A Hidden Markov Model approach to predicting yeast gene function from sequential gene expression data

Xutao Deng, Huimin Geng, Hesham H Ali

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Existing data mining tools can only achieve about 40% precision in function prediction of unannotated genes. We developed a gene function prediction tool based on profile Hidden Markov Models (HMMs). Each function class was modelled using a distinct HMM whose parameters were trained using yeast time-series gene expression profiles. Two structural variants of HMMs were designed and tested, each of them on 40 function classes. The highest overall prediction precision achieved was 67% using double-split HMM with leave-one-out cross-validation. We also attempted to generalise HMMs to dynamic Bayesian networks for gene function prediction using heterogeneous data sets.

Original languageEnglish (US)
Pages (from-to)263-273
Number of pages11
JournalInternational Journal of Bioinformatics Research and Applications
Volume4
Issue number3
DOIs
StatePublished - Jul 1 2008

Fingerprint

Hidden Markov models
Gene expression
Yeast
Genes
Yeasts
Gene Expression
Data Mining
Gene Regulatory Networks
Transcriptome
Bayesian networks
Data mining
Time series
Datasets

Keywords

  • Bioinformatics
  • Function prediction
  • Gene expression
  • HMM
  • Hidden Markov Model

ASJC Scopus subject areas

  • Biomedical Engineering
  • Health Informatics
  • Clinical Biochemistry
  • Health Information Management

Cite this

A Hidden Markov Model approach to predicting yeast gene function from sequential gene expression data. / Deng, Xutao; Geng, Huimin; Ali, Hesham H.

In: International Journal of Bioinformatics Research and Applications, Vol. 4, No. 3, 01.07.2008, p. 263-273.

Research output: Contribution to journalArticle

@article{b2940d3e18174bb3a44b2492492f3ae2,
title = "A Hidden Markov Model approach to predicting yeast gene function from sequential gene expression data",
abstract = "Existing data mining tools can only achieve about 40{\%} precision in function prediction of unannotated genes. We developed a gene function prediction tool based on profile Hidden Markov Models (HMMs). Each function class was modelled using a distinct HMM whose parameters were trained using yeast time-series gene expression profiles. Two structural variants of HMMs were designed and tested, each of them on 40 function classes. The highest overall prediction precision achieved was 67{\%} using double-split HMM with leave-one-out cross-validation. We also attempted to generalise HMMs to dynamic Bayesian networks for gene function prediction using heterogeneous data sets.",
keywords = "Bioinformatics, Function prediction, Gene expression, HMM, Hidden Markov Model",
author = "Xutao Deng and Huimin Geng and Ali, {Hesham H}",
year = "2008",
month = "7",
day = "1",
doi = "10.1504/IJBRA.2008.019574",
language = "English (US)",
volume = "4",
pages = "263--273",
journal = "International Journal of Bioinformatics Research and Applications",
issn = "1744-5485",
publisher = "Inderscience Enterprises Ltd",
number = "3",

}

TY - JOUR

T1 - A Hidden Markov Model approach to predicting yeast gene function from sequential gene expression data

AU - Deng, Xutao

AU - Geng, Huimin

AU - Ali, Hesham H

PY - 2008/7/1

Y1 - 2008/7/1

N2 - Existing data mining tools can only achieve about 40% precision in function prediction of unannotated genes. We developed a gene function prediction tool based on profile Hidden Markov Models (HMMs). Each function class was modelled using a distinct HMM whose parameters were trained using yeast time-series gene expression profiles. Two structural variants of HMMs were designed and tested, each of them on 40 function classes. The highest overall prediction precision achieved was 67% using double-split HMM with leave-one-out cross-validation. We also attempted to generalise HMMs to dynamic Bayesian networks for gene function prediction using heterogeneous data sets.

AB - Existing data mining tools can only achieve about 40% precision in function prediction of unannotated genes. We developed a gene function prediction tool based on profile Hidden Markov Models (HMMs). Each function class was modelled using a distinct HMM whose parameters were trained using yeast time-series gene expression profiles. Two structural variants of HMMs were designed and tested, each of them on 40 function classes. The highest overall prediction precision achieved was 67% using double-split HMM with leave-one-out cross-validation. We also attempted to generalise HMMs to dynamic Bayesian networks for gene function prediction using heterogeneous data sets.

KW - Bioinformatics

KW - Function prediction

KW - Gene expression

KW - HMM

KW - Hidden Markov Model

UR - http://www.scopus.com/inward/record.url?scp=47749103312&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=47749103312&partnerID=8YFLogxK

U2 - 10.1504/IJBRA.2008.019574

DO - 10.1504/IJBRA.2008.019574

M3 - Article

VL - 4

SP - 263

EP - 273

JO - International Journal of Bioinformatics Research and Applications

JF - International Journal of Bioinformatics Research and Applications

SN - 1744-5485

IS - 3

ER -