Bilayer Network Research Articles

A cell-free system for functional studies of small membrane proteins

Numerous small proteins have been discovered across all domains of life, among which many are hydrophobic and predicted to localize to the cell membrane. Based on a few that are well-studied, small membrane proteins are regulators involved in various biological processes, such as cell signaling, nutrient transport, drug resistance, and stress response. However, the function of most identified small membrane proteins remains elusive. Their small size and hydrophobicity make protein production challenging, hindering function discovery. Here, we combined a cell-free system with lipid sponge droplets and synthesized small membrane proteins in vitro. Lipid sponge droplets contain a dense network of lipid bilayers, which accommodates and extracts newly synthesized small membrane proteins from the aqueous surroundings. Using small bacterial membrane proteins MgrB, SafA, and AcrZ as proof of principle, we showed that the in vitro produced membrane proteins were functionally active, for example, modulating the activity of their target kinase as expected. The cell-free system produced small membrane proteins, including one from human, up to micromolar concentrations, indicating its high level of versatility and productivity. Furthermore, AcrZ produced in this system was used successfully for in vitro co-immunoprecipitations to identify interaction partners. This work presents a robust alternative approach for producing small membrane proteins, which opens a door to their function discovery in different domains of life.

Ultrasensitive and highly stretchable bilayer strain sensor based on bandage-assisted woven fabric with reduced graphene oxide and liquid metal

Flourishing flexible and wearable strain sensors still meet a challenge to realize ultrahigh sensitivity and wide working range simultaneously through a convenient strategy. Here, a strain sensor is designed based on natural rubber (NR) matrix and a bilayer structure composed of a highly stretchable liquid metal (LM) layer and a susceptible reduced graphene oxide (RGO)@carbonized bandage (CB) layer. By painting LM onto the surface of a graphene oxide (GO)-coated bandage template followed with a rapid flame pyrolysis, a dual conductive-layer woven fabric based on LM/RGO@CB was obtained. NR was then employed as an elastic substrate. Benefiting from the synergistic bilayer network and the high stretchability of the elastomer, the fabricated strain sensor exhibited ultrahigh sensitivity (gauge factor of up to 1.63 × 105), quite wide detection range (strain of 0 ∼ 515 %), low detection limit (0.25 %), fast response/recovery (135 ms/135 ms), and satisfactory durability (5700 cycles). The extraordinary strain-sensing performance allowed comprehensive monitoring of full-range (vigorous and subtle) human motions and physiological activities. This work provides a facile bilayer strategy for the preparation of high-performance strain sensors.

BIKAGCN: Knowledge-Aware Recommendations Under Bi-layer Graph Convolutional Networks

Recommender systems are a popular solution for the problem of information overload, offering personalized recommendations to users. Recent years, research has aimed to enhance recommender systems by employing knowledge graphs in conjunction with Graph convolutional network (GCN) to extract user and item features. Although GCN possess a great potential, they are still far from reaching their full capability in recommender systems. This paper introduces a novel approach—knowledge-aware recommendations under bi-layer graph convolutional networks (BIKAGCN) that combines attention and bi-layer GCNs to improve performance. The first layer of the BIKAGCN model trains embedding representations of users and items based on user-item interaction graphs. The second layer introduces a novel knowledge-aware layer of attention and graph convolutional network (KAGCN) layer that leverages both the first layer’s user-item embeddings and item knowledge graph embeddings. Experimental results on three publicly available datasets (MovieLens-20M, Last-FM, and Book-Crossing) demonstrate that BIKAGCN leads to significant performance improvements in recall@20 metric (14.41%, 8.86%, and 20.90%, respectively) compared to currently available state-of-the-art approaches. Moreover, the model maintains satisfactory performance in cold-start cases.The research provides some guidance for the direction of subsequent research on recommender systems.

Modeling the effect of observational social learning on parental decision-making for childhood vaccination and diseases spread over household networks.

In this paper, we introduce a novel model for parental decision-making about vaccinations against a childhood disease that spreads through a contact network. This model considers a bilayer network comprising two overlapping networks, which are either Erdős-Rényi (random) networks or Barabási-Albert networks. The model also employs a Bayesian aggregation rule for observational social learning on a social network. This new model encompasses other decision models, such as voting and DeGroot models, as special cases. Using our model, we demonstrate how certain levels of social learning about vaccination preferences can converge opinions, influencing vaccine uptake and ultimately disease spread. In addition, we explore how two different cultures of social learning affect the establishment of social norms of vaccination and the uptake of vaccines. In every scenario, the interplay between the dynamics of observational social learning and disease spread is influenced by the network's topology, along with vaccine safety and availability.

ABYOLOv4: improved YOLOv4 human object detection based on enhanced multi-scale feature fusion

The purpose of human object detection is to obtain the number of people and their position in images, which is one of the core problems in the field of machine vision. However, the high missing detection rate from small- and medium-sized human bodies due to the large variety of human scale in human object detection tasks still influences the performance of human object detection. To solve the above problem, this paper proposed an improved ASPP_BiFPN_YOLOv4 (ABYOLOv4) method to detect human object detection. In detail, Atrous Spatial Pyramid Pooling (ASPP) module was used to replace the original Spatial Pyramid Pooling module to increase the receptive field level of the network and improve the perception ability of multi-scale targets. Then, the original Path Aggregation Network (PANet) multi-scale fusion module was replaced by the self-built bi-layer bidirectional feature pyramid network (Bi-FPN). Meanwhile, a new feature was imported into the proposed model to reuse the mid- and low-level features, which could enhance the ability of the network to express the characteristics of small- and medium-sized targets. Finally, the standard convolution in Bi-FPN was replaced by depth-separable convolution to make the network achieve the balance of accuracy and the number of parameters. To identify the performance of the proposed ABYOLOv4 model, the human object detection experiment is carried out by using the public data set of VOC2007 and VOC2012, the improved YOLOv4 algorithm is 0.5% higher than the original AP algorithm, and the weight file size of the model is reduced by 45.3 M. The experimental results demonstrated that the proposed ABYOLOv4 network has higher accuracy and lower computational cost for human target detection.

3D-printed microfluidic-microwave device for droplet network formation and characterisation.

Microfluidic-microwave devices (MMDs) have emerged as precision tools for the rapid, accurate, sensitive, and non-invasive characterisation of liquids in low-volumes. However, the fabrication of MMDs remains a significant challenge. This is due to the complexities associated with integrating fluidic ducts and electronic components. Herein, we present a versatile and economical 3D-printing approach using ducts filled with liquid metal as an electrical conductor. Cyclic olefin copolymer, polylactic acid, and polypropylene were identified as printable dielectric materials for MMD fabrication. Substrates of 3D-printed cyclic olefin copolymer exhibited the lowest loss tangent (0.002 at 2.7 GHz), making them suitable materials for high-frequency microwave devices. Liquid metal, specifically gallium-indium eutectic, was injected into the printed ducts to form electrically conductive microwave structures. Exemplary MMDs operating at 2 GHz integrated split-ring microwave resonators that serve as sensitive detection geometries able to measure changes in dielectric properties, with droplet-forming fluidic junctions and flow channels. The performance of 3D-printed MMDs for microwave droplet sensing was comprehensively evaluated. These devices were used in the formation and characterisation of water-in-oil emulsions, constructing definable lipid-segregated droplet interface bilayer (DIB) networks. This work indicates the feasibility of using 3D-printed manifolds for the rapid prototyping of customised MMDs, and also demonstrates the potential of MMDs as new analytical research tools in droplet-based materials and biochemistry studies.

Accurate construction of monolayer, bilayer, sandwich bilayer, four-layer, multi-layer and chiral bilayer 2D pillararene-type supramolecular networks.

The accurate construction of mono-, bi- and multi-layer networks has been an important challenge, especially for bi- and multi-layer networks. Monolayer, bilayer, sandwich bilayer, four-layer, and multi-layer two-dimensional pillararene-type metal-organic coordination networks have been constructed from functionalized pillar[5]arene and pillar[6]arene by utilizing the coordination interaction of cobalt and copper ions and combining with temperature control and guest induction. These two-dimensional coordination networks exhibit the excellent plasticity of pillararenes and structural variety, which are characterized by X-ray single crystal diffraction and PXRD, confirming that pillararenes units can function as excellent tunable scaffolds for structural regulation. Two-dimensional chiral double-layer structure products are also constructed from R- and S-pillar[6]arene, which are obtained by high-performance liquid chromatography. Atomic force microscopic imaging confirms the thicknesses of these networks. Moreover, these networks also exhibit high iodine adsorption capacity in aqueous environments at ambient temperature. The monolayer, bilayer, sandwich bilayer, four-layer and multi-layer structures of the pillararene-type networks represent a new facile supramolecular self-assembly strategy and platform for designing more mono-, bi- and multi-layer two-dimensional nanomaterials and chiral two-dimensional double-layer structures provide a new method for the construction of more two-dimensional chiral polymers.

Mannan-oligosaccharides promote gut microecological recovery after antibiotic disturbance.

Antibiotic treatment often causes collateral damage to the gut microbiota, including changes in its diversity and composition. Dietary fiber helps maintain intestinal health, regulate short-chain fatty acids, and promote the recovery of the intestinal microbiome. However, it is currently unknown which specific plant-based dietary fiber is optimal as a dietary supplement for restoring the intestinal microbiota after antibiotic disturbance. Previously, we proposed predictive recovery-associated bacterial species (p-RABs) and identified the most important interventions. This study aimed to identify an optimal form of dietary fiber to recover the gut microbiome after antibiotic treatment. Therefore, we examined the types of dietary fibers associated with p-RABs through a p-RAB-metabolite bilayer network constructed from prior knowledge; we searched for dietary fiber that could provide nutritional support for Akkermansia muciniphila and Bacteroides uniformis. C57BL/6J mice were fed with 500 mg kg-1 of different types of dietary fibers daily for one week after being treated with ampicillin. The results showed that mannan-oligosaccharides could better promote the diversity of intestinal microbial growth, enhance the recovery of most genera, including Akkermansia and Bacteroides, and inhibit certain pathogenic bacteria, such as Proteus, compared to the other fiber types. Furthermore, mannan-oligosaccharides could regulate the levels of short-chain fatty acids, especially butyric acid. Functional predictions showed that starch metabolism, galactose metabolism, and the metabolism of other carbohydrates played key roles in the early recovery process. In conclusion, mannan-oligosaccharides could enhance the recovery of the intestinal microbiome after antibiotic treatment, offering valuable insights for targeted dietary strategies.

Adaptive convolutional dictionary learning for denoising seismocardiogram to enhance the classification performance of aortic stenosis

Background:Aortic stenosis (AS) is the most prevalent type of valvular heart disease (VHD), traditionally diagnosed using echocardiogram or phonocardiogram. Seismocardiogram (SCG), an emerging wearable cardiac monitoring modality, is proved to be feasible in non-invasive and cost-effective AS diagnosis. However, SCG waveforms acquired from patients with heart diseases are typically weak, making them more susceptible to noise contamination. While most related researches focus on motion artifacts, sensor noise and quantization noise have been mostly overlooked. These noises pose additional challenges for extracting features from the SCG, especially impeding accurate AS classification. Method:To address this challenge, we present a convolutional dictionary learning-based method. Based on sparse modeling of SCG, the proposed method generates a personalized adaptive-size dictionary from noisy measurements. The dictionary is used for sparse coding of the noisy SCG into a transform domain. Reconstruction from the domain removes the noise while preserving the individual waveform pattern of SCG. Results:Using two self-collected SCG datasets, we established optimal dictionary learning parameters and validated the denoising performance. Subsequently, the proposed method denoised SCG from 50 subjects (25 AS and 25 non-AS). Leave-one-subject-out cross-validation (LOOCV) was applied to 5 machine learning classifiers. Among the classifiers, a bi-layer neural network achieved a moderate accuracy of 90.2%, with an improvement of 13.8% from the denoising. Conclusions:The proposed sparsity-based denoising technique effectively removes stochastic sensor noise and quantization noise from SCG, consequently improving AS classification performance. This approach shows promise for overcoming instrumentation constraints of SCG-based diagnosis.

Numerical investigation of the Poisson's ratio of amorphous graphene sheets

AbstractThe one‐atom‐thick allotrope of graphite, C4 is called monolayer graphene and was discovered in 2004. It is well known for its fantastic electro‐mechanical properties. While transverse contraction and Poisson's ratio were previously studied only for homogeneous or crystalline two‐dimensional (2D) mono‐ and bilayer network structures under uniaxial tensile stress, we now numerically investigate these material properties for amorphous or non‐crystalline monolayer graphene. Here, we pay special attention to the influence of the network heterogeneity. We find the stress‐strain correlation in good agreement with the literature, and we find auxetic behavior, that is, negative Poisson's ratio, over a wide range within the elastic regime.

Inter-layer propagation of spiral waves: Effects of time-varying defect blocks and magnetic flows

The variation of magnetic flux with distance and time is relatively common in nature. This paper investigates the inter-layer propagation of spiral waves in bi-layer Hindmarsh-Rose networks under gradient magnetic field and defect blocks motion, with the following conclusions: (1) Magnetic flux slows down the propagation of waves. (2) For defect blocks moving linearly or away from each other, spiral waves appear in the network, while the magnetic flux suppresses the spiral wave formation. (3) For defect blocks close to or merging with each other, the magnetic flow can only convert spiral waves in the network into “eye-like waves”. These findings imply that the propagation of spiral waves is sensitive to the action site of magnetic flux, which may be important for scar tissue (e.g., cardiac tissue with poor postoperative recovery). Moreover, three statistical parameters defined to detect waveform evolution in excitable media show good performance.

Characterization of the physiochemical properties, microstructure, and molecular interactions of a novel rice-pea protein gel.

The application of rice and pea proteins in food production is limited due to their undesirable processing performance. The objective of this research was to develop a novel rice-pea protein gel using alkali-heat treatment. This gel had a higher solubility, stronger gel strength, better water retention capacity, and denser bilayer network. This is due to the alkali-heat induced modifications for the secondary structures of proteins (i.e., a decrease in the α-helix, and an increase in the β-sheets) and the interactions between protein molecules. The network structure of gel was more compact by adding 2% and 4% alkali-heat rice protein (AH-RP). This resulted in a stable double-layer network structure of gel. Adding 4% AH-RP significantly improved the hardness and elasticity of gel. This gel will have a good potential use for being the ingredient to produce the functional foods and meat analogs.

Two new Cd(II)-based coordination polymers: Dual sensor for sulfamethazine antibiotic and ferric ion

The precise combination of basic functionality and luminescence property can deliver multifunctional coordination polymers (CPs) with multifarious applications such as sensitive and selective detection of antibiotics and metal ions sensing. Herein, two new Cd(II)-based CPs having formula [Cd2(L)(bpy)2(H2O)2·4H2O] (1) and [Cd2(L)(phen)(H2O)] (2) have been synthesized and characterized using mixed ligand strategy by solvothermal reaction of Cd(II) salts using 1,1′-biphenyl-2,3,3′,5′-tetracarboxylic acid (H4L) as the main ligand and 2,2′-bipyridyl (bpy) as co-ligand for CP 1 and 1,10-phenanthroline (phen) as co-ligand for CP 2. The CP 1 possesses (4,4)-connected sql topology, while CP 2 displays 2D bilayer network composed of ring-like Cd2(OCO)6 subunits. The sensing studies indicated that both 1 and 2 can behave as luminescent sensor for sensitive and selective detection of sulfamethazine antibiotic with remarkable quenching and low limit of detection. Additionally, both CPs also displayed selective sensing for Fe3+ ions. The possible alleviation in the luminescence intensities of 1 and 2 in presence of antibiotic have been addressed using theoretical calculations which indicated that charge transfer as well as the weak interaction between antibiotic and CP may be responsible for the decline in the emission intensity.

Sponge-phase Lipid Droplets as Synthetic Organelles: An Ultrafast Study of Hydrogen Bonding and Interfacial Environments.

Bottom-up design of biomimetic organelles has gained recent attention as a route towards understanding the transition between non-living matter and life. Despite various artificial lipid membranes being developed, the specific relations between lipid structure, composition, interfacial properties, and morphology are not currently understood. Sponge-phase droplets contain dense, nonlamellar lipid bilayer networks that capture the complexities of the endoplasmic reticulum (ER), making them ideal artificial models of such organelles. Here, we combine ultrafast two-dimensional infrared (2D IR) spectroscopy and molecular dynamics simulations to investigate the interfacial H-bond networks in sponge-phase droplets composed of glycolipid and nonionic detergents. In the sponge phase, the interfacial environments are more hydrated and water molecules confined to the nanometer-scale aqueous channels in the sponge phase exhibit dynamics that are significantly slower compared to bulk water. Surfactant configurations and microscopic phase separation play a dominant role in determining membrane curvature and slow dynamics observed in the sponge phase. The studies suggest that H-bond networks within the nanometer-scale channels are disrupted not only by confinement but also by the interactions of surfactants, which extend 1-2 nm from the bilayer surface. The results provide a molecular-level description for controlling phase and morphology in the design of synthetic lipid organelles.

Characterization of anomalous diffusion through convolutional transformers

The results of the Anomalous Diffusion Challenge (AnDi Challenge) (Muñoz-Gil G et al 2021 Nat. Commun. 12 6253) have shown that machine learning methods can outperform classical statistical methodology at the characterization of anomalous diffusion in both the inference of the anomalous diffusion exponent α associated with each trajectory (Task 1), and the determination of the underlying diffusive regime which produced such trajectories (Task 2). Furthermore, of the five teams that finished in the top three across both tasks of the AnDi Challenge, three of those teams used recurrent neural networks (RNNs). While RNNs, like the long short-term memory network, are effective at learning long-term dependencies in sequential data, their key disadvantage is that they must be trained sequentially. In order to facilitate training with larger data sets, by training in parallel, we propose a new transformer based neural network architecture for the characterization of anomalous diffusion. Our new architecture, the Convolutional Transformer (ConvTransformer) uses a bi-layered convolutional neural network to extract features from our diffusive trajectories that can be thought of as being words in a sentence. These features are then fed to two transformer encoding blocks that perform either regression (Task 1 1D) or classification (Task 2 1D). To our knowledge, this is the first time transformers have been used for characterizing anomalous diffusion. Moreover, this may be the first time that a transformer encoding block has been used with a convolutional neural network and without the need for a transformer decoding block or positional encoding. Apart from being able to train in parallel, we show that the ConvTransformer is able to outperform the previous state of the art at determining the underlying diffusive regime (Task 2 1D) in short trajectories (length 10–50 steps), which are the most important for experimental researchers.

Bilayer network spreading dynamics driven by community structure and activity

Epidemic outbreaks in the real world are often accompanied by rapid information diffusion, which will change individual behavior patterns and affect the spread of epidemics. The community phenomenon in human society will also have an important influence on the spread of epidemics. The above factors to construct a new bilayer network are considered in this work. The activity-driven model is used to generate time-varying online information contact layer network and offline physical contact layer network. The information diffusion of individual online contact layer is used to affect the epidemic spreading dynamics of offline physical contact layer, and the individual mobility factor is used to control the community structure characteristics. In order to obtain the spreading dynamic equation of the whole network and analyze the spreading threshold of the network effectively, the microscopic Markov chain (MMC) approach is improved and extended to time-varying networks. Experimental verification of Monte Carlo simulations shows that the proposed method is highly accurate in predicting epidemic outbreak thresholds. The results show that individual mobility has no effect on the epidemic outbreak threshold, but it will affect the final number of infections in each community. The greater the individual contact capability of the online contact layer, the smaller the individual contact capability of the offline contact layer that can effectively suppress the epidemic spread. The above findings can present an important reference for effectively preventing and controlling the epidemic transmission in the real world.

Identifying Tumor-Associated Genes from Bilayer Networks of DNA Methylation Sites and RNAs.

Network theory has attracted much attention from the biological community because of its high efficacy in identifying tumor-associated genes. However, most researchers have focused on single networks of single omics, which have less predictive power. With the available multiomics data, multilayer networks can now be used in molecular research. In this study, we achieved this with the construction of a bilayer network of DNA methylation sites and RNAs. We applied the network model to five types of tumor data to identify key genes associated with tumors. Compared with the single network, the proposed bilayer network resulted in more tumor-associated DNA methylation sites and genes, which we verified with prognostic and KEGG enrichment analyses.

Laminin matrix adhesion regulates basal mammary epithelial cell identity.

Mammary epithelium is a bilayered ductal network composed of luminal and basal epithelial cells, which together drive the growth and functional differentiation of the gland. Basal mammary epithelial cells (MECs) exhibit remarkable plasticity and progenitor activity that facilitate epithelial expansion. However, their activity must be tightly regulated to restrict excess basal cell activity. Here, we show that adhesion of basal cells to laminin α5-containing basement membrane matrix, which is produced by luminal cells, presents such a control mechanism. Adhesion to laminin α5 directs basal cells towards a luminal cell fate, and thereby results in a marked decrease of basal MEC progenitor activity in vitro and in vivo. Mechanistically, these effects are mediated through β4-integrin and activation of p21 (encoded by CDKN1A). Thus, we demonstrate that laminin matrix adhesion is a key determinant of basal identity and essential to building and maintaining a functional multicellular epithelium.

Power Grid Dispatching Optimization Assistant Decision-making Model Based on Deep Learning

At present, the power grid dispatching assistant decision-making model relies on the model-driven solution method, which leads to too long a calculation time. Therefore, an auxiliary decision-making model for power grid dispatching optimization based on a deep learning algorithm is proposed. The objective function of power grid dispatching optimization is defined for the power grid connected with multiple renewable energy sources. Relying on cloud computing technology, the power grid dispatching and monitoring data are obtained and saved as different knowledge systems. The long and short-term memory network is selected from the deep learning field, and the Bi-LSTM network structure unit is designed. The scheduling decision model is constructed using the bi-layer Bi-LSTM neural network, and the optimal scheduling decision is obtained by deep learning of the historical scheduling data. The calculation results show that the longest time of the model is 0.04s. The shortest time is only 0.02s, which improves the efficiency of auxiliary decision-making and enhances the stability of power grid operation to a certain extent.

SPCMLMI: A structural perturbation-based matrix completion method to predict lncRNA-miRNA interactions.

Accumulating evidence indicated that the interaction between lncRNA and miRNA is crucial for gene regulation, which can regulate gene transcription, further affecting the occurrence and development of many complex diseases. Accurate identification of interactions between lncRNAs and miRNAs is helpful for the diagnosis and therapeutics of complex diseases. However, the number of known interactions of lncRNA with miRNA is still very limited, and identifying their interactions through biological experiments is time-consuming and expensive. There is an urgent need to develop more accurate and efficient computational methods to infer lncRNA-miRNA interactions. In this work, we developed a matrix completion approach based on structural perturbation to infer lncRNA-miRNA interactions (SPCMLMI). Specifically, we first calculated the similarities of lncRNA and miRNA, including the lncRNA expression profile similarity, miRNA expression profile similarity, lncRNA sequence similarity, and miRNA sequence similarity. Second, a bilayer network was constructed by integrating the known interaction network, lncRNA similarity network, and miRNA similarity network. Finally, a structural perturbation-based matrix completion method was used to predict potential interactions of lncRNA with miRNA. To evaluate the prediction performance of SPCMLMI, five-fold cross validation and a series of comparison experiments were implemented. SPCMLMI achieved AUCs of 0.8984 and 0.9891 on two different datasets, which is superior to other compared methods. Case studies for lncRNA XIST and miRNA hsa-mir-195-5-p further confirmed the effectiveness of our method in inferring lncRNA-miRNA interactions. Furthermore, we found that the structural consistency of the bilayer network was higher than that of other related networks. The results suggest that SPCMLMI can be used as a useful tool to predict interactions between lncRNAs and miRNAs.