Artificial Intelligence to combat Antimicrobial resistance

A powerful antibiotic that kills some of the most dangerous drug-resistant bacteria in the world has been discovered using artificial intelligence. Halicin, one among these promising antibiotics is effective against a broad range of resistant strains. Antibiotic resistance among pathogens is a growing concern among the science world and society. Antibiotic resistance occurs when bacteria mutate and evolve to adapt to the alternative mechanisms that antimicrobial drugs use to kill them. With no new antibiotics to tackle this growing resistance, approximately 10 million lives around the world could be at risk each year from infections by 2050, as the Cameron government’s O’Neill report warned.

Halicin discovered by AI at Massachusetts Institute of Technology (MIT), kills some of the most dangerous strains of bacteria known. This drug acts in a way, different from how existing antibacterials act, which increases its chances of being effective. It’s the first of its kind; discovered through an “in silico” model that is used to design new drugs based on what it has learned about chemical structures that enable drugs to kill bacteria. Tests have shown that the drug wipes out dangerous strains including Enterobacteriaceae and Acinetobacter baumannii, which are high priority pathogens that the World Health Organization (WHO) ranks as critical for new antibiotics to target. As per the information available, this drug stands to be one of the more powerful drugs till date that show a remarkable activity against a broad range of resistant strains. Tests on bacteria collected from patients showed that halicin killed Mycobacterium tuberculosis, the bug that causes tuberculosis (TB), and strains of Enterobacteriaceae that are resistant to carbapenems- a group of antibiotics that are considered the last resort for such infections. Halicin also cleared Clostridium difficile and multidrug-resistant Acinetobacter baumannii infections in mice.

The discovery of new antibiotics was aided by training the machine using a “deep learning” algorithm to identify the molecules that are capable of killing bacteria. To achieve this, the machine was fed with atomic and molecular features of nearly 2,500 drugs and natural compounds against Escherichia coli. Once this algorithm was learned, it was subjected to a working library of almost 6,000 compounds to identify potential antibiotics. Jonathan Stokes, the first author, said that the whole process of identification by the machine took only a few hours to give these satisfactory results. Keeping in mind these results, the scientists behind this study are preparing for a study with a larger number of drugs. With the success of this study, better health for the people can be assured as many of the incurable diseases will soon have a solution.

Paper link: https://doi.org/10.1016/j.cell.2020.01.021

Written by Jyoti Sharma
MSc Science and Technology Communication
CSIR-NISCAIR

Instagram link: https://instagram.com/episteme2020?igshid=14o31j9kxmzmt

Leave a Reply

%d