Topology Prediction Research Articles

A Structural-Similarity Conditional GAN Method to Generate Real-Time Topology for Shell–Infill Structures

Topology optimization (TO) can generate innovative conceptual configurations with shell–infill geometric features by distributing materials optimally within the design domain. However, physics-based topology optimization methods require repeated finite element analysis and variable updating, in which expensive computational cost limits their applications in wider industrial fields, especially for topology optimization for shell–infill structures. Fortunately, the arising of the data-based topology optimization method using deep learning has paved the way to realize real-time topology prediction for shell–infill structures. In this work, a novel and differentiable structural similarity (SSIM) loss function is introduced into the conditional generative adversarial network (cGAN) to construct the SSIM-cGAN model, and the single-channel coding strategy of initial condition is proposed to simplify the inputs of the deep learning model. SSIM-cGAN can generate shell–infill structures in real time after training with a small-scale dataset. The results generated by SSIM-cGAN and cGAN were put together for comparison, demonstrating that the shell–infill structure generated by SSIM-cGAN has lower error than cGAN, and the shell layer and porous infill structures are more integrated.

DTSEA: A network-based drug target set enrichment analysis method for drug repurposing against COVID-19

The Coronavirus Disease 2019 (COVID-19) pandemic is still wreaking havoc worldwide. Therefore, the urgent need for efficient treatments pushes researchers and clinicians into screening effective drugs. Drug repurposing may be a promising and time-saving strategy to identify potential drugs against this disease. Here, we developed a novel computational approach, named Drug Target Set Enrichment Analysis (DTSEA), to identify potent drugs against COVID-19. DTSEA first mapped the disease-related genes into a gene functional interaction network, and then it used a network propagation algorithm to rank all genes in the network by calculating the network proximity of genes to disease-related genes. Finally, an enrichment analysis was performed on drug target sets to prioritize disease-candidate drugs. It was shown that the top three drugs predicted by DTSEA, including Ataluren, Carfilzomib, and Aripiprazole, were significantly enriched in the immune response pathways indicating the potential for use as promising COVID-19 inhibitors. In addition to these drugs, DTSEA also identified several drugs (such as Remdesivir and Olumiant), which have obtained emergency use authorization (EUA) for COVID-19. These results indicated that DTSEA could effectively identify the candidate drugs for COVID-19, which will help to accelerate the development of drugs for COVID-19. We then performed several validations to ensure the reliability and validity of DTSEA, including topological analysis, robustness analysis, and prediction consistency. Collectively, DTSEA successfully predicted candidate drugs against COVID-19 with high accuracy and reliability, thus making it a formidable tool to identify potential drugs for a specific disease and facilitate further investigation.

Graph neural networks and implicit neural representation for near-optimal topology prediction over irregular design domains

This paper proposes a deep neural network-based topology optimization acceleration method for irregular design domains that predicts (near-)optimal topologies. A topology optimization problem is a complex non-Euclidean data, which can be embedded in a graph form, and a graph neural network encodes it to Euclidean data such as vectors and matrices. The encoded information is applied to a multi-layer perceptron-based implicit neural representation model, and the multi-layer perceptron approximator predicts the compliance optimal material distribution. The prediction performance of the proposed encoder-approximator architecture is evaluated for several topology optimization problems. The trained network provides 96.6% compliance accuracy, except for 8.0% of the outliers. The two criteria have been investigated to estimate potential outliers, and post-optimization can resolve the outlier within fewer iterations than the original optimization.

Optimal Cluster-Based Topology and Deep LSTM-Based Prediction Method for Data Reduction in IoT

In the Wireless Sensor Network (WSN), the data prediction approach is needed to attain data effectively by diminishing node energy consumption. Hence, in this research, Water cycle Fire Fly Optimization and Deep Long Short-Term Memory (WCFO+ Deep LSTM) approach is employed for aggregation and reduction of data. The processes involved in the developed method are node simulation, cluster-based topology construction, routing tree construction, and data aggregation. Initially, IoT nodes are simulated in the network environment. The cluster-based topology construction is made using the WCFO algorithm. The WCFO is developed by the integration of the Firefly Optimization Algorithm (FOA) and Water cycle algorithm (WCA). The cluster-based topology is constructed by considering the objective function that includes the parameters, including distance, delay, link quality, and energy. After that, the routing process is performed using developed WCFO approach for constructing a routing tree and estimating the optimal path. Finally, the Deep LSTM is trained by the proposed WCFO algorithm, which is utilized for executing data reduction and data aggregation process with minimum energy consumption. The devised WCFO+ Deep LSTM approach achieved better performance in terms of prediction error, delay, energy, and Packet delivery ratio (PDR) with values 0.029, 0.001[Formula: see text]s, 0.161[Formula: see text]J and 99.054%, respectively.

Learning Optimal Topology for Ad-Hoc Robot Networks

In this letter, we synthesize a data-driven method to predict the optimal topology of an ad-hoc robot network. This problem is technically a multi-task classification problem. However, we divide it into a class of multi-class classification problems that can be more efficiently solved. For this purpose, we first compose an algorithm to create ground-truth optimal topologies associated with various configurations of a robot network. This algorithm incorporates a complex collection of optimality criteria that our learning model successfully manages to learn. This model is an stacked ensemble whose output is the topology prediction for a particular robot. Each stacked ensemble instance constitutes three low-level estimators whose outputs will be aggregated by a high-level boosting blender. Applying our model to a network of 10 robots displays over 80% accuracy in the prediction of optimal topologies corresponding to various configurations of the cited network.

Time-Series Representation Learning in Topology Prediction for Passive Optical Network of Telecom Operators

The passive optical network (PON) is widely used in optical fiber communication thanks to its low cost and low resource consumption. However, the passiveness brings about a critical problem that it requires manual work to identify the topology structure, which is costly and prone to bringing noise to the topology logs. In this paper, we provide a base solution firstly introducing neural networks for such problems, and based on that solution we propose a complete methodology (PT-Predictor) for predicting PON topology through representation learning on its optical power data. Specifically, we design useful model ensembles (GCE-Scorer) to extract the features of optical power with noise-tolerant training techniques integrated. We further implement a data-based aggregation algorithm (MaxMeanVoter) and a novel Transformer-based voter (TransVoter) to predict the topology. Compared with previous model-free methods, PT-Predictor is able to improve prediction accuracy by 23.1% in scenarios where data provided by telecom operators is sufficient, and by 14.8% in scenarios where data is temporarily insufficient. Besides, we identify a class of scenarios where PON topology does not follow a strict tree structure, and thus topology prediction cannot be effectively performed by relying on optical power data alone, which will be studied in our future work.

Secondary and Topological Structural Merge Prediction of Alpha-Helical Transmembrane Proteins Using a Hybrid Model Based on Hidden Markov and Long Short-Term Memory Neural Networks.

Alpha-helical transmembrane proteins (αTMPs) play essential roles in drug targeting and disease treatments. Due to the challenges of using experimental methods to determine their structure, αTMPs have far fewer known structures than soluble proteins. The topology of transmembrane proteins (TMPs) can determine the spatial conformation relative to the membrane, while the secondary structure helps to identify their functional domain. They are highly correlated on αTMPs sequences, and achieving a merge prediction is instructive for further understanding the structure and function of αTMPs. In this study, we implemented a hybrid model combining Deep Learning Neural Networks (DNNs) with a Class Hidden Markov Model (CHMM), namely HDNNtopss. DNNs extract rich contextual features through stacked attention-enhanced Bidirectional Long Short-Term Memory (BiLSTM) networks and Convolutional Neural Networks (CNNs), and CHMM captures state-associative temporal features. The hybrid model not only reasonably considers the probability of the state path but also has a fitting and feature-extraction capability for deep learning, which enables flexible prediction and makes the resulting sequence more biologically meaningful. It outperforms current advanced merge-prediction methods with a Q4 of 0.779 and an MCC of 0.673 on the independent test dataset, which have practical, solid significance. In comparison to advanced prediction methods for topological and secondary structures, it achieves the highest topology prediction with a Q2 of 0.884, which has a strong comprehensive performance. At the same time, we implemented a joint training method, Co-HDNNtopss, and achieved a good performance to provide an important reference for similar hybrid-model training.

Unveiling Hidden Zeolitic Imidazolate Frameworks Guided by Intuition-Based Geometrical Factors.

Herein, synthesizable candidate topologies to form zeolitic imidazolate frameworks (ZIFs) are efficiently identified from over 2000000 hypothetical structures in zeolite databases, using structural descriptors extracted from known ZIFs. A combination of intuition-based structural descriptors, such as ring patterns, node numbers, and TOT bridging angles (T = tetrahedral metal nodes in zeolites and ZIFs), is used as data filters to eliminate topologies infeasible for ZIF formation. Carefully chosen structural descriptors facilitate the prediction of plausible ZIF topologies. To investigate potential applications as porous ZIFs, this work performs hydrogen adsorption screening and suggested notable target ZIFs. The collection of new plausible ZIFs, derived from the combined descriptors, will be a structural blueprint for synthetic chemists.

Probabilistic Network Topology Prediction for Active Planning: An Adaptive Algorithm and Application

This article tackles the problem of active planning to achieve cooperative localization for multirobot systems under measurement uncertainty in GNSS-limited scenarios. Specifically, we address the issue of accurately predicting the probability of a future connection between two robots equipped with range-based measurement devices. Due to the limited range of the equipped sensors, edges in the network connection topology will be created or destroyed as the robots move with respect to one another. Accurately predicting the future existence of an edge, given imperfect state estimation and noisy actuation, is therefore a challenging task. An adaptive power series expansion (or APSE) algorithm is developed based on current estimates and control candidates. Such an algorithm applies the power series expansion formula of the quadratic positive form in a normal distribution. Finite-term approximation is made to realize the computational tractability. Further analyses are presented to show that the truncation error in the finite-term approximation can be theoretically reduced to a desired threshold by adaptively choosing the summation degree of the power series. Several sufficient conditions are rigorously derived as the selection principles. Finally, extensive simulation results and comparisons, with respect to both single and multirobot cases, validate that a formally computed and therefore more accurate probability of future topology can help improve the performance of active planning under uncertainty.

General features of transmembrane beta barrels from a large database.

Large datasets for particular classes of proteins can contribute a birds-eye view that is more accurate than that gathered from studying particular exemplars. We used co-evolution data in the form of predicted contact maps to create a large, high-quality database of transmembrane β-barrels (TMBB). We classify sequences for our database using IsItABarrel, an approach combining topology prediction and rule-based TMBB classification. By applying simple feature detection on generated contact maps, our method achieves 96% balanced accuracy when discriminating this structural fold from other protein classes. Our database contains almost two million bacterial TMBB proteins--significantly larger and more accurate than previous datasets. We assessed the TMBB-content of 2,959 proteomes of bacterial organisms across several taxonomic categories and find tremendous variance among TMBB-containing organisms with some having as many as 300 TMBBs (6.79% of proteome) and others as few as 10 (0.27% of proteome). The distribution of the lengths of the TMBBs is suggestive of previously hypothesized duplication events. In addition, we find that the C-terminal β-signal varies among different classes of bacteria and type of protein though it is most commonly GHyGHyGφ+φ. The database is available through our IsItABarrel webapp and we anticipate that its quality and size will serve as a useful resource where high quality TMBB sequence data is required.

Proteome-Wide Detection and Annotation of Receptor Tyrosine Kinases (RTKs): RTK-PRED and the TyReK Database.

Receptor tyrosine kinases (RTKs) form a highly important group of protein receptors of the eukaryotic cell membrane. They control many vital cellular functions and are involved in the regulation of complex signaling networks. Mutations in RTKs have been associated with different types of cancers and other diseases. Although they are very important for proper cell function, they have been experimentally studied in a limited range of eukaryotic species. Currently, there is no available database for RTKs providing information about their function, expression, and interactions. Therefore, the identification of RTKs in multiple organisms, the documentation of their characteristics, and the collection of related information would be very useful. In this paper, we present a novel RTK detection pipeline (RTK-PRED) and the Receptor Tyrosine Kinases Database (TyReK-DB). RTK-PRED combines profile HMMs with transmembrane topology prediction to identify and classify potential RTKs. Proteins of all eukaryotic reference proteomes of the UniProt database were used as input in RTK-PRED leading to a filtered dataset of 20,478 RTKs. Based on the information collected for these RTKs from multiple databases, the relational TyReK database was created.

An investigation on the wake flow of a generic ship using IDDES: The effect of computational parameters

The study of the ship airwake is critical as it explores the effect of unsteady wake flow on helicopter operations. This paper studied the near-wake flow topology of a generic ship at Re = 8 × 104 to assess the capability of a hybrid RANS/LES (Reynolds-averaged Navier–Stokes/large-eddy simulation) approach, known as IDDES (improved delayed detached-eddy simulation). The impact of computational parameters, including the mesh grid, residual level, time-step size, momentum discretization scheme and transient formulation, on the wake flow was investigated. The numerical results were validated by using the previous experimental data and LES results. The results show that except of the mesh resolution all other computational parameters varied in the current study do not have significant effect on the global drag forces, but showing large differences on the prediction of the local wake flow structures. This point has not been evidently reported in the previous work. The finer grid resolution is sufficient to produce an accurate qualitative and quantitative prediction of the flow structures, while using a poor grid resolution (coarse mesh) leads to inaccurate prediction of the flow topology. The recommended parameters for the time-step size (7.5 × 10−5s) and residual level (1 × 10−4) provide sufficient accuracy of wake predictions, showing good agreement with reference studies. For the convective term of the momentum equation in IDDES, the bounded central differencing scheme is proposed for its discretization, while the bounded second-order implicit is proposed to be adopted as the transient formulation.

Net-clipping as a top-down approach for the prediction of topologies of MOFs built from reduced-symmetry linkers.

Reticular materials constructed from regular molecular building blocks (MBBs) have been widely explored in the past three decades. Recently, there has been increasing interest in the assembly of novel, intricate materials using less-symmetric ligands; however, current methods for predicting structure are not amenable to this increased complexity. To address this gap, we propose herein a generalised version of the net-clipping approach for anticipating the topology of metal-organic frameworks (MOFs) assembled from organic linkers and different polygonal and polyhedral MBBs. It relies on the generation of less-symmetric nets with less-connected linkers, via the rational deconstruction of more-symmetric and more-connected linkers in edge-transitive nets. We applied our top-down strategy to edge-transitive nets containing 4-c tetrahedral, 6-c hexagonal, 8-c cubic or 12-c hexagonal prism linkers, envisaging the formation of 102 derived and 46 clipped nets. Among these, we report 33 new derived nets (icn7-icn39) and 6 new clipped nets (icn1-icn6). Importantly, the feasibility of using net-clipping to anticipate clipped nets is supported by literature examples and new experimental additions. Finally, we suggest and illustrate that net-clipping can be extended to less-regular, non-edge transitive nets as well as to covalent-organic frameworks (COFs), thus opening new avenues for the rational design of new reticular materials exhibiting unprecedented topologies.

New insights into the pathogenicity of TMEM165 variants using structural modeling based on AlphaFold 2 predictions

TMEM165 is a Golgi protein playing a crucial role in Mn2+ transport, and whose mutations in patients are known to cause Congenital Disorders of Glycosylation. Some of those mutations affect the highly-conserved consensus motifs E-φ-G-D-[KR]-[TS] characterizing the CaCA2/UPF0016 family, presumably important for the transport of Mn2+ which is essential for the function of many Golgi glycosylation enzymes. Others, like the G>R304 mutation, are far away from these motifs in the sequence. Until recently, the classical membrane protein topology prediction methods were unable to provide a clear picture of the organization of TMEM165 inside the cell membrane, or to explain in a convincing manner the impact of patient and experimentally-generated mutations on the transporter function of TMEM165. In this study, AlphaFold 2 was used to build a TMEM165 model that was then refined by molecular dynamics simulation with membrane lipids and water. This model provides a realistic picture of the 3D protein scaffold formed from a two-fold repeat of three transmembrane helices/domains where the consensus motifs face each other to form a putative acidic cation-binding site at the cytosolic side of the protein. It sheds new light on the impact of mutations on the transporter function of TMEM165, found in patients and studied experimentally in vitro, formerly and within this study. More particularly and very interestingly, this model explains the impact of the G>R304 mutation on TMEM165’s function. These findings provide great confidence in the predicted TMEM165 model whose structural features are discussed in the study and compared to other structural and functional TMEM165 homologs from the CaCA2/UPF0016 family and the LysE superfamily.

Accurate prediction of topology of composite plates via machine learning and propagation of elastic waves

Composites are artificial or natural materials that exhibit extraordinary material properties over homogenous materials made of one of their constituents. Detecting the microstructure of unknown solid composites is still challenging for engineers due to the complexity and nonlinear dynamic response of the materials. The emerging of machine learning (ML) in solving complex problems has enabled scientists to find solutions for problems that were impossible before. In this article, we proposed a methodology using ML and propagation of elastic waves to accurately determine the topology of binary composite plates. Results of this study indicate that a short period of flexural elastic waves propagated through composites can efficiently collect the information of the microstructures of materials and multiple independent RF models can be leveraged to predict configurations of composites accurately by learning from the topologies of plates and their corresponding output elastic waves. The average prediction accuracy of the ML model on 8×8 composite plates can reach 96% using 100,000 samples which only occupy 1/2×1014 of all the possible combinations. This approach is the first attempt on this topic and could be modified for applications such as composite characterization, geology, and archaeology.

Expression profiling and characterization of a G-Box binding protein, B12Dg, from pearl millet

ObjectivesThis study aims to characterize the B12Dg from pearl millet and assess its expression in response to abiotic stress and plant hormones during germination. MethodscDNA of B12Dg has been sequenced, the deduced protein has been characterized through in silico tools, and B12Dg expression has been examined after treatment by the plant hormones and under abiotic stresses. ResultsThe sequenced cDNA for PgB12Dg consisted of 314 base pairs and had an open reading frame of 261 base pairs encoded for 87 amino acids. The deduced PgB12Dg protein comprised a B12D domain from 7 to 74 amino acids and a transmembrane helix. Strong conservation was observed among the PgB12Dg protein and its orthologues from monocots. Transmembrane topology prediction and subcellular localization suggested that the deduced PgB12Dg protein accommodates transmembrane helix spanning from 10 to 32 amino acids embedded in the inner mitochondrion membrane. Expression analysis by real-time PCR indicated that PgB12Dg was upregulated in roots but downregulated in seeds from 3 day-old-seedlings under drought and cold stress. At the same time, no significant up or downregulation was shown under salt stress. Treatment by plant hormones showed upregulation by gibberellic acid and downregulation by abscisic acid in one-day-old germinated seeds. ConclusionsThe regulation of PgB12Dg by plant hormones and abiotic stress in pearl millet during germination can help to understand the transcriptional regulation mechanisms of the PgB12Dg during plant growth under stress.

Numerical predictions of flow topology and heat transfer in a square duct with staggered V-ribs

Performance augmentation of a square duct heat exchanger by a passive technique is numerically investigated. V-pattern ribs are used as vortex turbulators to enhance heat exchanger performance. The V-rib arrangement is designed with two important factors: 1. to increase the heat transfer coefficient by disturbing thermal boundary layers and to expedite fluid mixing and 2. to remain or decline the friction loss across the ribbed duct when compared with a typical in-line V-rib arrangement. The effects of rib heights (b/H = 0.05–0.20), rib pitch (P/H = 1–2) and flow paths (+x and -x) on fluid structure and thermal characteristics are studied within a laminar flow regime. Numerical analysis using the finite volume method (FVM) is chosen to predict the fluid structure and thermal mechanisms within the ribbed duct. Validating topics: smooth duct validation and grid independence, are firstly investigated. Simulation results are analyzed in forms of fluid structure and thermal behaviors. Performance assessment (thermal enhancement factor, friction factor ratio and Nusselt number ratio) within the tested section are also concluded. The simulation results show that the best TEF is found to be around 4.5, while the maximum Nusselt number ratio is around 19.39.

A mesoscale model for the micromechanical study of gels

Gels are comprised of polymer networks swelled by some interstitial solvent. They are under wide investigation by material scientists and engineers for their broad applicability in fields ranging from adhesives to tissue engineering. Gels’ mechanical properties greatly influence their efficacy in such applications and are largely dictated by their underlying microstructures and constituent-scale properties. Yet predictively mapping the local-to-global property functions of gels remains difficult due - in part - to the complexity introduced by solute-solvent interactions. We here introduce a novel, discrete mesoscale modeling method that preserves local solute concentration-dependent gradients in osmotic pressure through the Flory-Huggins mixing parameter, χ . The iteration of the model used here replicates gels fabricated from telechelically crosslinked star-shaped polymers and intakes χ , macromer molecular weight ( M w ), crosslink functionality ( f ), and as-prepared solute concentration ( ϕ * ) as its inputs, all of which are analogues to the control parameters of experimentalists. Here we demonstrate how this method captures solvent-dependent homogenization ( χ ≤ 0.5 ) or phase separation ( χ > 0.5 ) of polymer suspensions in the absence of phenomenological pairwise potentials. We then demonstrate its accurate, ab initio prediction of gel topology, isotropic swelling mechanics, and uniaxial tensile stress for a 10k tetra-PEG gel. Finally, we use the model to predict trends in the mechanical response and failure of multi-functional PEG-based gels over a range of M w and f , while investigating said trends’ micromechanical origins. The model predicts that increased crosslink functionality results in higher initial chain stretch (as measured at the equilibrated swollen state) for gels of the same underlying chain length, which improves modulus and failure stress but decreases failure strain and toughness.

Topology Prediction and Structural Controllability Analysis of Complex Networks Without Connection Information

In this article, we consider complex networks without connection information. The absence of global structure induces a great obstacle in the structural controllability analysis of these networks. Thus, a topology predicting method based on the connection probability matrix is proposed to provide a global structure for structural controllability analysis in this article. Furthermore, the modified principles of predicting global network topologies are established to acquire a more accurate global connection relationship. Eventually, the drive node set of these networks is determined by predicting global topologies. The accuracy of the proposed topology predicting method is verified by numerical simulations in the context of artificial networks and real networks. The results reveal that the global topology and structural controllability of complex networks with large scale and high edge density could be accurately predicted by utilizing the proposed method.

IMMAT: Mesh Reconstruction from Single View Images by Medial Axis Transform Prediction

The representation of a 3D shape is a key element for capturing the overall structure as well as the local details. In this paper, we propose to predict a mesh representation of the Medial Axis Transform (called medial mesh) as an intermediate representation with our IMMAT framework, to reconstruct the 3D shape from a single view image. Because the MAT contains the skeleton topology and local thickness information, it not only enhances the ability to reconstruct topologically complex shapes but also better preserves the local details with its thickness control. The framework consists of three modules. The Image2Sphere module predicts the medial spheres inside the shape surface and the Topology Prediction module predicts the topological relationship (skeleton) between the predicted spheres. Then the MAT Smoothing module smooths the predicted MAT and fine-tunes the coordinates and radii of the spheres by graph convolution. The final 3D surface can be reconstructed by converting the predicted MAT to an implicit surface through CSG operation and then extracting the boundary surface through Marching Cubes. Experimental results show that our method outperforms the state-of-the-art methods both quantitatively and qualitatively on the reconstruction task.