Model the relationship between gene expression and TFBSs using a simplified neural network with Bayesian variable selection

Xiaobo Zhou, Kuang Yu Liu, Guangqin Li, Stephen Wong

Research output: Contribution to journalConference article

Abstract

Although numerous computational methods consider the identification of individual transcription factor binding sites (TFBSs), very few focus on the interactions between these sites. In this study, we study the relationship between transcription factor binding sites and microarray gene expression data. A probit regression with one linear term plus nonlinear (it is actually a simplified neural network) is used to build a predictive model of outcome of interest (either gene expression ratios or up- and down-regulations) using these transcription factor binding sites. This issue is related to the more general problem of expression prediction in which we want to find small subsets of TFBSs to be used as predictors of possible co-expressed genes and those genes do share some DNA regulatory motifs. Given some maximum number of predictors to be used, a full search of all possible predictor sets is prohibitive. This paper considers Bayesian variable selection for prediction using the nonlinear probit model (or simplified neural network). We applied this nonlinear model with Bayesian motif selection on one gene expression data set. These TFs demonstrated intricate regulatory roles either as a family or as individual members and our analysis created plausible hypotheses for combinatorial interaction among TFBSs.

Original languageEnglish (US)
Pages (from-to)719-724
Number of pages6
JournalLecture Notes in Computer Science
Volume3498
Issue numberIII
DOIs
StatePublished - 2005
EventSecond International Symposium on Neural Networks: Advances in Neural Networks - ISNN 2005 - Chongqing, China
Duration: May 30 2005Jun 1 2005

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint Dive into the research topics of 'Model the relationship between gene expression and TFBSs using a simplified neural network with Bayesian variable selection'. Together they form a unique fingerprint.

Cite this