Antimicrobial resistance (AMR) is one of the world’s urgent public health problems. New therapeutics and strategies are needed to contend with the growing number of resistant superbugs. Now, scientists at the Queensland University of Technology (QUT) and the QUT Center for Microbiome Research report they have discovered almost a million potential sources of antibiotics in the natural world. The researchers used machine learning to identify 863,498 promising antimicrobial peptides.

Their findings are published in Cell in an article titled, “Discovery of antimicrobial peptides in the global microbiome with machine learning.”

“There is an urgent need for new methods for antibiotic discovery,” explained Luis Pedro Fragao Bento Coelho, associate professor and a researcher at the QUT Center for Microbiome Research.

“It is one of the top public health threats, killing 1.27 million people each year.”

Without intervention, it is estimated that AMR could cause up to 10 million deaths per year by 2050.

“Using artificial intelligence to understand and harness the power of the global microbiome will hopefully drive innovative research for better public health outcomes,” he said.

Associate professor Luis Pedro Fragao Bento Coelho has explored the global microbiome to find peptides that could be used as antibiotics.  [QUT]

The team verified the machine predictions by testing 100 laboratory-made peptides against clinically significant pathogens. They found 79 disrupted bacterial membranes and 63 specifically targeted antibiotic-resistant bacteria such as Staphylococcus aureus and Escherichia coli.

“Moreover, some peptides helped to eliminate infections in mice; two in particular reduced bacteria by up to four orders of magnitude,” Coelho said.

In a preclinical model, tested on infected mice, treatment with these peptides produced results similar to the effects of polymyxin B—a commercially available antibiotic that is used to treat meningitis, pneumonia, sepsis, and urinary tract infections.

More than 60,000 metagenomes (a collection of genomes within a specific environment), which together contained the genetic makeup of over one million organisms, were analyzed to get these results. They came from sources across the globe including marine and soil environments, and human and animal guts.

The resulting AMPSphere—a comprehensive database comprising these novel peptides—has been published as a publicly available, open-access resource for new antibiotic discovery.

Coelho’s research was conducted as part of his ARC Future Fellowship through the QUT School of Biomedical Science, in collaboration with the Cesar de la Fuente laboratory at the University of Pennsylvania, Fudan University, the European Molecular Biology Laboratory, and APC Microbiome Ireland.

The post Machine Learning Digs Up a Treasure Trove of Antimicrobial Peptides appeared first on GEN – Genetic Engineering and Biotechnology News.

Source