Leiden Early Drug Discovery & Development

Finding and valorizing new antibiotics using AI

Antibiotics are a class of medicine most people take for granted. But pathogenic bacteria are becoming more and more resistant to our antibiotics, and this poses a great challenge for future treatments. There is thus a great societal need to identify new molecules that can address new targets and be developed into drugs that fight multidrug-resistant microbes.

Bacteria are an untapped source of antibiotic molecules

Genome sequencing of hundreds of thousands of bacteria has revealed that they boast an enormous untapped capacity for making yet-unknown natural products, which may be a rich source of antibiotics. However, many molecules that are identified are closely related to ones that are already known, and it has been difficult to identify pathways towards novel classes of such molecules. One particular promising source of antimicrobial molecules is constituted by peptides.

LED3 algorithms discover genetic patterns

As there is little financial incentive for pharmaceutical companies to develop new antibiotics, the public sector has a major role in this, and we at LED3 have stepped up. We recently achieved a breakthrough in using artificial intelligence to recognize generic patterns associated with genes encoding the production of these peptides, thus allowing the systematic discovery of dozens of new classes of peptide natural products, which constitute a great potential source of new antibiotics. 

The value of interdisciplinary drug discovery

One of these molecules, pristinin, was structurally characterized and found to be a member of a new class. This highlights the power of the combination of state-of-the-art computational genome mining, metabolomics and genetic techniques available within LED3. In the near future, we will use this technology to discover additional classes of molecules, and we will extend this to include new machine learning technology.