Large-scale regulatory network analysis from microarray data: modified Bayesian network learning and association rule mining

Zan Huang, Jiexun Li, Hua Su, George S Watts, Hsinchun Chen

Research output: Contribution to journalArticle

37 Citations (Scopus)

Abstract

We present two algorithms for learning large-scale gene regulatory networks from microarray data: a modified information-theory-based Bayesian network algorithm and a modified association rule algorithm. Simulation-based evaluation using six datasets indicated that both algorithms outperformed their unmodified counterparts, especially when analyzing large numbers of genes. Both algorithms learned about 20% (50% if directionality and relation type were not considered) of the relations in the actual models. In our empirical evaluation based on two real datasets, domain experts evaluated subsets of learned relations with high confidence and identified 20-30% to be "interesting" or "maybe interesting" as potential experiment hypotheses.

Original languageEnglish (US)
Pages (from-to)1207-1225
Number of pages19
JournalDecision Support Systems
Volume43
Issue number4
DOIs
StatePublished - Aug 2007

Fingerprint

Association Learning
Association rules
Bayesian networks
Microarray Analysis
Microarrays
Electric network analysis
Genes
Information Theory
Gene Regulatory Networks
Information theory
Network learning
Microarray
Association rule mining
Network analysis
Network Analysis
Association Rules
Bayesian Networks
Learning

Keywords

  • Association rules
  • Bayesian networks
  • Genetic regulatory networks
  • Microarray

ASJC Scopus subject areas

  • Management Information Systems
  • Information Systems
  • Information Systems and Management

Cite this

Large-scale regulatory network analysis from microarray data : modified Bayesian network learning and association rule mining. / Huang, Zan; Li, Jiexun; Su, Hua; Watts, George S; Chen, Hsinchun.

In: Decision Support Systems, Vol. 43, No. 4, 08.2007, p. 1207-1225.

Research output: Contribution to journalArticle

@article{e78117e32a254f74b19375fd42d612e5,
title = "Large-scale regulatory network analysis from microarray data: modified Bayesian network learning and association rule mining",
abstract = "We present two algorithms for learning large-scale gene regulatory networks from microarray data: a modified information-theory-based Bayesian network algorithm and a modified association rule algorithm. Simulation-based evaluation using six datasets indicated that both algorithms outperformed their unmodified counterparts, especially when analyzing large numbers of genes. Both algorithms learned about 20{\%} (50{\%} if directionality and relation type were not considered) of the relations in the actual models. In our empirical evaluation based on two real datasets, domain experts evaluated subsets of learned relations with high confidence and identified 20-30{\%} to be {"}interesting{"} or {"}maybe interesting{"} as potential experiment hypotheses.",
keywords = "Association rules, Bayesian networks, Genetic regulatory networks, Microarray",
author = "Zan Huang and Jiexun Li and Hua Su and Watts, {George S} and Hsinchun Chen",
year = "2007",
month = "8",
doi = "10.1016/j.dss.2006.02.002",
language = "English (US)",
volume = "43",
pages = "1207--1225",
journal = "Decision Support Systems",
issn = "0167-9236",
publisher = "Elsevier",
number = "4",

}

TY - JOUR

T1 - Large-scale regulatory network analysis from microarray data

T2 - modified Bayesian network learning and association rule mining

AU - Huang, Zan

AU - Li, Jiexun

AU - Su, Hua

AU - Watts, George S

AU - Chen, Hsinchun

PY - 2007/8

Y1 - 2007/8

N2 - We present two algorithms for learning large-scale gene regulatory networks from microarray data: a modified information-theory-based Bayesian network algorithm and a modified association rule algorithm. Simulation-based evaluation using six datasets indicated that both algorithms outperformed their unmodified counterparts, especially when analyzing large numbers of genes. Both algorithms learned about 20% (50% if directionality and relation type were not considered) of the relations in the actual models. In our empirical evaluation based on two real datasets, domain experts evaluated subsets of learned relations with high confidence and identified 20-30% to be "interesting" or "maybe interesting" as potential experiment hypotheses.

AB - We present two algorithms for learning large-scale gene regulatory networks from microarray data: a modified information-theory-based Bayesian network algorithm and a modified association rule algorithm. Simulation-based evaluation using six datasets indicated that both algorithms outperformed their unmodified counterparts, especially when analyzing large numbers of genes. Both algorithms learned about 20% (50% if directionality and relation type were not considered) of the relations in the actual models. In our empirical evaluation based on two real datasets, domain experts evaluated subsets of learned relations with high confidence and identified 20-30% to be "interesting" or "maybe interesting" as potential experiment hypotheses.

KW - Association rules

KW - Bayesian networks

KW - Genetic regulatory networks

KW - Microarray

UR - http://www.scopus.com/inward/record.url?scp=34250013587&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34250013587&partnerID=8YFLogxK

U2 - 10.1016/j.dss.2006.02.002

DO - 10.1016/j.dss.2006.02.002

M3 - Article

AN - SCOPUS:34250013587

VL - 43

SP - 1207

EP - 1225

JO - Decision Support Systems

JF - Decision Support Systems

SN - 0167-9236

IS - 4

ER -