【摘要】 利用二级结构端点附近残基的特异性分布特征,应用支持向量机(SVM)预测二级结构端点并得出蛋白质上各位置为端点的可能性分值.根据预测出的端点可能性分值对目前的二级结构预测工具进行端点校正.对psiSVM和PSIPRED这两种工具进行了校正和测试,结果表明校正后端点预测精度相对于psiSVM提高约10%,相对于PSIPRED提高近5%,同时整体预测精度也有一定程度的提高.更多还原

【Abstract】 Based on the residue position-specific performance,we built the prediction modes by SVM(support vector machine) to predict the probability of one position to be secondary structure ends(i.e.helix/strand start/finish positions).Then an end-position modification scheme was developed to correct the ends prediction results of exist secondary structure prediction tools. In this study,we tested the modification scheme on psiSVM and PSIPRED,and the results show that compared with psiSVM and PSIPRED,the ends prediction accuracy after the modification improved about 10% and 5% respectively,and the whole prediction accuracy also get some improve.更多还原