Researchers have counted nearly a million potential sources of antibiotics in nature. 

An international research team, which includes Associate Professor Luis Pedro Coelho, a computational biologist at QUT, used machine learning to identify 863,498 promising antimicrobial peptides (AMPs). 

These small molecules can kill or inhibit the growth of infectious microbes.

The discovery comes at a crucial time, as the global community faces an escalating threat from antimicrobial resistance (AMR). 

Current antibiotics are increasingly ineffective against superbugs, making new methods for antibiotic discovery essential. 

“There is an urgent need for new methods for antibiotic discovery,” says Professor Coelho, a researcher at the QUT Centre for Microbiome Research. 

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

Without intervention, experts estimate that AMR could result in up to 10 million deaths annually 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,” Professor Coelho said.

The team validated the machine learning predictions by synthesising 100 peptides in the lab and testing them against clinically significant pathogens. 

Of these, 79 were found to disrupt bacterial membranes, and 63 specifically targeted antibiotic-resistant bacteria such as golden staph and E. coli. 

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

In preclinical models, these peptides showed results comparable to polymyxin B, a commercially available antibiotic used to treat various serious infections including meningitis, pneumonia, sepsis, and urinary tract infections.

The research involved analysing over 60,000 metagenomes, encompassing the genetic material of more than one million organisms from diverse environments such as marine and soil habitats, as well as human and animal guts. 

The researchers used the data to create AMPSphere, a comprehensive and publicly accessible database of novel peptides for new antibiotic discovery.