基于房室模型和WOA-BP神经网络的药效组分配比计算方法  

Calculation of pharmacodynamic component ratios based on thecompartment model and WOA-BP neural networks

作　　者：杨婕妤 李勇[1] 向诚[2] 何子懿 YANG JieYu;LI Yong;XIANG Cheng;HE ZiYi(Faculty of Information Engineering and Automation;Faculty of Life Science and Technology,Kunming University of Science and Technology,Kunming 650550,China)

机构地区：[1]昆明理工大学信息工程与自动化学院,昆明650550 [2]昆明理工大学生命科学与技术学院,昆明650550

出　　处：《北京化工大学学报（自然科学版）》2023年第6期94-104,共11页Journal of Beijing University of Chemical Technology(Natural Science Edition)

基　　金：国家自然科学基金(82160787)。

摘　　要：基于房室模型建立了中药药效组分作用于机体的时-量关系,推导出药效组分的血药浓度所对应的体内药量比例;利用鲸鱼优化算法(WOA)优化的反向传播(BP)神经网络建立了药效组分的血药浓度与机体整体反应的量-效关系模型;利用WOA-BP神经网络的反向传播过程调整药物效应,构建了新的量-效关系,并结合时-量关系确定药效变化后的药效组分配比。以辅助苯巴比妥抑制癫痫的青阳参(Cynanchum otophyllum)皂苷M1和M2的药效组分配比分析为例进行实证研究,初始的M1和M2配比(给药剂量比)为2∶1,利用建立的基于房室模型和WOA-BP神经网络的药效组分配比计算方法,确定在药效提升5%后M1和M2的配比为2.26∶1,此时与初始M1和M2配比时的药效相比,青阳参抑制癫痫的效果增强。本模型能够在有限的实验条件下快速确定药效提高时所对应的药效组分配比,为中药的配伍研究提供了一种新方法。Based on the compartment model,the time-dose relationship of the effective components of traditional Chinese medicine acting on the body has been established,and the ratio of dosage in the body corresponding to the blood concentration of the effective components has been deduced.The dose-effect relationship model between the blood concentration of the active components and the overall response of the body was established by using a whale optimization algorithm(WOA)optimized back propagation(BP)neural network.A new dose-effect relationship was established by using the backpropagation process of the WOA⁃BP neural network to adjust the drug effect,and the ratio of the pharmacodynamic components after the change of efficacy was determined by combining the time-dose relationships.An empirical study was carried out by analysis of the ratio of the pharmacodynamic components of qingyangshen(Cynanchum otophyllum)saponins M1 and M2,which are known to assist phenobarbital in inhibi⁃ting epilepsy.The initial dosage ratio of M1 to M2 was 2∶1.The ratio of M1 to M2 was determined to be 2.26∶1 af⁃ter the efficacy was increased by 5%by using the established method based on the compartment model and the WOA-BP neural network.As a result,compared with the efficacy of the initial M1 and M2 ratio,the inhibition of epilepsy by qingyangshen was enhanced.The model can quickly determine the corresponding pharmacodynamic component ratio when the efficacy is improved under limited experimental conditions and provides a new method for compatibility studies of traditional Chinese medicine.

关 键 词：组分配比 房室模型 BP神经网络 鲸鱼优化算法(WOA) 青阳参 癫痫 

分 类 号：R917[医药卫生—药物分析学]