User Tools

Site Tools


Generating co-expression networks based on MEM data

Category:Data Analysis
Supervisor:Raivo Kolde
Abstract:The goal is to find a good way to generate a network using MEM data and possibly MEM itself. The evaluation of the network generation method can be done studying various network properties, sub- networks and their annotations as given by Graphweb and overlaps with other known biological net- works.


  • Study MEM data repository and MEM command line interface to learn how to make produce networks.
  • Try different similarity metrics, ways to combine information and similarity thresholds to generate the networks.
  • Characterize obtained networks by standard network properties, like node degree distribution, …
  • Compare networks in Graphweb, to see how functional annotation of networks depends on the parameters used when building it
  • Calculate the overlap between the obtained and available PPI networks.


  • Adler et al. Mining for coexpression across hundreds of datasets using novel rank aggregation and visualization methods. Genome Biol. (2009) vol.10 (12) pp. R139
  • Reimand et al. GraphWeb: mining heterogeneous biological networks for gene modules with functional significance. Nucleic Acids Res. (2008) pp.
topics/bib/kolde2011a.txt · Last modified: 2011/10/19 10:31 by swen