Search

Multivariate time series anomaly detection has garnered significant attention in fields such as IT operations, finance, medicine, and industry. However, a key challenge lies in the fact that anomaly patterns often exhibit multi-scale temporal variations, which existing detection models often fail to capture effectively. This limitation significantly impacts detection accuracy. To address this issue, we propose the MFAM-AD model, which combines the strengths of convolutional neural networks (CNNs) and bi-directional long short-term memory (Bi-LSTM). The MFAM-AD model is designed to enhance anomaly detection accuracy by seamlessly integrating temporal dependencies and multi-scale spatial features. Specifically, it utilizes parallel convolutional layers to extract features across different scales, employing an attention mechanism for optimal feature fusion. Additionally, Bi-LSTM is leveraged to capture time-dependent information, reconstruct the time series and enable accurate anomaly detection based on reconstruction errors. In contrast to existing algorithms that struggle with inadequate feature fusion or are confined to single-scale feature analysis, MFAM-AD effectively addresses the unique challenges of multivariate time series anomaly detection. Experimental results on five publicly available datasets demonstrate the superiority of the proposed model. Specifically, on the datasets SMAP, MSL, and SMD1-1, our MFAM-AD model has the second-highest F1 score after the current state-of-the-art DCdetector model. On the datasets NIPS-TS-SWAN and NIPS-TS-GECCO, the F1 scores of MAFM-AD are 0.046 (6.2%) and 0.09 (21.3%) higher than those of DCdetector, respectively(the value ranges from 0 to 1). These findings validate the MFAMAD model’s efficacy in multivariate time series anomaly detection, highlighting its potential in various real-world applications.

Multivariate time series anomaly detection is a crucial data mining technique with a wide range of applications in areas such as IT applications. Currently, the majority of anomaly detection methods for time series data rely on unsupervised approaches due to the rarity of anomaly labels. However, in real-world scenarios, obtaining a limited number of anomaly labels is feasible and affordable. Effective usage of these labels can offer valuable insights into the temporal characteristics of anomalies and play a pivotal role in guiding anomaly detection efforts. To improve the performance of multivariate time series anomaly detection, we proposed a novel deep learning model named EDD (Encoder-Decoder-Discriminator) that leverages limited anomaly samples. The EDD model innovatively integrates a graph attention network with long short term memory (LSTM) to extract spatial and temporal features from multivariate time series data. This integrated approach enables the model to capture complex patterns and dependencies within the data. Additionally, the model skillfully maps series data into a latent space, utilizing a carefully crafted loss function to cluster normal data tightly in the latent space while dispersing abnormal data randomly. This innovative design results in distinct probability distributions for normal and abnormal data in the latent space, enabling precise identification of anomalous data. To evaluate the performance of our EDD model, we conducted extensive experimental validation across three diverse datasets. The results demonstrate the significant superiority of our model in multivariate time series anomaly detection. Specifically, the average F1-Score of our model outperformed the second-best method by 2.7% and 73.4% in both evaluation approaches, respectively, highlighting its superior detection capabilities. These findings validate the effectiveness of our proposed EDD model in leveraging limited anomaly samples for accurate and robust anomaly detection in multivariate time series data.

Abstract Siamese networks have attracted wide attention in visual tracking due to their competitive accuracy and speed. However, the existing Siamese trackers usually leverage a fixed linear aggregation of feature maps, which does not effectively fuse the different layers of features with attention. Besides, most of Siamese trackers calculate the similarity between the template and the search region through a cross‐correlation operation between the features of the last blocks from the two branches, which might introduce the redundant noise information. In order to solve these problems, this study proposes a novel Siamese visual tracking method via cross‐layer calibration fusion, termed SiamCCF. An attention‐based feature fusion module is employed using local attention and non‐local attention to fuse the features from the deep and shallow layers, so as to capture both local details and high‐level semantic information. Moreover, a cross‐layer calibration module can use the fused features to calibrate the features of the last network blocks and build the cross‐layer long‐range spatial and inter‐channel dependencies around each spatial location. Extensive experiments demonstrate that the proposed method has achieved competitive tracking performance compared with state‐of‐the‐art trackers on challenging benchmarks, including OTB100, OTB2013, UAV123, UAV20L, and LaSOT.

Pulmonary fibrosis (PF) is an irreversible lung disease that is characterized by excessive scar tissue with a poor median survival rate of 2–3 years. The inhibition of transforming growth factor-β receptor type-I (TGF-β RI) by an appropriate drug may provide a promising strategy for the treatment of this disease. Polygonum cuspidatum (PC) is a well-known traditional Chinese herbal medicine which has an anti-PF effect. Accordingly, a combination of high resolution mass spectrometry with an in silico strategy was developed as a new method to search for potential chemical ingredients of PC that target the TGF-β RI. Based on this strategy, a total of 24 ingredients were identified. Then, absorption, distribution, metabolism, and excretion (ADME)-related properties were subsequently predicted to exclude compounds with potentially undesirable pharmacokinetics behaviour. Molecular docking studies on TGF-β RI were adopted to discover new PF inhibitors. Eventually, a compound that exists in PC known as resveratrol was proven to have excellent biological activity on TGF-β RI, with an IC50 of 2.211 μM in vitro. Furthermore, the complex formed through molecular docking was tested via molecular dynamics simulations, which revealed that resveratrol had strong interactions with residues of TGF-β RI. This study revealed that resveratrol has significant potential as a treatment for PF due to its ability to target TGF-β RI. In addition, this research demonstrated the exploration of natural products with excellent biological activities toward specific targets via high resolution mass spectrometry in combination with in silico technology is a promising strategy for the discovery of novel drugs.

This study proposes a progressive building facade detection method based on structure priors to effectively detect building facades from massive array InSAR spatial points with noise. First, the proposed method projects the initial array of InSAR three-Dimensional (3D) points on the ground to produce connected regions that correspond to building facades and then progressively detects potential line segments in each connected region under the guidance of structure priors. Furthermore, the proposed method generates building facades according to the detected line segments and their corresponding 3D points. In this process, the line segment detection space of the current connected region is constructed based on line segments detected in its neighboring connected regions, thereby improving the overall efficiency and reliability of the current line segment detection. Experimental results confirm that the proposed method can efficiently produce more reliable building facades from a massive array of InSAR 3D points with noise, overcoming several difficulties (such as low efficiency and inferior reliability) encountered in traditional multi-model fitting methods.

Electricity theft has significant impact on the power grids in terms of generating non-technical losses, which eventually degrading the power quality and minimizing the outfitted profit. In this paper, we proposed a hybrid approach based on deep learning and support vector machine for the detection of energy theft to facilitate and assess energy supplier companies to eliminate the issue of insufficient power, irregular power expenditure and ineffective electricity monitoring. A deep convolutional neural network is proposed for the feature learning using smart meters data in different times, varying from hours to days. Extracted features were further used to train support vector machine, which classify the features in two categories as theft and non-theft. Furthermore, a dropout layer is introduced in convolutional neural network model to avoid over fitting issues. Several careful experiments were carried out on real time customers smart meter data and the results validate the effectiveness of the proposed method in terms of accuracy and less detection error.

Using appropriate classification and recognition technology can help physicians make clinical diagnoses and decisions more effectively as a result of the ongoing development of artificial intelligence technology in the medical field. There are currently a number of issues with the detection of common pediatric dermatoses, including the challenge of image collection, the low resolution of some collected images, the intra-class variability and inter-class similarity of disease symptoms, and the mixing of disease symptom detection results. To resolve these problems, we first introduced the Random Online Data Augmentation and Selective Image Super-Resolution Reconstruction (RDA-SSR) method, which successfully avoids overfitting in training, to address the issue of the small dataset and low resolution of collected images, increase the number of images, and improve the image quality. Second, for the issue of an imbalance between difficult and simple samples, which is brought on by the variation within and between classes of disease signs during distinct disease phases. By increasing the loss contribution of hard samples for classification on the basis of the cross-entropy, we propose the DK_Loss loss function for two-stage object detection, allowing the model to concentrate more on the learning of hard samples. Third, in order to reduce redundancy and improve detection precision, we propose the Fliter_nms post-processing method for the intermingling of detection results based on the NMS algorithm. We created the CPD-10 image dataset for common pediatric dermatoses and used the Faster R-CNN network training findings as a benchmark. The experimental results show that the RDA-SSR technique, while needing a similar collection of parameters, can improve mAP by more than 4%. Furthermore, experiments were conducted over the CPD-10 dataset and PASCAL VOC2007 dataset to evaluate the effectiveness of DK_Loss over the two-stage object detection algorithm, and the results of cross-entropy loss-function-based training are used as baselines. The findings demonstrated that, with DK_Loss taken into account, its mAP is 1–2% above the baseline. Furthermore, the experiments confirmed that the Fliter_nms post-processing method can also improve model precision.

Schisandra chinensis, a widely used Chinese herbal medicine, was considered as central nervous system (CNS) drug for years. Both ethanol extracts (EES) and water extracts (WES) of it were applied clinically. Unfortunately, the difference of their efficacy and even effective material foundation of S. chinensis remains obscure. In this study, to explore the active constituents of S. chinensis, we compared pharmacodynamics and chemical profiles in vitro/in vivo of EES/WES for the first time using multiple chemical analysis, pharmacological and data processing approaches. It was proved that there was no significant difference in the anti-depressive effects between WES and EES. However, the contents of most components in vitro and in plasma were higher in EES than those in WES, which was unconvincing for their similar efficacy. Therefore, we further explored components of S. chinensis targeted onto brain and the results showed that 5 lignans were identified with definite absorptivity respectively both in EES and WES caused by the limitation of blood−brain barrier. Moreover, bioinformatic analysis predicted their anti-depressive action. Above all, the systematic strategy screened 5 brain-targeted effective substances of S. chinensis and it was suggested that exploring the components into nidi would promote the studies on herbs effective material basis. KEY WORDS: Schisandra chinensis, Ethanol extracts, Water extracts, Pharmacodynamics, Chemical profiles, Components into nidi, Effective material basis, Traditional Chinese medicine

A sensitive, fast and comprehensive method for the quality assessment of Semen Ziziphi Spinosae (SZS) standard decoction with characterization of its chemical components was developed and validated. UPLC-Q/TOF-MS/MS system was used to identify thirty-six chemical components of SZS standard decoction which included nucleosides, phenolic acids, alkaloids, and flavonoids. Furthermore, a UPLC-PDA method was validated to simultaneously determine adenosine, protocatechuic acid, magnoflorine, catechin, protocatechin, vicenin II, spinosin, kaempferol-3-rutinoside, and 6'''-feruloylspinosin which represent four species of characteristic compounds. The qualitative method had been validated according to Chinese Pharmacopoeia (2015 edition) in terms of lineary, repeatability, recovery and stability for all analytes, with the results showing good precision, accuracy and stability. In conclusion, the method using UPLC combined with MS and PDA provided a novel way for the standardization and identification of SZS standard decoction, and also offered a basis for qualitative analysis and quality assessment of the preparations for SZS standard decoction. Keywords: Identification, Semen Ziziphi Spinosae (SZS), Standard decoction, Quality assessment, UPLC, TOF-MS/MS

The excessive expression of reactive oxygen species is closely connected to many diseases. Considerable studies have demonstrated dandelion as well as its ingredients exhibited antioxidant activity. However, specific material basis reflecting the antioxidant activity has not been comprehensively investigated. In this study, a spectrum–effect relationship study on dandelion between fingerprinting and antioxidant activity was analyzed in detail, while a UHPLC quantification method developed and completely validated for simultaneous determination of active ingredients in dandelion. With the establishment of dandelion fingerprints of different regions, 24 common peaks were characterized. The classic FRAP method and ABTS methods were then used to detect their antioxidant activity. Partial least squares regression analysis, bivariate correlation analysis and grey correlation method were used to accomplish the spectrum–effect relationship. Eventually, the ingredients with antioxidant activity which could be considered as candidate quality markers of dandelion were discovered through spectrum–effect relationship analysis. The six compounds including caftaric acid, chlorogenic acid, caffeic acid, chicoric acid, isochlorogenic acid A, and isochlorogenic acid C were quantitatively determined. The developed UHPLC assay method was accurate, precise, and reliable. The study has elucidated the antioxidant material basis of dandelion and provided a scientific basis for the quality control of dandelion.

Background: Astaxanthin from natural sources is typically esterified with fatty acids; hence, it must be hydrolyzed to remove esters before identification and quantification by conventional HPLC. Alkaline-catalyzed saponification and enzyme-catalyzed enzymolysis are the most commonly used de-esterification methods. However, information on the efficiency and isomerization during de-esterification of natural astaxanthin esters by these two methods remains scarce. Therefore, we conducted two HPLC-based experiments to determine which method is better for hydrolyzing astaxanthin esters. Results: To assess the effect of enzymolysis (0.67 U/mL cholesterol esterase, at 37°C) and saponification (0.021 M NaOH, at 5°C) conditions on free astaxanthin recovery and destruction or structural transformation of astaxanthin, we varied the total treatment time across a range of 195 min. The results showed that enzymolysis and saponification were complete in 60 min and 90 min, respectively. After complete hydrolysis, the maximum free astaxanthin recovery obtained by enzymolysis was 42.6% more than that obtained by saponification. The identification of by-products, semi-astacene and astacene, during the process of saponification also indicated that a more severe degradation of astaxanthin occurred during saponification. Moreover, the composition of astaxanthin isomers during saponification was similar to that of the isomers during enzymolysis between 30 min and 75 min (all-trans:9-cis:13-cis = 21:3:1, approximately) but dramatically changed after 90 min, whereas the composition in the enzymolysis treatment remained relatively stable throughout. Conclusion: Compared with saponification, enzymolysis with cholesterol esterase was recommended as a more accurate method for de-esterification of natural astaxanthin esters for further qualitative and quantitative HPLC analysis.How to cite: Su F, Xu H, Yang N, et al. Hydrolytic efficiency and isomerization during the process of de-esterifying natural astaxanthin esters using saponification and enzymolysis. Electron J Biotechnol 2018;34. https://doi.org/10.1016/j.ejbt.2018.05.002. Keywords: Alkaline saponification, Astacene, Astaxanthin esters, Astaxanthin hydrolysis, Carotenoid, Cholesterol esterase, De-esterifying astaxanthin, Enzymolysis, Hydrolytic efficiency, Saponification, Semi-astacene

Epidemiological, cross-sectional, and prospective studies have suggested that insomnia, Alzheimer’s disease (AD) and depression are mutually interacting conditions and frequently co-occur. The monoamine and amino acid neurotransmitter systems in central nervous system were involved in the examination of neurobiological processes of this symptom complex. However, few studies have reported systematic and contrastive discussion of different neurotransmitters (NTs) changing in these neurological diseases. Thus, it is necessary to establish a reliable analytical method to monitoring NTs and their metabolite levels in rat brain tissues for elucidating the differences in pathophysiology of these neurological diseases. A rapid, sensitive and reliable LC-MS/MS method was established for simultaneous determination of the NTs and their metabolites, including tryptophan (Trp), tyrosine (Tyr), serotonin (5-HT), 5-hydroxyindolacetic acid (5-HIAA), dopamine (DA), acetylcholine (ACh), norepinephrine (NE), glutamic acid (Glu), and γ-aminobutyric acid (GABA) in rat brain tissues. The mobile phase consisting of methanol and 0.01% formic acid in water was performed on an Inertsil EP C18 column, and the developed method was validated well. Results demonstrated that there were significant differences for 5-HT, DA, NE, Trp, Tyr and ACh between model and control group in all three models, and a Bayes linear discriminant function was established to distinguish these three kinds of nervous system diseases by DA, Tyr and ACh for their significant differences among control and three model groups. It could be an excellent strategy to provide perceptions into the similarity and differentia of mechanisms from the point of NTs’ changing in brain directly and a new method to distinguish insomnia, depression and AD from view of essence.

Headache is a common episodic or chronic neurologic disorder. Treatment options and diagnosis are restricted by an incomplete understanding of disease pathology and the lack of diagnostic markers. Wu-Zhu-Yu decoction (WZYD), a traditional Chinese medicine (TCM) formula containing four TCM herbs, is commonly used in the treatment of headache in China. To deeply understand more about headache and investigate the pain-relief mechanism of WZYD, a comprehensive metabolomics study combined with multivariate data processing strategy was carried out. An LC-high resolution mass spectrometry-based metabolomics approach was applied to characterize metabolic biomarker candidates. Multiple pattern recognition including principal component analysis-discriminant analysis, partial least squares-discriminant analysis and hierarchical cluster analysis were used to determine groups and confirm important variables. A total of 17 potential biomarkers were characterized and related metabolic pathways were identified. The study demonstrated that the established metabolomics strategy is a powerful approach for investigating the mechanism of headache attack and WZYD. In addition, the approach may highlight biomarkers and metabolic pathways and can capture subtle metabolite changes from headache, which may lead to an improved mechanism understanding of central nervous system diseases and TCM treatment.

Yeast two-hybrid (Y2H) methods are powerful tools for detecting protein-protein interactions. The traditional Y2H method has been widely applied to screen novel protein interactions since it was established two decades ago. The high false-positive rate of the traditional method drove the development of modified Y2H systems. Here, we describe a novel Y2H system using zinc-finger nucleases (ZFNs). ZFNs contain two functional domains, a zinc-finger DNA-binding domain (ZFP) and a non-specific nuclease domain (FokI). In this system, the bait is expressed as a fusion protein with a specific ZFP, and the prey is fused to the FokI. A reporter vector is designed such that the ZFN target site disrupts the Gal4 open reading frame. By transforming the three plasmids into a yeast strain (AH109), the interaction between the bait and prey proteins reconstitutes ZFN function and generates the double-strand break (DSB) on its target site. The DNA DSB repair restores Gal4 function, which activates the expression of the four reporter genes. We used p53-SV40LT interacting proteins to prove the concept. In addition, 80% positive rate was observed in a cDNA screening test against WDSV orfA protein. Our results strongly suggested that this Y2H system could increase screening reliability and reproducibility, and provide a novel approach for interactomics research.

DNA binding domain of the transcription activator-like effectors (TALEs) from Xanthomonas sp. consists of tandem repeats that can be rearranged according to a simple cipher to target new DNA sequences with high DNA-binding specificity. This technology has been successfully applied in varieties of species for genome engineering. However, assembling long TALE tandem repeats remains a big challenge precluding wide use of this technology. Although several new methodologies for efficiently assembling TALE repeats have been recently reported, all of them require either sophisticated facilities or skilled technicians to carry them out. Here, we described a simple and efficient method for generating customized TALE nucleases (TALENs) and TALE transcription factors (TALE-TFs) based on TALE repeat tetramer library. A tetramer library consisting of 256 tetramers covers all possible combinations of 4 base pairs. A set of unique primers was designed for amplification of these tetramers. PCR products were assembled by one step of digestion/ligation reaction. 12 TALE constructs including 4 TALEN pairs targeted to mouse Gt(ROSA)26Sor gene and mouse Mstn gene sequences as well as 4 TALE-TF constructs targeted to mouse Oct4, c-Myc, Klf4 and Sox2 gene promoter sequences were generated by using our method. The construction routines took 3 days and parallel constructions were available. The rate of positive clones during colony PCR verification was 64% on average. Sequencing results suggested that all TALE constructs were performed with high successful rate. This is a rapid and cost-efficient method using the most common enzymes and facilities with a high success rate.

Polyamines, one of the most important kind of biomarkers in cancer research, were investigated in order to characterize different cancer types. An integrative approach which combined ultra-high performance liquid chromatography—tandem mass spectrometry detection and multiple statistical data processing strategies including outlier elimination, binary logistic regression analysis and cluster analysis had been developed to discover the characteristic biomarkers of lung and liver cancer. The concentrations of 14 polyamine metabolites in biosamples from lung (n = 50) and liver cancer patients (n = 50) were detected by a validated UHPLC-MS/MS method. Then the concentrations were converted into independent variables to characterize patients of lung and liver cancer by binary logic regression analysis. Significant independent variables were regarded as the potential biomarkers. Cluster analysis was engaged for further verifying. As a result, two values was discovered to identify lung and liver cancer, which were the product of the plasma concentration of putrescine and spermidine; and the ratio of the urine concentration of S-adenosyl-l-methionine and N-acetylspermidine. Results indicated that the established advanced method could be successfully applied to characterize lung and liver cancer, and may also enable a new way of discovering cancer biomarkers and characterizing other types of cancer.

The Wu-Zhu-Yu decoction is a traditional Chinese medicine formula for the treatment of headache. To reveal its material basis, a rapid and reliable liquid chromatography-high resolution mass spectrometry method was established for comprehensive profiling of the chemical ingredients in the Wu-Zhu-Yu decoction. The method was used on a quadrupole time-of-flight mass spectrometer along with an advanced data processing procedure consisting of mass accuracy screening, mass defect filtering and fragment filtering. After eliminating interference with a filtering approach, the MS data profiling was made more distinct and accurate. With the optimized conditions of only 35 min LC separation and single sample injection of each positive or negative ion mode, a total of 168 compounds were characterized, including 23 evodiamine and its analogous alkaloids, 12 limonoids, 17 gingerols, 38 ginsenosides, 15 flavonoids, 16 organic acids, 14 alkaloids, 5 saponins, 3 2,2-dimethylchromenes and 25 other compounds. The fragmentation patterns of representative compounds were illustrated as well. Integrative qualitative analysis of the Wu-Zhu-Yu decoction by high resolution mass spectrometry was accomplished and reported for the first time. The study demonstrated that the established method was a powerful and reliable strategy for comprehensive detection and would be widely applicable for identification of complicated components from herbal prescriptions, and may provide a basis for chemical analysis of other complex mixtures.