An analysis of chemical ingredients network of Chinese herbal formulae for the treatment of coronary heart disease

Abstract

As a complex system, the complicated interactions between chemical ingredients, as well as the potential rules of interactive associations among chemical ingredients of traditional Chinese herbal formulae are not yet fully understood by modern science. On the other hand, network analysis is emerging as a powerful approach focusing on processing complex interactive data. By employing network approach in selected Chinese herbal formulae for the treatment of coronary heart disease (CHD), this article aims to construct and analyze chemical ingredients network of herbal formulae, and provide candidate herbs, chemical constituents, and ingredient groups for further investigation. As a result, chemical ingredients network composed of 1588 ingredients from 36 herbs used in 8 core formulae for the treatment of CHD was produced based on combination associations in herbal formulae. In this network, 9 communities with relative dense internal connections are significantly associated with 14 kinds of chemical structures with P<0.001. Moreover, chemical structural fingerprints of network communities were detected, while specific centralities of chemical ingredients indicating different levels of importance in the network were also measured. Finally, several distinct herbs, chemical ingredients, and ingredient groups with essential position in the network or high centrality value are recommended for further pharmacology study in the context of new drug development.

Fig 1. A diagram of network construction.

Fig 2. Chemical ingredients network.

Note: Node represents chemical ingredient, while node size corresponds to the weighted degree centrality of a chemical ingredient. Each node is colored according to community membership, which was determined using a modularity optimization method. The strength of the lines corresponds to the frequency of the combination into sampled formulae between two chemical ingredients. Vertex positions determined using spectral graph analytic methods according to the normalized Laplacian so that chemical ingredients that are strongly interconnected positioned nearer to each other. Boxes illustrate outstanding nodes and show dominant structural types of chemical ingredients characterizing communities. The numbers in the parentheses in the boxes indicate how many percentages of chemical ingredients are contained by representative herbs of specific communities. Representative herbs are top one or two Chinese medicines which contain more than 50% chemical ingredients in specific communities. Square brackets specify chemical ingredients at the distant positions by showing their chemical names, chemical structural types, and corresponding herbs.

Fig 3. Chemical ingredients fingerprint.

Note: G1 Alkaloids; G2 Amino acid; G3 Fatty acid; G4 Flavonoids and its glycosides; G5 Glycosides; G6 Phenylpropanoids and its derivates; G7 Protein and Enzyme; G8 Quinones; G9 Saccharides and its derivates; G10 Steroids and its glycosides; G11 Stilbenoids; G12 Terpenoids and its derivates; G13 Triterpenoids and its glycosides; G14 Volatile oils.

Fig 4. Ingredient ranking by different centralities.

Note: The NA represents chemical ingredient whose number of CAS is unavailable in SCIfinder.