DeepION: A Deep Learning-Based Low-Dimensional Representation Model of Ion Images for Mass Spectrometry Imaging.

Mass spectrometry imaging (MSI) is a high-throughput imaging technique capable of the qualitative and quantitative in situ detection of thousands of ions in biological samples. Ion image representation is a technique that produces a low-dimensional vector embedded with significant spectral and spatial information on an ion image, which further facilitates the distance-based similarity measurement for the identification of colocalized ions. However, given the low signal-to-noise ratios inherent in MSI data coupled with the scarcity of annotated data sets, achieving an effective ion image representation for each ion image remains a challenge. In this study, we propose DeepION, a novel deep learning-based method designed specifically for ion image representation, which is applied to the identification of colocalized ions and isotope ions. In DeepION, contrastive learning is introduced to ensure that the model can generate the ion image representation in a self-supervised manner without manual annotation. Since data augmentation is a crucial step in contrastive learning, a unique data augmentation strategy is designed by considering the characteristics of MSI data, such as the Poisson distribution of ion abundance and a random pattern of missing values, to generate plentiful ion image pairs for DeepION model training. Experimental results of rat brain tissue MSI show that DeepION outperforms other methods for both colocalized ion and isotope ion identification, demonstrating the effectiveness of ion image representation. The proposed model could serve as a crucial tool in the biomarker discovery and drug development of the MSI technique.

A BEST classification system of large to giant congenital melanocytic nevi based on expert consensus and distribution characteristics.

BACKGROUND: Large to giant congenital melanocytic nevi (LGCMN) significantly decrease patients' quality of life, but the inaccuracy of current classification system makes their clinical management challenging. OBJECTIVES: To improve and extend the existing LGCMN 6B/7B classification systems by developing a novel LGCMN classification system based on a new phenotypic approach to clinical tool development. METHODS: Three hundred and sixty-one LGCMN cases were categorized into four subtypes based on anatomic site: bonce (25.48%), extremity (17.73%), shawl (19.67%) and trunks (37.12%) LGCMN. A 'BEST' classification system of LGCMN was established and validated by a support vector machine classifier combined with the 7B system. RESULTS: The most common LGCMN distributions were on bonce and trunks (bathing trunk), whereas breast/belly and body LGCMN were exceptionally rare. Sexual dimorphism characterized distribution, with females showing a wider range of lesions in the genital area. Nearly half of the patients with bathing trunk LGCMN exhibited a butterfly-like distribution. Approximately half of the LGCMN with chest involvement did not have nipple-areola complex involvement. Abdomen, back and buttock involvement was associated with the presence of satellite nevi (r = 0.558), and back and buttock involvement was associated with the presence of nodules (r = 0.364). CONCLUSIONS: The effective quantification of a standardized anatomical site provides data support for the accuracy of the 6B/7B classification systems. The simplified BEST classification system can help establish a LGCMN clinical database for exploration of LGCMN aetiology, disease management and prognosis prediction.

Accurate Classification and Prediction of Acute Myocardial Infarction through an ARMD Procedure.

The risk stratification of acute myocardial infarction (AMI) patients is of prime importance for clinical management and prognosis assessment. Thus, we propose an ensemble machine learning analysis procedure named ADASYN-RFECV-MDA-DNN (ARMD) to address sample-unbalanced problems and enable stratification and prediction of AMI outcomes. The ARMD analysis procedure was applied to the NMR data of sera from 534 AMI-related subjects in four categories with an extremely imbalanced sample proportion. Firstly, the adaptive synthetic sampling (ADASYN) algorithm was used to address the issue of the original sample imbalance. Secondly, the recursive feature elimination with cross-validation (RFECV) processing and random forest mean decrease accuracy (RF-MDA) algorithm was performed to identify the differential metabolites corresponding to each AMI outcome. Finally, the deep neural network (DNN) was employed to classify and predict AMI events, and its performance was evaluated by comparing the four traditional machine learning methods. Compared with the other four machine learning models, DNN presented consistent superiority in almost all of the model parameters including precision, f1-score, sensitivity, specificity, area under the receiver operating characteristic curve (AUC), and classification accuracy, highlighting the potential of deep learning in classification and stratification of clinical diseases. The ARMD analysis procedure was a practical analysis tool for supervised classification and regression modeling of clinical diseases.

No Correlation between Plasma GPNMB Levels and Multiple System Atrophy in Chinese Cohorts.

BACKGROUND: Glycoprotein nonmetastatic melanoma protein B (GPNMB) has been demonstrated to mediate pathogenicity in Parkinson's disease (PD) through interactions with α-synuclein, and plasma GPNMB tended to be a novel biomarker for PD. OBJECTIVE: The goal of this study was to investigate whether plasma GPNMB could act as a potential biomarker for the clinical diagnosis and severity monitoring of multiple system atrophy (MSA), another typical synucleinopathy. METHODS: Plasma GPNMB levels in patients with MSA, patients with PD, and healthy control subjects (HCs) were quantified using enzyme-linked immunosorbent assays. RESULTS: A total of 204 patients with MSA, 65 patients with PD, and 207 HCs were enrolled. The plasma GPNMB levels in patients with MSA were similar to those in HCs (P = 0.251) but were significantly lower than those in patients with PD (P = 0.003). Moreover, there was no significant correlation detected between the plasma GPNMB levels and disease severity scores of patients with MSA. CONCLUSIONS: No evidence was detected for the biomarker potential of plasma GPNMB in MSA. © 2023 International Parkinson and Movement Disorder Society.

Differentiation and prognostic stratification of acute myeloid leukemia by serum-based spectroscopy coupling with metabolic fingerprints.

Acute myeloid leukemia (AML) is a heterogeneous disease characterized by complex molecular and cytogenetic abnormalities. New approaches to predict the prognosis of AML have increasingly attracted attention. There were 98 non-M3 AML cases and 48 healthy controls were enrolled in the current work. Clinically routine assays for cytogenetic and molecular genetic analyses were performed on the bone marrow samples of patients with AML. Meanwhile, metabolic profiling of these AML subjects was also performed on the serum samples by combining Ag nanoparticle-based surface-enhanced Raman spectroscopy (SERS) with proton nuclear magnetic resonance (NMR) spectroscopy. Although most of the routine biochemical test showed no significant differences between the M0-M2 and M5 groups, the metabolic profiles were significantly different either between AML subtypes or between prognostic risk subgroups. Specific SERS bands were screened to serve as potential markers for AML subtypes. The results demonstrated that the classification models for M0-M2 and M5 shared two bands (i.e., 1328 and 741 cm-1 ), all came from nucleic acid signals. Furthermore, Metabolic profiles provided various differential metabolites responsible for different AML subtypes, and we found altered pathways mainly included energy metabolism like glycolysis, pyruvate metabolism, and metabolisms of nucleic acid bases as well as specific amino acid metabolisms. It is concluded that integration of SERS and NMR provides the rational and could be reliable to reveal AML differentiation, and meanwhile lay the basis for experimental and clinical practice to monitor disease progression and prognostic evaluation.

Profiling of Homocysteine Metabolic Pathway Related Metabolites in Plasma of Diabetic Mellitus Based on LC-QTOF-MS.

Background: Homocysteine (Hcy) has been found to be closely related to the occurrence of diabetes mellitus (DM) and is considered as one of the risk factors of DM. However, Hcy alone is not enough as a factor to predict DM, and our study analyzed and determined the relationship between the main metabolites involved in the Hcy metabolic pathway and DM. Methods: A total of 48 clinical samples were collected, including 18 health control samples and 30 DM samples. All standards and samples were detected by LC-QTOF-MS. Multivariate statistical analysis and k-means cluster analysis were performed to screen and confirm the metabolites significantly correlated with DM. Results: A total of 13 metabolites of the Hcy metabolic pathway were detected in the samples. The content of Hcy, cysteine, taurine, pyridoxamine, methionine, and choline were significantly increased in the DM group (p < 0.05). Hcy, choline, cystathionine, methionine, and taurine contributed significantly to the probabilistic principal component analysis (PPCA) model. The odds ratios (OR) of Hcy, cysteine, taurine, methionine, and choline were all greater than one. K-means cluster analysis showed that the Hcy, taurine, methionine, and choline were significantly correlated with the distribution of glucose values (divided into four levels: 10.5−11.7 mmol/L, 7.7−9.7 mmol/L, 6.0−6.9 mmol/L, and 5.0−5.9 mmol/L, respectively). Conclusion: Hcy, taurine, methionine, and choline can be used as risk factors for diabetes diagnosis and are expected to be used for the assessment of diabetes severity.

Metabolomic analyses redefine the biological classification of pancreatic cancer and correlate with clinical outcomes.

Pancreatic ductal adenocarcinoma (PDAC) is characterized by high heterogeneity, and the postoperative prognosis of different patients often varies greatly. Therefore, the classification of pancreatic cancer patients and precise treatment becomes particularly important. In our study, 1 H NMR spectroscopy was used to analyze the 76 PDAC serum samples and identify the potential metabolic subtypes. The metabolic characteristics of each metabolic subtype were screened out and the relationship between metabolic subtype and the long-term prognosis was further identified. The clinical stages of PDAC did not show the metabolic differences at the serum metabolomic level. And three metabolic subtypes, basic, choline-like and amino acid-enriched types, were defined by the hierarchical cluster analysis of the serum metabolites and the disturbed metabolic pathways. The characteristic metabolites of each PDAC subtype were identified, and the metabolite model was established to distinguish the PDAC patients in the different subtypes. Among the three metabolic subtypes, choline-like type displayed better long-term prognosis compared to the other two types of patients. Metabolic subtypes are of clinical importance and are closer to expressing the heterogeneity in the actual life activities of pancreatic cancer than molecular typing. The excavation of metabolic subtypes based on this will be more in line with clinical reality and more promising to guide clinical precision individualization treatment.

Metabolic Interactions and Differences between Coronary Heart Disease and Diabetes Mellitus: A Pilot Study on Biomarker Determination and Pathogenesis.

Comprehensive understanding of plasma metabotype of diabetes mellitus (DM), coronary heart disease (CHD), and especially diabetes mellitus with coronary heart disease (CHDDM) is still lacking. In this work, the plasma metabolic differences and links of DM, CHD, and CHDDM patients were investigated by the strategy of comparative metabolomics based on 1H NMR spectroscopy combined with network analysis for revealing their metabolic differences. A total of 17 metabolites are related to three diseases, among which valine, alanine, leucine, isoleucine, and N-acetyl-glycoprotein are positively correlated with CHD and CHDDM (odds ratios (OR) > 1). The trimethylamine oxide, glycerol, lactose, indoleacetate, and scyllo-inositol are closely related to the development of DM to CHDDM (OR > 1), and indoleactate (OR: 1.06, 95% confidence interval (CI): 1.01-1.12) and lactose (OR: 2.46, 95% CI: 1.67-3.25) are particularly prominent in CHDDM. We identified three multi-biomarkers types that were significantly associated with glycosylated hemoglobin (HbA1C) at baseline. All diseases demonstrated dysregulated glycolysis/gluconeogenesis and amino acid biosynthesis pathway. In addition, enrichment in tryptophan metabolism observed in CHDDM, enrichment in inositol phosphate metabolism observed in DM, and the metabolites related to microbiota metabolism were dysregulated in both DM and CHDDM. The comparative metabolomics strategy of multi-diseases offers a new perspective in disease-specific markers and pathogenic pathways.

Application of pathways activity profiling to urine metabolomics for screening Qi-tonifying biomarkers and metabolic pathways of honey-processed Astragalus.

Honey-processed Astragalus, a widely used Qi-tonifying and immunomodulating herb in traditional Chinese medicine, has strengthened the tonic effects and achieved fewer side effects compared with astragali radix in clinical application. Here, we focus on Qi-tonifying biomarkers and pathways of honey-processed Astragalus using urine metabolomics that provide the basis for building the linkage between metabolites in rat urine and its symptoms. The spleen Qi deficiency model group, normal group, astragali radix group, and honey-processed Astragalus group were implemented to evaluate Qi-tonifying effects. Twelve potential biomarkers were screened by multivariate statistical analysis by using ultra high performance liquid chromatography with quadrupole time-of-flight mass spectrometry. Furthermore, pathways activity profiling showed unique pathways that are primarily involved in tryptophan metabolism, tricarboxylic acids cycle, and methionine metabolism. The results demonstrated that metabolomics coupled with pathway activity profiling were promising tools. It might serve as a novel methodological clue to systematically dissect the underlying Qi-tonifying mechanism of honey-processed Astragalus.

Structure Characterization of Honey-Processed Astragalus Polysaccharides and Its Anti-Inflammatory Activity In Vitro.

Honey-processed Astragalus is a dosage form of Radix Astragalus mixed with honey by a traditional Chinese medicine processing method which strengthens the tonic effect. Astragalus polysaccharide (APS), perform its immunomodulatory effects by relying on the tonic effect of Radix Astragalus, therefore, the improved pharmacological activity of honey-processed Astragalus polysaccharide (HAPS) might be due to structural changes during processing. The molecular weights of HAPS and APS were 1,695,788 Da, 2,047,756 Da, respectively, as determined by high performance gel filtration chromatography combined with evaporative light scattering detection (HPGFC-ELSD). The monosaccharide composition was determined by ultra-performance liquid chromatogram quadrupole time-of-flight mass spectrometry (UPLC/ESI-Q-TOF-MS) after pre-column derivatization with 1-phenyl-3-methyl-5-pyrazolone (PMP). The results showed that the essential components were mannose, glucose, xylose, arabinose, glucuronic acid and rhamnose, is molar ratios of 0.06:28.34:0.58:0.24:0.33:0.21 and 0.27:12.83:1.63:0.71:1.04:0.56, respectively. FT-IR and NMR analysis of HAPS results showed the presence of uronic acid and acetyl groups. The anti-inflammatory activities of HAPS were more effective than those of APS according to the NO contents and the expression of IFN-γ, IL-1ß, IL-22 and TNF-α in lipopolysaccharide (LPS)-induced RAW264.7 cells. This findings suggest that the anti-inflammatory and bioactivity improvement might be associated with molecular structure changes, bearing on the potential immunomodulatory action.

[Identification of biomarkers in urine of rats with spleen Qi deficiency and biological significance].

To identify biomarkers for spleen Qi deficiency by analyzing small molecule metabolites in urine, in order to expound the relationship between biomarkers and metabolic pathways. The spleen Qi deficiency model was established through dietary restriction and overstrain. All of the rats received D-xylose absorption experiment and blood routine test. Urine samples were collected in the next day. The urine samples were analyzed using UPLC-Q-TOF-MS to obtain the dataset of urine metabolic group. Principal component analysis (PCA), orthogonal partialleast squares-discriminant analysis (OPLS-DA) and other multivariate statistical methods were employed to evaluate the quality of the dataset and screen out potential biomarkers of spleen Qi deficiency. The results of D-xylose absorption and blood routine demonstrated that the spleen Qi deficiency model was successfully established. In positive ion mode and negative ion mode, PCA and OPLS-DA score plots could clearly distinguish model group and blank group. According to S-plot of OPLS-DA, VIP value, t-test and area under receiver operating characteristic curve (ROC), 24 biomarkers, including phenylalanine, succinic acid, aconitic acid, isocitrate acid, betaine, kynurenine, indole, creatine, creatinine, orotic acid, xanthine, and xanthurenic acid, were identified as associated with the spleen Qi deficiency, mainly involving energy metabolism, amino acid metabolism, tryptophan metabolism, purine metabolism and pyrimidine metabolism. Urine metabolomics method combined with online software package for data processing and analysis metabolic pathway can provide new methods and ideas for studies for spleen Qi deficiency and other traditional Chinese medicine symptoms.

Microbial community characterization, activity analysis and purifying efficiency in a biofilter process.

The growth and metabolism of microbial communities on biologically activated carbon (BAC) play a crucial role in the purification of drinking water. To gain insight into the growth and metabolic characteristics of microbial communities and the efficiency of drinking water treatment in a BAC filter, we analyzed the heterotrophic plate count (HPC), phospholipid, dehydrogenase, metabolic function and water quality parameters during start-up and steady-state periods. In the start-up process of the filter with natural biofilm colonization, the variation in heterotrophic plate count levels was S-curved. The total phospholipid level was very low during the first 5 days and reached a maximum value after 40 days in the filter. The activity of dehydrogenase gradually increased during the first 30 days and then reached a plateau. The functional diversity of the microbial community in the filter increased, and then reached a relatively stable level by day 40. After an initial decrease, which was followed by an increase, the removal rate of NH4(+)-N and COD(Mn) became stable and was 80% and 28%, respectively, by day 40. The consumption rate of dissolved oxygen reached a steady level after 29 days, and remained at 18%. At the steady operation state, the levels of HPC, phospholipid, dehydrogenase activity and carbon source utilization had no significant differences after 6 months compared to levels measured on day 40. The filter was shown to be effective in removing NH4(+)-N, NO2(-)-N, COD(Mn), UV254, biodegradable dissolved organic carbon and trace organic pollutants from the influent. Our results suggest that understanding changes in the growth and metabolism of microorganisms in BAC filter could help to improve the efficiency of biological treatment of drinking water.

Integration of metabolomics and network pharmacology for enhancing mechanism understanding and medication combination recommendation for diabetes mellitus and diabetic nephropathy.

With the increasing prevalence of diabetes mellitus (DM) and diabetic nephropathy (DN), effective treatment is particularly important for the recovery of patients. However, the currently approved drugs are usually tailored to clinical symptoms and no mechanism-targeted drugs are available. In this study, the combination of metabolomics and network pharmacology was applied to provide reasonable medication combination regimens to meet the different clinical needs for the targeted treatment of DM and DN. An NMR-based metabolomic strategy was applied to identify the potential urinary biomarkers of DM or/and DN, while network pharmacology was used to identify the therapy targets of DM and DN by intersecting the targets of diseases and currently approved drugs. According to the enriched signaling pathways using the potential biomarkers and the therapy targets, the specific medication combinations were recommended for the specific clinical demands in terms of hypoglycemic, hypertensive, and/or lipid-lowering. For DM, 17 potential urinary biomarkers and 12 disease-related signaling pathways were identified, and 34 combined medication regimens related to hypoglycemia, hypoglycemia, and hypertension, and hypoglycemia, hypertension, and lipid-lowering were administered. For DN, 22 potential urinary biomarkers and 12 disease-related signaling pathways were identified, and 21 combined medication regimens related to hypoglycemia, hypoglycemia, and hypertension were proposed. Molecular docking was used to verify the binding ability, docking sites, and structure of the drug molecules to target proteins. Moreover, an integrated biological information network of the drug-target-metabolite-signaling pathways was constructed to provide insights into the underlined mechanism of DM and DN as well as clinical combination therapy.

Recoupled-STOCSY-based co-expression network analysis to extract phenotype-driven metabolite modules in NMR-based metabolomics dataset.

Nuclear magnetic resonance (NMR)-based metabolomics study usually involves spectral preprocessing, identification of biomarkers and interpretation of biological processes and pathogenesis, however, the traditional procedure is bound to inborn defects. In this study, a new analytical frame was proposed to assist spectral alignment and dimensionality reduction, screen the differential metabolites and get biological explanation of the metabolic network by combing weighted gene co-expression network analysis (WGCNA) and recoupled statistical total correlation spectroscopy (RSTOCSY). The performance of RSTOCSY-based WGCNA method was evaluated by the NMR dataset of serum from coronary heart disease with diabetes mellitus (CHDDM) patients. The statistical recoupling of variables (SRV) was successfully used to categorize the whole dataset into a number of superclusters of signals and served to spectral alignment, and its effectiveness was confirmed by the wine dataset with a larger spectral drift. Three phenotype-driven metabolite modules related to CHDDM were identified from the dataset by WGCNA, and 22 metabolites were further identified from the three modules according to the metabolic correlations within or between modules, and 40 significant metabolic correlations were observed from the intra- and inter-metabolites in the 2D pseudospectrum. These modules involve amino acid metabolism, microbial metabolism and glucose metabolism, and their analysis of metabolite network diffusion revealed a new discovery that the ferroptosis pathway is related to CHDDM. This RSTOCSY-based WGCNA approach provides an effective analysis workflow for information recovery and structure identification of metabolites and improving interpretability and understanding of the disease pathogenesis.

Cross-comparative metabolomics reveal sex-age specific metabolic fingerprints and metabolic interactions in acute myocardial infarction.

The elucidation of metabolic perturbations and gender-age-specific metabolic characteristics associated with acute myocardial infarction (AMI) is essential for clinical risk stratification and disease management. A comprehensive cross-comparative metabolomics analysis was performed on the sera from 445 healthy controls, 347 AMI patients without cardiovascular disease (CVD), 79 AMI with CVD (AMICVD) patients including 27 deaths. Machine-learning-based integrated biomarker profiling and global network analysis were used to create a multi-biomarker for distinguishing the different AMI outcomes. The changes of most metabolites were dependent on AMI, but gender and age also give additional contributions to the changes of histidine, malonate, O-acetyl-glycoprotein and trimethylamine N-oxide. The altered metabolic pathways included gut dysbiosis, increased amino acid metabolism, glucose metabolism and ketone metabolism, and inactivation of tricarboxylic acid cycle. Enhanced histidine metabolism and microbiota dysbiosis may be one of the key factors during the developing of AMI into AMICVD. For the differential diagnosis of AMI events, three sets of specific multi-biomarkers provided relatively high accuracy with the areas under the curve more than 0.8 and hazard ratio more than 1 in the discovery set, and the results were reproduced and confirmed by the validation set. First use of cross-comparative metabolomics and machine-learning-based integrated biomarker analysis gives great capability to discriminate the different AMI outcomes. Also, the multi-biomarkers seem to be a valid and accurate auxiliary diagnosis biomarker in addition to standard stratification based on clinical parameters.

An integrating strategy for serum metabolomics and microarray analysis to expand the understanding of diet-induced obesity.

Obesity is a key component of metabolic syndrome and is precipitated by complex interactions between multiple environmental and genetic factors. The integration of multi-level bioinformation is needed to understand the altered endogenous molecule and metabolic mechanisms. In this study, an integrated analytical strategy was proposed by combining microarray data from a gene expression omnibus database and in vitro serum metabolomic data to unearth bioinformation associated with cafeteria diet induced obesity. In the diet induced obese rats, 23 genes and 9 metabolites showed significant changes, in which the increased levels of alanine, lactate and lactate dehydrogenase B (Ldhb) and the decreased levels of citrate and pyruvate indicated an enhanced glycolysis and a disordered Krebs cycle. Furthermore, the closeness centrality of Slc27a2, Apobr, alanine and histidine in the correlations network of pathways, genes and metabolites was 0.5036, 0.5111, 0.5702, and 0.5352, respectively. These close links between metabolites and genes would be highly useful to assess the degree of obesity and to understand the developmental mechanism of obesity. The pathway enrichment analysis of genes and metabolites proved that a disturbed glucose metabolism and biosynthesis of amino acids are typical metabolic features of cafeteria-induced obesity. The metabolomics combined with microarray data not only could identify the biomarkers, but also would be beneficial to the follow-up research of obesity treatment, especially providing a methodological basis for the study of other diseases.

A metabonomics and lipidomics based network pharmacology study of qi-tonifying effects of honey-processed Astragalus on spleen qi deficiency rats.

Honey-processed Astragalus is a dosage form of radix Astragali processed with honey, which is deemed to contain better qi-tonifying effects in traditional Chinese medicine theroy. Our previous study has demonstrated that honey-processed Astragalus exhibited a better effect on reinforcing qi (vital energy) and immune improvement toward spleen qi deficiency compared with radix Astragali. However, the detailed mechanisms related to qi-tonifying effects of honey-processed Astragalus is still unclear. In this study, we evaluated the qi-tonifying effects of honey-processed Astragalus on spleen qi deficiency rats and predicted the mechanisms by aggregating metabonomics, lipidomics and network pharmacology. The results revealed that body weights, symptom scores, the levels of red blood cell, white blood cell, lymphocyte, spleen and thymus indexes, and three cytokines (TNF-α, IL-6, IFN-γ) in honey-processed Astragalus treated rats were improved in comparison with spleen qi deficiency rats. In parallel, based on the 26 biomarkers screened in metabonomics and lipidomics, we inferred that glycerophospholipid metabolism significantly regulated in pathway analysis was connected with qi-tonifying effects. Moreover, the network pharmacology analysis concluded that the compounds targets of honey-processed Astragalus CDK2, NOS3, MAPK14, PTGS1 and PTGS2 interacted with markers targets PLA2G(s) family and LYPLA1 could be responsible for regulation of glycerophospholipid metabolism to develop qi-tonifying effects. What's more, the above processes were possibly through VEGF signaling and MAPK signaling pathways.

Multi-fingerprint profiling combined with chemometric methods for investigating the quality of Astragalus polysaccharides.

It is a challenge to ascertain the quality of polysaccharides due to their complex chemical structure; therefore, multi-fingerprint profiling was used to investigate the quality of Astragalus polysaccharides (APS) harvested from Inner Mongolia (NM) and Gansu (GS) with the help of chemometric analysis. Additionally, FT-IR and 1H NMR were applied to characterize the chemical structure of the harvested APS. The spectral fingerprinting results indicated that APS had reduced similarity when they were from different origins. Further, PCA showed that NM and GS could be distinguished and that the main differences from the loading plots were in the absorption intensity of carbonyls and H1 signals of Galp and ß-glucose. Moreover, UPLC/Q-TOF-MS fingerprints were established based on the monosaccharide composition of the APS. The concentration of monosaccharides and results of cluster analysis indicated that GlcA might be an indicator that can be used to distinguish NM and GS. Overall, this multiple fingerprint method was stable, comprehensive and valid for monitoring APS quality.

Screening and Identification of the Metabolites of Water Extracts of Raw and Honey-Processed Astragalus in Rat Urine Based on UHPLC/ESI-Q-TOF-MS and Multivariate Statistical Analysis.

Radix Astragali is a famous traditional Chinese medicine and honey-processed Astragalus is a product of Radix Astragali acquired by honey-processing. These two products are widely utilized to treat various diseases. In this study, we screened bioactive components and metabolites of raw and honey-processed Astragalus in rat urine by ultra-performance liquid chromatography equipped with electrospray ionization/quadrupole time-of-flight mass spectrometry (UHPLC/ESI-Q-TOF-MS) combined with multivariate statistical analysis. In total, 62 compounds, including 7 parent compounds and 55 metabolites, were detected and 11 metabolites were characterized for the first time. The identified metabolites indicated that the metabolic reactions of Astragalus in rats included hydroxylation, glucuronidation, deglucosidation, monomethylation, demethylation, sulfation, hydrogenation, and dehydroxylation. The metabolic pathways of raw and honey-processed Astragalus in rat urine also were clarified. Through multivariate statistical analysis of the data of the raw and honey-processed Astragalus groups, we found that 20 compounds were differential components and that 1 metabolite only existed in the honey-processed Astragalus group. The differences in these ingredients between these two groups might provide the basis for interpreting the biologic activity differences in traditional Chinese medicine treatments. Graphical Abstract.