The goal of this project is to develop new methods that help biomedical scientists in the analysis of their data.
Technological advances in clinical measurement devices based on sequencing, imaging, and wearables promise to accurately diagnose diseases more accurate and in their earliest stages when they can be readily treated.
Machine learning is central to this vision of personalized medicine, where each individual is monitored based on their medical history, as well as their own genetic and environmental disease risk.
With the size and complexity of the data mounting, computational and statistical analysis is becoming a bottleneck.
We develop new methods from artificial intelligence and machine learning to analyze large volumes of complex data that are derived from clinical measurements and derive new scientific knowledge in an increasingly automated fashion and aid the medical experts in data interpretation.
In order to enable us to interpret the disease risk of an individual, we need to develop models that view the risk of an individual relative to a healthy population, and relative to individuals who have a disease.
This project not only focuses on methods development in machine learning (ML) and statistics to make this vision become a reality, but also aims on developing such models that ultimately will serve as an empirical footing for personalized preventative medicine.
Accurate statistical models as we plan to develop that take into account confounding are essential to determine robust associations and to derive precise risk models for diseases in the presence of environment, lifestyle, medication, and molecular measurements.