Scored protein-protein interaction to predict subcellular localizations for yeast using diffusion kernel Conference

Mondal, AM, Hu, J. (2013). Scored protein-protein interaction to predict subcellular localizations for yeast using diffusion kernel . Lecture Notes in Computer Science, 8251 LNCS 647-655. 10.1007/978-3-642-45062-4_91

cited authors

  • Mondal, AM; Hu, J

abstract

  • Network-based protein localization prediction is explored utilizing the protein-protein interaction score along with the network connectivity. Score-based diffusion kernel is introduced to solve the problem. Four different PPI networks, namely, co-expressed PPI, Genetic PPI, Physical PPI, and scored PPI are used for analysis. Our investigation shows that PPI score does have positive impact in predicting subcellular protein localization. At high average PPI score of 891, performance accuracy ranges from 0.78 for 'punctate composite' to 0.93 for 'nucleolus' and at low average PPI score of 169, performance accuracy ranges from 0.60 for 'cytoplasm' to 0.83 for 'mitochondrion'. © Springer-Verlag 2013.

authors

publication date

  • December 1, 2013

published in

Digital Object Identifier (DOI)

start page

  • 647

end page

  • 655

volume

  • 8251 LNCS