Determining semantically related significant genes

Research output: Contribution to journalArticlepeer-review

20 Scopus citations

Abstract

GO relation embodies some aspects of existence dependency. If GO term x is existence-dependent on GO term y, the presence of y implies the presence of x. Therefore, the genes annotated with the function of the GO term y are usually functionally and semantically related to the genes annotated with the function of the GO term x. A large number of gene set enrichment analysis methods have been developed in recent years for analyzing gene sets enrichment. However, most of these methods overlook the structural dependencies between GO terms in GO graph by not considering the concept of existence dependency. We propose in this paper a biological search engine called RSGSearch that identifies enriched sets of genes annotated with different functions using the concept of existence dependency. We observe that GO term x cannot be existence-dependent on GO term y, if x and y have the same specificity (biological characteristics). After encoding into a numeric format the contributions of GO terms annotating target genes to the semantics of their lowest common ancestors (LCAs), RSGSearch uses microarray experiment to identify the most significant LCA that annotates the result genes. We evaluated RSGSearch experimentally and compared it with five gene set enrichment systems. Results showed marked improvement.

Original languageBritish English
Article number6868276
Pages (from-to)1119-1130
Number of pages12
JournalIEEE/ACM Transactions on Computational Biology and Bioinformatics
Volume11
Issue number6
DOIs
StatePublished - 1 Nov 2014

Keywords

  • gene ontology
  • gene set enrichment analysis
  • related genes
  • semantic similarity
  • Semantically related genes

Fingerprint

Dive into the research topics of 'Determining semantically related significant genes'. Together they form a unique fingerprint.

Cite this