Goldammer/Wolfien Lab: Data Science & AI
Vision
"In order to answer medical questions and gain a better understanding of underlying disease mechanisms, we apply and continuously develop computer-aided, AI-based approaches for data analysis."
The application of computational approaches in life sciences and medicine is becoming increasingly important. The underlying data is diverse, ranging from routine clinical and laboratory data to high-resolution biosignals and RNA sequencing data. For data analysis, the data must be harmonised and adapted to standardised data formats (e.g. OMOP). The development of modularised and automated workflows is therefore a key factor. Various tools and methods are used, ranging from classical statistical methods and network analysis to deep learning and large language models. In addition, the methods developed are applied and validated in interdisciplinary collaborations in the fields of oncology, rare diseases, inflammation, cardiology, sleep medicine and others. The work of the research area thus considerably facilitates the application of automated approaches in the clinical environment and therefore contributes significantly to the goals of the Medical Informatics Initiative by supporting improved diagnosis, prevention and therapy.
Research areas
Central research topics of the lab are:
AI & Machine Learning in Medicine
- Development of models for classification, detection and clustering of patient data
- Analysis of Common Data Models for parameter prioritization
- Generation of synthetic data based on tabular patient information
- Explainability of models (XAI)
Development of Data Analysis Workflows
- Analysis of bulk, single-cell and spatially-resolved gene expression data
- Harmonization and analysis of biosignals
- Processing and integration of multi-omics data using network modeling
- Visualization of clinical data
- Configuration and automation of machine learning processes
Selected projects
- ACRIBIS — Personalized risk assessment for cardiovascular diseases
- GeMTeX — Automated indexing of medical texts for research
- MiHUBx — Medical Informatics Hub in Saxony, digital workflow integration in personalized oncology (work package 6)
- PM4Onco — Personalized Medicine for Oncology
- SATURN — Smart physician portal for patients with unclear diseases
-
Somnolink — Connected sleep data and decision support along the patient path for better care of Obstructive Sleep Apnea
Courses Offered (every semester)
- Complex Practical Course "Medical Informatics I": Medical Decision Support Systems in the Master's/Diploma program in Computer Science at the Technical University of Dresden
- Elective Course " Competence Training in Medical Data Science" for medical students at the Medical Faculty of TU Dresden
- Leadership of seminar groups in the subject "Cross-sectional Area 1: Epidemiology, Medical Biometry and Medical Informatics" for medical students and staff of the Medical Faculty of TU Dresden
Group leader
© MF/Stephan Wiegand
Research fellow
NameMs Dr.-Ing. Miriam Goldammer
© MF/Stephan Wiegand
Research fellow
NameMr Dr.-Ing. Markus Wolfien