目的 建立金银花药材在抗菌活性方面质量优劣的评价模型。方法 采用琼脂打洞扩散法研究21批金银花抗菌活性，建立活性数据库，采用HPLC测定21批金银花药材指纹图谱，建立化学成分群数据库。采用偏最小二乘法（PLS）分析金银花药材活性和化学成分的相关性，计算每批金银花药材的谱效关系（YZL）值（正系数相关峰峰面积与对应系数乘积之和/负系数相关峰峰面积与对应系数乘积之和），对YZL值进行正态拟合并界定，确定判断金银花药材优劣的YZL区间值。另取10批金银花药材对该YZL区间值进行验证。结果 从金银花药材指纹图谱中提取了14个共有峰，采用PLS法建立了金银花药材活性与各共有峰峰面积的回归模型，得到各峰的标准化回归系数。当YZL值≥ 2.53时，为抗菌优质金银花，当YZL值≤ 1.79时，为抗菌劣质金银花。10批金银花药材验证了该结果。结论 YZL值评价金银花药材在抗菌活性方面的质量优劣具有重现性。

Objective To establish the model for evaluating the quality of Lonicera japonica in antimicrobial activity. Methods The active database was established on the basis of antimicrobial activity of 21 batches of L. japonica by using the agar drifting diffusion bioassay. Similarly, the chemical composition database was obtained by detecting the fingerprints of 21 batches of L. japonica from HPLC. The correlation between the activity and chemical composition was analyzed by partial least squares (PLS) method. Then, the YZL interval values were determined to evaluate the quality of L. japonica by normal fitting and probability calculation method. The results were verified by the other 10 batches of L. japonica. Results Fourteen common peaks were obtained from the fingerprints of 21 batches of L. japonica. The standardized regression coefficients of each common peak were got by establishing regression model of the relationship between peak area and activity by PLS method. After normal fitting and defining YZL values, L. japonica was decided as high quality at YZL value ≥ 2.53, on the contrary, as poor quality at YZL value ≤ 1.79. The results were verified by 10 batches of L. japonica. Conclusion The YZL values could provide a repeatable evaluation methodology for the quality control of L. japonica in antimicrobial activity.

国家"重大新药创制"科技重大专项（2013ZX09402203）