Fuzzy systems modeling for protein-protein interaction prediction in Saccharomyces cerevisie

Sakhinah Abu Bakar, Javid Taheri, Albert Y. Zomaya

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Most of the biological functions are mediated by protein-protein interactions in the organism. If one of these interactions behaves improperly, it may lead to a disease. Therefore, the study of protein-protein interactions is very important to improve our understanding of diseases and can provide the basis for new therapeutic approaches. Although, there are no concrete properties in predicting protein-protein interactions, it is known from experimentally determined protein-protein interactions that interacting proteins have a high probability to share similar functions, cellular roles and sub-cellular locations. If two proteins have similar functions, they will theoretically share similar three-dimensional structures as well. Therefore, it is believed that if two proteins have similar secondary structures, they will also have similar three-dimensional structures and consequently share similar functions. As a result they will interact with each other. However, if these proteins have similar sequence, they do not always have similar secondary structures and consequently similar three-dimensional structures and functions. Based on these theories, we predict the interacting proteins in Saccharomyces cerevisie (baker's yeast) from the information of their secondary structures using computational method. This paper proposes multiple independent fuzzy systems for predicting protein-protein interactions from the similarity of protein secondary structures. Our method consists of two main stages: (1) similarity score computation, and (2) similarity classification. The first stage involves three steps: (1) Multiple-sequence alignment (MSA) - finding multiple-sequence alignment for every family groups of proteins in Saccharomyces cerevisie, (2) Secondary structure prediction (SSP) - predicting secondary structure of aligned proteins sequence using secondary structure prediction tool called SSpro, and (3) Similarity measurement (Sim) - computing similarity scores of predicted secondary structures for every possible proteins pairs based on the number of three conformational states: helix (H), sheet (E), and coil (C). In the classification stage, N multiple independent first order Sugeno Fuzzy Systems are generated to model the behavior of similarity scores of all possible proteins pairs to classify the interacting and non-interacting pairs; here N is the number of protein. Every system determines initial rules based on the clusters information obtained from the fuzzy clustering method. We employ principal component analysis in every system to compress the dimension of input data. Our model has been trained and tested using 1029 proteins with already known 2965 positive interactions of Saccharomyces cerevisie (baker's yeast). This proposed model achieves good accuracy when compared with experimentally determined proteins interactions from the Database of Interacting Proteins.

Original languageEnglish
Title of host publication18th World IMACS Congress and MODSIM09 International Congress on Modelling and Simulation: Interfacing Modelling and Simulation with Mathematical and Computational Sciences, Proceedings
Pages782-788
Number of pages7
Publication statusPublished - 2009
Externally publishedYes
Event18th World IMACS Congress and International Congress on Modelling and Simulation: Interfacing Modelling and Simulation with Mathematical and Computational Sciences, MODSIM09 - Cairns, QLD
Duration: 13 Jul 200917 Jul 2009

Other

Other18th World IMACS Congress and International Congress on Modelling and Simulation: Interfacing Modelling and Simulation with Mathematical and Computational Sciences, MODSIM09
CityCairns, QLD
Period13/7/0917/7/09

Fingerprint

Fuzzy Modeling
Protein-protein Interaction
Fuzzy systems
System Modeling
Fuzzy Systems
Secondary Structure
Proteins
Protein
Prediction
Multiple Sequence Alignment
Structure Prediction
Yeast
Three-dimensional
Interaction
Fuzzy Clustering
Protein Structure
Protein Sequence
Helix
Coil
Clustering Methods

Keywords

  • Fuzzy system modeling
  • Protein-protein interaction prediction
  • Secondary structures

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Computational Mathematics
  • Modelling and Simulation

Cite this

Abu Bakar, S., Taheri, J., & Zomaya, A. Y. (2009). Fuzzy systems modeling for protein-protein interaction prediction in Saccharomyces cerevisie. In 18th World IMACS Congress and MODSIM09 International Congress on Modelling and Simulation: Interfacing Modelling and Simulation with Mathematical and Computational Sciences, Proceedings (pp. 782-788)

Fuzzy systems modeling for protein-protein interaction prediction in Saccharomyces cerevisie. / Abu Bakar, Sakhinah; Taheri, Javid; Zomaya, Albert Y.

18th World IMACS Congress and MODSIM09 International Congress on Modelling and Simulation: Interfacing Modelling and Simulation with Mathematical and Computational Sciences, Proceedings. 2009. p. 782-788.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abu Bakar, S, Taheri, J & Zomaya, AY 2009, Fuzzy systems modeling for protein-protein interaction prediction in Saccharomyces cerevisie. in 18th World IMACS Congress and MODSIM09 International Congress on Modelling and Simulation: Interfacing Modelling and Simulation with Mathematical and Computational Sciences, Proceedings. pp. 782-788, 18th World IMACS Congress and International Congress on Modelling and Simulation: Interfacing Modelling and Simulation with Mathematical and Computational Sciences, MODSIM09, Cairns, QLD, 13/7/09.
Abu Bakar S, Taheri J, Zomaya AY. Fuzzy systems modeling for protein-protein interaction prediction in Saccharomyces cerevisie. In 18th World IMACS Congress and MODSIM09 International Congress on Modelling and Simulation: Interfacing Modelling and Simulation with Mathematical and Computational Sciences, Proceedings. 2009. p. 782-788
Abu Bakar, Sakhinah ; Taheri, Javid ; Zomaya, Albert Y. / Fuzzy systems modeling for protein-protein interaction prediction in Saccharomyces cerevisie. 18th World IMACS Congress and MODSIM09 International Congress on Modelling and Simulation: Interfacing Modelling and Simulation with Mathematical and Computational Sciences, Proceedings. 2009. pp. 782-788
@inproceedings{343d635ddec64afebaaf0f2b4c7a4e4f,
title = "Fuzzy systems modeling for protein-protein interaction prediction in Saccharomyces cerevisie",
abstract = "Most of the biological functions are mediated by protein-protein interactions in the organism. If one of these interactions behaves improperly, it may lead to a disease. Therefore, the study of protein-protein interactions is very important to improve our understanding of diseases and can provide the basis for new therapeutic approaches. Although, there are no concrete properties in predicting protein-protein interactions, it is known from experimentally determined protein-protein interactions that interacting proteins have a high probability to share similar functions, cellular roles and sub-cellular locations. If two proteins have similar functions, they will theoretically share similar three-dimensional structures as well. Therefore, it is believed that if two proteins have similar secondary structures, they will also have similar three-dimensional structures and consequently share similar functions. As a result they will interact with each other. However, if these proteins have similar sequence, they do not always have similar secondary structures and consequently similar three-dimensional structures and functions. Based on these theories, we predict the interacting proteins in Saccharomyces cerevisie (baker's yeast) from the information of their secondary structures using computational method. This paper proposes multiple independent fuzzy systems for predicting protein-protein interactions from the similarity of protein secondary structures. Our method consists of two main stages: (1) similarity score computation, and (2) similarity classification. The first stage involves three steps: (1) Multiple-sequence alignment (MSA) - finding multiple-sequence alignment for every family groups of proteins in Saccharomyces cerevisie, (2) Secondary structure prediction (SSP) - predicting secondary structure of aligned proteins sequence using secondary structure prediction tool called SSpro, and (3) Similarity measurement (Sim) - computing similarity scores of predicted secondary structures for every possible proteins pairs based on the number of three conformational states: helix (H), sheet (E), and coil (C). In the classification stage, N multiple independent first order Sugeno Fuzzy Systems are generated to model the behavior of similarity scores of all possible proteins pairs to classify the interacting and non-interacting pairs; here N is the number of protein. Every system determines initial rules based on the clusters information obtained from the fuzzy clustering method. We employ principal component analysis in every system to compress the dimension of input data. Our model has been trained and tested using 1029 proteins with already known 2965 positive interactions of Saccharomyces cerevisie (baker's yeast). This proposed model achieves good accuracy when compared with experimentally determined proteins interactions from the Database of Interacting Proteins.",
keywords = "Fuzzy system modeling, Protein-protein interaction prediction, Secondary structures",
author = "{Abu Bakar}, Sakhinah and Javid Taheri and Zomaya, {Albert Y.}",
year = "2009",
language = "English",
isbn = "9780975840078",
pages = "782--788",
booktitle = "18th World IMACS Congress and MODSIM09 International Congress on Modelling and Simulation: Interfacing Modelling and Simulation with Mathematical and Computational Sciences, Proceedings",

}

TY - GEN

T1 - Fuzzy systems modeling for protein-protein interaction prediction in Saccharomyces cerevisie

AU - Abu Bakar, Sakhinah

AU - Taheri, Javid

AU - Zomaya, Albert Y.

PY - 2009

Y1 - 2009

N2 - Most of the biological functions are mediated by protein-protein interactions in the organism. If one of these interactions behaves improperly, it may lead to a disease. Therefore, the study of protein-protein interactions is very important to improve our understanding of diseases and can provide the basis for new therapeutic approaches. Although, there are no concrete properties in predicting protein-protein interactions, it is known from experimentally determined protein-protein interactions that interacting proteins have a high probability to share similar functions, cellular roles and sub-cellular locations. If two proteins have similar functions, they will theoretically share similar three-dimensional structures as well. Therefore, it is believed that if two proteins have similar secondary structures, they will also have similar three-dimensional structures and consequently share similar functions. As a result they will interact with each other. However, if these proteins have similar sequence, they do not always have similar secondary structures and consequently similar three-dimensional structures and functions. Based on these theories, we predict the interacting proteins in Saccharomyces cerevisie (baker's yeast) from the information of their secondary structures using computational method. This paper proposes multiple independent fuzzy systems for predicting protein-protein interactions from the similarity of protein secondary structures. Our method consists of two main stages: (1) similarity score computation, and (2) similarity classification. The first stage involves three steps: (1) Multiple-sequence alignment (MSA) - finding multiple-sequence alignment for every family groups of proteins in Saccharomyces cerevisie, (2) Secondary structure prediction (SSP) - predicting secondary structure of aligned proteins sequence using secondary structure prediction tool called SSpro, and (3) Similarity measurement (Sim) - computing similarity scores of predicted secondary structures for every possible proteins pairs based on the number of three conformational states: helix (H), sheet (E), and coil (C). In the classification stage, N multiple independent first order Sugeno Fuzzy Systems are generated to model the behavior of similarity scores of all possible proteins pairs to classify the interacting and non-interacting pairs; here N is the number of protein. Every system determines initial rules based on the clusters information obtained from the fuzzy clustering method. We employ principal component analysis in every system to compress the dimension of input data. Our model has been trained and tested using 1029 proteins with already known 2965 positive interactions of Saccharomyces cerevisie (baker's yeast). This proposed model achieves good accuracy when compared with experimentally determined proteins interactions from the Database of Interacting Proteins.

AB - Most of the biological functions are mediated by protein-protein interactions in the organism. If one of these interactions behaves improperly, it may lead to a disease. Therefore, the study of protein-protein interactions is very important to improve our understanding of diseases and can provide the basis for new therapeutic approaches. Although, there are no concrete properties in predicting protein-protein interactions, it is known from experimentally determined protein-protein interactions that interacting proteins have a high probability to share similar functions, cellular roles and sub-cellular locations. If two proteins have similar functions, they will theoretically share similar three-dimensional structures as well. Therefore, it is believed that if two proteins have similar secondary structures, they will also have similar three-dimensional structures and consequently share similar functions. As a result they will interact with each other. However, if these proteins have similar sequence, they do not always have similar secondary structures and consequently similar three-dimensional structures and functions. Based on these theories, we predict the interacting proteins in Saccharomyces cerevisie (baker's yeast) from the information of their secondary structures using computational method. This paper proposes multiple independent fuzzy systems for predicting protein-protein interactions from the similarity of protein secondary structures. Our method consists of two main stages: (1) similarity score computation, and (2) similarity classification. The first stage involves three steps: (1) Multiple-sequence alignment (MSA) - finding multiple-sequence alignment for every family groups of proteins in Saccharomyces cerevisie, (2) Secondary structure prediction (SSP) - predicting secondary structure of aligned proteins sequence using secondary structure prediction tool called SSpro, and (3) Similarity measurement (Sim) - computing similarity scores of predicted secondary structures for every possible proteins pairs based on the number of three conformational states: helix (H), sheet (E), and coil (C). In the classification stage, N multiple independent first order Sugeno Fuzzy Systems are generated to model the behavior of similarity scores of all possible proteins pairs to classify the interacting and non-interacting pairs; here N is the number of protein. Every system determines initial rules based on the clusters information obtained from the fuzzy clustering method. We employ principal component analysis in every system to compress the dimension of input data. Our model has been trained and tested using 1029 proteins with already known 2965 positive interactions of Saccharomyces cerevisie (baker's yeast). This proposed model achieves good accuracy when compared with experimentally determined proteins interactions from the Database of Interacting Proteins.

KW - Fuzzy system modeling

KW - Protein-protein interaction prediction

KW - Secondary structures

UR - http://www.scopus.com/inward/record.url?scp=80053011231&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=80053011231&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9780975840078

SP - 782

EP - 788

BT - 18th World IMACS Congress and MODSIM09 International Congress on Modelling and Simulation: Interfacing Modelling and Simulation with Mathematical and Computational Sciences, Proceedings

ER -