Antimicrobial resistance (AMR) is one of the world's most pressing public health issues. New treatments and strategies are needed to address the rise in resistant superbugs. Now, scientists from Queensland University of Technology (QUT) and the QUT Microbiome Research Centre report they have discovered the potential source of nearly one million antibiotics in nature. Using machine learning, the researchers identified 863,498 promising antimicrobial peptides.
The results of their research are cell The article is titled “Discovery of antimicrobial peptides in global microbiomes using machine learning.”
“New methods for antibiotic discovery are urgently needed,” explains Luis Pedro Fragao Bento Coello, associate professor and researcher at the Queensland University of Technology's Centre for Microbiome Research.
“It is one of the greatest public health threats, killing 1.27 million people each year.”
It is estimated that without intervention, AMR could cause up to 10 million deaths per year by 2050.
“We hope that using artificial intelligence to understand and harness the power of the global microbiome will drive innovative research that will improve public health outcomes,” he said.
The team tested the machine's predictions by testing 100 lab-made peptides against clinically important pathogens. They found that 79 different bacterial membranes were disrupted and 63 different antibiotic-resistant bacteria were targeted. Staphylococcus aureus and E. coli.
“Moreover, several of the peptides helped to eliminate infection in mice, with two in particular reducing bacteria by up to four orders of magnitude,” Coelho said.
In preclinical models tested on infected mice, treatment with these peptides produced results similar to those of polymyxin B, a commercially available antibiotic used to treat meningitis, pneumonia, sepsis and urinary tract infections.
To obtain these results, more than 60,000 metagenomes (a collection of genomes in a specific environment), containing the genetic makeup of more than one million organisms, were analysed from sources around the world, including marine and soil environments, and human and animal guts.
As a result, a comprehensive database containing these novel peptides, AMPSphere, has been released as a publicly available open-access resource for the discovery of new antibiotics.
Coelho's research was conducted as part of an ARC Future Fellowship through the Queensland University of Technology's School of Biomedical Sciences, in collaboration with Cesar de la Fuente's lab at the University of Pennsylvania, Fudan University, the European Molecular Biology Laboratory and APC Microbiome Ireland.
