CAIMed Group 1b: AI and Bioinformatics

To advance our understanding of diseases and their personalized treatment, the identification of genetic risk factors and their molecular signaling pathways, as well as the development of predictive models for disease progression and severity, are of the utmost importance. At MHH, existing and planned patient cohorts with state-of-the-art (single-cell) multi-omics data are available. Our junior research group focuses on the pre-processing (integration) of molecular data to generate standardized, high-quality datasets for the analyses conducted by other CAIMed junior research groups. The aim is to evaluate these data at scale using innovative AI-based methods. This includes identification of factors that correlate with disease severity and progression in order to predict individual responses to diseases and treatments, thereby establishing a molecular basis for the stratification of patient groups.

Currently, the group is engaged in the evaluation of datasets for the prediction of long-COVID, within the framework of the BMFTR-funded projects AID-PAIS and FEDCOV, which work on data from several German COVID-19 cohorts and the UK-Biobank. In addition, datasets from other chronic viral infections and associated fatigue syndromes (e.g., ME/CFS), such as those arising from chronic hepatitis C virus infections, are being integrated.

The overarching goal is to facilitate the translation of these mathematical models into clinical treatment, diagnostic, and predictive applications as a critical first step toward individualized prevention and therapy.

Head

Dr Mohamad Ballan, Dr Maximilian Schieck and Dr Sebastian Klein

Department head

Research focus

Personalized medicine is gaining increasing importance in medical research, as personalized treatment approaches tailored to the specific needs and genetic profiles of individual patients can lead to optimized outcomes. In this context, diverse datasets play a central role—ranging from genetic information to other multi-omics data and clinical data. These data come from various sources and often encompass different formats and structures, making their integration a significant challenge. Currently, there is often a lack of efficient methods to meaningfully combine and analyze these heterogeneous datasets.

However, this research group, which brings together experts from the fields of life sciences, computer science, AI, and medicine, provides an ideal foundation to address this complexity. Our goal is to examine the datasets for both their clinical and methodological rigor. AI methods will then be used to integrate the resulting datasets and make them available for larger AI algorithms and models for data processing (e.g. other groups within CAIMed like Group 1c).