interestingengineering.com
AI helps find nearly a million potential antibiotic sources in nature
Aman Tripathi
5–7 minutes
Machine learning reveals almost a million potential antibiotics hidden in the genetic material found in the natural world.
Updated: Jun 06, 2024 10:02 AM EST
Associate Professor Luis Pedro Coelho used AI to uncover a vast reservoir of natural antimicrobial peptides.
Queensland University of Technology
In a significant leap forward in the fight against antibiotic resistance, an international research team has harnessed the power of artificial intelligence (AI) to uncover almost a million new potential antibiotic sources in the natural world.
Led by Associate Professor Luis Pedro Coelho, a computational biologist from the Queensland University of Technology (QUT), the research team utilized machine learning to analyze over 60,000 metagenomes—vast collections of genetic material from diverse environments, including soil, oceans, and the human gut.
“We used datasets that were publicly available. They had generally been collected for other purposes,” Professor Coelho told
Interesting Engineering when asked about the collection and utilization of metagenomic data from diverse environments.
Validity of the findings
The innovative AI-powered approach allowed the team to identify a staggering 863,498 promising antimicrobial peptides, small molecules with the potential to kill or inhibit the growth of harmful bacteria.
To validate their findings, the researchers tested 100 laboratory-made
peptides against clinically significant pathogens.
“We tested some peptides that were particularly promising, but, importantly, we also tested some peptides chosen completely at random from our dataset (to provide an unbiased estimate of how good our dataset is),” Professor Coelho told
IE.
Effective against serious infections
The test results revealed that 79 peptides could disrupt bacterial membranes and 63 specifically targeted antibiotic-resistant bacteria like Staphylococcus aureus and Escherichia coli.
These results are particularly encouraging, as they suggest that these peptides could be effective against some of the most dangerous and difficult-to-treat infections.
Further preclinical trials in mice showed that two of the peptides were remarkably potent, reducing bacteria by up to four orders of
magnitude.
The effects were comparable to those of polymyxin B, a commonly used antibiotic, demonstrating the potential of these newly discovered peptides for therapeutic applications.
Better than traditional antibiotics
“What we are measuring is the decrease in bacterial load after topical application of the candidate peptides and comparing to a positive control (which is a clinically-used antibiotic),” stated Professor Coelho.
He highlighted the peptides’ potential for a much more targeted spectrum of activity.
“For conventional antibiotics, even ‘narrow spectrum’ antibiotics still target a large number of organisms, while peptides potentially be much more specific,” Coelho stated.
He also explained how these peptides are much better than traditional antibiotics.
“The fact that antibiotics disrupt the normal gut microbiome should be seen as a negative side effect of taking them. Thus, potentially reducing this side effect could be a large benefit,” asserted Professor Coelho while explaining the benefits of the peptides.
Creation of public database for new antibiotic discoveries
To facilitate further research and development, the team has created the
AMPSphere, a publicly accessible database containing all the identified peptides.
This valuable resource is expected to accelerate the discovery of new antibiotics and contribute to the ongoing battle against antimicrobial resistance.
“At the moment, we are continuing to explore the basic scientific aspects. What we’d really like to be able to do better is predict and potentially tune the specificity of the peptides,” Coelho concluded.
“We are also trying to develop approaches to be able to screen candidates faster.”
The researchers aim to save millions of lives
By harnessing the power of AI to explore the vast genetic diversity of the natural world, researchers have opened up a wealth of possibilities for developing life-saving drugs.
The findings of the study come with a renewed focus on combatting antimicrobial resistance (AMR) as humanity deals with the growing number of superbugs resistant to current drugs.
RECOMMENDED ARTICLES
AMR is a global health crisis, with the World Health Organization warning that it could lead to 10 million deaths annually by 2050 if left unchecked.
The development of new antibiotics is crucial to combat this threat, as many existing drugs are becoming increasingly ineffective against resistant
bacteria.
