The joint project Deep-iAMR deals with the topic of antibiotic resistance (AMR). These are spreading more and more around the world and represent a global threat to the world population. With an ever-growing number of infectious diseases caused by microorganisms, many antibiotics developed so far have already lost their effectiveness. Effective prevention or treatment is no longer possible or only possible with great medical effort. It is therefore urgently necessary to develop new antibiotics or to identify new starting points for antibiotics through intensive molecular research - for example by elucidating resistance mechanisms.

This is where Deep-iAMR comes in: As part of the project, so-called artificial neural networks and deep learning concepts are to be optimized. In the future, they will be able to differentiate and classify antibiotic resistance mechanisms within newly sequenced bacterial genomes and identify potential new antibiotic targets.

It is planned to combine various omics datasets with clinical, phenotypic and genome-based information of well-characterized multidrug-resistant E.coli bacteria. This information is intended to be used to teach the artificial neural network, which - trained in this way - will then be able to recognize and analyze new resistance patterns based on similarities. The aim is to use the network to identify more and more complex connections in the development of antibiotic resistance in the future.

Publications: doi: 10.1016/j.csbj.2022.03.007; doi: 10.3390/antibiotics11111611; doi: 10.1093/bioinformatics/btab681 

Contact: Univ.-Prof. Dr. Dominik Heider

Funding reference number: :31L0209B