Mining conserved regions by an improved protein structural encoding method
Date Issued
2008
Date
2008
Author(s)
Yang, Chia-Jui
Abstract
Analysis of protein structure were mainly divided into two aspects – global structure and local structure，especially the latter correlated closely with analysis of protein function. Most biologists supposed when some frequent patterns reveal in certain protein structure group, it may have some meanings of protein function or evolution in these regions, biologists usually name these regions “conserved regions”. Unfortunately it is very time-consuming when we want to find these conserved regions in a huge database of protein structure, and therefore how to use technology of data mining to solve this problem has become a hot thesis of bioinformatics.In this paper, we use concept of NRS (neighborhood residues sphere) to record distribution of amino acid residue of protein local structure. In order to cluster similar local structure quickly, we encoded every protein local structure to 1-Dimension information. Through heuristic experiments and discussions, we verified accuracy of every encoding method. Further we applied encoding method to mine possible conserved regions which may catalyze in enzyme structure classification database. Finally we also discussed the issue of flexibility and stability of global structure based on this structure encoding method scheme.
Subjects
conserved region
protein structure comparison
encoding
indexing
geometric hashing
Type
thesis
File(s)![Thumbnail Image]()
Loading...
Name
ntu-97-R95525053-1.pdf
Size
23.53 KB
Format
Adobe PDF
Checksum
(MD5):6617b5baa7d6157592764d9b452c42d1