Mining protein-protein interaction networks: Denoising effects

Elisabetta Marras, Enrico Capobianco

Research output: Contribution to journalArticlepeer-review

2 Scopus citations


A typical instrument to pursue analysis in complex network studies is the analysis of the statistical distributions. They are usually computed for measures which characterize network topology, and are aimed at capturing both structural and dynamics aspects. Protein-protein interaction networks (PPIN) have also been studied through several measures. It is in general observed that a power law is expected to characterize scale-free networks. However, mixing the original noise cover with outlying information and other system-dependent fluctuations makes the empirical detection of the power law a difficult task. As a result the uncertainty level increases when looking at the observed sample; in particular, one may wonder whether the computed features may be sufficient to explain the interactome. We then address noise problems by implementing both decomposition and denoising techniques that reduce the impact of factors known to affect the accuracy of power law detection.

Original languageEnglish (US)
Article numberP01006
JournalJournal of Statistical Mechanics: Theory and Experiment
Issue number1
StatePublished - 2009
Externally publishedYes


  • Genomic and proteomic networks
  • Network dynamics

ASJC Scopus subject areas

  • Statistical and Nonlinear Physics
  • Statistics and Probability
  • Statistics, Probability and Uncertainty


Dive into the research topics of 'Mining protein-protein interaction networks: Denoising effects'. Together they form a unique fingerprint.

Cite this