Prediction Of Structures Research Articles

Prediction of long-term prestress loss for prestressed concrete cylinder structures using machine learning

The long-term prestress loss caused by shrinkage and creep of concrete and stress relaxation of prestressed tendons has significant effects on the sealability and safety of prestressed concrete cylinder structures such as nuclear reactor containments and liquified natural gas (LNG) tanks. By utilizing machine learning (ML) techniques, this study aims to establish an intelligent approach for the long-term prestress loss prediction of concrete cylinder structures. Firstly, based on the Infrastructure Technology Institute of Northwestern University (NU-ITI) database of concrete shrinkage and creep performance, the explicit expressions are presented for concrete shrinkage and creep function using genetic programming (GP); Moreover, the concrete constitutive model is incorporated into a general finite-element software package based on the ABAQUS UMAT platform. Then finite element analysis (FEA) models are established and calibrated based on the existing long-term prestress loss tests of prestressed concrete beams. In addition to the experimental results in the literature, the numerical results of the FEA model are also used to form the database of the long-term prestress losses for concrete cylinder structures. Finally, three prediction models of long-term prestress loss are proposed by utilizing the artificial neural network (ANN), one-dimensional convolutional neural network (1D CNN) and genetic programming (GP). Compared with the measured results of nuclear containments in practical engineering, the ML based prediction models are demonstrated to be accurate and efficient in evaluating the long-term prestress loss for prestressed concrete cylinder structures.

Signatures of Life Detected in Images of Rocks Using Neural Network Analysis Demonstrate New Potential for Searching for Biosignatures on the Surface of Mars.

Microorganisms play a role in the construction or modulation of various types of landforms. They are especially notable for forming microbially induced sedimentary structures (MISS). Such microbial structures have been considered to be among the most likely biosignatures that might be encountered on the martian surface. Twenty-nine algorithms have been tested with images taken during a laboratory experiment for testing their performance in discriminating mat cracks (MISS) from abiotic mud cracks. Among the algorithms, neural network types produced excellent predictions with similar precision of 0.99. Following that step, a convolutional neural network (CNN) approach has been tested to see whether it can conclusively detect MISS in images of rocks and sediment surfaces taken at different natural sites where present and ancient (fossil) microbial mat cracks and abiotic desiccation cracks were observed. The CNN approach showed excellent prediction of biotic and abiotic structures from the images (global precision, sensitivity, and specificity, respectively, 0.99, 0.99, and 0.97). The key areas of interest of the machine matched well with human expertise for distinguishing biotic and abiotic forms (in their geomorphological meaning). The images indicated clear differences between the abiotic and biotic situations expressed at three embedded scales: texture (size, shape, and arrangement of the grains constituting the surface of one form), form (outer shape of one form), and pattern of form arrangement (arrangement of the forms over a few square meters). The most discriminative components for biogenicity were the border of the mat cracks with their tortuous enlarged and blistered morphology more or less curved upward, sometimes with thin laminations. To apply this innovative biogeomorphological approach to the images obtained by rovers on Mars, the main physical and biological sources of variation in abiotic and biotic outcomes must now be further considered.

Seismic response prediction of structures based on Runge-Kutta recurrent neural network with prior knowledge

In the seismic analysis of structural systems, dynamic response prediction is an essential problem and is significant in every stage during the structural life cycle. Conventionally, response analysis is carried out by numerical analysis. However, when the structural parameter is unknown, the establishment of a numerical model will be difficult. Enlightened by the Runge-Kutta (RK) numerical algorithm, this paper proposes a novel recurrent neural network named Runge-Kutta recurrent neural network (RKRNN) to realize the seismic response prediction. A partition training strategy is formulated to train the proposed neural network and to improve the efficiency of training. The proposed model can be trained by using a limited number of samples. Three numerical examples are utilized to validate the feasibility of RKRNN model including a linear three degrees of freedom (DOFs) system, a nonlinear single DOF system with Bouc-Wen hysteresis, and a numerical reinforced concrete bridge model. Additionally, the site monitoring data from a real-world bridge is utilized to further validate the proposed network. The results show that the proposed RKRNN model can effectively and efficiently predict the structural response under seismic load and exhibits robustness to noise, with good potential for applications in engineering practice.

Identification and comprehensive analysis of circRNA-miRNA-mRNA regulatory networks in osteoarthritis.

Osteoarthritis (OA) is a common orthopedic degenerative disease, leading to high disability in activities of daily living. There remains an urgent need to identify the underlying mechanisms and identify new therapeutic targets in OA diagnosis and treatment. Circular RNAs (circRNAs) play a role in the development of multiple diseases. Many studies have reported that circRNAs regulate microRNAs (miRNAs) through an endogenous competitive mechanism. However, it remains unclear if an interplay between circRNAs, miRNAs, and target genes plays a deeper regulatory role in OA. Four datasets were downloaded from the GEO database, and differentially expressed circRNAs (DECs), differentially expressed miRNAs (DEMs), and differentially expressed genes (DEGs) were identified. Functional annotation and pathway enrichment analysis of DEGs and DECs were carried out to determine the main associated mechanism in OA. A protein-protein network (PPI) was constructed to analyze the function of, and to screen out, hub DEGs in OA. Based on the artificial intelligence prediction of protein crystal structures of two hub DEGs, TOP2A and PLK1, digitoxin and oxytetracycline were found to have the strongest affinity, respectively, with molecular docking. Subsequently, overlapping DEMs and miRNAs targeted by DECs obtained target DEMs (DETMs). Intersection of DEGs and genes targeted by DEMs obtained target DEGs (DETGs). Thus, a circRNA-miRNA-mRNA regulatory network was constructed from 16 circRNAs, 32 miRNAs, and 97 mRNAs. Three hub DECs have the largest number of regulated miRNAs and were verified through in vitro experiments. In addition, the expression level of 16 DECs was validated by RT-PCR. In conclusion, we constructed a circRNA-miRNA-mRNA regulatory network in OA and three new hub DECs, hsa_circ_0027914, hsa_circ_0101125, and hsa_circ_0102564, were identified as novel biomarkers for OA.

A New General Framework for Response Prediction of Composite Structures Based on Digital Twin with Three Effective Error Correction Strategies

A New General Framework for Response Prediction of Composite Structures Based on Digital Twin with Three Effective Error Correction Strategies

Mapping the FtsQBL divisome components in bacterial NTD pathogens as potential drug targets.

Cytokinesis is an essential process in bacterial cell division, and it involves more than 25 essential/non-essential cell division proteins that form a protein complex known as a divisome. Central to the divisome are the proteins FtsB and FtsL binding to FtsQ to form a complex FtsQBL, which helps link the early proteins with late proteins. The FtsQBL complex is highly conserved as a component across bacteria. Pathogens like Vibrio cholerae, Mycobacterium ulcerans, Mycobacterium leprae, and Chlamydia trachomatis are the causative agents of the bacterial Neglected Tropical Diseases Cholera, Buruli ulcer, Leprosy, and Trachoma, respectively, some of which seemingly lack known homologs for some of the FtsQBL complex proteins. In the absence of experimental characterization, either due to insufficient resources or the massive increase in novel sequences generated from genomics, functional annotation is traditionally inferred by sequence similarity to a known homolog. With the advent of accurate protein structure prediction methods, features both at the fold level and at the protein interaction level can be used to identify orthologs that cannot be unambiguously identified using sequence similarity methods. Using the FtsQBL complex proteins as a case study, we report potential remote homologs using Profile Hidden Markov models and structures predicted using AlphaFold. Predicted ortholog structures show conformational similarity with corresponding E. coli proteins irrespective of their level of sequence similarity. Alphafold multimer was used to characterize remote homologs as FtsB or FtsL, when they were not sufficiently distinguishable at both the sequence or structure level, as their interactions with FtsQ and FtsW play a crucial role in their function. The structures were then analyzed to identify functionally critical regions of the proteins consistent with their homologs and delineate regions potentially useful for inhibitor discovery.

Fatigue life prediction of lattice structures using a strain-life method based on an equivalent solid model

In this study, an efficient fatigue life estimation approach is proposed based on an equivalent solid model, along with two stress concentration factors, instead of a lattice model to simplify simulations for the strain-life method. One of the stress concentration factors accounts for the stress concentrations at the strut joints in lattice structures without any manufacturing effect, while the other accounts for the stress concentrations due to surface defects on the struts introduced by the additive manufacturing process. A simple solid cell, for which the dimensions are the same as those of the lattice cell, is first used to estimate the fatigue life of the bulk material used in the additive manufacturing process using the Brown–Miller strain-life method. Combinations of two stress concentration factors were then utilized to predict the fatigue life of the lattice structure from the equivalent solid model. Using this approach, computationally expensive lattice structure simulations were eliminated, and the fatigue life of the lattice structure was predicted directly from the bulk material fatigue life estimation. The proposed approach was applied to different lattice types made of Ti-6Al-4V alloy using the laser-powder bed fusion additive manufacturing process. The fatigue life results estimated using the proposed approach were in good agreement with the experimental fatigue life results reported in the literature.

Family-wide analysis of integrin structures predicted by AlphaFold2

Recent advances in protein structure prediction using AlphaFold2, known for its high efficiency and accuracy, have opened new avenues for comprehensive analysis of all structures within a single protein family. In this study, we evaluated the capabilities of AphaFold2 in analyzing integrin structures. Integrins are heterodimeric cell surface receptors composed of a combination of 18 α and 8 β subunits, resulting in a family of 24 different members. Both α and β subunits consist of a large extracellular domain, a short transmembrane domain, and typically, a short cytoplasmic tail. Integrins play a pivotal role in a wide range of cellular functions by recognizing diverse ligands. Despite significant advances in integrin structural studies in recent decades, high-resolution structures have only been determined for a limited subsets of integrin members, thus limiting our understanding of the entire integrin family. Here, we first analyzed the single-chain structures of 18 α and 8 β integrins in the AlphaFold2 protein structure database. We then employed the newly developed AlphaFold2-multimer program to predict the α/β heterodimer structures of all 24 human integrins. The predicted structures show a high level of accuracy for the subdomains of both α and β subunits, offering high-resolution structure insights for all integrin heterodimers. Our comprehensive structural analysis of the entire integrin family unveils a potentially diverse range of conformations among the 24 members, providing a valuable structure database for studies related to integrin structure and function. We further discussed the potential applications and limitations of the AlphaFold2-derived integrin structures.

Prediction of concrete structures and Selection of input factor by digital hammering inspection using machine learning

Prediction of concrete structures and Selection of input factor by digital hammering inspection using machine learning

Crystal-structure simulation of molecular semiconductors: brickwork-related crystal structures of methylthiolated peri-condensed polycyclic aromatic hydrocarbons.

Despite the critical importance to carrier transport properties, studies on the control and prediction of crystal structures of molecular semiconductors have not been well-matured. To tackle this issue, we have developed "in silico crystallization" (ISC) protocols for simulating the brickwork (BW) crystal structures of methylchalcogenolated polycyclic aromatic hydrocarbons (PAHs). In this study, by carefully analyzing a BW-related polymorph of experimental crystal structures, an inclined brickwork (iBW) structure, we further extend the ISC protocol to simulate various BW-related crystal structures including iBW structures. Rational conditional branching in the simulation not only makes it possible to simulate eight polymorph candidates of methylchalcogenolated PAHs but also helps understand the relationship between the polymorphs. Furthermore, the relative favorability of each polymorphic candidate, i.e., the likelihood of the appearance among the polymorph candidates, can also be evaluated.

Electron density mapping of boron clusters via convolutional neural networks to augment structure prediction algorithms.

Determination and prediction of atomic cluster structures is an important endeavor in the field of nanoclusters and thereby in materials research. To a large extent the fundamental properties of a nanocluster are mainly governed by its molecular structure. Traditionally, structure elucidation is achieved using quantum mechanics (QM) based calculations that are usually tedious and time consuming for large nanoclusters. Various structural prediction algorithms have been reported in the literature (CALYPSO, USPEX). Although they tend to accelerate the structure exploration, they still require the aid of QM based calculations for structure evaluation. This makes the structure prediction process quite a computationally expensive affair. In this paper, we report on the creation of a convolutional neural network model, which can give relatively accurate energies for the ground state of nanoclusters from the promolecule density on the fly and could thereby be utilized for aiding structure prediction algorithms. We tested our model on dataset consisting of pure boron nanoclusters of varying sizes.

Study of the binding of ΔFN3.1 fragments of the Bifidobacterium longum GT15 with TNFα and prevalence of domain-containing proteins in groups of bacteria of the human gut microbiota.

Aim: This study is mainly devoted to determining the ability of ∆FN3.1 protein fragments of Bifidobacterium (B.) longum subsp. longum GT15, namely two FN3 domains (2D FN3) and a C-terminal domain (CD FN3), to bind to tumor necrosis factor-alpha (TNF-α). Methods: Fragments of the fn3 gene encoding the 2D FN3 and CD FN3 were cloned in Escherichia (E.) coli. In order to assess the binding specificity between 2D FN3 and CD FN3 to TNFα, we employed the previously developed sandwich ELISA system to detect any specific interactions between the purified protein and any of the studied cytokines. The trRosetta software was used to build 3D models of the ∆FN3.1, 2D FN3, and CD FN3 proteins. The detection of polymorphism in the amino acid sequences of the studied proteins and the analysis of human gut-derived bacterial proteins carrying FN3 domains were performed in silico. Results: We experimentally showed that neither 2D FN3 nor CD FN3 alone can bind to TNFα. Prediction of the 3D structures of ΔFN3.1, 2D FN3, and CD FN3 suggested that only ΔFN3.1 can form a pocket allowing binding with TNFα to occur. Polymorphism analysis of amino acid sequences of ΔFN3.1 proteins in B. longum strains uncovered substitutions that can alter the conformation of the spatial structure of the ΔFN3.1 protein. We also analyzed human gut-derived bacterial proteins harboring FN3 domains which allowed us to differentiate between those containing motifs of cytokine receptors (MCRs) in their FN3 domains and those lacking them. Conclusion: Only the complete ∆FN3.1 protein can selectively bind to TNFα. Analysis of 3D models of the 2D FN3, CD FN3, and ΔFN3.1 proteins showed that only the ΔFN3.1 protein is potentially capable of forming a pocket allowing TNFα binding to occur. Only FN3 domains containing MCRs exhibited sequence homology with FN3 domains of human proteins.

Computational Methods for Protein Tertiary Structure Analysis.

Protein folding accuracy is fundamental to all cells. In spite of this, it is difficult to maintain the fidelity of protein synthesis and folding due to the fact that the implicit genetic and biochemical systems are inherently prone to error, which leads to the constant production of a certain amount of misfolded proteins. This problem is further compounded by genetic variation and the effects of environmental stress. To that end, the prediction of protein structures for tertiary protein structure analysis and prediction might be an ideal approach for the study of mutation effects in macromolecules and their complexes. With the development and accessibility to increasingly powerful computational systems, this type of study will enable a wide variety of opportunities for the creation of better-targeted peptide-based pharmacotherapy and prospects for precision medicine in future.

Prediction of effector protein structures from fungal phytopathogens enables evolutionary analyses

Elucidating the similarity and diversity of pathogen effectors is critical to understand their evolution across fungal phytopathogens. However, rapid divergence that diminishes sequence similarities between putatively homologous effectors has largely concealed the roots of effector evolution. Here we modelled the structures of 26,653 secreted proteins from 14 agriculturally important fungal phytopathogens, six non-pathogenic fungi and one oomycete with AlphaFold 2. With 18,000 successfully predicted folds, we performed structure-guided comparative analyses on two aspects of effector evolution: uniquely expanded sequence-unrelated structurally similar (SUSS) effector families and common folds present across the fungal species. Extreme expansion of lineage-specific SUSS effector families was found only in several obligate biotrophs, Blumeria graminis and Puccinia graminis. The highly expanded effector families were the source of conserved sequence motifs, such as the Y/F/WxC motif. We identified new classes of SUSS effector families that include known virulence factors, such as AvrSr35, AvrSr50 and Tin2. Structural comparisons revealed that the expanded structural folds further diversify through domain duplications and fusion with disordered stretches. Putatively sub- and neo-functionalized SUSS effectors could reconverge on regulation, expanding the functional pools of effectors in the pathogen infection cycle. We also found evidence that many effector families could have originated from ancestral folds conserved across fungi. Collectively, our study highlights diverse effector evolution mechanisms and supports divergent evolution as a major force in driving SUSS effector evolution from ancestral proteins.

Do metastable polymorphs always grow faster? Measuring and comparing growth kinetics of three polymorphs of tolfenamic acid.

The phenomenon of molecular crystal polymorphism is of central importance for all those industries that rely on crystallisation for the manufacturing of their products. Computational methods for the evaluation of thermodynamic properties of polymorphs have become incredibly accurate and a priori prediction of crystal structures is becoming routine. The computational study and prediction of the kinetics of crystallisation impacting polymorphism, however, have received considerably less attention despite their crucial role in directing crystallisation outcomes. This is mainly due to the lack of available experimental data, as nucleation and growth kinetics of polymorphs are generally difficult to measure. On the one hand, the determination of overall nucleation and growth kinetics through batch experiments suffers from unwanted polymorphic transformations or the absence of experimental conditions under which several polymorphs can be nucleated. On the other hand, growth rates of polymorphs obtained from measurements of single crystals are often only recorded along a few specific crystal dimensions, thus lacking information about overall growth and rendering an incomplete picture of the problem. In this work, we measure the crystal growth kinetics of three polymorphs (I, II and IX) of tolfenamic acid (TFA) in isopropanol solutions, with the intention of providing a meaningful comparison of their growth rates. First, we analyse the relation between the measured growth rates and the crystal structures of the TFA polymorphs. We then explore ways to compare their relative growth rates and discuss their significance when trying to determine which polymorph grows faster. Using approximations for describing the volume of TFA crystals, we show that while crystals of the metastable TFA-II grow the fastest at all solution concentrations, crystals of the metastable TFA-IX become kinetically competitive as the driving force for crystallisation increases. Overall, both metastable forms TFA-II and TFA-IX grow faster than the stable TFA-I.

Prediction of vortex structures in pulsatile flow through S-bend arterial geometry with different stenosis levels

Arterial stenosis poses a high cardiovascular risk, and clinical intervention is needed when these stenoses grow beyond a specific limit. The study of vortex dynamics in these diseased arteries can be beneficial to understand its severity. Therefore, in the present work, we have investigated the flow structures in an S-bend arterial geometry with different levels of stenosis using a sharp interface immersed boundary method. We have observed an onset of Kelvin-Helmholtz-type vortex roll-up for higher degrees of stenoses. Fluctuations in the wall shear stress are observed for higher stenosis degrees. However, these fluctuations depend on the position and length of the stenosis. Newtonian and non-Newtonian Carreau fluids predict similar vortex structures, although minor differences in the Kelvin-Helmholtz vortex structures and associated fluctuations are observed in the diastolic phase. The Newtonian fluid predicts a slightly longer low time-averaged wall shear stress (≤0.5 Pa) region immediately after the stenosis compared with the Carreau fluid in the 58 % blockage S-bend artery.

Self-organizing maps as a data-driven approach to elucidate the packing motifs of perylene diimide derivatives

The efficient classification or prediction of crystal structures into a small number of families of related structures can be extremely important in the design of materials with specific packing and properties.

RNA 3D Structure Comparison Using RNA-Puzzles Toolkit.

Computational modeling of RNA three-dimensional (3D) structure may help in unrevealing the molecular mechanisms of RNA molecules and in designing molecules with novel functions. An unbiased blind assessment to benchmark the computational modeling is required to understand the achievements and bottlenecks of the prediction, while a standard structure comparison protocol is necessary. RNA-Puzzles is a community-wide effort on the assessment of blind prediction of RNA tertiary structures. And RNA-Puzzles toolkit is a computational resource derived from RNA-Puzzles, which includes (i) decoy sets generated by different RNA 3D structure prediction methods; (ii) 3D structure normalization, analysis, manipulation, and visualization tools; and (iii) 3D structure comparison metric tools. In this chapter, we illustrate a standard RNA 3D structure prediction assessment protocol using the selected tools from RNA-Puzzles toolkit: rna-tools and RNA_assessment.

A Hybrid Particle Swarm Optimization-Tuning Algorithm for the Prediction of Nanoparticle Morphology from Microscopic Images

Aggregated nanoparticle structures are quite ubiquitous in aerosol and colloidal science, specifically in nanoparticle synthesis systems such as combustion processes where coagulation results in the formation of fractal-like structures. In addition to their size, morphology of the particles also plays a key role in defining various physicochemical properties. Electron microscopy based images are the most commonly used tools in visualizing these aggregates, and prediction of the 3-dimensional structures from the microscopic images is quite complex. Typically, 2-dimensional features from the images are compared to available structures in a database or regression equations are used to predict 3-dimensional morphological parameters including Fractal dimension and pre-exponential factor. In this study, we propose a combination of evolutionary algorithm and forward tuning model to predict the best fit 3-dimensional structures of aggregates from their projection images. 2-dimensional features from a projection image are compared to the candidate projections generated using FracVAL code and optimized using Particle Swarm Optimization to obtain the 3-dimensional structure of the aggregate. Various 3-dimensional properties including hydrodynamic diameter and mobility diameter of the retrieved structures are then compared with the properties of the aggregate used to form the candidate projection image, to test the suitability of the algorithm. Results show that the hybrid algorithm can closely predict the 3-dimensional structures from the projection images with less than 10% difference in the predicted 3-dimensional properties including mobility diameter and radius of gyration.

Omics and Remote Homology Integration to Decipher Protein Functionality.

In the recent years, several "omics" technologies based on specific biomolecules (from DNA, RNA, proteins, or metabolites) have won growing importance in the scientific field. Despite each omics possess their own laboratorial protocols, they share a background of bioinformatic tools for data integration and analysis. A recent subset of bioinformatic tools, based on available templates or remote homology protocols, allow computational fast and high-accuracy prediction of protein structures. The quickly predict of actually unsolved protein structures, together with late omics findings allow a boost of scientific advances in multiple fields such as cancer, longevity, immunity, mitochondrial function, toxicology, drug design, biosensors, and recombinant protein engineering. In this chapter, we assessed methodological approaches for the integration of omics and remote homology inferences to decipher protein functionality, opening the door to the next era of biological knowledge.