Zhao Chunyan, Zhang Haixia, Zhang Xiaoyun, Zhang Ruisheng, Luan Feng, Liu Mancang, Hu Zhide, Fan Botao
Department of Chemistry, Lanzhou University, Lanzhou, 730000, China.
Pharm Res. 2006 Jan;23(1):41-8. doi: 10.1007/s11095-005-8716-4. Epub 2006 Nov 30.
Development of reliable computational models to predict/classify milk-to-plasma (M/P) drug concentration ratio remains a challenging object. Support vector machine (SVM) method, as a new algorithm, was constructed to distinguish the potential risk of drugs to nursing infants.
Each drug was represented by a large pool of descriptors, of which five were found to be most important for constructing the predictive models. Next, two classification models, linear discriminant analysis (LDA) and SVM, were developed with bootstrapping validation based on the selected molecular descriptors.
The classification accuracy of training set and test set for SVM was 90.63 and 90.00%, respectively. The total accuracy for SVM was 90.48%, which was higher than that of LDA (77.78%). Comparison of the two methods shows that the performance of SVM was better than that of LDA, which implies that the SVM method is an effective tool in evaluating the risk of drugs when experimental M/P ratios have not been investigated.
开发可靠的计算模型以预测/分类乳-血（M/P）药物浓度比值仍然是一个具有挑战性的目标。支持向量机（SVM）方法作为一种新算法，被构建用于区分药物对哺乳婴儿的潜在风险。
每种药物由大量描述符表示，其中发现有五个对于构建预测模型最为重要。接下来，基于所选分子描述符，通过自助法验证开发了两种分类模型，即线性判别分析（LDA）和支持向量机（SVM）。
支持向量机训练集和测试集的分类准确率分别为90.63%和90.00%。支持向量机的总准确率为90.48%，高于线性判别分析（77.78%）。两种方法的比较表明，支持向量机的性能优于线性判别分析，这意味着在尚未研究实验性M/P比值时，支持向量机方法是评估药物风险的有效工具。
