Study On The Prediction Of Protein Super Secondary Structure And Mitochondrial Proteins Of Malaria Parasite

Protein function is inherently correlated with its structure. Therefore, the study of protein structure is a basic premise for the prediction of its function. How to predict the protein structure is one of the research topics in the life science, however, determination of protein structure purely using experimental approaches is time-consuming and expensive, moreover, experimenters also are unable to overcome technical difficulties. Thus, theoretical method is a key way for predicting the spatial structure of protein. Super secondary structure is a critical bridge between primary structure and three-dimensional structure. The prediction of protein super secondary structure has important research significance.In the study, firstly, we select 123 proteins with sequence similarity less than 30% between two protein sequences, then, we chose five kinds of simple super secondary structures and β-hairpin motif as data subsets. In the two data subsets, we used chemical shifts of six nuclei as features, and combined with a variety of algorithms to predict, respectively. Finally, quadratic discriminant analysis achieved best results, namely seven-fold cross-validation achieved the averaged sensitivity, specificity and the overall accuracy are 81.8%,95.19%,82.91%, respectively in the five kinds of simple protein super secondary structures. Three-fold cross-validation achieved the sensitivity, specificity, overall accuracy and Mathew’s correlation coefficient are 92%、94%、87%、0.85, respectively in the β-link and β-loop-β motifs. The results illustrate that chemical shift combine with quadratic discriminant analysis can effectively predict protein super secondary structure.Malaria has been one of the serious infection diseases caused by Plasmodium falciparum (P. Falciparum). Mitochondrial proteins of P. Falciparum are regarded as effective drug targets against malaria. In recent years, with the rapid development of bioinformatics, more and more theoretical predictions are devoted to research for accurating identify mitochondrial proteins of malaria parasite.In the study, we added firstly a novel parameter for predicting mitochondrial proteins of malaria parasite based on protein secondary structure compositions. We select p/M233 as data subset with sequence similarity less than 25% between two protein sequences. Firstly, we extracted three kinds of feature parameters, namely, three kinds of protein secondary structures compositions,20 amino acid compositions and 400 dipeptide compositions, and used the analysis of variance to screen 400 dipeptide compositions. Secondly, we adopted these features to predict mitochondrial proteins of malaria parasite by using support vector machine. Finally, we found that 1) the addition of protein secondary structure compositions can indeed improve the prediction accuracy. This result demonstrated that protein secondary structure parameter is a valid feature in the prediction of mitochondrial proteins of malaria parasite; 2) feature combination can improve the prediction results. Moreover, in the information redundancy, feature selection can reduce the dimension and simplify the calculation. We achieved the sensitivity of 98.16%, the specificity of 97.64% and overall accuracy of 97.88% with 0.957 of Mathew’s correlation coefficient by using 3 PSS +20 AAC +34 DC as feature in 15-fold cross-validation. This result is compared with that of the similar work, showing the superiority of our work.