A study on how AI and computer algorithms can help discover new antibiotics faster and more effectively was nominated Best Tech Idea of 2021 by the Dutch science magazine KIJK. Assistant Professor Marnix Medema carries out this project in collaboration with Leiden University. Anyone who is interested can vote for the Best Tech Idea of 2021 on KIJK’s website.
“I see this nomination as crucial support in our search for new antibiotics,” says Marnix Medema, Assistant Professor in Bioinformatics at Wageningen University & Research (WUR). Increasingly more bacterial species are growing resistant to known antibiotics. This creates a real need for new antibiotics, but the classical search method – based on cultures of bacteria or fungi that can potentially produce antibiotics – is not only very time-consuming but is also increasingly failing to deliver new antibiotics. Furthermore, bacteria in such cultures do not produce all the antibiotics they are genetically capable of producing. Smart software that can in one go scan all of a bacterium’s genes for their potential as antibiotics, offers a solution.
For this study, Medema is collaborating with Leiden University. Professor of Molecular Biotechnology Gilles van Wezel from the Institute of Biology Leiden (IBL) was already working on discovering new antibiotics in his laboratory. Collaboration with Medema’s research group in Wageningen was therefore a logical step; the Bioinformatics research group has much expertise to offer when it comes to building software that can scan through bacterial DNA, for example clusters of genes known for producing molecules with antibiotic effects.
“This can lead to the development of new medications,” says Medema. “We’ve created an algorithm for finding new routes for producing such molecules, which we’ve so far applied to one bacterial group. This has already led to the identification of 42 new types of gene clusters that can potentially lead to the discovery of new antibiotics. Isolating and characterising these molecules is still a lot of work, but we’ve already studied one molecule in detail.”
Follow-up research will continue to take place in collaboration with Leiden University, where Medema was appointed Visiting Professor in Theoretical Biology in May 2020. Together with Van Wezel, he supervised PhD student Alexander Kloosterman in carrying out the research. Last year, a new PhD student joined the project.
Medema has worked on developing algorithms for the large-scale acquisition of genetic information from plants, bacteria and fungi since 2015. For the purpose of developing new medicines such as antibiotics, but also to improve our understanding of the ecological interactions between these organisms. “Seven years ago, I developed the algorithm that formed the basis for this new method with my colleagues in the US.”
“This is super cool research”, he continues. “Thanks to Artificial Intelligence, we’ve discovered a completely new class of molecules.” And the research is of great importance for society. “Pharmaceutical companies are less interested in developing new antibiotics because of the high investments required and because doctors refrain as much as possible from using antibiotics in an attempt to prevent resistance from developing. This is why academic research plays such a crucial role in this context.”
With a view to its societal significance, the WUR and Leiden University study is made available as open science. Sharing all research data allows other researchers to also apply this search method to find new antibiotics. “And there are other possibilities too. Our method can also be used to develop nutritional supplements or cancer drugs, or to decipher the role of microbial molecules in nature, for example in protection against plant diseases.”
Vote for the best idea
The KIJK longlist for Best Tech Idea of 2021 includes twenty candidates. You can vote for one of the ideas until 5 October.