The intrinsic geometric structure of protein-protein interaction networks for protein interaction prediction

Yi Fang, Mengtian Sun, Guoxian Dai, Karthik Ramani

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Recent developments in the high-throughput technologies for measuring protein-protein interaction (PPI) have profoundly advanced our ability to systematically infer protein function and regulation. To predict PPI in a net-work, we develop an intrinsic geometry structure (IGS) for the network, which exploits the intrinsic and hidden relationship among proteins in the network through a heat diffusion process. We apply our approach to publicly available PPI network data for the evaluation of the performance of PPI prediction. Experimental results indicate that, under different levels of the missing and spurious PPIs, IGS is able to robustly exploit the intrinsic and hidden relationship for PPI prediction with a higher sensitivity and specificity compared to that of recently proposed methods.

Original languageEnglish (US)
Title of host publicationIntelligent Computing in Bioinformatics - 10th International Conference, ICIC 2014, Proceedings
PublisherSpringer Verlag
Pages487-493
Number of pages7
ISBN (Print)9783319093291
DOIs
StatePublished - 2014
Event10th International Conference on Intelligent Computing, ICIC 2014 - Taiyuan, China
Duration: Aug 3 2014Aug 6 2014

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8590 LNBI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other10th International Conference on Intelligent Computing, ICIC 2014
CountryChina
CityTaiyuan
Period8/3/148/6/14

Keywords

  • Diffusion Geometry
  • PPI Network
  • protein function prediction

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint Dive into the research topics of 'The intrinsic geometric structure of protein-protein interaction networks for protein interaction prediction'. Together they form a unique fingerprint.

Cite this