Title :
Predicting βαβ motifs based on SVM by using the ID and MS values
Author :
Lixia Sun ; Xiuzhen Hu ; Shaobo Li
Author_Institution :
Coll. of Sci., Inner Mongolia Univ. of Technol., Huhhot, China
Abstract :
From the Protein Data Bank (PDB), we screened out 1635 sequences with identity <;25% and resolution <; 3.0 Å. Each sequence contains at least one βαβ motif. This new dataset contains 4277 βαβ motifs and 3366 non-βαβ motifs. We fixed sequence length for βαβ motifs. By using the parameters with increment of diversity (ID) values, matrix scoring (MS) values and amino acids component to express the information of sequence, a support vector machine algorithm for predicting βαβ motifs was proposed. The overall accuracy and Matthew´s correlation coefficient of 5-fold cross-validation achieved 77.7% and 0.527.
Keywords :
bioinformatics; molecular biophysics; molecular configurations; proteins; support vector machines; ID value; MS value; Protein Data Bank; SVM; beta-alpha-beta motif prediction; diversity increment; fixed sequence length; matrix scoring; protein sequences; sequence information; support vector machine algorithm; βαβ motifs; Support Vector Machine algorithm; amino acids component; increment of diversity; scoring matrix;
Conference_Titel :
Biomedical Engineering and Informatics (BMEI), 2012 5th International Conference on
Conference_Location :
Chongqing
Print_ISBN :
978-1-4673-1183-0
DOI :
10.1109/BMEI.2012.6513166
