AI & Data Science

The research of our team has a methodological as well as an application-oriented component, where method development is typically driven by specific questions arising in applications (for instance, in the pharma industry). Currently, these applications essentially cover

• better drug targets:
  • AI methods for drug target prioritization
  • AI methods for adverse event prediction
• precision medicine – the right drug for the right patient:
  • AI methods for prediction of disease risk, disease subtypes/strata, disease progression, treatment response
• better trials:
  • enrichment trials
  • synthetic controls
  • digital twins

To address the highly complex issues that emerge in our different applications, a broad range of AI and data science techniques is needed (covering various types of neural network architectures, Bayesian learning, Bayesian Networks, kernel methods, boosting, and others). At the same time, off-the-shelf solutions rarely provide satisfactory results. Hence, a significant proportion of our work goes into the adaptation, development, and design of AI and data science techniques that are tailored to solve a particular application problem. During the last years, our method developments have specifically covered

• hybrid AI/knowledge infusion: Methods that infuse human knowledge (knowledge graphs or differential equations) into data-driven machine learning models
• (generative) modeling of multivariate time series, including approaches to deal with missing values
• models that deal with multiple data modalities across biological scales.

We have a long-standing experience with different types of data:

• -omics
• longitudinal clinical studies
• various types of real-world data:
  • structured and unstructured electronic health record (EHR) and claims data
  • digital “biomarkers”: data coming from digital device technologies, e.g., gait sensors, smartphone applications

Our services cover the entire value chain in translational biomedical research in the biotech and pharmaceutical industry. We offer companies tailor-made solutions in AI and data mining, for example through contract research. This way we are able to address different needs of our customers and to add temporary resources and know-how to their internal projects.

• de Jong, J., Emon, M. A., Wu, P., Karki, R., Sood, M., Godard, P., ... & Fröhlich, H. (2019). Deep learning for clustering of multivariate clinical patient trajectories with missing values. Giga Science, 8(11), giz134.
  https://doi.org/10.1093/gigascience/giz134
• Khanna, S., Domingo-Fernández, D., Iyappan, A., Emon, M. A., Hofmann-Apitius, M., & Fröhlich, H. (2018). Using multi-scale genetic, Neuoimaging and clinical data for predicting Alzheimer’s disease and reconstruction of relevant biological mechanisms. Scientific reports, 8(1),11173.
  https://www.nature.com/articles/s41598-018-29433-3
  
• Philipp Wendland, Colin Birkenbihl, Marc Gomez-Freixa, Meemansa Sood, Maik Kschischo, Holger Fröhlich (2022). Generation of realistic synthetic data using Multimodal Neural Ordinary Differential Equations. npj Digital Medicine volume 5, Article number: 122.
  https://www.nature.com/articles/s41746-022-00666-x
  
• Johann de Jong, Ioana Cutcutache, Matthew Page, Sami Elmoufti, Cynthia Dilley, Holger Fröhlich, Martin Armstrong (2021). Towards realizing the vision of precision medicine: AI based prediction of clinical drug response. Brain, Volume 144, Issue 6.
  https://academic.oup.com/brain/article/144/6/1738/6178276
  
• Birkenbihl, C., Salimi, Y., Fröhlich, H. (2021). Unraveling the heterogeneity in Alzheimer's disease progression across multiple cohorts and the implications for data-driven disease modeling. Alzheimer's & Dementia 2022; 18: 251– 261. 
  https://doi.org/10.1002/alz.12387