摘要
目的提出一种全排序模板匹配算法,用于中药色谱指纹图谱的色谱峰匹配和指纹图谱相似度计算。方法以高效液相色谱法(HPLC)测定丹参药材中丹参酮ⅡA为例,分别采集样品溶液的色谱图,对色谱峰进行匹配,在原有算法基础上,提出一种全排序模板匹配算法,这种算法不以任何一张实际的指纹图谱为模板,而是根据所有图谱上的所有色谱峰的保留时间进行升序排序、聚类,依从色谱峰最接近、同一组匹配峰中不包含两个同源峰、待匹配的峰不超过匹配阈值等原则。结果这种方法克服了原算法中的某些错配、漏配等缺陷,得到了较满意的匹配结果。结论此算法为以保留时间为依据的色谱峰匹配算法找到了新的切入点。
Objective: To introduce a new template peak matching algorithm for calculating the similarity of chromatographic fingerprint of traditional Chinese herbs. Methods: Tanshinone ⅡA in S. miltiorrhiza Bge of different batches were determined by HPLC and the chromatograms of them were obtained. We designated a new peak matching algorithm based on the previous algorithms, which employed a certain real chromatographic fingerprint as their template. In the new algorithm, we arranged the retention times of chromatographic peaks of all chromatograms into an ascending order, forming a template. The sorting procedure complied with the following 2 rules. First, the same peak matches the nearest corresponding peak. Second, one peak does not appear twice in the same chromatogram. Results: The new algorithm avoided the shortcomings of previous algorithm, such as mismatching and missed matching, and obtained a satisfactory outcome. Conclusion: Our new algorithm provides a basis for improving the reliability of retention time-based peak matching algorithm.
出处
《第二军医大学学报》
CAS
CSCD
北大核心
2007年第2期183-187,共5页
Academic Journal of Second Military Medical University
基金
上海市科委重点项目(03DZ19548)
上海市科技发展基金(02419125).~~
关键词
全排序模板匹配算法
色谱指纹图谱
中草药
total sequencing template matching algorithm
chromatographic fingerprint
drugs, Chinese herbal