New class of Antibiotic discovered by Artificial Intelligence

New class of Antibiotic discovered by Artificial Intelligence

By, VAISHNAVI

Antibiotics are most common agents used in modern health care. Antibiotic resistance being the major reason for lack of pharmacological response contributing to several deaths. No new class of antibiotics were discovered over the past decade to combat the antibiotic resistance. Antimicrobial resistance was increasing with no advancement in discovery of new classes of antibiotics. Traditional discovery of clinical candidate requires at least 3 to 6 years. But within few hours, thousands or hundreds of thousands of preclinical candidates can be discovered with help of machines which accelerate the drug discovery process. So, Artificial Intelligence (AI) model was used for initiation of new antibiotic discovery. 

Initially 39,000 compounds were tested for its efficacy and toxicity in both Staphylococcus aureus and three types of human cells from liver, skeletal muscle and lungs. The results were used to train the AI model about the patterns in each compound’s chemical atoms and bonds. So now, the AI model was able to detect the efficacy of compound against Staphylococcus aureus and also its potential toxicity in human cells. 

Twelve Million compounds were screened by the team at laboratory of James Collins of the Broad Institute of the Massachusetts Institute of Technology and Harvard University using a type of AI known as Deep Learning. Further the chemical substructures were identified which revealed more specific properties of each compounds. By comparing those substructures, new classes of potential antibiotics were identified. Thereafter, 283 promising compounds were tested in mice to prove its efficacy in treating skin and thigh infections caused by resistant species, Methicillin-resistant Staphylococcus aureus (MRSA) which was considered as crucial pathogen to kill. Out of those, 2 non-toxic compounds were capable of killing both MRSA and Vancomycin-resistant Enterococci (VRE) in mouse models. 

Although more studies are needed to be performed for FDA approval of drug and its translation for clinical use. Systematic toxicity studies and Pre-IND (Investigational new drug) studies are yet to be performed for the drug to enter phase-I of clinical trial.