基于最小二乘支持向量机的大肠癌K-ras基因突变预测

Least squares support vector machine-based prediction of K-ras mutations in colorectal cancers

目的探讨利用最小二乘支持向量机模型对大肠癌K-ras基因突变进行预测的可行性。方法首先采用测序法检测90例大肠癌患者癌组织K-ras基因突变情况,继而选取特征变量用最小二乘支持向量机模型进行预测并与测序结果进行比较。结果重复100次的最小二乘支持向量机模型预测发现,训练集的准确率为100%,方差为0;检验集的准确率为79.4%,方差为4.51。结论应用最小二乘支持向量机预测模型预测大肠癌K-ras基因突变是可行的,有助于指导临床诊断、治疗和评价预后。

Objective To evaluate the feasibility of least squares support vector machine（LS-SVM） to predict the K-ras mutations in colorectal cancers.Methods Direct sequencing was first applied to detect the K-ras mutations in 90 colorectal cancers.The LS-SVM model was then constructed to predict the K-ras mutations in these colorectal cancers using selected characteristic variables for 100 times,compared with the direct sequencing results.Results The accuracy and the variance for training atlas were 100% and 0,respectively,while the accuracy and the variance for test atlas were 79.4% and 4.51,respectively.Conclusion LS-SVM is reliable for predicting the K-ras mutations in colorectal cancers,which would provide guidance for clinical diagnosis,therapy,and prognosis.