家蚕蛋白亚细胞定位预测模型的构建及其初步应用

Modeling and Preliminary Application of Sub-cellular Localization Prediction Model for Proteins of Bombyx Mori

为研究家蚕蛋白及家蚕杆状病毒蛋白的亚细胞定位并提高预测模型的特异性和准确率,构建了家蚕蛋白亚细胞定位预测模型,并将该模型初步应用于家蚕核型多角体病毒(BmNPV)P10蛋白的亚细胞定位预测中。结果表明,总体预测准确率在不分段时为60.6%,分两段、三段和四段时分别为78.9%、78.4%和80.6%;对BmNPV P10蛋白的预测结果为宿主细胞核,通过免疫细胞荧光实验对预测结果进行验证,结果表明预测结果与实际相符。因此,采用分段的方法能够提高预测准确率,且家蚕病毒蛋白可以利用其宿主家蚕蛋白亚细胞定位预测模型进行亚细胞定位。

In order to investigate the sub-cellular localization of proteins from Bombyx mori and Bmobyx mori nuclear polyhedrosis virus and improve the specificity and accuracy rate of prediction model,the authors developed a prediction model for sub-cellular localization of proteins in Bombyx mori and preliminarily applied it to predict Bmobyx mori nuclear polyhedrosis virus protein P10.The results indicate that the overall accuracy rate of the prediction model is 60.6% when the protein sequence is not segmented.When the protein sequence is divided into two,three and four segments,the overall accuracy rates are 78.9%,78.4%and 80.6%respectively.The prediction result of BmNPV P10 protein is nucleus of its host.The prediction result was verified through immune cell fluorescence experiment.The results show the prediction result conforms to the actual condition.Therefore,segmentation method can improve prediction accurate.In addition,sub-cellular localization can be made for proteins from Bmobyx mori by utilization of its host Bmobyx mori sub-cellular localization prediction model.