A rank-based statistical test for measuring synergistic effects between two gene sets

Bioinformatics. 2011 Sep 1;27(17):2399-405. doi: 10.1093/bioinformatics/btr382. Epub 2011 Jun 23.

Abstract

Motivation: Due to recent advances in high-throughput technologies, data on various types of genomic annotation have accumulated. These data will be crucially helpful for elucidating the combinatorial logic of transcription. Although several approaches have been proposed for inferring cooperativity among multiple factors, most approaches are haunted by the issues of normalization and threshold values.

Results: In this article, we propose a rank-based non-parametric statistical test for measuring the effects between two gene sets. This method is free from the issues of normalization and threshold value determination for gene expression values. Furthermore, we have proposed an efficient Markov chain Monte Carlo method for calculating an approximate significance value of synergy. We have applied this approach for detecting synergistic combinations of transcription factor binding motifs and histone modifications.

Availability: C implementation of the method is available from http://www.hgc.jp/~yshira/software/rankSynergy.zip.

Contact: [email protected]

Supplementary information: Supplementary data are available at Bioinformatics online.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Binding Sites
  • Breast Neoplasms / genetics
  • Cell Line, Tumor
  • Female
  • Gene Expression*
  • Genomics / methods
  • Histones / metabolism
  • Humans
  • Markov Chains*
  • Models, Statistical
  • Monte Carlo Method*
  • Nucleotide Motifs
  • Promoter Regions, Genetic
  • Statistics, Nonparametric
  • Transcription Factors / metabolism*

Substances

  • Histones
  • Transcription Factors