Industrial Soft Sensor Optimized by Improved PSO: A Deep Representation-Learning Approach.

Soft sensors based on deep learning approaches are growing in popularity due to their ability to extract high-level features from training, improving soft sensors' performance. In the training process of such a deep model, the set of hyperparameters is critical to archive generalization and reliability. However, choosing the training hyperparameters is a complex task. Usually, a random approach defines the set of hyperparameters, which may not be adequate regarding the high number of sets and the soft sensing purposes. This work proposes the RB-PSOSAE, a Representation-Based Particle Swarm Optimization with a modified evaluation function to optimize the hyperparameter set of a Stacked AutoEncoder-based soft sensor. The evaluation function considers the mean square error (MSE) of validation and the representation of the features extracted through mutual information (MI) analysis in the pre-training step. By doing this, the RB-PSOSAE computes hyperparameters capable of supporting the training process to generate models with improved generalization and relevant hidden features. As a result, the proposed method can generate more than 16.4% improvement in RMSE compared to another standard PSO-based method and, in some cases, more than 50% improvement compared to traditional methods applied to the same real-world nonlinear industrial process. Thus, the results demonstrate better prediction performance than traditional and state-of-the-art methods.

MS-Driven Metabolic Alterations Are Recapitulated in iPSC-Derived Astrocytes.

OBJECTIVE: Astrocytes play a significant role in the pathology of multiple sclerosis (MS). Nevertheless, for ethical reasons, most studies in these cells were performed using the Experimental Autoimmune Encephalomyelitis model. As there are significant differences between human and mouse cells, we aimed here to better characterize astrocytes from patients with MS (PwMS), focusing mainly on mitochondrial function and cell metabolism. METHODS: We obtained and characterized induced pluripotent stem cell (iPSC)-derived astrocytes from three PwMS and three unaffected controls, and performed electron microscopy, flow cytometry, cytokine and glutamate measurements, gene expression, in situ respiration, and metabolomics. We validated our findings using a single-nuclei RNA sequencing dataset. RESULTS: We detected several differences in MS astrocytes including: (i) enrichment of genes associated with neurodegeneration, (ii) increased mitochondrial fission, (iii) increased production of superoxide and MS-related proinflammatory chemokines, (iv) impaired uptake and enhanced release of glutamate, (v) increased electron transport capacity and proton leak, in line with the increased oxidative stress, and (vi) a distinct metabolic profile, with a deficiency in amino acid catabolism and increased sphingolipid metabolism, which have already been linked to MS. INTERPRETATION: Here we describe the metabolic profile of iPSC-derived astrocytes from PwMS and validate this model as a very powerful tool to study disease mechanisms and to perform non-invasive drug targeting assays in vitro. Our findings recapitulate several disease features described in patients and provide new mechanistic insights into the metabolic rewiring of astrocytes in MS, which could be targeted in future therapeutic studies. ANN NEUROL 2022;91:652-669.

Sirtuins as pharmacological targets in neurodegenerative and neuropsychiatric disorders.

Histone deacetylases (HDACs) are enzymes that regulate several processes, such as transcription, cell proliferation, differentiation and development. HDACs are classified as either Zn2+ -dependent or NAD+ -dependent enzymes. Over the years, experimental and clinical evidence has demonstrated that HDAC modulation is a critical process in neurodegenerative and psychiatric disorders. Nevertheless, most of the studies have focused on the role of Zn2+ -dependent HDACs in the development of these diseases, although there is growing evidence showing that the NAD+ -dependent HDACs, known as sirtuins, are also very promising targets. This possibility has been strengthened by reports of decreased levels of NAD+ in CNS disorders, which can lead to alterations in sirtuin activation and therefore result in increased pathology. In this review, we discuss the role of sirtuins in neurodegenerative and neuropsychiatric disorders as well the possible rationale for them to be considered as pharmacological targets in future therapeutic interventions. LINKED ARTICLES: This article is part of a themed issue on Building Bridges in Neuropharmacology. To view the other articles in this section visit http://onlinelibrary.wiley.com/doi/10.1111/bph.v179.8/issuetoc.

Bifidobacterium Strains Present Distinct Effects on the Control of Alveolar Bone Loss in a Periodontitis Experimental Model.

Periodontitis is an inflammatory disease induced by a dysbiotic oral microbiome. Probiotics of the genus Bifidobacterium may restore the symbiotic microbiome and modulate the immune response, leading to periodontitis control. We evaluated the effect of two strains of Bifidobacterium able to inhibit Porphyromonas gingivalis interaction with host cells and biofilm formation, but with distinct immunomodulatory properties, in a mice periodontitis model. Experimental periodontitis (P+) was induced in C57Bl/6 mice by a microbial consortium of human oral organisms. B. bifidum 1622A [B+ (1622)] and B. breve 1101A [B+ (1101)] were orally inoculated for 45 days. Alveolar bone loss and inflammatory response in gingival tissues were determined. The microbial consortium induced alveolar bone loss in positive control (P + B-), as demonstrated by microtomography analysis, although P. gingivalis was undetected in oral biofilms at the end of the experimental period. TNF-α and IL-10 serum levels, and Treg and Th17 populations in gingiva of SHAM and P + B- groups did not differ. B. bifidum 1622A, but not B. breve 1101A, controlled bone destruction in P+ mice. B. breve 1101A upregulated transcription of Il-1ß, Tnf-α, Tlr2, Tlr4, and Nlrp3 in P-B+(1101), which was attenuated by the microbial consortium [P + B+(1101)]. All treatments downregulated transcription of Il-17, although treatment with B. breve 1101A did not yield such low levels of transcripts as seen for the other groups. B. breve 1101A increased Th17 population in gingival tissues [P-B+ (1101) and P + B+ (1101)] compared to SHAM and P + B-. Administration of both bifidobacteria resulted in serum IL-10 decreased levels. Our data indicated that the beneficial effect of Bifidobacterium is not a common trait of this genus, since B. breve 1101A induced an inflammatory profile in gingival tissues and did not prevent alveolar bone loss. However, the properties of B. bifidum 1622A suggest its potential to control periodontitis.

Industrial Semi-Supervised Dynamic Soft-Sensor Modeling Approach Based on Deep Relevant Representation Learning.

Soft sensors based on deep learning have been growing in industrial process applications, inferring hard-to-measure but crucial quality-related variables. However, applications may present strong non-linearity, dynamicity, and a lack of labeled data. To deal with the above-cited problems, the extraction of relevant features is becoming a field of interest in soft-sensing. A novel deep representative learning soft-sensor modeling approach is proposed based on stacked autoencoder (SAE), mutual information (MI), and long-short term memory (LSTM). SAE is trained layer by layer with MI evaluation performed between extracted features and targeted output to evaluate the relevance of learned representation in each layer. This approach highlights relevant information and eliminates irrelevant information from the current layer. Thus, deep output-related representative features are retrieved. In the supervised fine-tuning stage, an LSTM is coupled to the tail of the SAE to address system inherent dynamic behavior. Also, a k-fold cross-validation ensemble strategy is applied to enhance the soft-sensor reliability. Two real-world industrial non-linear processes are employed to evaluate the proposed method performance. The obtained results show improved prediction performance in comparison to other traditional and state-of-art methods. Compared to the other methods, the proposed model can generate more than 38.6% and 39.4% improvement of RMSE for the two analyzed industrial cases.

[Description of an outbreak of lepidopterism (dermatitis associated with contact with moths) among sailors in Salvador, State of Bahia]. / Descrição de um surto de lepidopterismo (dermatite associada ao contato com mariposas) entre marinheiros, ocorrido em Salvador, Estado da Bahia.

An occurrence of pruritic papular dermatitis among the whole crew of a Filipino commercial ship in Salvador, State of Bahia, was associated with contact with Hylesia moths. This unusual type of dermatitis is caused by the bristles (flechettes) on the moths' bodies. Reporting on such cases serves to warn about possible similar situations.