Combining PSSM and physicochemical feature for protein structure prediction with support vector machine

International Symposium on Bioinformatics, Chemometrics and Metabolomics

生物科学;化学

Kurniawan, I.^1 ; Haryanto, T.^1 ; Hasibuan, L.S.^1 ; Agmalaro, M.A.^1

Department of Computer Science, Bogor Agricultural University, Indonesia^1

关键词: Physicochemical features;    Position specific scoring matrix;    Protein secondary structure;    Protein structure prediction;    Quaternary structure;    Support vector machine models;    Tertiary structures;    Three-dimensional structure;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/835/1/012006/pdf
DOI  :  10.1088/1742-6596/835/1/012006

学科分类：生物科学（综合）

来源: IOP

【 摘 要 】

Protein is one of the giant biomolecules that act as the main component of the organism. Protein is formed from building blocks namely amino acids. Hierarchically, the structure of protein is divided into four levels: primary, secondary, tertiary, and quaternary structure. Protein secondary structure is formed by amino acid sequences that would form three-dimensional structures and have information about the tertiary structure and function of proteins. This study used 277,389 protein residue data from enzyme categories. Position-specific scoring matrix (PSSM) profile and physicochemical are used for features. This study developed support vector machine models to predict the protein secondary structure by recognizing patterns of amino acid sequences. The Q3 results showed that the best scores obtained are 93.16% from the dataset that has 260 features with the radial kernel. Combining PSSM and physicochemical feature additions can be used for prediction.