This article is part of the supplement: The ISIBM International Joint Conferences on Bioinformatics, Systems Biology and Intelligent Computing (IJCBS)
Feature amplified voting algorithm for functional analysis of protein superfamily
1 Department of Computer Science, National Tsing Hua University, 101, Section 2 Kuang Fu Road, Hsinchu, Taiwan 300
2 Macronix International Co., Ltd., 16, Li-Hsin Road, Science Park, Hsinchu, Taiwan 300
3 Department of Computer Science and Information Engineering, Chang Gung University, 259 Wen Hwa 1st Road, Kwei Shan, Taoyuan 333, Taiwan.
Citation and License
BMC Genomics 2010, 11(Suppl 3):S14 doi:10.1186/1471-2164-11-S3-S14Published: 1 December 2010
Identifying the regions associated with protein function is a singularly important task in the post-genomic era. Biological studies often identify functional enzyme residues by amino acid sequences, particularly when related structural information is unavailable. In some cases of protein superfamilies, functional residues are difficult to detect by current alignment tools or evolutionary strategies when phylogenetic relationships do not parallel their protein functions. The solution proposed in this study is Feature Amplified Voting Algorithm with Three-profile alignment (FAVAT). The core concept of FAVAT is to reveal the desired features of a target enzyme or protein by voting on three different property groups aligned by three-profile alignment method. Functional residues of a target protein can then be retrieved by FAVAT analysis. In this study, the amidohydrolase superfamily was an interesting case for verifying the proposed approach because it contains divergent enzymes and proteins.
The FAVAT was used to identify critical residues of mammalian imidase, a member of the amidohydrolase superfamily. Members of this superfamily were first classified by their functional properties and sources of original organisms. After FAVAT analysis, candidate residues were identified and compared to a bacterial hydantoinase in which the crystal structure (1GKQ) has been fully elucidated. One modified lysine, three histidines and one aspartate were found to participate in the coordination of metal ions in the active site. The FAVAT analysis also redressed the misrecognition of metal coordinator Asp57 by the multiple sequence alignment (MSA) method. Several other amino acid residues known to be related to the function or structure of mammalian imidase were also identified.
The FAVAT is shown to predict functionally important amino acids in amidohydrolase superfamily. This strategy effectively identifies functionally important residues by analyzing the discrepancy between the sequence and functional properties of related proteins in a superfamily, and it should be applicable to other protein families.