The prevalence of networked sensors and actuators in many real-world systems such as smart buildings, factories, power plants, and data centres generate substantial amounts of multivariate time series data for these systems. Many of these cyber-physical systems (CPSs) are engineered for mission-critical tasks and are thus targets for cyber-attacks. The rich sensor data can be continuously monitored for intrusion events through anomaly detection. However, conventional threshold-based anomaly detection methods are inadequate due to the dynamic complexities of these systems, while supervised machine learning methods are unable to exploit the large amounts of data due to the lack of labeled data. On the other hand, current unsupervised machine learning approaches have not fully exploited the spatial-temporal correlation and other dependencies amongst the multiple variables (sensors/actuators) in the system for detecting anomalies. Most of the current techniques also employed simple comparison between the present states and predicted normal ranges for anomaly detection, which can be inadequate given the highly dynamic behaviors of the systems. In this work, we propose an unsupervised multivariate anomaly detection method based on Generative Adversarial Networks (GANs), using the Long-Short-Term-Memory Recurrent Neural Networks (LSTM-RNN) as the base models (namely, the generator and discriminator) in the GAN framework to capture the temporal correlation of time series distributions. Instead of treating each data stream independently, our proposed Multivariate Anomaly Detection with GAN (MAD-GAN) framework considers the entire variable set concurrently to capture the latent interactions amongst the variables. We also fully exploit both the generator and discriminator produced by the GAN, using a novel anomaly score called DR-score to detect anomalies by discrimination and reconstruction. We have tested our proposed MAD-GAN using two recent datasets collected from real world CPS: the Secure Water Treatment (SWaT) and the Water Distribution (WADI) datasets. Our experimental results showed that the proposed MAD-GAN is effective in reporting anomalies caused by various cyberintrusions compared in these complex real-world systems.
AimsPhospholipase A2 receptor (PLA2R) and thrombospondin type-1 domain-containing 7A (THSD7A) were identified as pathogenic antigens in patients with membranous nephropathy (MN). Notably, PLA2R is detected in few patients with malignancy-associated MN, and a high incidence of cancer is reported in patients with THSD7A-associated MN. Therefore, the roles of PLA2R and THSD7A in malignancy-associated MN must be clarified.MethodsSerum anti-PLA2R antibodies and glomerular PLA2R staining were assessed in 36 patients with malignancy-associated MN, followed by examination of serum anti-THSD7A antibodies and glomerular THSD7A. THSD7A staining in cancer tissues was also assessed in 9 of the 36 patients.ResultsTwelve (33%) of 36 patients were positive for both glomerular PLA2R and serum anti-PLA2R antibodies, one of whom had enhanced glomerular THSD7A staining. Two patients were positive for either glomerular PLA2R or serum anti-PLA2R antibody. All these patients had IgG4-dominant deposits in glomeruli. Among the 22 (61%) patients who were double negative for glomerular PLA2R and serum anti-PLA2R antibodies, 17 of 20 (85%) had IgG1-dominant deposits in glomeruli, and 2 (9.1%) were positive for glomerular THSD7A staining. Serum anti-THSD7A antibody was not detected in any of the 36 patients. Among the nine patients with available cancer tissues, positive staining of THSD7A in the cancer tissues was observed in five (56%) patients, and one showed enhanced glomerular staining of THSD7A.ConclusionsScreening of glomerular PLA2R antigen and serum anti-PLA2R antibodies is necessary in patients with malignancy-associated MN, whereas the incidence of glomerular THSD7A antigen or circulating anti-THSD7A antibodies is uncommon.
BackgroundThrombotic microangiopathy (TMA) in the kidney is a histopathologic lesion that occurs in a number of clinical settings and is often associated with poor renal prognosis. The standard test for the diagnosis of TMA is the renal biopsy; noninvasive parameters such as potential biomarkers have not been developed.MethodsWe analyzed routine parameters in a cohort of 220 patients with suspected TMA and developed a diagnostic laboratory panel by logistic regression. The levels of candidate markers were validated using an independent cohort (n = 46), a cohort of systemic lupus erythematosus (SLE) (n = 157) and an expanded cohort (n = 113), as well as 9 patients with repeat biopsies.ResultsOf the 220 patients in the derivation cohort, 51 patients with biopsy-proven TMA presented with a worse renal prognosis than those with no TMA (P = 0.002). Platelet and L-lactate dehydrogenase (LDH) levels showed an acceptable diagnostic value of TMA (AUC = 0.739 and 0.756, respectively). A panel of 4 variables - creatinine, platelets, ADAMTS13 (a disintegrin and metalloprotease with thrombospondin type 1 repeats 13) activity and LDH - can effectively discriminate patients with TMA (AUC = 0.800). In the validation cohort, the platelet and LDH levels and the 4-variable panel signature robustly distinguished patients with TMA. The discrimination effects of these three markers were confirmed in patients with SLE. Moreover, LDH levels and the 4-variable panel signature also showed discrimination values in an expanded set. Among patients undergoing repeat biopsy, increased LDH levels and panel signatures were associated with TMA status when paired evaluations were performed. Importantly, only the 4-variable panel was an independent prognostic marker for renal outcome (hazard ratio = 3.549; P<0.001).ConclusionsThe noninvasive laboratory diagnostic panel is better for the early detection and prognosis of TMA compared with a single parameter, and may provide a promising biomarker for clinical application.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.