基于GO的蛋白质亚细胞定位方法研究

Study of Prediction Method for Protein Subcellular Localization Using GO

蛋白质的亚细胞定位对蛋白质功能的研究有重要的意义。蛋白质合成后只有被转运到特定的亚细胞中,才能有效地发挥其功能。选择GO信息作为蛋白质的特征信息,并采用模糊K近邻(Fuzzy KNN)算法对蛋白质亚细胞定位进行预测。预测结果显示,GO信息可以提高亚细胞定位的准确率;而采用Jackknife方法,准确率达到73%。

Protein subcellular localization plays an important role in researching the function of protein.It will work effectively only when transported to right sub-cell after it was compounded.GO information as the feature vector of protein was used to predict protein subcellular localization,with Fuzzy KNN as classifiers.As the prediction results showed,the accuracy was improved by using GO information;with the method of Jackknife,the accuracy was 73%.