The second research cluster is particularly interested in the classification of observations in order to automate tedious and expensive tasks. The spread of some neurodegenerative diseases can for instance be identified based on images. At the moment, this classification is entirely performed by lab assistants. By manually training classification algorithms however, a new machine learning pipeline built on Support Vector Machines could automatically classify image data to support diagnostic treatment approaches.

• Cellular morphologies in neurodegeneration

Data from clinical studies lack the proper form to be directly included in computational frameworks. It is necessary to develop suitable pipelines for mining, selection and integration of a variety of data types, ranging from doctor’s comments, prescriptions, laboratory results and even genomic information. Text recognition techniques, semantic relation analysis and correlation detection applied to cohort data allows determination of risk factors for individual patients.

• Biomarker detection in Clinical Cohort Data using Machine Learning
• Gene regulatory networks and exploitation for cellular reprogramming/regeneration