Artificial Intelligence to weed out antibiotic-resistant bacteria


Applying a machine-learning algorithm, scientists at MIT have developed a potent antibiotic that is adequate to kill a lot of lethal disease-causing bacteria that are even resilient to most well-known efficient antibiotics.

How do an antibiotic works?

An antibiotic is one that inhibits the growth of microorganisms by targeting the cell wall, nucleic acid or protein synthesis that is quintessential for the growth of microorganisms.

How is the novel antibiotic compound different from the existing ones?

Antibiotic resistance rises when bacterium mutates and well-evolves to the mechanisms that antimicrobial drugs used to kill them. However, this antibiotic has been effective in blocking all the ways bacteria can use to prevent themselves.

How did AI discover the antibiotic?

To develop new antibiotics, scientists trained a “deep learning algorithm” to identify the kind of molecules that kill bacteria. To do this, they provided the program info on the molecular and atomic features of nearly 2500 drugs and judge to what extent are they able to block the growth of E.coli. The algorithm focused on compounds that seemed promising. Scientists describe the way they treated many drug-resistant infections with halicin, a compound initially developed to treat diabetes. Test showed that halicin is efficient in killing Mycobacterium tuberculosis, the causative agent for Tuberculosis, strains of Enterobacteriaceae that are resistant to carbapenems and also multidrug-resistant Acinetobacter baumannii. The treatment functioned against each species they tried, with the exemption of a lung pathogen, Pseudomonas aeruginosa,

Mechanism halicin follows to kill a bacterium

Halicin kills bacteria by destroying the ability of the microorganism to uphold an electrochemical gradient through their cell membranes which is essential to yield ATP, the energy currency of the cell, without which the cell probably dies. After halicin, the scientists planned to test more than 100 million molecules designated from the ZINC15 database.

Leave a Reply

Your email address will not be published. Required fields are marked *

Share via
Copy link