一种基于R语言的大孔吸附树脂预测模型的构建

Construction of a Prediction Model for Macroporous Resin Based on R Language

针对精准筛选高选择性分离中药多糖的大孔吸附树脂时间冗长、操作繁琐的现状,该文构建了一种快速筛选大孔吸附树脂的预测模型。采用R语言建立大孔吸附树脂参数、葡聚糖分子量与吸附量之间的多源信息模型,利用多源信息模型快速预测和筛选对中药多糖具有最佳吸附效果的大孔吸附树脂类型。所建多源信息模型的相关系数(r2)为0.9012,P值为0.000782,均方根误差(RMSE)为19.89。采用枸杞多糖对构建的多源信息模型进行验证,利用该模型预测LX-T19、LX-T81、LX-20、LX-1180和D101型大孔吸附树脂对枸杞多糖的吸附效果,结果发现最佳大孔吸附树脂为LX-20,其吸附量为117.40 mg·g-1,验证实验得到其实际吸附量为111.23 mg·g-1。结果表明,所建立的多源信息预测模型准确、可靠,对精准、快速筛选分离纯化中药多糖的大孔吸附树脂具有参考价值。

Aiming at the current situation of long time and cumbersome operation in precise screening of macroporous resins with high selective separation performance for polysaccharides of traditional Chinese medicine,a predictive model for the rapid screening of macroporous resins was constructed in this paper.A multi-source information model for the parameters of a macroporous adsorption resin,the molecular weight of glucan and the adsorption capacity was established by using R language,which was used to quickly predict and screen the type of macroporous adsorption resin with the optimal adsorption effect for polysaccharides of traditional Chinese medicine.The correlation coefficient(r2)of the established multi-source information model is 0.9012,the P value is 0.000782 and the root mean square error(RMSE)is 19.89,which was verified with lycium barbarum polysaccharide,and was used to predict the adsorption effects of LX-T19,LX-T81,LX-20,LX-1180 and D101 type macroporous adsorption resins on lycium barbarum polysaccharides.Results showed that LX-20 is the best macroporous adsorption resin,with its adsorption capacity of 117.40 mg·g-1,while its actual adsorption capacity by the verification experiment is 111.23 mg·g-1.The established multi-source information prediction model is accurate and reliable,and has a reference value for the precise and rapid screening of macroporous adsorption resins used for purification of polysaccharides in Chinese medicine.