Home // ADVCOMP 2014, The Eighth International Conference on Advanced Engineering Computing and Applications in Sciences // View article
Authors:
Hui-Ling Huang
Hua-Chin Lee
Phasit Charoenkwan
Wen-Lin Huang
Li-Sun Shu
Shinn-Ying Ho
Keywords: bioluminescent proteins; feature selection; fuzzy rules; genetic algorithm; knowledge acquisition; physicochemical properties.
Abstract:
New applications of using bioluminescent proteins (BLPs) are constantly increasing in a variety of research fields such as protein engineering of using single-cell bioluminescent organisms to determine how animals move through water. In this study, we propose a knowledge acquisition method for characterizing BLPs and understanding their functions using a compact set of fuzzy rules. The rule set was obtained by designing an if-then fuzzy-rule-based bioluminescent protein classifier (named iFBPC) with physicochemical properties as input features. In designing iFBPC, feature selection, membership function design, and fuzzy rule base generation are all simultaneously optimized using an intelligent genetic algorithm (IGA). We used the same benchmark dataset for comparisons used in existing SVM-based prediction methods BLProt and PBLP using 100 and 15 features of physicochemical properties, respectively. The classifier iFBPC has two fuzzy rules (one for BLP and the other for non-BLP) and four physicochemical properties with test accuracy of 74.82% where BLProt and PBLP have accuracies of 80.06% and 81.79%, respectively. The four physicochemical properties are structures, protein linkers, nucleation, and membrane proteins in the AAindex database. The analysis of characterizing BLPs was conducted based on knowledge of the fuzzy rule base.
Pages: 24 to 29
Copyright: Copyright (c) IARIA, 2014
Publication date: August 24, 2014
Published in: conference
ISSN: 2308-4499
ISBN: 978-1-61208-354-4
Location: Rome, Italy
Dates: from August 24, 2014 to August 28, 2014