目的 基于超分子“印迹模板”理论，划分、整合六味地黄浓缩丸（Liuwei Dihuang Concentrated Pills，LDCP）指纹图谱，将其表征为物质单元并进行谱量学质量控制与评价研究。方法 采用匹配频数统计矩法划分、整合与表征50批LDCP指纹图谱物质单元，根据朗伯-比尔定律将各物质单元总峰面积与出膏率进行多元线性回归建立谱量学方程，进行谱量学研究。结果 50批LDCP的UPLC指纹图谱被划分为35个物质单元（A₁～A₃₅），经多元线性回归得到35个物质单元（A₁～A₃₅）总峰面积与出膏率的谱量学方程：P_T＝23.390＋0.041 06 A₁－9.100×10⁻³ A₂＋0.014 68 A₃＋0.027 98 A₄－0.033 61 A₅－0.042 25 A₆－6.608×10⁻³ A₇－0.025 90 A₈－0.145 60 A₉＋0.165 00 A₁₀－0.027 50 A₁₁＋3.408×10⁻³ A₁₂－0.021 03 A₁₃－1.070×10⁻³ A₁₄＋2.833×10⁻⁵ A₁₅－8.774×10⁻³ A₁₆＋0.018 52 A₁₇－1.882 9×10⁻³ A₁₈＋0.023 61 A₁₉＋8.566×10⁻³ A₂₀＋0.013 94 A₂₁－5.894×10⁻³ A₂₂－0.012 27 A₂₃－0.014 91 A₂₄＋1.792×10⁻³ A₂₅－1.571×10⁻⁴ A₂₆－3.942×10⁻³ A₂₇－0.054 80 A₂₈＋0.083 15 A₂₉＋0.119 30 A₃₀－0.060 71 A₃₁－0.083 42 A₃₂＋0.014 96 A₃₃－2.989×10⁻³ A₃₄＋0.063 17 A₃₅（r＝0.915）。结论 该方法能以较高准确度划分、整合并表征LDCP指纹图谱物质单元，且保留原指纹图谱整的总量统计矩特性，其谱量学方程能较好的预测LDCP的出膏率与平均出膏率，可为LDCP的质量控制提供新思路与新方法。

Objective Based on the theory of supramolecular “imprinting template”, the fingerprint of Liuwei Dihuang Concentrated Pills (六味地黄丸浓缩丸, LDCP) was divided and integrated, and it was characterized as a material unit, and the spectralquality control and evaluation reaearch were conducted.Methods The matching frequency statistical moment method was used to divide, integrate and characterize the material units of 50 batches of LDCP fingerprints. According to Lambert-Beer’s law, the total peak area of each material unit and the paste yield were subjected to multiple linear regression to establish a spectral equation for spectral research. Results The UPLC fingerprints of 50 batches of LDCP were divided into 35 material units (A₁ − A₃₅). The spectral equations of the total peak area and the paste yield of the 35 material units (A₁ − A₃₅) were obtained by multiple linear regression: P_T = 23.390 + 0.041 06 A₁ − 9.100 × 10⁻³ A₂ + 0.014 68 A₃ + 0.027 98 A₄ − 0.033 61 A₅ − 0.042 25 A₆ − 6.608 × 10⁻³ A₇ − 0.025 90 A₈ − 0.145 60 A₉ + 0.165 00 A₁₀ − 0.027 50 A₁₁ + 3.408 × 10⁻³ A₁₂ − 0.021 03 A₁₃ − 1.070 × 10⁻³ A₁₄ + 2.833 × 10⁻⁵ A₁₅ − 8.774 × 10⁻³ A₁₆ + 0.018 52 A₁₇ − 1.882 9 × 10⁻³ A₁₈ + 0.023 61 A₁₉ + 8.566 × 10⁻³ A₂₀ + 0.013 94 A₂₁ − 5.894 × 10⁻³ A₂₂ − 0.012 27 A₂₃ − 0.014 91 A₂₄ + 1.792 × 10⁻³ A₂₅ − 1.571 × 10⁻⁴ A₂₆ − 3.942 × 10⁻³ A₂₇ − 0.054 80 A₂₈ + 0.083 15 A₂₉ + 0.119 30 A₃₀ − 0.060 71 A₃₁ − 0.083 42 A₃₂ + 0.014 96 A₃₃ − 2.989 × 10⁻³ A₃₄ + 0.063 17 A₃₅ (r = 0.915). Conclusion This method can divide, integrate and characterize the LDCP fingerprint material units with high accuracy, and retain the total statistical moment characteristics of the original fingerprint. Its spectral equation can better predict the paste rate of LDCP and the average paste rate, which can provide new ideas and methods for the quality control of LDCP.

国家自然科学基金资助项目（82274215）；湖南中医药大学校级科研项目（2022XJZKC006）；校级研究生创新课题（2022CX75）；湖南中医药大学科研基金项目（2019XJJJ024）；湖南中医药大学2022年大学生创新创业训练计划（20220701）