not know, and the viruses found in this way are often an accidental find. But artificial intelligence helps to get around this problem. As noted by Jie Ren, a computational biologist at the University of Southern California in Los Angeles,
“Previously, people did not have a way to quickly and thoroughly study viruses. But now we have the tools to find them. We have trained the neural network to identify the genetic sequence of viruses from the unusual Inoviridae family. These viruses live in bacteria and change the behavior of the host. “
Experts collected a vast array of data on the genome, and AI recovered more than 10,000 genes of Inoviridae and grouped them. About 6000 fragments found were previously unknown to science. Moreover, each such fragment can exist separately from the others. As the experts say, this discovery can become a new tool for studying the diversity of viruses on our planet.