Evolving Hidden Markov Models for protein secondary structure prediction

Kyoung Jae Won; Thomas Hamelryck; Adam Prügel-Bennett; Anders Krogh

Evolving Hidden Markov Models for protein secondary structure prediction

Kyoung Jae Won^*, Thomas Hamelryck, Adam Prügel-Bennett, Anders Krogh

^*Corresponding author for this work

10 Citations (Scopus)

Abstract

New results are presented for the prediction of secondary structure information for protein sequences using Hidden Markov Models (HMMs) evolved using a Genetic Algorithm (GA). We achieved a Q ₃ measure of 75% using one of the most stringent data set ever used for protein secondary structure prediction. Our results beat the best hand-designed HMM currently available and are comparable to the best known techniques for this problem. A hybrid GA incorporating the Baum-Welch algorithm was used. The topology of the HMM was restricted to biologically meaningful building blocks. Mutation and crossover operators were designed to explore this space of topologies.

Original language	English
Title of host publication	2005 IEEE Congress on Evolutionary Computation, IEEE CEC 2005. Proceedings
Number of pages	8
Volume	1
Publication date	31 Oct 2005
Pages	33-40
ISBN (Print)	0780393635, 9780780393639
Publication status	Published - 31 Oct 2005
Event	2005 IEEE Congress on Evolutionary Computation, IEEE CEC 2005 - Edinburgh, Scotland, United Kingdom Duration: 2 Sept 2005 → 5 Sept 2005

Conference

Conference	2005 IEEE Congress on Evolutionary Computation, IEEE CEC 2005
Country/Territory	United Kingdom
City	Edinburgh, Scotland
Period	02/09/2005 → 05/09/2005

Cite this

@inproceedings{8963cbb28805441ba8bbf726e86b35cf,

title = "Evolving Hidden Markov Models for protein secondary structure prediction",

abstract = "New results are presented for the prediction of secondary structure information for protein sequences using Hidden Markov Models (HMMs) evolved using a Genetic Algorithm (GA). We achieved a Q 3 measure of 75% using one of the most stringent data set ever used for protein secondary structure prediction. Our results beat the best hand-designed HMM currently available and are comparable to the best known techniques for this problem. A hybrid GA incorporating the Baum-Welch algorithm was used. The topology of the HMM was restricted to biologically meaningful building blocks. Mutation and crossover operators were designed to explore this space of topologies.",

author = "Won, {Kyoung Jae} and Thomas Hamelryck and Adam Pr{\"u}gel-Bennett and Anders Krogh",

year = "2005",

month = oct,

day = "31",

language = "English",

isbn = "0780393635",

volume = "1",

pages = "33--40",

booktitle = "2005 IEEE Congress on Evolutionary Computation, IEEE CEC 2005. Proceedings",

note = "2005 IEEE Congress on Evolutionary Computation, IEEE CEC 2005 ; Conference date: 02-09-2005 Through 05-09-2005",

}

TY - GEN

T1 - Evolving Hidden Markov Models for protein secondary structure prediction

AU - Won, Kyoung Jae

AU - Hamelryck, Thomas

AU - Prügel-Bennett, Adam

AU - Krogh, Anders

PY - 2005/10/31

Y1 - 2005/10/31

N2 - New results are presented for the prediction of secondary structure information for protein sequences using Hidden Markov Models (HMMs) evolved using a Genetic Algorithm (GA). We achieved a Q 3 measure of 75% using one of the most stringent data set ever used for protein secondary structure prediction. Our results beat the best hand-designed HMM currently available and are comparable to the best known techniques for this problem. A hybrid GA incorporating the Baum-Welch algorithm was used. The topology of the HMM was restricted to biologically meaningful building blocks. Mutation and crossover operators were designed to explore this space of topologies.

AB - New results are presented for the prediction of secondary structure information for protein sequences using Hidden Markov Models (HMMs) evolved using a Genetic Algorithm (GA). We achieved a Q 3 measure of 75% using one of the most stringent data set ever used for protein secondary structure prediction. Our results beat the best hand-designed HMM currently available and are comparable to the best known techniques for this problem. A hybrid GA incorporating the Baum-Welch algorithm was used. The topology of the HMM was restricted to biologically meaningful building blocks. Mutation and crossover operators were designed to explore this space of topologies.

UR - http://www.scopus.com/inward/record.url?scp=27144535913&partnerID=8YFLogxK

M3 - Article in proceedings

AN - SCOPUS:27144535913

SN - 0780393635

SN - 9780780393639

VL - 1

SP - 33

EP - 40

BT - 2005 IEEE Congress on Evolutionary Computation, IEEE CEC 2005. Proceedings

T2 - 2005 IEEE Congress on Evolutionary Computation, IEEE CEC 2005

Y2 - 2 September 2005 through 5 September 2005

ER -

Evolving Hidden Markov Models for protein secondary structure prediction

Abstract

Conference

Other files and links

Fingerprint

Cite this