In Nature Communications, a group of scientists from the University of Warsaw described an innovative artificial intelligence model that proposed promising antibacterial substances.

As the researchers point out, the increasing drug resistance of bacteria poses a growing global threat. Some experts even fear that humanity will soon become defenceless against bacteria. One class of substances that can help are peptides - protein-like molecules produced by living organisms. Some are designed to destroy microbes.

These molecules have a huge advantage, namely that bacteria are extremely slow to acquire resistance to them. Unfortunately, as the scientists from the University of Warsaw emphasize, it has not yet been possible to find one that is superior to conventional antibiotics in treating bacterial infections.

The team led by Professor Ewa Szczurek from the Faculty of Mathematics, Mechanics and Computer Science of the University of Warsaw, approached this challenge in an unconventional way, with the help of artificial intelligence. The researchers used their own, innovative intelligent algorithms. The system called HydrAMP generated previously unknown peptide sequences with potent antimicrobial activity.

Sequences, because peptides consist of sequentially arranged smaller molecules - amino acids. The computer generated new strings of amino acids after being fed data on the existing ones.

“Since we have a large enough number of such sequences, we can teach our model to distinguish random sequences from peptide coding sequences. We also have data on the activity of known peptides. The model can learn this too,” explains Paulina Szymczak, a doctoral candidate at the Interdisciplinary Doctoral School and first author of the paper published inNature Communications, quoted on the University of Warsaw website.

Professor Wojciech Kamysz’s team from the Medical University of Gdansk synthesized HydrAMP-generated peptides and tested them on bacterial strains and red blood cells to verify their safety. The results are promising.

“We discovered fifteen completely new active peptides. Among them is a peptide called Varsavian, which shows very promising activity against dangerous bacteria such as methicillin-resistant Staphylococcus aureus,” reports Professor Szczurek

What's more, the AI developed by her team can improve on previously known peptides. A peptide that was completely ineffective against bacteria was converted into a peptide with a very promising activity profile without making it toxic.

The scientists point out that they have developed the first model that can modify known peptides in this way. A patent application for the discovered molecules has already been submitted to the Patent Office; they may now be tested in clinical trials and used as drugs in the future.

Find out more in the source article. (PAP)

Author: Marek Matacz

mat/ bar/ kap/

tr. RL