Applying systems biology to validate prescientific quackery is tooth fairy science

Although I did not coin it, I frequently use the term “quackademic medicine” because, unfortunately, there’s a lot of quackademic medicine around. Although regular readers know what the term means, i always feel obligated to briefly explain what quackademic medicine is, for the benefit of any newbie who might happen upon this blog. Basically, it is a term used to describe an increasingly common and alarming phenomenon, the infiltration of rank quackery and pseudoscience into medical academia. You might think it impossible or unbelievable, but it’s anything but. Beginning around a quarter of a century ago, alternative medicine started finding its way into medical schools and academic medical centers, first in the form of “complementary and alternative medicine” (CAM) and then later renamed “integrative medicine.” The intent of renaming it “integrative medicine” was to portray the idea of “integrating” the “best of both worlds” (alternative medicine and conventional medicine) as though they were equals, but in reality all integrative medicine does is to integrate quackery with real medicine; that is, when it isn’t busy co-opting perfectly science-based modalities, such as exercise, nutrition, and lifestyle changes, as somehow being “alternative” and “integrated” into medicine.

One of the most pernicious effects of quackademic medicine is not just how it places the imprimatur of real medicine on sheer quackery like acupuncture, much of traditional Chinese medicine (TCM), reflexology, naturopathy, homeopathy, and the like. That’s bad enough. What’s just as bad, though, is how it distorts research, wastes money, and encourages research into treatments that are so implausible that it is unethical to subject man or beast to them in the name of science. I’ve listed many examples before, the first example coming to mind being the clinical trials of homeopathy for infectious childhood diarrhea in Third World countries, acupuncture for menopausal hot flashes, or acupuncture for in vitro fertilization. Then there are the many experiments subjecting rodents to various quackery, such as this one torturing rats in the name of pseudoscience by shocking them with electroacupuncture.

A while ago, I noted that the open access journal, PLoS One seemed to have a penchant for publishing horrible studies of quackery, such as one purporting to identify anatomic correlates of acupuncture meridians and functional MRI (fMRI) studies of laser acupuncture. Of course, it’s not just PLoS one, but a lot of journals, even journals within the Nature stable of journals, such as Scientific Reports. I’m referring to an article that its authors, Sun et al, entitled To Unveil the Molecular Mechanisms of Qi and Blood through Systems Biology-Based Investigation into Si-Jun-Zi-Tang and Si-Wu-Tang formulae. It might better have been entitled Disguising pharmacognosy research using crude plant extracts as evidence for qi and the five element theory of medicine.

Basically, this study uses various bioinformatics databases of compounds and their molecular targets in the cell, does a bunch of analyses that use the known effects of various ingredients in the TCM compounds to predict what molecular signaling pathways they affect. They used TCMID Cytoscape thusly:

Since the TCM formulae are normally composed of several medicinal herbs, and each herb normally has many ingredients, and each ingredient has a lot of targets, a formula is a complex biologic active network. Fortunately, along with the rapid development of life science and computer science, a variety of computational tools and bioinformatic database have been developed to facilitate the analysis of a large number of genes associated with complex ingredients of TCM formulae2, which provide opportunities to predict potential pharmacological actions of TCM formulae and clarify complex molecular mechanisms of formulae and theories of TCM. Based on primary biomolecular databases, e.g. Kyoto Encyclopedia of Genes and Genomes (KEGG,, HPRD4, PDB5, TTD6, OMIM7, Drug-Bank8, STITCH9 and ChEMBL10, a lot of TCM-related databases have been developed, such as TCMID11, HIT12, TCM [email protected], TCMGeneDIT14, TCM-ID15, TCMSP16 and CHMIS-C17. These TCM-related databases complement each other to provide information on complex interactions of TCM-active ingredient-gene-disease2. Among these TCM-related databases, TCMID ( contains 3,791 diseases, 47,000 prescriptions, 8,159 herbs, 6,828 drugs, 25,210 compounds and 17,521 related targets, which facilitates the study of interactions between formula, ingredient, gene and disease to uncover the molecular biological mechanisms of TCM. Meanwhile, there are several network analysis tools for biological functionality of TCM-related network analysis, such as Cytoscape18,19. More than 150 specialized plugins integrated in Cytoscape can be used to import and map existing interaction data cataloged in public databases2, such as ClueGO20, BioGrid Plugin21 and MiMI22. ClueGO integrates Gene Ontology (GO) terms as well as KEGG/BioCarta pathways to create functionally organized GO/pathway term network and analyze one or compare two lists of genes and comprehensively visualizes functionally grouped terms.

I’ll translate. Basically, it is now possible to use biomolecular databaes, databases of molecular targets and their interactions with various molecules, and sophisticated bioinformatics tools like Cytoscape to import and map known interaction data from public databases and use network analyses to model and identify likely signaling pathways affected by the molecules. This list can be shaped and narrowed down by other databases cataloguing targets and signaling pathways known to be involved with specific diseases and conditions. All of this has become fairly standard in drug development, both in pharma and academia. It’s basically science.

But what are these sophisticated bioinformatics techniques being used for in this study. Here’s where it goes off the deep end. Now, I frequently refer to studies like this as “tooth fairy science.” It’s a term coined by Harriet Hall to describe science purporting to measure the properties of a phenomenon whose existence has not been established. Basically, you can study how much money the tooth fairy leaves each night, but you haven’t established the existence of the tooth fairy. Similarly, you can study the “effects” of homeopathy, but you haven’t established that homeopathy is anything more than water. Believe it or not, this isn’t tooth fairy science, as you will see. However, it is something almost as bad.

First, take a look at this figure to see how Sun et al set up their experiment (click to embiggen):

CAM systems biology

Yes, on the left, there is “Qi deficiency” on the left and “blood deficiency” on the right. Both are TCM diagnoses, not diagnoses that have anything to do with biology or medical reality, given that qi is a term to describe the “life energy.” I sometimes wonder if qi deficiency is caused by excessive use by the alien healing machine in Babylon 5 to heal others (that’s a joke only real Babylon 5 geeks will get). Be that as it may, on the one side, you have a diagnosis based on a prescientific concept of how the body works in an ancient Eastern medical system resembling ancient European humoral theory treated with an herbal remedy called Si-Jun-Zi-Tang (SJZT) based on that same prescientific understanding. On the other side you have a different diagnosis based on a prescientific concept of how the body works in an ancient Eastern medical system resembling ancient European humoral theory treated with another herbal remedy called Si-Wu-Tang (SWT) based on that same prescientific understanding. Each is then run through TCMID Cytoscape to find what is described as an herb-ingredient gene network. The two are compared using ClueGO KEGG analysis of predicted targets of SWT and SJZT. It’s not important that you understand what that is other than that it’s a computer analysis. You then get fancy network analysis charts like this, in which red nodes represent terms of SWT; green nodes represent terms (pathways) of SJZT; grey nodes represent common terms of the two formulae (again, click to embiggen):

Systems biology of woo

A systems biology map of woo masquerading as pharmacognosy.

It’s not important that you know exactly what this chart means. All it shows is basically how the two herbs affect different signaling pathways. Basically, this is nothing but pharmacognosy (natural products pharmacology) using very “dirty” natural products (raw herbs with active compounds not isolated or otherwise purified). The authors then use this to try to convince you that the pharmacognosy investigations performed in this paper somehow validate the prescientific mystical beliefs of TCM, but only after first trying to convince you that TCM foresaw how modern medicine would work:

Combination therapy is the major feature of TCM, which is increasingly recognized by modern western medicine, such as cocktail therapy for HIV27 and the opinion shifting from targeting a single disease-causing molecule to the pursuit of combination therapies that comprise more than one active ingredient28. According to the symptoms of patients, different kinds of Chinese medicines are combined to form formulae to improve clinical efficacy1. such as SWT and SJZT have been used to rectify blood deficiency and Qi deficiency respectively for about 1,000 years.

This is, of course, utter nonsense. First, it invokes the false dichotomy of “Eastern” versus “Western” medicine. There is no such thing. There is scientific medicine, which knows no “East” or “West.” There is also quackery, which, unfortunately, also knows no “East” or “West.” (After all, homeopathy is “Western” medicine, having been invented, as it was, in Germany by Samuel Hahnemann over 200 years ago.) That’s just a pet peeve of mine, though. Scientific medicine has always used combination therapy, whether antibiotics to try to eliminate an infection before resistance evolves, the many, many multidrug combinations used in the chemotherapy of cancer, or even something as mundane as treating hypertension with more than one drug. There is nothing unusual or new about this. While it is true that scientific medicine did at one time tend to look for single genes that could be targeted, my perception is that the reason that changed was not so much a revelation that those ancient Chinese were right after all but rather as a result of the increasingly sophisticated technology that permits the study of gene networks and changes in the expression of every gene in the genome simultaneously instead of the levels of individual genes. TCM advocates would have you believe that TCM was there first, looking at individuals in a “holistic fashion,” but that’s just nonsense. TCM came by its beliefs because humans didn’t know any better and didn’t have the tools at the time to know better.

I used to call the co-opting of modern genomic techniques as “woo-omics,” to go along with genomics, metabolomics, and all the other “omics” in modern medical science, and TCM is not alone in this. This study is a perfect example of woo-omics or woo systems analysis. It’s perfectly fine as pharmacognosy, although most scientists in drug development would not be too thrilled with the use of crude extracts, but it is no evidence that the precepts of TCM have anything to do with reality. None of this stops the authors from arguing just that:

Blood deficiency normally manifests anaemia, vertigo, heart palpitations and menstrual discomfort. The SWT formula has effects on stimulating hematopoiesis in bone marrow, anti-coagulant, vasodilatation and sedative, so it can be used to treat anemia, bone formation dysmenorrhea and other estrogen-related diseases. Qi deficiency normally manifests lack of strength, body function decline and decreased disease resistance, and so on. The SZJT formula has effects on regulating granulocyte macrophage colony-stimulating factor secretion, enhancing phagocytosis of macrophages, recovering cAMP signal pathway and recovery of intestinal microflora. However, the mechanisms of the pharmacological action of SWT and SJZT have not yet been clarified.


In conclusion, SWT with the functions of influencing amino acid and carbohydrate metabolism is significant different from SJZT with the actions of influencing neuroendocrine system by affecting excitatory synapses (serotonergic synapse) to regulate thyroid hormone, ovarian steroidogenesis, prolactin and oxytocin secrction and can regulate corrective hormones secretion and promoting vascular smooth muscle contraction. The common effects of SWT and SJZT are regulating the functional activities of the hypothalamic pituitary adrenal and gonadal axis by affect inhibitory synapses (GABAergic synapse) to stimulate estrogen and steroid hormones secretion and strengthening the ability of anti-infection. All the differences and common pathways also reflect the characteristics of blood deficiency and Qi deficiency, and the molecular mechanism of blood and Qi of TCM.

Molecular mechanism. You keep using that term. I do not think it means what you think it means.

Seriously, their argument boils down to being able to vaguely relate some of the molecular pathways apparently affected by SWT and SJZT to the TCM diagnoses of qi deficiency and blood deficiency and using that vague relation to claim that the precepts of TCM are valid and that this study has provided evidence in support of those precepts. This is a feature, not a bug, of studies like this, and this is not the first such study. Not by a long shot.

Hmmm. Maybe I was wrong. Maybe this is tooth fairy science, after all. Strike that. There’s no “maybe” about it.