The aim of this project is to create (hierarchical) association networks. Methodology will be developed for two different types of data. Construction of association networks from dynamic data (using the time series measurements to find associations) is one objective. The second objective focuses on identification of changes (for instance due to treatment) in association networks, both for networks based on static data and dynamic time series data.
In the first stage of this project most efforts have been put into the analysis of lipidomics data from a study that aimed to monitor treatment effects from dynamic time series.
A first exploratory analysis of the data resulted in the identification of high-dimensional effects of treatment and visualization of important time trajectories present in the lipidomics profiles. A manuscript for publication of this exploratory analysis of this study is almost finished. This first exploratory step was necessary to understand the difficulties imposed by the experimental design and evaluate important treatment and time effects. The next step will cover the inference of an association network, based on time series data, identifying modules (clusters) and the relations between them.