# LegionProfiler: a computational tool for the identification of virulence factors and classification of Legionella pneumophila serogroup 1 isolates

## TL;DR

LegionProfiler is a tool that identifies virulence factors in Legionella pneumophila to classify isolates and improve pneumonia diagnosis and treatment.

## Contribution

LegionProfiler introduces an automated method for classifying Legionella isolates based on virulence factor domains.

## Key findings

- LegionProfiler identifies virulence factor protein domains in genome assemblies of Legionella isolates.
- The tool classifies isolates into high- or low-virulence groups based on their potential to cause severe pneumonia.
- LegionProfiler is available as a web service and can be run locally using Docker.

## Abstract

Legionella pneumophila has significantly contributed to multiple cases of pneumonia with a high rate of mortality globally. Its ability to exploit host mechanisms through several expressed virulence factors poses challenges for diagnosis, treatment, and outbreak control. To address this, we developed LegionProfiler, a computational tool that swiftly identifies virulence factor protein domains within genome assemblies of Legionella pneumophila serogroup 1 isolates and classifies them into high- or low-virulence groups. LegionProfiler automates the probing of genome assemblies for virulence-associated protein domains and determines the isolate’s potential to cause severe pneumonia infection. The LegionProfiler workflow is made available through a user-friendly interface to enhance technical control of infectious sources and adds important insights to the general epidemiology of clinical isolates. It could also support the development of targeted therapeutic strategies that will improve patient treatment.

LegionProfiler is freely accessible as a web service at https://legionprofiler.uni-muenster.de, and can also be run locally in a Docker container. The source code can be found at https://imigitlab.uni-muenster.de/heiderlab/legionprofiler or at Zenodo (DOI: 10.5281/zenodo.15592325).