Addressing the Computational Challenges of Personalized Medicine

Summary

Personalized medicine (PM) is understood as a non-traditional medical approach, in which patients are stratified based on their disease subtype, disease risk, disease prognosis or treatment response using specialized diagnostic tests. High promises for the whole health care sector are associated with PM, and correspondingly the topic has received a lot of attention during the last years. PM is tightly connected to and dependent on computational sciences (computer science, mathematical modeling, computational statistics, bioinformatics). Currently, shortcomings of computational methodology constitute an important bottleneck for PM, which hinders full realization.

The goal of the planned seminar was to bring together an international and interdisciplinary group of experts in different computational science disciplines in order to discuss, how some of the major existing computational challenges could be better addressed in the future, namely:

1. how to enhance prediction performance of computational models for PM
2. how to improve their interpretability
3. how to validate and implement them in practice

The seminar joined together expertise that is usually scattered across different disciplines. The seminar had a strict focus on computational methodology, but few selected non-computational scientists closed the gap to the application field.