Search

Recent years have witnessed a surge of interest in learning representations of graph-structured data, with applications from social networks to drug discovery. However, graph neural networks, the machine learning models for handling graph-structured data, face significant challenges when running on conventional digital hardware, including the slowdown of Moore’s law due to transistor scaling limits and the von Neumann bottleneck incurred by physically separated memory and processing units, as well as a high training cost. Here we present a hardware–software co-design to address these challenges, by designing an echo state graph neural network based on random resistive memory arrays, which are built from low-cost, nanoscale and stackable resistors for efficient in-memory computing. This approach leverages the intrinsic stochasticity of dielectric breakdown in resistive switching to implement random projections in hardware for an echo state network that effectively minimizes the training complexity thanks to its fixed and random weights. The system demonstrates state-of-the-art performance on both graph classification using the MUTAG and COLLAB datasets and node classification using the CORA dataset, achieving 2.16×, 35.42× and 40.37× improvements in energy efficiency for a projected random resistive memory-based hybrid analogue–digital system over a state-of-the-art graphics processing unit and 99.35%, 99.99% and 91.40% reductions of backward pass complexity compared with conventional graph learning. The results point to a promising direction for next-generation artificial intelligence systems for graph learning.

The eastern oyster Crassostrea virginica is a major aquaculture species for the USA. The sustainable development of eastern oyster aquaculture depends upon the continued improvement of cultured stocks through advanced breeding technologies. The Eastern Oyster Breeding Consortium (EOBC) was formed to advance the genetics and breeding of the eastern oyster. To facilitate efficient genotyping needed for genomic studies and selection, the consortium developed two single-nucleotide polymorphism (SNP) arrays for the eastern oyster: one screening array with 566K SNPs and one breeders' array with 66K SNPs. The 566K screening array was developed based on whole-genome resequencing data from 292 oysters from Atlantic and Gulf of Mexico populations; it contains 566,262 SNPs including 47K from protein-coding genes with a marker conversion rate of 48.34%. The 66K array was developed using best-performing SNPs from the screening array, which contained 65,893 oyster SNPs including 22,984 genic markers with a calling rate of 99.34%, a concordance rate of 99.81%, and a much-improved marker conversion rate of 92.04%. Null alleles attributable to large indels were found in 13.1% of the SNPs, suggesting that copy number variation is pervasive. Both arrays provided easy identification and separation of selected stocks from wild progenitor populations. The arrays contain 31 mitochondrial SNPs that allowed unambiguous identification of Gulf mitochondrial genotypes in some Atlantic populations. The arrays also contain 756 probes from 13 oyster and human pathogens for possible detection. Our results show that marker conversion rate is low in high polymorphism species and that the two-step process of array development can greatly improve array performance. The two arrays will advance genomic research and accelerate genetic improvement of the eastern oyster by delineating genetic architecture of production traits and enabling genomic selection. The arrays also may be used to monitor pedigree and inbreeding, identify selected stocks and their introgression into wild populations, and assess the success of oyster restoration.

Doxorubicin (DOX) is a chemotherapeutic drug for a variety of malignancies, while its application is restricted by the cardiovascular toxic effects characterized by oxidative stress. Ferroptosis is a novel iron-dependent regulated cell death driven by lipid peroxidation. Our study aimed to investigate the role of Elabela (ELA) in DOX-induced oxidative stress and ferroptosis. In cultured rat aortic adventitial fibroblasts (AFs), stimulation with DOX dramatically induced cytotoxicity with reduced cell viability and migration ability, and enhanced lactate dehydrogenase (LDH) activity. Importantly, ELA and ferrostatin-1 (Fer-1) mitigated DOX-mediated augmentation of reactive oxygen species (ROS) in rat aortic AFs, accompanied by upregulated levels of Nrf2, SLC7A11, GPX4, and GSH. In addition, ELA reversed DOX-induced dysregulation of apoptosis- and inflammation-related factors including Bax, Bcl2, interleukin (IL)-1β, IL6, IL-10, and CXCL1. Intriguingly, knockdown of Krüppel-like factor 15 (KLF15) by siRNA abolished ELA-mediated alleviation of ROS production and inflammatory responses. More importanly, KLF15 siRNA impeded the beneficial roles of ELA in DOX-pretreated rat aortic AFs by suppressing the Nrf2/SLC7A11/GPX4 signaling. In conclusion, ELA prevents DOX-triggered promotion of cytotoxicity, and exerts anti-oxidative and anti-ferroptotic effects in rat aortic AFs via activation of the KLF15/GPX4 signaling, indicating a promising therapeutic value of ELA in antagonizing DOX-mediated cardiovascular abnormality and disorders.

Previous studies have shown mitochondrial dysfunction in various acute kidney injuries and chronic kidney diseases. Lipoic acid exerts potent effects on oxidant stress and modulation of mitochondrial function in damaged organ. In this study we investigated whether alpha lipoamide (ALM), a derivative of lipoic acid, exerted a renal protective effect in a type 2 diabetes mellitus mouse model. 9-week-old db/db mice were treated with ALM (50 mg·kg

This study involves isothermal kinetic simulation to evaluate the parameters of inhibition conditions for

The preparation of antifungal-functionalized silica nanoparticles by covalent attachment of several silylated derivatives based on the topical antifungal agent Micozanole is described. Grafting and co-condensation procedures are used to obtain mesoporous or dense nanoparticles. Cotton fabrics have been coated with these antifungal-functionalized silica nanoparticles under ultrasonic conditions. The characterization of the functionalized nanoparticles and coated cotton fabrics is performed by microscopic and spectroscopic techniques. The antifungal activity of all the modified cotton textiles is tested against Trichophyton mentagrophytes CECT 2783, Aspergillus brasiliensis CECT 2091, Penicillium chrysogenum CECT 2307 and Candida albicans CECT 1001, resulting to be highly dependent on the microorganism. Remarkably, all of the modified fabrics are active against Candida albicans and Trichophyton mentagrophytes, with up to 75 and 90% effectiveness, respectively. High expectations arise for topical cutaneous applications in wound dressings (bandages, gauzes, strips). Graphical abstract

Mycoviruses are widespread in all major taxonomic groups of filamentous fungi. Previous research has indicated that mycoviruses are associated with the phytopathogenic fungus Botryosphaeria dothidea. In this study, three distinct double-stranded RNA viruses were detected in B. dothidea strain YCLYY11 isolated from a leaf spot of longan (Dimocarpus longana). The results of BLAST analysis revealed that the predicted amino acid sequences of those viruses were similar to those of Botryosphaeria dothidea chrysovirus 1, Botryosphaeria dothidea partitivirus 1, and an apparent novel victorivirus. Sequencing and analysis of the complete genome of the novel victorivirus indicated it is 5218 bp in length and contains two open reading frames (ORFs) that overlap at the tetranucleotide AUGA. BLASTp analysis of the proteins encoded by ORF1 and ORF2 showed that they were most similar to the coat protein and RNA-dependent RNA polymerase of Sphaeropsis sapinea RNA virus 2 (81.37% and 74.09% identical, respectively). A phylogenetic tree showed that the novel virus clustered together with victoriviruses and was separate from members of the other four genera of the family Totiviridae. Based on its genome structure and the results of phylogenetic analysis, we propose that this novel victorivirus should be named "Botryosphaeria dothidea victorivirus 3". This is also the first report of these three mycoviruses coinfecting a strain of B. dothidea.

Non-volatile computing-in-memory macros that are based on two-dimensional arrays of memristors are of use in the development of artificial intelligence edge devices. Scaling such systems to three-dimensional arrays could provide higher parallelism, capacity and density for the necessary vector–matrix multiplication operations. However, scaling to three dimensions is challenging due to manufacturing and device variability issues. Here we report a two-kilobit non-volatile computing-in-memory macro that is based on a three-dimensional vertical resistive random-access memory fabricated using a 55 nm complementary metal–oxide–semiconductor process. Our macro can perform 3D vector–matrix multiplication operations with an energy efficiency of 8.32 tera-operations per second per watt when the input, weight and output data are 8, 9 and 22 bits, respectively, and the bit density is 58.2 bit µm–2. We show that the macro offers more accurate brain MRI edge detection and improved inference accuracy on the CIFAR-10 dataset than conventional methods.

This study aimed to uncover the microRNA and messenger RNA (miRNA/mRNA) interactions in the pathophysiological process of calcified aortic valve disease (CAVD) of the human aortic valve. RNA sequencing of six selected samples (3 healthy control samples vs. 3 CAVD samples) was performed to obtain mRNA and miRNA sequences, and differential expression (DE) analysis of miRNA and mRNAs was performed. To build a CAVD-specific miRNA-mRNA interactome, the upregulated mRNAs and downregulated miRNAs were selected, followed by the establishment of inverse DE of mRNA-miRNA co-expression network based on Pearson's correlation coefficient using miRanda in the R language software. Subsequently, pathway enrichment analysis was performed to elucidate CAVD-related pathways that were likely mediated by miRNA regulatory mechanisms. In addition, miRNAs with an mRNA correlation greater than 0.9 in the co-expression network were selected for anti-calcification verification in a CAVD cellular model. We identified 216 mRNAs (99 downregulated and 117 upregulated) and 602 miRNAs (371 downregulated and 231 upregulated) that were differentially expressed between CAVD and healthy aortic valves. After applying Pearson's correlation toward miRNA-mRNA targets, a regulatory network of 67 miRNAs targeting 76 mRNAs was created. The subsequent pathway enrichment analysis of these targeted mRNAs elucidated that genes within the focal adhesion pathway are likely mediated by miRNA regulatory mechanisms. The selected hsa-miR-629-3p and TAGLN pair exhibited anti-calcification effects on osteogenic differentiation-induced human aortic valve interstitial cells (hVICs). On integrating the miRNA and mRNA sequencing data for healthy aortic valves and those with CAVD, the CAVD-associated miRNA-mRNA interactome and related pathways were elucidated. Additional cell function data demonstrated anti-calcification effects of the selected hsa-miR-629-3p targeting TAGLN, validating that it is a potential therapeutic target for inhibiting CAVD.

Voluntary contribution has become the only source of donor lungs in China since 2015. To elaborate the outcomes of patients awaiting lung transplantation (LTx) after the implementation of donation after brain death, we performed a retrospective study that encompassed 205 patients with end-stage lung disease who registered for LTx at Shanghai Pulmonary Hospital from January 1, 2015 to January 1, 2021. A total of 180 patients were enrolled in the study. The median waiting time was 1.25 months. Interstitial lung disease (ILD) (103/180, 57.2%) and chronic obstructive pulmonary disease (COPD) (56/180, 31.1%) were the most common diseases in our study population. The mean pulmonary artery pressure (mPAP) of patients in the died-waiting group was higher than that of the survivors (53.29±21.71 mmHg vs. 42.11±18.58 mmHg, P=0.002). The mortality of patients with ILD (34/103, 33.00%) was nearly twice that of patients with COPD (10/56, 17.86%) while awaiting LTx (P=0.041). In the died-waiting group, patients with ILD had a shorter median waiting time than patients with COPD after being listed (0.865 months vs. 4.720 months, P=0.030). ILD as primary disease and mPAP35 mmHg were two significant independent risk factors for waitlist mortality, with hazard ratios (HR) of 3.483 (95% CI 1.311-9.111; P=0.011) and 3.500 (95% CI 1.435-8.536; P=0.006). Hence, LTx is more urgently needed in patients with ILD and pulmonary hypertension.

In this paper, we study the following fractional Choquard equation with critical or supercritical growth $$\begin{aligned} \ (-\Delta )^su+V(x)u=f(x,u)+\lambda \left[ |x|^{-\mu }*|u|^p\right] p|u|^{p-2}u, \quad x \in {\mathbb {R}}^N, \end{aligned}$$ ( - Δ ) s u + V ( x ) u = f ( x , u ) + λ | x | - μ ∗ | u | p p | u | p - 2 u , x ∈ R N , where $$0 0 < s < 1 , $$(-\Delta )^s$$ ( - Δ ) s denotes the fractional Laplacian of order s, $$N>2s$$ N > 2 s , $$0 0 < μ < 2 s and $$p\ge 2_{\mu ,s}^*:=\frac{2N-\mu }{N-2s}$$ p ≥ 2 μ , s ∗ : = 2 N - μ N - 2 s , which is the critical exponent in the sense of Hardy-Littlewood-Sobolev inequality. Under some suitable conditions, we prove that the equation admits a nontrivial solution for small $$\lambda >0$$ λ > 0 by variational methods, which extends results in Bhattarai in J. Differ. Equ. 263, 3197–3229 (2017).