Two databases have been constructed to facilitate applications of cheminformatics and molecular modeling to medicinal plants. The first contains data on known chemical constituents of 240 commonly used Chinese herbs, the other contains information on target specificities of bioactive plant compounds. Structures are available for all compounds. In the case of the Chinese herbal constituents database, further details include trivial and systematic names, compound class and skeletal type, botanical and Chinese (pinyin) names of associated herb(s), CAS registry number, chirality, pharmacological and toxicological information, and chemical references. For the bioactive plant compounds database, details of molecular target(s), IC50 and related measures, and associated botanical species are given. For Chinese herbs, approximately 7000 unique compounds are listed, though some are found in more than one herb, the total number for all herbs being 8264. For bioactive plant compounds, 2597 compounds active against 78 molecular targets are covered. Statistical relationships within and between the two databases are explored.

Random Forest, a form of multiple decision trees, has been used to screen a database of Chinese herbal constituents for potential inhibitors against several therapeutically important molecular targets. These comprise cyclic adenosine 3′-5′-monophosphate phosphodiesterases, protein kinase A, cyclooxygenases, lipoxygenases, aldose reductase, and three HIV targets-integrase, protease, and reverse transcriptase. In addition, compounds were identified which may inhibit the expression of inducible nitric oxide synthase and/or nitric oxide production in vivo. A total of 240 Chinese herbs containing 8264 compounds were screened in silico, including many used on a regular basis in traditional Chinese medicine. Active compounds were selected from another database of 2597 phytochemicals and related natural products with known target affinities and covered a wide range of structural classes. Random Forest was found to perform well, even on highly unbalanced data characteristic of ligand-based screening where the compounds to be screened are far more numerous than the number of active compounds used in training. Despite a conservative screening protocol, a wide variety of compounds from Chinese herbs were hit. Of particular interest were the relatively large number of herbs predicted to inhibit multiple targets, as well as a number which appeared to contain inhibitors of the same target from different phytochemical classes. The latter point to the possibility that individual species may make use of alternative phytochemical strategies in target inhibition. A literature search provided evidence to support 83 herb-target predictions.

The available databases that catalogue information on traditional Chinese medicines are reviewed in terms of their content and utility for in-silico research on Chinese herbal medicines, as too are the various protein database resources, and the software available for use in such studies. The software available for bioinformatics and 'omics studies of Chinese herbal medicines are summarised, and a critical evaluation given of the various in-silico methods applied in screening Chinese herbal medicines, including classification trees, neural networks, support vector machines, docking and inverse docking algorithms. Recommendations are made regarding any future in-silico studies of Chinese herbal medicines.

Distribution patterns of 8411 compounds from 240 Chinese herbs were analyzed in relation to the herbal categories of traditional Chinese medicine (TCM), using Random Forest (RF) and self-organizing maps (SOM). RF was used first to construct TCM profiles of individual compounds, which describe their affinities for 28 major herbal categories, while simultaneously minimizing the level of noise associated with the complex array of diverse phytochemicals found in herbs from each category. Profiles were then reduced and visualized with SOM. The distribution of 10 major phytochemical classes, in relation to TCM profile, was delineated with SOM-Ward clustering. These classes comprised aliphatics, alkaloids, simple phenolics, lignans, quinones, polyphenols (flavonoids and tannins), and mono-, sesqui-, di-, and triterpenes (including sterols). Highly distinctive patterns of association between phytochemical class and TCM profile were revealed, suggesting that a strong phytochemical basis underlies the traditional language of Chinese medicine. Maps trained after random permutation of herbs assigned to each category were, by contrast, devoid of feature, providing additional evidence for the significance of these associations. Most classes were split into relatively few clusters, and further analysis revealed that simple descriptors, comprising skeletal type, molecular weight, and calculated log P, were in most cases able to readily discriminate within-class clusters. Relationships between TCM profile and predicted activities, relating to therapeutically important molecular targets, were explored and indicate that ethnopharmacological data could play an important role in pharmaceutical prospecting from Chinese herbs as well as identifying links between Chinese and Western medicine.

No abstract