The Deep-Legion project deals with Legionnaires' disease, a particularly virulent form of pneumonia. The aim is to identify so-called virulence factors - properties that determine the pathogenic effect - of the human pathogenic bacterium Legionella pneumophila.

Legionella is a difficult bacteria to culture, underdiagnosed, and inherently resistant to routine antibiotic treatment. This is where Deep-Legion comes in: As part of the project, the virulence factors of the intracellular pathogen are to be determined using Deep Learning approaches based on existing molecular biological and clinical data. The aim is to predict their origin and function and then validate these predictions in vitro.

The plan is to use computer-assisted Deep Learning methods to carry out automated analyzes and comparisons of DNA and protein sequences in different Legionella strains and to correlate them with the clinical virulence potential of these strains. The virulence factors identified in this way are intended to provide new insights into the biological understanding of the individual Legionella strains. With this information, the current diagnostic and treatment options for patients with Legionella pneumonia could be significantly expanded and improved. In the event of successful routine application in clinical practice, it is also planned to apply the developed Deep Learning methods to other human pathogens.
 

Contact: Dr. rer. nat. Oluwafemi Abimbola Sarumi

Funding reference number: : 031L0288A and 031L0288C