基于数据挖掘的抗炎抗肿瘤中药药性特征分析

Analysis of properties of anti⁃inflammatory and anti⁃tumor traditional Chinese herbal medicines based on data mining

目的基于数据挖掘分析抗炎抗肿瘤中药的分布特征及药性特征。方法以《中药学》中记载的481味临床常用中药为研究对象,通过文本数据归纳及中英文数据库文献检索,筛选出具有抗炎、抗肿瘤作用的中药,建立数据库,采用频数分析、药类性味归经分析、共现分析对抗炎抗肿瘤中药药性特征进行挖掘,运用KH Coder 3.0软件将文本数据中的词频、词性、词间关系等特征进行可视化分析。结果(1)共筛选出抗炎、抗肿瘤、抗炎且抗肿瘤中药分别为368味、321味和286味。(2)3组药物分布特征相似,均分布于21种药类;药性上寒凉与温热占比相当;药味以苦、甘、辛味较多;归经以肝、肺、胃、脾、肾经为主;性味组合苦寒、苦温和辛温较多。(3)可视化分析发现,3组中药属性间存在不同的聚类关系。结论抗炎中药与抗肿瘤中药重合度高,二者存在多条可能相关的共同作用通路;炎-癌转化不同阶段可能与中药证治特征存在相关性。

Objective To analyze the distribution and pharmacological properties of anti-inflammatory and anti-tumor traditional Chinese herbal medicines based on data mining.Methods Totally 481 clinically used traditional Chinese herbs recorded in Chinese Materia Medica were included for research,and Chinese herbs with anti-inflammatory and anti-tumor effects were screened out through text data analysis and literature retrieval from Chinese and English databases.The database was established,and frequency analysis,analysis of herbal property,flavor and meridian tropism,and co-occurrence analysis were used to find the characteristics of anti-inflammatory and anti-tumor herbs.KH Coder 3.0 was used to visualize and analyze the features of word frequency,part of speech,and inter-word relationships in the text data.Results①A total of 368 anti-inflammatory herbs,321 anti-tumor herbs and 286 anti-inflammatory and anti-tumor herbs were screened out.②The classifications of herbs in the 3 groups were similar as each group involves 21 herbal categories.The proportion of herbs with cool/cold property was comparable to that of herbs with warm/heat property.Most herbs had bitter,sweet and/or pungent flavors and entered the liver,lung,stomach,spleen,and/or kidney meridians.The property-flavor combinations of bitter-cold,bitter-warm and pungent-warm were more in frequency.③Visualization analysis revealed that there were different clustering correlations among the three groups of herbs.Conclusions There is a high degree of overlap between anti-inflammatory herbs and anti-tumor herbs,and there are multiple potentially relevant common pathways between the two groups of herbs.The different stages of inflammation-cancer transformation and syndrome-treatment characteristics of herbal medicines may be correlated.