🦠 AI discovers nearly 1 million new antibiotics – progress in the fight against antibiotic resistance

• Researchers used AI to analyze genetic data from tens of thousands of bacteria and other organisms.
• Nearly one million potential antibiotic compounds were identified, with 79 percent able to kill at least one microbe.
• AI has drastically accelerated the process of discovering new antibiotic candidates.

Algorithm finds new molecules

Researchers used an algorithm to scan “the entire microbial diversity on Earth” and found nearly one million new molecules, according to César de la Fuente, a professor at the University of Pennsylvania.

De la Fuente, who leads the Machine Biology Group, explains that the algorithm can sort through enormous amounts of information much faster than traditional methods like collecting water and soil samples. This is especially important because microbes are everywhere, from the ocean to the human gut.

The study is the largest antibiotic discovery effort to date.

The researchers gathered genomes and metagenomes from publicly available databases and searched for DNA sequences that could have antimicrobial activity. Of the molecules predicted to have antimicrobial activity, 100 were synthesized in the lab to test their effectiveness against bacteria.

Seventy-nine percent of the tested molecules could kill at least one microbe, indicating their potential as antibiotics.

Antimicrobial resistance, which includes antibiotic resistance, caused over 1.2 million deaths in 2019 and could increase to 10 million deaths annually by 2050, according to the World Health Organization.

Used AMPSphere, which is open to all

AMPSphere is an open and accessible resource that provides detailed information on the predicted antimicrobial peptides, including their sequences, original genes, and biochemical properties. By including data from various environments, AMPSphere offers insights into the peptides’ evolutionary origins and their specific adaptations to different habitats.

Only a small portion of the identified peptides have been found in existing databases, suggesting that AMPSphere contains many previously unknown sequences.

Faster results with AI

This is an example of how AI and machine learning can be used to sift through large amounts of data and find valuable results, something that would have taken humans a very long time to achieve.

We will see similar effects in many other fields.

4o