014

Title:

A novel similarity measure in gene co-expression network

Discipline: Systems Biology, Computer Science

Presenter:

Niloofar Aghaieabiane

Abstract:

Gene co-expression networks (GCNs) provide an effective way of modeling and studying complex biological systems. Working with GCNs involves two steps: (i) constructing the network based on measures of pairwise similarity between genes, (ii) identifying biologically meaningful modules of genes in the network (clustering). The quality of the identified modules is in principle dependent on the network construction step. Therefore, defining an appropriate measure for gene similarity has a paramount role in extracting significant biological knowledge from the networks. In this study, we propose a new method for constructing GCN networks which introduces to standard methodologies a pre-processing of the raw data, and a new measure of gene similarity based on a geometric rather than a statistical approach. When compared against the state-of-the-art methods, our approach produces highly more significant clusters of genes.

Author(s):

Niloofar Aghaieabiane, Ioannis Koutis

Funding Acknowledgements:

Computer Science Department