Neural networks to improve coronavirus and influenza vaccines

Researchers from the bioinformatics group of the Russian Institute of Artificial Intelligence AIRI have created a program that determines the binding sites of antibodies to proteins of viruses and bacteria.

The creators called this program SEMA (Spatial Epitope Modeling). with Artificial intelligence – spatial modeling of nipples using artificial intelligence.

An epitope or antigenic determinant located on the surface of a virus is called an antibody attracting region and serves as a “landing pad” for antibodies.

According to the researchers, the analyzed proteins are loaded into the program as a chain of amino acids or a three-dimensional structure. The SEMA program determines which amino acid residues make up epitopes and how well they interact with antibodies. This information can then be used to select the most promising protein fragments to add to the vaccine, as well as to search for new therapeutic antibodies.

“Today, a number of different tools have been developed for studying proteins in the context of their interaction with antibodies,” says Tatyana Shashkova, senior researcher at the AIRI Institute, in an interview with the Izvestia newspaper, “and we are trying to create a suitable basis for studying proteins using modern artificial intelligence models , since the current version of SEMA is its beginning.”

And he adds that getting results from working with this tool is difficult and time consuming.

Scientists believe that this tool, developed in close cooperation with scientists from the Gamalia Center, will be useful to biotechnological institutes and companies working in the field of creating vaccines and drugs for the treatment of viral and bacterial diseases.

Source: Izvestia newspaper.