Novel technological solutions for rapid diagnostics of infection & AMR

Antibiotic and antimicrobial resistance (AMR) is a huge global health concern. It has been predicted that AMR will be the primary cause of 10 million yearly deaths worldwide by 2050.

01.06.2023–31.05.2025

HINN

Project goal

To use new technologies in combination with AI and ML to create a system that can detect and characterize resistant pathogens much faster ( < 5 hours rather than 2-4 days) than the current clinical methods. This would help to reduce the spread of antimicrobial resistance and can potentially save millions of human and animal lives in the long run.

About the project

Current methods for detecting resistant bacteria are time-consuming. This contributes to inappropriate use of antibiotics and the spread of AMR.

Project objectives

Use AI / ML combined with novel technologies to characterize pathogenic bacteria and their susceptibility profiles in less than 5 hours by:

Development of novel algorithms for taxonomic characterization of sequencing data
Investigating explainability of Deep Learning models for pathogen classification from quantitative phase microscopy (QPM) images
Novel AI / ML model development for prediction of minimum inhibitory concentration (MIC) of antibiotics from bacterial whole genome sequencing (WGS) data
Building models to investigate the national and regional burden of AMR

Figure 3 description of project goals as described in the text

Results

All research results in the project are available from the CRISTIN project page.

Project News

Preprint available: Real-time Taxonomic Characterization of Long-read Mixed-species Sequencing Samples in Sorted Motif Distance Space: Voyager” - downloadable at Cristin

Software

Voyager: Tool for Real-time Taxonomic Characterization of Long-read Mixed-species Sequencing Samples in Sorted Motif Distance Space
Language: C++ / Java
Repository Link