ADVERTISEMENT

Seeing the Medicine in Chinese Herbs through the Random Forest

Data mining of Chinese herbal compounds targets HIV/AIDs and other diseases
Chinese traditional herb ginkgo



© GUY CALI/CORBIS
Over the past several years, scientific journals have been abuzz with reports of the chemical constituents of Chinese herbs used in traditional medicine. Think ginseng, Ginkgo biloba and ginger, to name a few.

A group of researchers at King's College London decided to use a computer screening to construct a single database both to catalogue the chemical makeup of 240 species of herb and to indicate which target enzymes and receptors implicated in diseases—such as HIV/AIDS, cancer and Alzheimer's disease—those components may be able to regulate.

"The motivation for the cataloguing and the provision of tools for the data mining is really to provide a way for perhaps not the herbs themselves, but the purified constituents [that] might be discovered as new therapeutics of this disease or that disease," says King's College pharmacologist David Barlow, who credits the study's co-author, PhD student Thomas Ehrman, with the idea.

Barlow's group used a screening algorithm called Random Forest, which is a type of decision tree, to compile its database. The algorithm involves each entity being screened with a random set of questions—in this case, mostly about the herb's constituents—to tease out which of the biological targets it could possibly effect. "If you take a set of 8,000-plus constituents from herbs and get their chemical details and feed them through the system, you essentially classify them one by one and say, 'This one has a fingerprint appropriate for this target and, therefore, may have this use,'" Barlow explains.

The targets the researchers chose fell into five categories: cell signal regulators, implicated in cancer, asthma and depression; nitrous oxide overproduction or overexpression, which is associated with the hardening of arteries and inflammation as well as Huntington's and Alzheimer's diseases; cyclooxygenase and lipooxygenase, two targets of anti-inflammatory agents that are tied to Alzheimer's, cancer and arthritis; aldose reductase, an enzyme responsible for complications from diabetes, such as eye disorders; and the viral enzymes HIV-1 integrase, protease and reverse transcriptase, all implicated in catalyzing the HIV virus's life cycle. Of the 240 herbs sampled, 62 percent of them were found to have constituents that could be useful in treating one of these targets. Fifty-three percent of the plants may be able to tackle more than one disease.

Barlow says that, going forward, the study's third co-author, Peter Hylands, a natural products chemist, will likely try to isolate, purify and screen some of these constituents from the herbs—such as flavonoids, which produce yellow or blue pigments in plants, and terpenoids, plant fats known for antibacterial qualities. Clinical studies of their effects could then be studied in animal or human models.

Another interesting result of the study, notes Barlow, is comparing what the screening found with uses of some of these herbs in Chinese traditional medicine. He says that many of the plants shown to have aldose reductase inhibitors are already prescribed for cataracts, and phosphodiesterase enzyme inhibitors, which scientists believe could be useful for asthma, obesity and depression, are already being utilized in "yin tonics" to treat the stomach, heart and lungs.

"One might see a link there between certainly the obesity," Barlow says, "and possibly asthma as well with the kind of uses these herbs are put to in traditional medicine."

Rights & Permissions
Share this Article:

Comments

You must sign in or register as a ScientificAmerican.com member to submit a comment.
Scientific American Back To School

Back to School Sale!

12 Digital Issues + 4 Years of Archive Access just $19.99

Order Now >

X

Email this Article

X