Applied bioinformatics and artificial intelligence

Getting the most out of big data for biomedical translation

Hands typing on a laptop. Overlaid are digital image elements such as binary code and abstract graphics.
© tippapatt, adobe.stock.com

The availability of large amounts of data has revolutionized research in the life sciences in the past few years, offering a wide range of opportunities for knowledge gain and future applications. By combining the disciplines of mathematics, computer science, medicine and life sciences, bioinformatics has made it possible to store, categorize, analyze, evaluate and visualize biological data and to simulate biochemical processes. 

Application areas at Fraunhofer ITEM

Left image column: microscopic representation of airways opened to different degrees. Right image column: graphical representation of the same airways but as computed models.
© Fraunhofer ITEM
Airway recognition in microscopy data by means of neural networks. Based on a microscopic image of an airway (input image, left column), neural networks allow prediction of the airway lumen (right column).

At Fraunhofer ITEM, researchers develop methods and possibilities for the preparation, analysis and visualization of biomedical data, as well as data models and data analysis pipelines. The focus of our research is on the mapping of cellular and regulatory processes and their translation into applications for humans. Bioinformatics methods are used, for example, for personalized tumor therapy to develop optimized testing strategies and for research on RNAs as diagnostic biomarkers and therapeutic targets. For personalized therapies or for patient stratification, the knowledge gained from big data is key to identifying adequate treatment strategies. Stratification also plays a major role for hazard and risk assessment of chemicals, nanomaterials, and environmental exposure, as the sensitivity to noxious agents differs between subpopulations.

In addition, the Fraunhofer researchers are using bioinformatics and artificial intelligence to advance towards intelligent image data analysis and are further developing this technology, so as to optimize the analysis of histological images and support clinical processes. 

Bioinformatics: recent projects and highlights

 

Smart health: project "PrivacyUmbrella"

Fraunhofer ITEM researchers have developed a multi-task federated learning software as part of the PrivacyUmbrella project, implementing knowledge distillation.

 

Threshold of Toxicological Concern

The TTC concept: Some carcinogenic substances do not directly damage DNA, raising the question as to whether such carcinogens should be subject to other regulatory limits.

DigitaLung: Web app analyzes lung sounds

 

Project »Genealyzer«

Web application for the analysis and comparison of gene expression data.

Project MyDeepLearn

End-to-end web application for image analysis using neural networks.

Project archive

Here you can find more projects sorted by our research and development competences.