Protein fold recognition using segmentation-based feature extraction model

Authors:
Abdollah Dehzangi;Abdul Sattar
Affiliations:
Institute for Integrated and Intelligent Systems (IIIS), Griffith University, Brisbane, Australia, National ICT Australia (NICTA), Brisbane, Australia;Institute for Integrated and Intelligent Systems (IIIS), Griffith University, Brisbane, Australia, National ICT Australia (NICTA), Brisbane, Australia
Venue:
ACIIDS'13 Proceedings of the 5th Asian conference on Intelligent Information and Database Systems - Volume Part I
Year:
2013

Citing 7
Cited 0

Random Forests

Machine Learning
Ensemble classifier for protein fold pattern recognition

Bioinformatics
Data Mining: Practical Machine Learning Tools and Techniques, Second Edition (Morgan Kaufmann Series in Data Management Systems)

Data Mining: Practical Machine Learning Tools and Techniques, Second Edition (Morgan Kaufmann Series in Data Management Systems)
PFRES

Bioinformatics
Boosting Methods for Protein Fold Recognition: An Empirical Comparison

BIBM '08 Proceedings of the 2008 IEEE International Conference on Bioinformatics and Biomedicine
Protein Fold Prediction Problem Using Ensemble of Classifiers

ICONIP '09 Proceedings of the 16th International Conference on Neural Information Processing: Part II
Research Article: A protein fold classifier formed by fusing different modes of pseudo amino acid composition via PSSM

Computational Biology and Chemistry

Quantified Score

Hi-index	0.00

Visualization

Abstract

Protein Fold recognition (PFR) is considered as an important step towards protein structure prediction. It also provides significant information about general functionality of a given protein. Despite all the efforts have been made, PFR still remains unsolved. It is shown that appropriately extracted features from the physicochemical-based attributes of the amino acids plays crucial role to address this problem. In this study, we explore 55 different physicochemical-based attributes using two novel feature extraction methods namely segmented distribution and segmented density. Then, by proposing an ensemble of different classifiers based on the AdaBoost.M1 and Support Vector Machine (SVM) classifiers which are diversely trained on different combinations of features extracted from these attributes, we outperform similar studies found in the literature for over 2% for the PFR task.