Network Pharmacology and Other Modern Methods in TCM Research
What network pharmacology is, how it is used to study Chinese herbal formulas, its known weaknesses, and why its computer predictions are hypotheses that still need lab and clinical testing.
Content last updated . This page has not been medically reviewed.
In one paragraph
Network pharmacology is a computer-based way of mapping which chemicals in a medicine might act on which proteins and pathways in the body. Researchers often use it on Chinese herbal formulas, which contain many herbs and many compounds. It is a way to generate hypotheses. A network diagram is not a clinical trial, and it does not show that a formula works in people.Evidence labels: preclinical (lab or computer findings), hypothesis generation, established (human trials).
Where the idea comes from
A 2025 review in the journal Chinese Medicine says the concept was introduced in 2007 by the British pharmacologist Andrew Hopkins, who proposed that drugs act through several targets within biological networks rather than one. [1] A 2013 article by Li and Zhang described a "TCM network pharmacology" approach that shifts from "one target, one drug" to "network target, multi-components", with methods to predict the targets of herbal compounds and to interpret combined effects of formulas. The authors argue it could help move TCM from experience-based toward evidence-based medicine. [2] That is the authors' aim and view, not a demonstrated result.What a typical study does
Based on the 2025 review, a typical study:- Lists the compounds in an herb or formula from public databases, which also hold predicted target information.
- Filters them, often by computed properties such as oral bioavailability and drug-likeness.
- Uses network analysis to pick out key components and targets, building an "herbs, components, targets, diseases" network. [1]
A 2019 review notes that researchers have conducted both computational and experimental network studies of herbal formulas; the computer step on its own is only a prediction. [3]
Known limits
| Problem | What the source says |
|---|---|
| Missing evaluation framework | The 2025 review says the absence of an evaluation framework is a critical factor undermining reliability; a "Guidelines for Network Pharmacology Evaluation Methods" was released in 2021. [1] |
| Data collection | Databases lack standardized data collection and processed herbs are insufficiently considered. [1] |
| Which compounds matter | Screening by computed properties often disregards how much of each compound is actually in the formula, so the compounds picked may not be the main or most relevant ones. [1] |
| Database quality | A 2019 review that compared TCM databases using 10 well-known herbs concluded they are powerful and comprehensive but need further improvements. [3] |
| Reliability of outcomes | The 2025 review states that questions persist about the reliability of study outcomes. [1] |
Because of these limits, two studies of the same formula can predict different targets. A predicted target is a lead to be tested, not a finding.
What it can and cannot tell you
- Can: suggest which compounds or pathways deserve laboratory study, help plan quality-control markers, and organize large amounts of data. [1]
- Cannot: show a formula is safe, show a dose, show benefit in people, or support the claim that a traditional pattern (such as "dampness") corresponds to a molecular network. Those need clinical trials with comparators; see Placebo, sham controls and blinding in TCM trials.
How to read a headline
If a news story says "scientists discovered how a Chinese formula works" and the study is network pharmacology, check whether it includes animal or human data, whether the formula was chemically analyzed, and whether a trial exists. Our guide How to read a TCM study lists these questions. For the difference between traditional frameworks and biomedical explanations see Traditional frameworks vs biomedical explanations.Do not change or stop prescribed medicines because of a computer-predicted herbal effect. Ask a clinician or pharmacist about any herbal product, particularly if you take other medicines.
Frequently asked questions
Is network pharmacology proof that a herbal formula works?
No. It predicts possible targets from databases. Human benefit requires clinical trials. [1]
Why is it popular in TCM research?
Its "multi-component, multi-target" idea matches how formulas are described, and its authors present it as a way to study formulas as a whole. [2]
Who invented network pharmacology?
The 2025 review credits Andrew Hopkins with introducing the concept in 2007. [1]
Are there standards for these studies?
A guideline for evaluation methods was released in 2021, but the 2025 review still describes the lack of an evaluation framework as a core reliability problem. [1]
Does it replace lab or clinical research?
No. The same review notes that traditional work on the material basis of efficacy used extraction, purification and experimental validation, and presents network pharmacology as a faster complement. [1]
References
- Zhai Y et al. Network pharmacology: a crucial approach in traditional Chinese medicine research. Chin Med 2025. Read 2026-10-22.
- Li S, Zhang B. Traditional Chinese medicine network pharmacology: theory, methodology and application. Chin J Nat Med 2013. Read 2026-10-22.
- Zhang R et al. Network Pharmacology Databases for Traditional Chinese Medicine: Review and Assessment. Front Pharmacol 2019. Read 2026-10-22.