Send to

Choose Destination
Planta Med. 2014 Aug;80(12):1036-44. doi: 10.1055/s-0034-1382951. Epub 2014 Aug 21.

Feature selection for the identification of antitumor compounds in the alcohol total extracts of Curcuma longa.

Author information

Key Laboratory of Systems Bioengineering, Ministry of Education and Tianjin Key Laboratory of Biological and Pharmaceutical Engineering, Department of Pharmaceutical Engineering, School of Chemical Engineering and Technology, Tianjin University, Tianjin, China.


Antitumor activity has been reported for turmeric, the dried rhizome of Curcuma longa. This study proposes a new feature selection method for the identification of the antitumor compounds in turmeric total extracts. The chemical composition of turmeric total extracts was analyzed by gas chromatography-mass spectrometry (21 ingredients) and high-performance liquid chromatography-mass spectrometry (22 ingredients), and their cytotoxicity was detected through an 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay against HeLa cells. A support vector machine for regression and a generalized regression neural network were used to research the composition-activity relationship and were later combined with the mean impact value to identify the antitumor compounds. The results showed that six volatile constituents (three terpenes and three ketones) and seven nonvolatile constituents (five curcuminoids and two unknown ingredients) with high absolute mean impact values exhibited a significant correlation with the cytotoxicity against HeLa cells. With the exception of the two unknown ingredients, the identified 11 constituents have been reported to exhibit cytotoxicity. This finding indicates that the feature selection method may be a supplementary tool for the identification of active compounds from herbs.

[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Georg Thieme Verlag Stuttgart, New York
Loading ...
Support Center