Application of Data Fusion in Traditional Chinese Medicine: A Review.

Traditional Chinese medicine is characterized by numerous chemical constituents, complex components, and unpredictable interactions among constituents. Therefore, a single analytical technique is usually unable to obtain comprehensive chemical information. Data fusion is an information processing technology that can improve the accuracy of test results by fusing data from multiple devices, which has a broad application prospect by utilizing chemometrics methods, adopting low-level, mid-level, and high-level data fusion techniques, and establishing final classification or prediction models. This paper summarizes the current status of the application of data fusion strategies based on spectroscopy, mass spectrometry, chromatography, and sensor technologies in traditional Chinese medicine (TCM) in light of the latest research progress of data fusion technology at home and abroad. It also gives an outlook on the development of data fusion technology in TCM analysis to provide references for the research and development of TCM.

[Manufacturing classification system for oral solid dosage forms of traditional Chinese medicines (â ¡): classification of tablets disintegration behavior].

In this paper, 50 batches of representative traditional Chinese medicine tablets were selected and the disintegration time was examined with the method in Chinese Pharmacopoeia. The disintegration time and disintegration phenomenon were recorded, and the dissolution behaviors of water-soluble and ultraviolet-absorbent components during the disintegration process of tablets were characterized by self-control method. The results revealed that coating type and raw material type influenced the disintegration time of tablets. It was found that only 4% of traditional Chinese medicine tablets had obvious fragmentation during the disintegration process, while 96% of traditional Chinese medicine tablets showed gradual dissolution or dispersion. Furthermore, according to the disintegration speed, disintegration phenomenon, and whether the cumulative dissolution of measured components was > 90% at complete disintegration, a disintegration behavior classification system(DBCS) was created for the regular-release traditional Chinese medicine tablets. As a result, the disintegration behaviors of 50 batches of traditional Chinese medicine tablets were classified into four categories, i.e. ⅠA_2, ⅠB_1, ⅡB_1, and ⅡB_2. traditional Chinese medicine tablets(Class I) with disintegration time ≤ 30 min were defined to be rapid in disintegration, which can be the objective of optimization or improvement of Chinese herbal extract(semi extract) tablets. Different drug release models were used to fit the dissolution curve of traditional Chinese medicine tablets with gradual dissolution or dispersion phenomenon(i.e. Type B tablets). The results showed that the dissolution curves of water-soluble components in the disintegration process conformed to the zero order kinetics and the Ritger-Peppas model. It could be inferred that the disintegration mechanisms of type B tablets were a combination of dissolution controlled and swelling controlled mechanisms. This study contributes to understanding the disintegration behavior of traditional Chinese medicine tablets, and provides a reference for the design and improvement of disintegration performance of traditional Chinese medicine tablets.

[Comparison of alkaloids in Aconiti Kusnezoffii Radix, Aconiti Radix, and Aconiti Lateralis Radix Praeparata based on UHPLC-Q-Exactive Orbitrap MS/MS].

UHPLC-Q-Exactive Orbitrap MS/MS was used to systematically analyze and compare the alkaloids in Aconiti Kusnezoffii Radix, Aconiti Radix, and Aconiti Lateralis Radix Praeparata. After the samples were pretreated in the solid-phase extraction cartridges, 0.1% ammonium hydroxide(A)-acetonitrile(B) was used for gradient elution. The LC-MS method for characterization of alkaloids in the three herbal medicines was established in ESI positive ion mode to collect high resolution MS data of reference substances and samples. On the basis of the information of reference substance cracking behavior, retention time, accurate molecular mass, and related literature, a total of 155 alkaloids were identified in Aconiti Kusnezoffii Radix, Aconiti Radix, and Aconiti Lateralis Radix Prae-parata. Specifically, 130, 127, and 92 alkaloids were identified in Aconiti Kusnezoffii Radix, Aconiti Radix, and Aconiti Lateralis Radix Praeparata, respectively. Monoester alkaloids and amino-alcohol alkaloids were dominant in the three herbal medicines, and the alkaloids in Aconiti Kusnezoffii Radix and Aconiti Radix were similar. This paper can provide a reference for elucidating the pharmacological effects and clinical application differences of the three herbal medicines produced from plants of Aconitum.

[Evaluation and classification of dissolution behavior and capability of Chinese medicine granules based on an inline turbidity sensor].

In this paper, the inline turbidity sensor technology was used to quantify the turbidity of the solution during the dissolution of Chinese medicine granules. The probe measurement position and the magnetic stirring speed were optimized. As a result, the stirring speed was 400 r·min~(-1), and the probe position was at 1/4 of the diameter of the beaker. The measurement results were accurate and reliable. Totally 105 batches of commercially available Chinese medicine granules were collected and dissolved according to the requirements of the Chinese Pharmacopoeia. At the time point of 5 min, 57 batches of granules were completely dissolved, and the corresponding turbidity values ranged between 0-70 FTU; 32 batches of granules showed a slight turbidity, and the corresponding turbidity values ranged between 70-350 FTU; 14 batches of granule solution were turbid, and the corresponding turbidity values ranged between 350-2 000 FTU; two batches of granule solution were heavily turbid, and the corresponding turbidity values were >2 000 FTU. Among the above results, the number of batches in line with the pharmacopoeia dissolution requirement was 84.76%, and the dissolution of some granules still needed to be improved. The turbidity sensor recorded the change curve of turbidity value over time(solubility behavior curve). The degree of important of disintegration and dissolution during the dissolution process showed disintegration > dissolution, disintegration≈dissolution, disintegration < dissolution. The dissolution behavior of the granules can be classified into three categories. The analysis of the mechanism in the process of granule solubility provides a basis for product process improvement.

[Mechanism of "herb soaking with exact amount of water" during moistening process of ginseng based on needle pressure sensor].

In this study, the texture analyzer acupuncture pressure sensor was used to objectively characterize the "herb soaking with exact amount of water" for moistening process of ginseng. The single factor rotation experiment was used to investigate the effects of puncture speed, puncture depth and puncture site on puncture force and work. According to ginseng processing method in Chinese Pharmacopoeia, ginseng medicinal materials with diameters of about 1 cm and 2 cm were selected, and puncture experiments were carried out at the set measurement time to determine the hardness, work and water absorption of the ginseng moistening process. The endpoint threshold for the ginseng softening process was determined and verified. To reflect the actual internal conditions of the ginseng softening process, the puncture depth was preferably 70%, and the puncture speed was 30 mm·min~(-1). In the ginseng moistening process, the softening hardness and the puncture work were in accordance with the first-order kinetic equation y=a×exp(-k×x). The 0 h initial hardness a of 1 cm and 2 cm ginseng herbs were 289.8 N and 1 227 N, and the rate constants K were 0.149 4 N·h~(-1) and 0.100 7 N·h~(-1), respectively. After the ginseng was completely softened, the force required for puncture was 10 N, which can be used as the standard for "drug penetration". At this time, the water absorption rate of ginseng was 70%-100%. The softening time of ginseng with a diameter of 1 cm was about 20-22 h, and the softening time of ginseng with a diameter of 2 cm was about 40-46 h. A needle-type pressure sensor was used to accurately determine the end point of the softening process of ginseng and reduce the loss of active ingredients. The study results provide reference for the softening process kinetics and the process intelligent monitoring of other dried roots and rhizomes.

[Key technologies and applications of industrial big data in manufacturing of Chinese medicine].

Along with the striding of the Chinese medicine(CM) manufacturing toward the Industry 4.0, some digital factories have accumulated lightweight industrial big data, which become part of the enterprise assets. These digital assets possess the possibility of solving the problems within the CM production system, like the Sigma gap and the poverty of manufacturing knowledge. From the holistic perspective, a three-tiered architecture of CM industrial big data is put forward, and it consists of the data integration layer, the data analysis layer and the application scenarios layer. In data integration layer, sensing of CM critical quality attributes is the key technology for big data collection. In data analysis and mining layer, the self-developed iTCM algorithm library and model library are introduced to facilitate the implementation of the model lifecycle methodologies, including process model development, model validation, model configuration and model maintenance. The CM quality transfer structure is closely related with the connection mode of multiple production units. The system modeling technologies, such as the partition-integration modeling method, the expanding modeling method and path modeling method, are key to mapping the structure of real manufacturing system. It is pointed out that advance modeling approaches that combine the first-principles driven and data driven technologies are promising in the future. At last, real-world applications of CM industrial big data in manufacturing of injections, oral solid dosages, and formula particles are presented. It is shown that the industrial big data can help process diagnosis, quality forming mechanism interpretations, real time release testing method development and intelligent product formulation design. As renewable resources, the CM industrial big data enable the manufacturing knowledge accumulation and product quality improvement, laying the foundation of intelligent manufacturing.

Optimal Selection of Incoming Materials from the Inventory for Achieving the Target Drug Release Profile of High Drug Load Sustained-Release Matrix Tablet.

In the pharmaceutical process, raw material (including APIs and excipients) variability can be delivered to the final product, and lead to batch-to-batch and lot-to-lot variances in its quality, finally impacting the efficacy of the drug. In this paper, the Panax notoginseng saponins (PNS) sustained-release matrix tablet was taken as the model formulation. Hydroxypropyl methylcellulose with the viscosity of 4000 mPa·s (HPMCK4M) from different vendors and batches were collected and their physical properties were characterized by the SeDeM methodology. The in-vitro dissolution profiles of active pharmaceutical ingredients (APIs) from matrix tablets made up of different batches HPMC K4M displayed significant variations. Multi-block partial least squares (MB-PLS) modeling results further demonstrated that physical properties of excipients played dominant roles in the drug release. In order to achieve the target drug release profile with respect to those far from the criteria, the optimal selection method of incoming materials from the available was established and validated. This study provided novel insights into the control of the input variability of the process and amplified the application of the SeDeM expert system, emphasizing the importance of the physical information of the raw materials in the drug manufacturing process.

[UHPLC-HRMS metabonomics to study effect of sulfur-fumigated Ophiopogonis Radix on endogenous metabolites in rats].

The project was launched to analyze the effects of sulfur-fumigated Ophiopogonis Radix on endogenous metabolites in rats by metabonomics. The preparation method of sulfur-fumigated Ophiopogonis Radix in laboratory was established. Then the blood samples of SD rats in blank group,Ophiopogonis Radix extract group and sulfur-fumigated Ophiopogonis Radix extract group were investigated by UHPLC-Q-Exactive. The differential metabolites were screened and identified by PCA(principal component analysis),OPLSDA(orthogonal partial least squares discriminant analysis) and variable importance projection(VIP),and the metabolic pathways were analyzed. Finally,a total of 15 potential biomarkers were identified. Compared with the samples of Ophiopogonis Radix extract group,sulfur-fumigated Ophiopogonis Radix mainly affected the biosynthesis and metabolism of amino acids in normal rats. Its mechanism may be related to the biosynthesis of phenylalanine,tyrosine,tryptophan and aminoacyl-tRNA as well as the metabolism of phenylalanine and tryptophan. Based on UHPLC-HRMS metabonomics,this paper discussed the effects of sulfur-fumigated Ophiopogonis Radix on endogenous metabolites in rats,which provided an idea for the metabolic study of other sulfur-fumigated traditional Chinese medicines.

Setting up multivariate specifications on critical raw material attributes to ensure consistent drug dissolution from high drug-load sustained-release matrix tablet.

The purpose of this study was to describe the raw material variability that influenced the in-vitro dissolution behavior of high drug-load sustained-release matrix tablet and to ensure the consistent quality of the final product. The Panax notoginseng saponins (PNS) - hydroxypropyl methylcellulose - anhydrous lactose - magnesium stearate (57:20:23:0.5%, w/w) was used as the model formulation. PNS extract powders with lot-to-lot and source-to-source differences were collected to cover the common cause variations and their physicochemical properties were characterized by the chromatographic fingerprints and the SeDeM expert system. It was found that the release behavior of active pharmaceutical ingredients (APIs) in PNS from different batches exhibited considerable variations. Latent variable modeling results demonstrated that the physical properties of raw materials played major roles in predicting the drug dissolution. PNS extracts with high specific surface area, the width of particle size distribution and hygroscopicity or low moisture content led to an increase in drug release. In order to perform efficient pass/fail judgments for incoming new materials, multivariate specifications of critical material attributes (CMAs) were established and the multivariate design space in line with the quality by design (QbD) principles was explored to achieve the release target.

[Application of quality by design in granulation process for Ginkgo leaf tablet (⠢)： process control strategy based on design space].

In this paper, under the guidance of quality by design (QbD) concept, the control strategy of the high shear wet granulation process of the ginkgo leaf tablet based on the design space was established to improve the process controllability and product quality consistency. The median granule size (D50) and bulk density (Da) of granules were identified as critical quality attributes (CQAs) and potential critical process parameters (pCPPs) were determined by the failure modes and effect analysis (FMEA). The Plackeet-Burmann experimental design was used to screen pCPPs and the results demonstrated that the binder amount, the wet massing time and the wet mixing impeller speed were critical process parameters (CPPs). The design space of the high shear wet granulation process was developed within pCPPs range based on the Box-Behnken design and quadratic polynomial regression models. ANOVA analysis showed that the P-values of model were less than 0.05 and the values of lack of fit test were more than 0.1, indicating that the relationship between CQAs and CPPs could be well described by the mathematical models. D50 could be controlled within 170 to 500 µm, and the bulk density could be controlled within 0.30 to 0.44 g•cm⁻³ by using any CPPs combination within the scope of design space. Besides, granules produced by process parameters within the design space region could also meet the requirement of tensile strength of the ginkgo leaf tablet..

[Continual improvement of quantitative analytical method development of Panax notogineng saponins based on quality by design].

This study is aimed to propose a continual improvement strategy based on quality by design (QbD). An ultra high performance liquid chromatography (UPLC) method was developed to accomplish the method transformation from HPLC to UPLC of Panax notogineng saponins (PNS) and achieve the continual improvement of PNS based on QbD, for example. Plackett-Burman screening design and Box-Behnken optimization design were employed to further understand the relationship between the critical method parameters (CMPs) and critical method attributes (CMAs). And then the Bayesian design space was built. The separation degree of the critical peaks (ginsenoside Rg1 and ginsenoside Re) was over 2.0 and the analysis time was less than 17 min by a method chosen from the design space with 20% of the initial concentration of the acetonitrile, 10 min of the isocratic time and 6%•min⁻¹ of the gradient slope. At last, the optimum method was validated by accuracy profile. Based on the same analytical target profile (ATP), the comparison of HPLC and UPLC including chromatograph method, CMA identification, CMP-CMA model and system suitability test (SST) indicated that the UPLC method could shorten the analysis time, improve the critical separation and satisfy the requirement of the SST. In all, HPLC method could be replaced by UPLC for the quantity analysis of PNS.

Identification of Metabolites of 6'-Hydroxy-3,4,5,2',4'-pentamethoxychalcone in Rats by a Combination of Ultra-High-Performance Liquid Chromatography with Linear Ion Trap-Orbitrap Mass Spectrometry Based on Multiple Data Processing Techniques.

In this study, an efficient strategy was established using ultra-high-performance liquid chromatography coupled with linear ion trap-Orbitrap mass spectrometry (UHPLC-LTQ-Orbitrap MS) to profile the in vivo metabolic fate of 6'-hydroxy-3,4,5,2',4'-pentamethoxychalcone (PTC) in rat urine and feces. The UHPLC-LTQ-Orbitrap method combines the high trapping capacity and MS(n) scanning function of the linear ion trap along with accurate mass measurements within 5 ppm and a resolving power of up to 30,000 over a wider dynamic range compared to many other mass spectrometers. In order to reduce the potential interferences of endogenous substances, the post-acquisition processing method including high-resolution extracted ion chromatogram (HREIC) and multiple mass defect filters (MMDF) were developed for metabolite detection. As a result, a total of 60 and 35 metabolites were detected in the urine and feces, respectively. The corresponding in vivo reactions such as methylation, hydroxylation, hydrogenation, decarbonylation, demethylation, dehydration, methylation, demethoxylation, sulfate conjugation, glucuronide conjugation, and their composite reactions were all detected in this study. The result on PTC metabolites significantly expanded the understanding of its pharmacological effects, and could be targets for future studies on the important chemical constituents from herbal medicines.

[Near infrared spectrum monitoring of scale up process for macroporous resin purification of Grardeniae extract].

To establish a fast detection method during the purifying process of the extracts from Grardeniae using macroporous resin based on near infrared spectroscopy. First, the ethanol eluent was collected from the purification process of small size sample; and near infrared (NIR) spectrum was collected. Then the content of the geniposide was determined by HPLC method, and partial least squares (PLS) method was used to establish the quantitative model to predict the content of geniposide by NIR spectrum. This model was used to supervise the changes of geniposide concentrations in ethanol eluent during medium scale process. Experimental results showed that the NIR small scale model can accurately predict the concentrations of geniposide in the production process of medium scale. However, with the proceeding of batch processes, the prediction performance of the model was decreased, so model updating method was employed to maintain the model. After twice updates, the NIR quantitative model can accurately predict the concentrations of the geniposide during medium scale process. Therefore, through model updates, the established NIR quantitative model can be applied in different scales of macroporous resin purification processes, to improve the data utilization efficiency of small scale process and save the cost of rebuilding the quantitative model of medium scale.

Sample Size Requirements of a Pharmaceutical Material Library: A Case in Predicting Direct Compression Tablet Tensile Strength by Latent Variable Modeling.

The material library is an emerging, new data-driven approach for developing pharmaceutical process models. How many materials or samples should be involved in a particular application scenario is unclear, and the impact of sample size on process modeling is worth discussing. In this work, the direct compression process was taken as the research object, and the effects of different sample sizes of material libraries on partial least squares (PLS) modeling in the prediction of tablet tensile strength were investigated. A primary material library comprising 45 materials was built. Then, material subsets containing 5 × i (i = 1, 2, 3, …, 8) materials were sampled from the primary material library. Each subset underwent sampling 1000 times to analyze variations in model fitting performance. Both hierarchical sampling and random sampling were employed and compared, with hierarchical sampling implemented with the help of the tabletability classification index d. For each subset, modeling data were organized, incorporating 18 physical properties and tableting pressure as the independent variables and tablet tensile strength as the dependent variable. A series of chemometric indicators was used to assess model performance and find important materials for model training. It was found that the minimum R2 and RMSE values reached their maximum, and the corresponding values were kept almost unchanged when the sample sizes varied from 20 to 45. When the sample size was smaller than 15, the hierarchical sampling method was more reliable in avoiding low-quality few-shot PLS models than the random sampling method. Two important materials were identified as useful for building an initial material library. Overall, this work demonstrated that as the number of materials increased, the model's reliability improved. It also highlighted the potential for effective few-shot modeling on a small material library by controlling its information richness.

Rapid identification and determination of adulteration in medicinal Arnebiae Radix by combining near infrared spectroscopy with chemometrics.

The medicinal Arnebia Radix (AR) is one of widely-used Chinese herbal medicines (CHMs), usually adulterated with non-medicinal species that seriously compromise the quality of AR and affect patients' health. Detection of these adulterants is usually performed by using expensive and time-consuming analytical instruments. In this study, a rapid, non-destructive, and effective method was proposed to identify and determine the adulteration in the medicinal AR by near-infrared (NIR) spectroscopy coupled with chemometrics. 37 batches of medicinal AR samples originated from Arnebia euchroma (Royle) Johnst., 11 batches of non-medicinal AR samples including Onosma paniculatum Bur. et Franch and Arnebia benthamii (Wall. ex G. Don) Johnston, and 72 batches of adulterated AR samples were characterized by NIR spectroscopy. The data driven-soft independent modeling by class analogy (DD-SIMCA) and partial least squares-discriminant analysis (PLS-DA) were separately used to differentiate the authentic from adulterated AR samples. Then the PLS and support vector machine (SVM) were applied to predict the concentration of the adulteration in the adulterated AR samples, respectively. As a result, the classification accuracies of DD-SIMCA and PLS-DA models were 100% for the calibration set, and 96.7% vs. 100% for the prediction set. Moreover, the relative prediction deviation (RPD) values of PLS models reached 11.38 and 7.75 for quantifying two adulterants species, which were obviously superior to the SVM models. It can be concluded that the NIR spectroscopy coupled with chemometrics is feasible to identify the authentic from adulterated AR samples and quantify the adulteration in adulterated AR samples.

Synergistic effect of Euphorbia kansui stir-fried with vinegar and bile acids on malignant ascites effusion through modulation of gut microbiota.

Background: Toxic Euphorbia kansui (EK) is employed to treat malignant ascites effusion (MAE). EK stir-fried with vinegar (VEK) has been demonstrated to reduce toxicity due to its preserved water-expelling effect. This was demonstrated to be correlated with gut microbiota. Therein, bile acids (BAs) have a bidirectional relationship with the gut microbiota. Therefore, the aim of this study is to explore whether BA-mediated gut microbiota influences the water-expelling effect of VEK against MAE. Methods: The MAE rat model was established by intraperitoneal injection of Walker-256 tumor cells. A reliable simultaneous method for the determination of 15 bile acids in rat feces using ultra-high-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) was established and applied to analyze the fecal BAs in rats treated with VEK. The screened BA was then administered to VEK-treated MAE rats. The water-expelling effect was evaluated using histopathological analysis, biochemical examination, inflammatory factors in ascites, urine volume, ascites amount, and intestinal aquaporin expression. The microbial composition was determined using 16S rRNA sequencing, and the contents of bile acids were finally measured. Results: VEK decreased the content of fecal deoxycholic acid (DCA), lithocholic acid (LCA), and taurocholic acid (TCA) while increasing the content of ursodeoxycholic acid (UDCA). VEK alleviated liver, stomach, and intestinal injuries; oxidative damage; and inflammation, which were further ameliorated with UDCA intervention. VEK alleviated MAE by increasing the fecal water content, urine volume, and AQP3 protein expression and decreasing the urine levels of Na+, K+, and Cl-. This was retained with the intervention of UDCA. UDCA and VEK regulated the BA metabolism disorder to a certain extent. Analysis of gut microbiota showed that VEK increased the abundance of Lactobacillus and decreased that of Prevotella_9 in MAE rats. The combined administration of UDCA and VEK showed a better modulation of the microbiota structure than that of VEK alone, and the effect of this administration reached closer to the reference state. Conclusion: The water-expelling effect of VEK did not directly depend on the BA-mediated gut microbiota. However, VEK and BAs had a synergistic effect on malignant ascites effusion through the regulation of the gut microbiota. These results provided a scientific basis for the reasonable usage of VEK and the novel combination treatment strategy of VEK and UDCA.

The interaction between polyphyllin I and SQLE protein induces hepatotoxicity through SREBP-2/HMGCR/SQLE/LSS pathway.

Polyphyllin I (PPI) and polyphyllin II (PII) are the main active substances in the Paris polyphylla. However, liver toxicity of these compounds has impeded their clinical application and the potential hepatotoxicity mechanisms remain to be elucidated. In this work, we found that PPI and PII exposure could induce significant hepatotoxicity in human liver cell line L-02 and zebrafish in a dose-dependent manner. The results of the proteomic analysis in L-02 cells and transcriptome in zebrafish indicated that the hepatotoxicity of PPI and PII was associated with the cholesterol biosynthetic pathway disorders, which were alleviated by the cholesterol biosynthesis inhibitor lovastatin. Additionally, 3-hydroxy-3-methy-lglutaryl CoA reductase (HMGCR) and squalene epoxidase (SQLE), the two rate-limiting enzymes in the cholesterol synthesis, selected as the potential targets, were confirmed by the molecular docking, the overexpression, and knockdown of HMGCR or SQLE with siRNA. Finally, the pull-down and surface plasmon resonance technology revealed that PPI could directly bind with SQLE but not with HMGCR. Collectively, these data demonstrated that PPI-induced hepatotoxicity resulted from the direct binding with SQLE protein and impaired the sterol-regulatory element binding protein 2/HMGCR/SQLE/lanosterol synthase pathways, thus disturbing the cholesterol biosynthesis pathway. The findings of this research can contribute to a better understanding of the key role of SQLE as a potential target in drug-induced hepatotoxicity and provide a therapeutic strategy for the prevention of drug toxic effects with similar structures in the future.

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.

A general procedure for establishing composite quality evaluation indices based on key quality attributes of traditional Chinese medicine.

Licorice, a medicinal herb and food flavor ingredient, has been widely used in traditional Chinese medicine (TCM) for the past 4000 years. In this study, we propose a new quality evaluation approach for licorice quality control based on the key quality attributes commonly used in TCM. The high quality of TCM formulations is ensured by verifying the genuine origin and implementing good agricultural and collection practices for each medicinal herb. In our study, the genuine production area, the harvest season, and the number of growth years were considered the key quality attributes of TCM. To ensure the representativeness of our analysis, we obtained a total of 158 licorice sample batches that differed in the number of growth years, the location of the production areas, and the season for harvesting. Initially, the 158 sample batches were subjected to ultra-high-performance liquid chromatography-quadrupole time-of-flight tandem mass spectrometry (UHPLC-QTOF-MS/MS). A preliminary screen identified 11 licorice compounds related to the three key quality attributes of TCM . An analysis by ultra-high-performance liquid chromatography coupled with triple quadrupole tandem mass spectrometry (UHPLC-TQ-MS/MS) verified the presence of 34 compounds in all licorice samples. These 34 compounds included the 11 compounds related to the three key quality attributes of the samples, along with other bioactive components identified in previous studies. After using UHPLC-TQ-MS/MS to assess the signal peak intensities of the 34 compounds, we selected 17 licorice compounds to establish sample content evaluation indices, which were determined by high-performance liquid chromatography at four different wavelengths in all 158 licorice sample batches. Finally, the screen identified nine compounds that were closely associated with the quality attributes of licorice based on principal component analysis (PCA) and orthogonal partial least squares discriminant analysis (OPLS-DA). Our results suggested that liquiritin and eight other compounds could be used as quality control indicators of licorice, which provided a foundation to establish the TCM quality composite evaluation index (TCM QCEI). In summary, this research concept can serve as a reference for research on quality markers and the evaluation of TCM.

Discovering Temporal Patterns in Longitudinal Nontargeted Metabolomics Data via Group and Nuclear Norm Regularized Multivariate Regression.

Temporal associations in longitudinal nontargeted metabolomics data are generally ignored by common pattern recognition methods such as partial least squares discriminant analysis (PLS-DA) and orthogonal partial least squares discriminant analysis (OPLS-DA). To discover temporal patterns in longitudinal metabolomics, a multitask learning (MTL) method employing structural regularization was proposed. The group regularization term of the proposed MTL method enables the selection of a small number of tentative biomarkers while maintaining high prediction accuracy. Meanwhile, the nuclear norm imposed into the regression coefficient accounts for the interrelationship of the metabolomics data obtained on consecutive time points. The effectiveness of the proposed method was demonstrated by comparison study performed on a metabolomics dataset and a simulating dataset. The results showed that a compact set of tentative biomarkers charactering the whole antipyretic process of Qingkailing injection were selected with the proposed method. In addition, the nuclear norm introduced in the new method could help the group norm to improve the method's recovery ability.