In silico Identification and Mechanism Exploration of Hepatotoxic Ingredients in Traditional Chinese Medicine

Construction of an explanatory model for predicting hepatotoxicity: a case study of the potentially hepatotoxic components of Gardenia jasminoides

It is well-known that the hepatotoxicity of drugs can significantly influence their clinical use. Despite their effective therapeutic efficacy, many drugs are severely limited in clinical applications due to significant hepatotoxicity. In response, researchers have created several machine learning-based hepatotoxicity prediction models for use in drug discovery and development. Researchers aim to predict the potential hepatotoxicity of drugs to enhance their utility. However, current hepatotoxicity prediction models often suffer from being unverified, and they fail to capture the detailed toxicological structures of predicted hepatotoxic compounds. Using the 56 chemical constituents of Gardenia jasminoides as examples, we validated the trained hepatotoxicity prediction model through literature reviews, principal component analysis (PCA), and structural comparison methods. Ultimately, we successfully developed a model with strong predictive performance and conducted visual validation. Interestingly, we discovered that the predicted hepatotoxic chemical constituents of Gardenia possess both toxic and therapeutic effects, which are likely dose-dependent. This discovery greatly contributes to our understanding of the dual nature of drug-induced hepatotoxicity.

Toxicity evaluation of processing Evodiae fructus based on intestinal microbiota

Background

With the development of healthcare services, drug efficacy, and safety have become the focus of drug use, and processing alters drug toxicity and efficacy, exploring the effects of processing on Evodiae fructus (EF) can guide the clinical use of drugs.

Methods

Fifty male Kunming mice were randomly divided into the control group (CCN), raw small-flowered EF group (CRSEF), raw medium-flowered EF group (CRMEF), processing small-flowered EF group (CPSEF), and processing medium-flowered EF group (CPMEF). The CRSEF, CRMEF, CPSEF, and CPMEF groups were gavaged with aqueous extracts of raw small-flowered EF dry paste (RSEF), medium-flowered EF dry paste (RMEF), processing small-flowered EF dry paste (PSEF) and processing medium-flowered EF dry paste (PMEF), respectively, for 21 days at 5 times the pharmacopeial dosage. Upon concluding the experiment, histopathological sections of liver and kidney tissues were examined. Additionally, levels of aspartate aminotransferase (AST), alanine aminotransferase (ALT), serum creatinine (SCr), and blood urea nitrogen (BUN) were determined. DNA from the intestinal contents of the mice was extracted, and 16S rRNA full-length high-throughput sequencing was performed.

Results

After fed EF 21 days, mice exhibited a decreasing trend in body weight. Comparative analysis with the CCN group revealed an upward trend in SCr, BUN, AST, and ALT levels in both CRSEF and CRMEF groups. The CRMEF group displayed notably elevated BUN and AST levels, with an observed increasing trend in Scr and ALT. Kidney sections unveiled cellular edema and considerable inflammatory cell infiltrates, whereas significant liver damage was not evident. Compared with CRSEF, Bun levels were significantly lower while AST levels were significantly higher in the CPMEF group. Additionally, the intestinal microbiota diversity and the relative abundance of Psychrobacter decreased significantly, and the relative abundance of Staphylococcus, Jeotgalicoccus, and Salinicoccus increased significantly in the CPMEF group. AST, ALT, and SCr were positively correlated with Staphylococcus, Jeotgalicoccus, and Salinicoccus.

Conclusion

In conclusion, PMEF significantly increased harmful bacteria (Staphylococcus, Jeotgalicoccus, and Salinicoccu) and decreased beneficial bacteria. SEF with 5 times the clinical dose showed nephrotoxicity and SEF nephrotoxicity decreased after processing, but EF hepatotoxicity was not significant, which may be due to insufficient dose concentration and time.

Hydroxychloroquine Attenuates hERG Channel by Promoting the Membrane Channel Degradation: Computational Simulation and Experimental Evidence for QT-Interval Prolongation with Hydroxychloroquine Treatment

INTRODUCTION

The coronavirus disease 2019 (COVID-19) pandemic has led to millions of confirmed cases and deaths worldwide and has no approved therapy. Currently, more than 700 drugs are tested in the COVID-19 clinical trials, and full evaluation of their cardiotoxicity risks is in high demand.

METHODS

We mainly focused on hydroxychloroquine (HCQ), one of the most concerned drugs for COVID-19 therapy, and investigated the effects and underlying mechanisms of HCQ on hERG channel via molecular docking simulations. We further applied the HEK293 cell line stably expressing hERG-wild-type channel (hERG-HEK) and HEK293 cells transiently expressing hERG-p.Y652A or hERG-p.F656A mutants to validate our predictions. Western blot analysis was used to determine the hERG channel, and the whole-cell patch clamp was utilized to record hERG current (IhERG).

RESULTS

HCQ reduced the mature hERG protein in a time- and concentration-dependent manner. Correspondingly, chronic and acute treatment of HCQ decreased the hERG current. Treatment with brefeldin A (BFA) and HCQ combination reduced hERG protein to a greater extent than BFA alone. Moreover, disruption of the typical hERG binding site (hERG-p.Y652A or hERG-p.F656A) rescued HCQ-mediated hERG protein and IhERG reduction.

CONCLUSION

HCQ can reduce the mature hERG channel expression and IhERG via enhancing channel degradation. The QT prolongation effect of HCQ is mediated by typical hERG binding sites involving residues Tyr652 and Phe656.

In silico toxicity studies of traditional Chinese herbal medicine: A mini review

With the availability of public databases that store compound-target/compound-toxicity information, and Traditional Chinese medicine (TCM) databases, in silico approaches are used in toxicity studies of TCM herbal medicine. Here, three in silico approaches for toxicity studies were reviewed, which include machine learning, network toxicology and molecular docking. For each method, its application and implementation e.g., single classifier vs. multiple classifier, single compound vs. multiple compounds, validation vs. screening, were explored. While these methods provide data-driven toxicity prediction that is validated in vitro and/or in vivo, it is still limited to single compound analysis. In addition, these methods are limited to several types of toxicity, with hepatotoxicity being the most dominant. Future studies involving the testing of combination of compounds on the front end i.e., to generate data for in silico modeling, and back end i.e., validate findings from prediction models will advance the in silico toxicity modeling of TCM compounds.

Identification of intrinsic hepatotoxic compounds in Polygonum multiflorum Thunb. using machine-learning methods

ETHNOPHARMACOLOGICAL RELEVANCE

Polygonum multiflorum Thunb. (PM) is a herb, extracts of which have been used as Chinese medicine for years. Although it is believed to be beneficial to the liver, heart, and kidneys, it causes idiosyncratic drug-induced liver injury (DILI).

AIM OF THE STUDY

We propose that the intrinsic DILI caused by natural products in PM (NPPM) is an important complementary mechanism to PM-related herb-induced liver injury, and aim to identify the ingredients with high DILI potential by machine learning methods.

MATERIALS AND METHODS

One hundred and ninety-seven NPPM were collected from the literature to identify the intrinsic hepatotoxic compounds. Additionally, a DILI-labeled dataset consisting of 2384 compounds was collected and randomly split into training and test sets. A diparametric optimization method was developed to tune the parameters of extended-connectivity fingerprints (ECFPs), Rdkit, and atom-pair fingerprints as well as those of machine-learning (ML) algorithms. Subsequently, K means were employed to cluster the NPPM that were predicted to have a high DILI risk. An in vitro cell-viability assay was performed using HepaRG cells to validate the prediction results.

RESULTS

ECFPs with the top 35% of features ranked by the F-value with support vector machine (SVM) yielded the best performance. The optimized SVM model achieved an accuracy of 0.761 and recall value of 0.834 on the test dataset. The silico screening for NPPM resulted in 47 ingredients with high DILI potential, which were clustered into six groups based on the elbow method. A representative subgroup that contained 21 ingredients, of which two dianthrones exhibited the lowest IC₅₀ value (0.7-0.9 μM) and anthraquinones showed moderate toxicity (15-25 μM), was constructed.

CONCLUSION

Using ML methods and in vitro screening, two classes of compounds, dianthrones and anthraquinones, were predicted and validated to have a high risk of DILI. The diparametric optimization method used in this study could provide a useful and powerful tool to screen toxicants for large datasets and is available at https://github.com/dreadlesss/Hepatotoxicity_predictor.

Deep learning and machine intelligence: New computational modeling techniques for discovery of the combination rules and pharmacodynamic characteristics of Traditional Chinese Medicine

It has been increasingly accepted that Multi-Ingredient-Based interventions provide advantages over single-target therapy for complex diseases. With the growing development of Traditional Chinese Medicine (TCM) and continually being refined of a holistic view, "multi-target" and "multi-pathway" integration characteristics of which are being accepted. However, its effector substances, efficacy targets, especially the combination rules and mechanisms remain unclear, and more powerful strategies to interpret the synergy are urgently needed. Artificial intelligence (AI) and computer vision lead to a rapidly expanding in many fields, including diagnosis and treatment of TCM. AI technology significantly improves the reliability and accuracy of diagnostics, target screening, and new drug research. While all AI techniques are capable of matching models to biological big data, the specific methods are complex and varied. Retrieves literature by the keywords such as "artificial intelligence", "machine learning", "deep learning", "traditional Chinese medicine" and "Chinese medicine". Search the application of computer algorithms of TCM between 2000 and 2021 in PubMed, Web of Science, China National Knowledge Infrastructure (CNKI), Elsevier and Springer. This review concentrates on the application of computational in herb quality evaluation, drug target discovery, optimized compatibility and medical diagnoses of TCM. We describe the characteristics of biological data for which different AI techniques are applicable, and discuss some of the best data mining methods and the problems faced by deep learning and machine learning methods applied to Chinese medicine.

Xiao-Yao-San protects against anti-tuberculosis drug-induced liver injury by regulating Grsf1 in the mitochondrial oxidative stress pathway

Background: Xiao-Yao-San (XYS) is a traditional Chinese prescription that regulates gastrointestinal function, improves mental and psychological abnormalities, and enhances liver function. However, the underlying mechanism of XYS for relieving anti-tuberculosis (AT) drug-induced liver injury is not clear.

Objective: The current study examined whether XYS alleviated the symptoms of AT drug-induced liver injury in mice via the mitochondrial oxidative stress pathway.

Methods: BALB/c male mice were randomly divided into four groups of 12 animals, including a control group, a model group, a 0.32 g/kg XYS group, and a 0.64 g/kg XYS group. The effect of XYS on the degree of liver injury was observed using haematoxylin and eosin staining (HE) and oil red O staining of pathological sections, biochemical parameters, and reactive oxygen species (ROS) levels. The protein expression of mitochondrial synthesis-related proteins and ferroptosis-related proteins was examined using Western blotting.

Results: XYS improved the pathological changes in liver tissue and reduced the level of oxidative stress in liver-injured mice. XYS increased the expression of mitochondrial synthesis-related proteins and reversed the expression of ferroptosis-related proteins. Knockdown of G-rich RNA sequence binding factor 1 (Grsf1) expression with Grsf1 shRNA blocked the protective effects of XYS in liver injury.

Conclusion: Our findings suggest that XYS alleviates AT drug-induced liver injury by mediating Grsf1 in the mitochondrial oxidative stress pathway.

DenovoProfiling: A webserver for de novo generated molecule library profiling

Various deep learning-based architectures for molecular generation have been proposed for de novo drug design. The flourish of the de novo molecular generation methods and applications has created a great demand for the visualization and functional profiling for the de novo generated molecules. An increasing number of publicly available chemogenomic databases sets good foundations and creates good opportunities for comprehensive profiling of the de novo library. In this paper, we present DenovoProfiling, a webserver dedicated to de novo library visualization and functional profiling. Currently, DenovoProfiling contains six modules: (1) identification & visualization module for chemical structure visualization and identify the reported structures, (2) chemical space module for chemical space exploration using similarity maps, principal components analysis (PCA), drug-like properties distribution, and scaffold-based clustering, (3) ADMET prediction module for predicting the ADMET properties of the de novo molecules, (4) molecular alignment module for three dimensional molecular shape analysis, (5) drugs mapping module for identifying structural similar drugs, and (6) target & pathway module for identifying the reported targets and corresponding functional pathways. DenovoProfiling could provide structural identification, chemical space exploration, drug mapping, and target & pathway information. The comprehensive annotated information could give users a clear picture of their de novo library and could guide the further selection of candidates for chemical synthesis and biological confirmation. DenovoProfiling is freely available at http://denovoprofiling.xielab.net.

Traditional Chinese Medicine Recognition Based on Target Detection

Traditional Chinese medicine (TCM) is widely used in China, but the large variety can easily lead to difficulties in visual identification. This study aims to evaluate the availability of target detection models to identify TCMs. We have collected images of 100 common TCMs in pharmacies, and use three current mainstream target detection models: Faster RCNN, SSD, and YOLO v5 to train the TCM dataset. By comparing the metrics of the three models, the results show that the YOLO v5 model has obvious advantages in the recognition of a variety of TCM, the mean average accuracy of the YOLO v5 is 94.33% and the FPS has reached 75, this model has a smaller number of parameters and solves the problem of detection and occlusion for small targets. Our experiments prove that the target detection technology has broad application prospects in the detection of TCM.

Comparison of Descriptor- and Fingerprint Sets in Machine Learning Models for ADME-Tox Targets

The screening of compounds for ADME-Tox targets plays an important role in drug design. QSPR models can increase the speed of these specific tasks, although the performance of the models highly depends on several factors, such as the applied molecular descriptors. In this study, a detailed comparison of the most popular descriptor groups has been carried out for six main ADME-Tox classification targets: Ames mutagenicity, P-glycoprotein inhibition, hERG inhibition, hepatotoxicity, blood–brain-barrier permeability, and cytochrome P450 2C9 inhibition. The literature-based, medium-sized binary classification datasets (all above 1,000 molecules) were used for the model building by two common algorithms, XGBoost and the RPropMLP neural network. Five molecular representation sets were compared along with their joint applications: Morgan, Atompairs, and MACCS fingerprints, and the traditional 1D and 2D molecular descriptors, as well as 3D molecular descriptors, separately. The statistical evaluation of the model performances was based on 18 different performance parameters. Although all the developed models were close to the usual performance of QSPR models for each specific ADME-Tox target, the results clearly showed the superiority of the traditional 1D, 2D, and 3D descriptors in the case of the XGBoost algorithm. It is worth trying the classical tools in single model building because the use of 2D descriptors can produce even better models for almost every dataset than the combination of all the examined descriptor sets.

Development of binary classification models for assessment of drug-induced liver injury in humans using a large set of FDA-approved drugs

Drug-induced liver injury (DILI) has been identified as one of the major causes for drugs withdrawn from the market, and even termination during the late stages of development. Therefore, it is imperative to evaluate the DILI potential of lead compounds during the research and development process. Although various computational models have been developed to predict DILI, most of which applied the DILI data were extracted from preclinical sources. In this investigation, the in silico prediction models for DILI were constructed based on 1140 FDA-approved drugs by using naïve Bayes classifier approach. The genetic algorithm method was applied for the molecular descriptors selection. Among these established prediction models, the NB-11 model based on eight molecular descriptors combined with ECFP_18 showed the best prediction performance for DILI, which gave 91.7% overall prediction accuracy for the training set, and 68.9% concordance for the external test set. Therefore, the established NB-11 prediction model can be used as a reliable virtual screening tool to predict DILI adverse effect in the early stages of drug design. In addition, some new structural alters for DILI were identified, which could be used for structural optimization in the future drug design by medicinal chemists.

iANP-EC: Identifying Anticancer Natural Products Using Ensemble Learning Incorporated with Evolutionary Computation

Cancer is one of the most deadly diseases that annually kills millions of people worldwide. The investigation on anticancer medicines has never ceased to seek better and more adaptive agents with fewer side effects. Besides chemically synthetic anticancer compounds, natural products are scientifically proved as a highly potential alternative source for anticancer drug discovery. Along with experimental approaches being used to find anticancer drug candidates, computational approaches have been developed to virtually screen for potential anticancer compounds. In this study, we construct an ensemble computational framework, called iANP-EC, using machine learning approaches incorporated with evolutionary computation. Four learning algorithms (k-NN, SVM, RF, and XGB) and four molecular representation schemes are used to build a set of classifiers, among which the top-four best-performing classifiers are selected to form an ensemble classifier. Particle swarm optimization (PSO) is used to optimise the weights used to combined the four top classifiers. The models are developed by a set of curated 997 compounds which are collected from the NPACT and CancerHSP databases. The results show that iANP-EC is a stable, robust, and effective framework that achieves an AUC-ROC value of 0.9193 and an AUC-PR value of 0.8366. The comparative analysis of molecular substructures between natural anticarcinogens and nonanticarcinogens partially unveils several key substructures that drive anticancerous activities. We also deploy the proposed ensemble model as an online web server with a user-friendly interface to support the research community in identifying natural products with anticancer activities.

Chemical Distance Measurement and System Pharmacology Approach Uncover the Novel Protective Effects of Biotransformed Ginsenoside C-Mc against UVB-Irradiated Photoaging

Long-term exposure to ultraviolet light induces photoaging and may eventually increase the risk of skin carcinogenesis. Rare minor ginsenosides isolating from traditional medicine Panax (ginseng) have shown biomedical efficacy as antioxidation and antiphotodamage agents. However, due to the difficulty of component extraction and wide variety of ginsenoside, the identification of active antiphotoaging ginsenoside remains a huge challenge. In this study, we proposed a novel in silico approach to identify potential compound against photoaging from 82 ginsenosides. Specifically, we calculated the shortest distance between unknown and known antiphotoaging ginsenoside set in the chemical space and applied chemical structure similarity assessment, drug-likeness screening, and ADMET evaluation for the candidates. We highlighted three rare minor ginsenosides (C-Mc, Mx, and F2) that possess high potential as antiphotoaging agents. Among them, C-Mc deriving from American ginseng (Panax quinquefolius L.) was validated by wet-lab experimental assays and showed significant antioxidant and cytoprotective activity against UVB-induced photodamage in human dermal fibroblasts. Furthermore, system pharmacology analysis was conducted to explore the therapeutic targets and molecular mechanisms through integrating global drug-target network, high quality photoaging-related gene profile from multiomics data, and skin tissue-specific expression protein network. In combination with in vitro assays, we found that C-Mc suppressed MMP production through regulating the MAPK/AP-1/NF-κB pathway and expedited collagen synthesis via the TGF-β/Smad pathway, as well as enhanced the expression of Nrf2/ARE to hold a balance of endogenous oxidation. Overall, this study offers an effective drug discovery framework combining in silico prediction and in vitro validation, uncovering that ginsenoside C-Mc has potential antiphotoaging properties and might be a novel natural agent for use in oral drug, skincare products, or functional food.

Comprehensive assessment of side effects in COVID-19 drug pipeline from a network perspective

Currently, coronavirus disease 2019 (COVID-19), has posed an imminent threat to global public health. Although some current therapeutic agents have showed potential prevention or treatment, a growing number of associated adverse events have occurred on patients with COVID-19 in the course of medical treatment. Therefore, a comprehensive assessment of the safety profile of therapeutic agents against COVID-19 is urgently needed. In this study, we proposed a network-based framework to identify the potential side effects of current COVID-19 drugs in clinical trials. We established the associations between 116 COVID-19 drugs and 30 kinds of human tissues based on network proximity and gene-set enrichment analysis (GSEA) approaches. Additionally, we focused on four types of drug-induced toxicities targeting four tissues, including hepatotoxicity, renal toxicity, lung toxicity, and neurotoxicity, and validated our network-based predictions by preclinical and clinical evidence available. Finally, we further performed pharmacovigilance analysis to validate several drug-tissue toxicities via data mining adverse event reporting data, and we identified several new drug-induced side effects without labeling in Food and Drug Administration (FDA) drug instructions. Overall, this study provides forceful approaches to assess potential side effects on COVID-19 drugs, which will be helpful for their safe use in clinical practice and promoting the discovery of antiviral therapeutics against SARS-CoV-2.

Systems pharmacology-based approach to investigate the mechanisms of Danggui-Shaoyao-san prescription for treatment of Alzheimer's disease

Background

Alzheimer’s disease (AD) is the most common cause of dementia in the elderly, characterized by a progressive and irreversible loss of memory and cognitive abilities. Currently, the prevention and treatment of AD still remains a huge challenge. As a traditional Chinese medicine (TCM) prescription, Danggui-Shaoyao-san decoction (DSS) has been demonstrated to be effective for alleviating AD symptoms in animal experiments and clinical applications. However, due to the complex components and biological actions, its underlying molecular mechanism and effective substances are not yet fully elucidated.

Methods

In this study, we firstly systematically reviewed and summarized the molecular effects of DSS against AD based on current literatures of in vivo studies. Furthermore, an integrated systems pharmacology framework was proposed to explore the novel anti-AD mechanisms of DSS and identify the main active components. We further developed a network-based predictive model for identifying the active anti-AD components of DSS by mapping the high-quality AD disease genes into the global drug-target network.

Results

We constructed a global drug-target network of DSS consisting 937 unique compounds and 490 targets by incorporating experimental and computationally predicted drug–target interactions (DTIs). Multi-level systems pharmacology analyses revealed that DSS may regulate multiple biological pathways related to AD pathogenesis, such as the oxidative stress and inflammatory reaction processes. We further conducted a network-based statistical model, drug-likeness analysis, human intestinal absorption (HIA) and blood-brain barrier (BBB) penetration prediction to uncover the key ani-AD ingredients in DSS. Finally, we highlighted 9 key ingredients and validated their synergistic role against AD through a subnetwork.

Conclusion

Overall, this study proposed an integrative systems pharmacology approach to disclose the therapeutic mechanisms of DSS against AD, which also provides novel in silico paradigm for investigating the effective substances of complex TCM prescription.

A Systems Pharmacology Approach for Identifying the Multiple Mechanisms of Action for the Rougui-Fuzi Herb Pair in the Treatment of Cardiocerebral Vascular Diseases

Cardiocerebral vascular diseases (CCVDs) are the main reasons for high morbidity and mortality all over the world, including atherosclerosis, hypertension, myocardial infarction, stroke, and so on. Chinese herbs pair of the Cinnamomum cassia Presl (Chinese name, rougui) and the Aconitum carmichaelii Debx (Chinese name, fuzi) can be effective in CCVDs, which is recorded in the ancient classic book Shennong Bencao Jing, Mingyibielu and Thousand Golden Prescriptions. However, the active ingredients and the molecular mechanisms of rougui-fuzi in treatment of CCVDs are still unclear. This study was designed to apply a system pharmacology approach to reveal the molecular mechanisms of the rougui-fuzi anti-CCVDs. The 163 candidate compounds were retrieved from Traditional Chinese Medicine System Pharmacology Database and Analysis Platform (TCMSP). And 84 potential active compounds and the corresponding 42 targets were obtained from systematic model. The underlying mechanisms of the therapeutic effect for rougui-fuzi were investigated with gene ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis. Then, component-target-disease (C-T-D) and target-pathway (T-P) networks were constructed to further dissect the core pathways, potential targets, and active compounds in treatment of CCVDs for rougui-fuzi. We also constituted protein-protein in interaction (PPI) network by the reflect target protein of the crucial pathways against CCVDs. As a result, 21 key compounds, 8 key targets, and 3 key pathways were obtained for rougui-fuzi. Afterwards, molecular docking was performed to validate the reliability of the interactions between some compounds and their corresponding targets. Finally, UPLC-Q-Exactive-MS^E and GC-MS/MS were analyzed to detect the active ingredients of rougui-fuzi. Our results may provide a new approach to clarify the molecular mechanisms of Chinese herb pair in treatment with CCVDs at a systematic level.

A Computational Toxicology Approach to Screen the Hepatotoxic Ingredients in Traditional Chinese Medicines: Polygonum multiflorum Thunb as a Case Study

In recent years, liver injury induced by Traditional Chinese Medicines (TCMs) has gained increasing attention worldwide. Assessing the hepatotoxicity of compounds in TCMs is essential and inevitable for both doctors and regulatory agencies. However, there has been no effective method to screen the hepatotoxic ingredients in TCMs available until now. In the present study, we initially built a large scale dataset of drug-induced liver injuries (DILIs). Then, 13 types of molecular fingerprints/descriptors and eight machine learning algorithms were utilized to develop single classifiers for DILI, which resulted in 5416 single classifiers. Next, the NaiveBayes algorithm was adopted to integrate the best single classifier of each machine learning algorithm, by which we attempted to build a combined classifier. The accuracy, sensitivity, specificity, and area under the curve of the combined classifier were 72.798, 0.732, 0.724, and 0.793, respectively. Compared to several prior studies, the combined classifier provided better performance both in cross validation and external validation. In our prior study, we developed a herb-hepatotoxic ingredient network and a herb-induced liver injury (HILI) dataset based on pre-clinical evidence published in the scientific literature. Herein, by combining that and the combined classifier developed in this work, we proposed the first instance of a computational toxicology to screen the hepatotoxic ingredients in TCMs. Then Polygonum multiflorum Thunb (PmT) was used as a case to investigate the reliability of the approach proposed. Consequently, a total of 25 ingredients in PmT were identified as hepatotoxicants. The results were highly consistent with records in the literature, indicating that our computational toxicology approach is reliable and effective for the screening of hepatotoxic ingredients in Pmt. The combined classifier developed in this work can be used to assess the hepatotoxic risk of both natural compounds and synthetic drugs. The computational toxicology approach presented in this work will assist with screening the hepatotoxic ingredients in TCMs, which will further lay the foundation for exploring the hepatotoxic mechanisms of TCMs. In addition, the method proposed in this work can be applied to research focused on other adverse effects of TCMs/synthetic drugs.