AI-driven modeling for preventing post-acute infection syndromes
Leveraging artificial intelligence with mechanistic modeling and high-performance computing
Background
- Dependent on the pathogen, up to 30 % of infected persons suffer from post-acute infection syndromes (PAIS) for many months or years after infection. The mechanisms of their development are not well understood.1
- Previous infections or vaccinations can reduce the risk of developing PAIS:2
- Indirectly, by reducing the risk of infection and severe courses of the disease.
- Directly, by a higher reaction of the immune system.
1 Choutka, J., Jansari, V., Hornig, M. & Iwasaki, A. Unexplained post-acute infection syndromes. Nature Medicine 28, 911–923 (2022).
2 Mikolajczyk, R. et al. Likelihood of Post-COVID Condition in people with hybrid immunity; data from the German National Cohort (NAKO). Journal of Infection 89 (2024)
Goals
- Understand the key variables that are predictors of PAIS. Understand the impact of vaccinations on the risk of PAIS at the individual level.
- Find optimal vaccination strategies to reduce the number of cases and the severity of PAIS.
- Make efficient and scalable use of available resources.
Work Packages
WP1 – Individual immunity and lifestyle / comorbidities- based reaction model
The goal of WP1 is the development of an individual immunity and lifestyle/comorbiditiesbased reaction model integrating key variables derived from analysis on the NAKO and DigiHero data sets.
WP2 – AI-Simulation coupling: AI-based inference
The goal of WP2 is to provide the best-performing vaccination strategies to prevent the development of PAIS given the assumptions of the iteratively refined models. Furthermore, the second goal of WP2 is to provide freely accessible software and pipelines that can be reused by scientists and clinicians to infer estimations on parameters such as, e.g., antibody waning as soon as new data has been collected.
WP3 – Project coordination & Dissemination
Work package 3 is dedicated to project coordination and dissemination of results.
Project Members
Dr. Martin Kühn
Project Lead
Carlotta Gerstein
Project Member
Prof. Dr. Michael Meyer-Herrmann
Project Member
Dr. David Kerkmann
Project Member
Prof. Dr. Rafael Mikolajczyk
Project Member
Jonas Frost
Project Member
Direct external collaborators
Roy Gusinow
Publications
| Year | Title | Authors | Journal | Links |
|---|
Presentations
| Year | Title | Presenter | Venue | Links |
|---|---|---|---|---|
| 2025 | AIMS - AI-driven Modeling for preventing post-acute infection Syndromes | Martin Kühn | CompLS Statusseminar | |
| 2025 | AIMS - AI-driven Modeling for preventing post-acute infection Syndromes | Jonas Frost | CAIMed Meetup, Hannover |
Software
- MEmilio – The software framework MEmilio is used for modeling and simulation of infectious disease dynamics and currently extended for PAIS. Documentation Release v2.1.0
Contact
Dr. Martin KühnE-Mail: Martin.Kuehn@uni-bonn.dee
IRU Mathematics and Life Sciences, University of Bonn
Endenicher Allee 64
53115 Bonn
Germany
