Andrey Alexeyenko: Network enrichment analysis in biological research

Seminar, Analys: Especially Analysis on Networks

Using gene interaction networks that represent functional relations between biological macromolecules (mostly genes and proteins) has been a very popular approach in the modern biology. We will focus on building and using global interactomes, i.e. very large graphs that include thousands of nodes with high node degree values and have the following remarkable properties:

1. scale-free topology;
2. abundance of alternative network paths that might lead to the same response;
3. very high false positive and false negative rates when the topology was computationally reconstructed;
4. quick dissipation of information and impossibility of quantitative response modeling due to very high node degrees.

Our main focus is a methodological framework for the network enrichment analysis. It should test various research hypotheses while combining biological interpretability with rigorous statistics. I will also briefly present other areas of research: computational reconstruction of the global network using large-scale datasets, finding distinct structural units in this graph, and defining "dynamic", i.e. unstable domains of biological importance. The research field is relatively young, abounds with unsolved problems, and the presentation would be thus open to discussion and suggestions. On the other hand, our approaches might turn relevant in other contexts. Welcome!