论文部分内容阅读
目的探讨K最近邻(KNN)法在含纤连蛋白(FN)域蛋白质亚细胞定位中的应用价值。方法选取含FN域蛋白质80个(40个细胞外蛋白质和40个细胞内蛋白质),采用KNN法进行蛋白质亚细胞定位,并采用jack-knife检验法和5维交叉验证法检验样本的定位的准确率。结果 KNN法定位细胞内蛋白36个,细胞外蛋白35个。jackknife法检验KNN法蛋白质定位准确率为88.75%,5维交叉法验证其定位准确率为82.5%。结论利用KNN法可较准确的预测含FN域蛋白质的亚细胞位置。
Objective To investigate the value of K nearest neighbors (KNN) in protein subcellular localization of fibronectin (FN) domain. Methods Eighty FN domain proteins (40 extracellular proteins and 40 intracellular proteins) were selected. Protein subcellular localization was performed by KNN method. The accuracy of localization of the samples was tested by jack-knife test and 5-D cross-validation rate. Results KNN method located 36 intracellular proteins and 35 extracellular proteins. Jackknife test KNN protein localization accuracy of 88.75%, 5-D cross method validation of its positioning accuracy of 82.5%. Conclusion KNN method can be used to predict subcellular location of FN-containing proteins more accurately.