"Traditional medicines have been connected to efficacy in man for thousands of years (though admittedly often not in controlled clinical trial settings)"
Which to me doesn't really mean anything. Of course there has been a correlation between using these medicines and feeling better, why would anyone keep using a medicine they didn't think was working? But it's not helpful to ask if people thought a medicine worked, it's useful to ask if they actually do work. For that we need to do a controlled clinical trial. Although the authors admit that many TCM compounds fail phase II and III clinical trials because they don't work, they insist that these failures prove that chemical derivatives are therefore needed to improve clinical efficacy. In other words, we know that compound X doesn't really work, so we'll change it a little bit and see if compound X' works any better (this is a common technique in drug discovery, and if they want to pursue this path that's fine). Two sentences later, almost in the same breath that they admit a lack of clinical efficacy, they claim that nature has evolved a multitude of chemical compounds with desirable properties. They later claim that:
"...natural products as well as traditional medicines have been an undervalued resource of lead structures in the current practice of drug discovery."
A statement that I don't feel is supported by the literature. If anything natural product synthesis has been overvalued by many researchers. Natural product synthesis has been a pretty huge field of research. Even so, many of our most effective drugs on the market today are derived from natural products - Asprin, Digoxin, and Premarin to name a few. Obviously nature has something to offer us. The study even mentions Artesunate, a derivative of a compound originally used in TCM that is now the gold standard for treating malaria.
I'm not saying that this paper should have been rejected outright. Quite the opposite, the technique itself seems fine to me. The computational modeling of protein/compound interactions could allow researchers to quickly determine which compounds are worth studying further. Yes, TCM compounds were used in the study, but that's actually pretty unimportant to the study itself; the method could have been applied to any arbitrary list of compounds.
And I'm not saying that we should discount all alternative medicine. Artesunate (the malaria drug) is one example of why that would be foolish. We can't refuse to study TCM compounds based on principle; we don't know a priori where we'll find relevant compounds. However, I find it completely unnecessary to discuss alternative medicine to the degree that is found in this article. Entire paragraphs are dedicated to the idea of balance as defined by TCM and throughout the article you'll find mention of "synergistic medicines" and other alternative medicine ideas. The authors seem to take every opportunity they can to connect this study to alternative medicine, a connection that I don't think is warranted given that the journal's focus is computational modeling.
Furthermore, proof that some compounds used in TCM are effective does not imply that TCM as a whole is effective at all, which seems to be the general theme of the paper. It seems to me that the authors are trying to sneak alternative medicine in as legitimate science. The paper even contained a diagram like the one above, explaining the principles of balance in TCM - an addition I thought was completely unnecessary and unrelated. Sure, you could say it belongs in the introduction as historical background of the topic, but to me it detracts from the real purpose of the paper - the description of a novel computational method to screen for new target compounds. Leave out the nonsense and get back to the chemistry already.