噁唑烷酮类化合物的定量构效关系

Quantitative structure-activity relationship of oxazolidinone compounds

目的研究新型全合成抗菌药噁唑烷酮类化合物的定量构效关系。方法计算118个噁唑烷酮类化合物的量子化学和分子连接性指数等描述符,用偏最小二乘(PLS)对其进行筛选,得到9个对抗菌活性有重要影响的分子描述符,然后分别用PLS方法和改进的人工神经网络(ANN)方法建立这些描述符及其抗菌活性之间定量结构活性关系的模型。结果比较ANN模型与PLS模型的结果,证明了化合物结构与活性之间的非线性关系。结论所建立的定量构效关系模型能有效地进行噁唑烷酮类化合物的最小抑菌浓度值的预测,为新药开发提供新的参考和思路。

OBJECTIVE To study the quantitative structure - activity relationship （QSAR） of oxazolidinone which is a new class of total synthetic antibacterial agents. METHODS Quantum chemical and molecular connection indexs etc. were used as descriptors and 118 oxazolidinone compounds were computed, and partial least square （PLS） was used to reved which description variables were the most relevant to activity. Then, QSAR models were developed that linked nine selected molecular descriptors with their antibacterial activity by PLS and improved artificail neural network （ ANN）, respectively. RESULTS Comparing the results of the two methods indicated that the relationship between molecular structure and activity was nonlinear. CONCLUSSION The QSAR model built can effectively predict the MIC values of oxazolidinone that provide new reference and idea for developing new drugs.