Computational Methods Research Articles

Exploring the Value of Computational Methods for Metajournalistic Discourse: The Example of COVID-19 Reporting in Dutch Newspapers

Exploring the Value of Computational Methods for Metajournalistic Discourse: The Example of COVID-19 Reporting in Dutch Newspapers

Conditional computation in neural networks: Principles and research trends

This article summarizes principles and ideas from the emerging area of applying conditional computation methods to the design of neural networks. In particular, we focus on neural networks that can dynamically activate or de-activate parts of their computational graph conditionally on their input. Examples include the dynamic selection of, e.g., input tokens, layers (or sets of layers), and sub-modules inside each layer (e.g., channels in a convolutional filter). We first provide a general formalism to describe these techniques in an uniform way. Then, we introduce three notable implementations of these principles: mixture-of-experts (MoEs) networks, token selection mechanisms, and early-exit neural networks. The paper aims to provide a tutorial-like introduction to this growing field. To this end, we analyze the benefits of these modular designs in terms of efficiency, explainability, and transfer learning, with a focus on emerging applicative areas ranging from automated scientific discovery to semantic communication.

Accurate prediction of CDR-H3 loop structures of antibodies with deep learning.

Accurate prediction of the structurally diverse complementarity determining region heavy chain 3 (CDR-H3) loop structure remains a primary and long-standing challenge for antibody modeling. Here, we present the H3-OPT toolkit for predicting the 3D structures of monoclonal antibodies and nanobodies. H3-OPT combines the strengths of AlphaFold2 with a pre-trained protein language model and provides a 2.24 Å average RMSDCα between predicted and experimentally determined CDR-H3 loops, thus outperforming other current computational methods in our non-redundant high-quality dataset. The model was validated by experimentally solving three structures of anti-VEGF nanobodies predicted by H3-OPT. We examined the potential applications of H3-OPT through analyzing antibody surface properties and antibody-antigen interactions. This structural prediction tool can be used to optimize antibody-antigen binding and engineer therapeutic antibodies with biophysical properties for specialized drug administration route.

Radiomics in oncology: A mini-review of principles, applications and challenges

Radiomics, a quantitative approach to medical imaging, employs computational methods to extract features from the images, revealing hidden characteristics of specific regions. This emerging field leverages advanced techniques to analyze a spectrum of features from modalities, including computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography (PET) scans, aiming to decode tissue characteristics, disease progression, and treatment responses. The radiomics workflow integrates image acquisition, segmentation, feature selection, and data integration, utilizing advanced techniques such as deep learning, machine learning, and data mining. Radiomics demonstrates considerable potential in cancer detection and management, exhibiting high sensitivity and specificity in distinguishing between benign and malignant tumors and predicting outcomes. However, challenges such as imaging protocol variability, overfitting, and standardization requirements impede its broad clinical adoption. Innovations in multi-modal radiomics, deep learning, and genomics integration strive to mitigate these constraints. This review elucidates radiomics’ capabilities, current applications, benefits, challenges, and future directions in oncology.

In Silico Comparative Analysis of Ivermectin and Nirmatrelvir Inhibitors Interacting with the SARS-CoV-2 Main Protease

Exploring therapeutic options is crucial in the ongoing COVID-19 pandemic caused by SARS-CoV-2. Nirmatrelvir, which is a potent inhibitor that targets the SARS-CoV-2 Mpro, shows promise as an antiviral treatment. Additionally, Ivermectin, which is a broad-spectrum antiparasitic drug, has demonstrated effectiveness against the virus in laboratory settings. However, its clinical implications are still debated. Using computational methods, such as molecular docking and 100 ns molecular dynamics simulations, we investigated how Nirmatrelvir and Ivermectin interacted with SARS-CoV-2 Mpro(A). Calculations using density functional theory were instrumental in elucidating the behavior of isolated molecules, primarily by analyzing the frontier molecular orbitals. Our analysis revealed distinct binding patterns: Nirmatrelvir formed strong interactions with amino acids, like MET49, MET165, HIS41, HIS163, HIS164, PHE140, CYS145, GLU166, and ASN142, showing stable binding, with a root-mean-square deviation (RMSD) of around 2.0 Å. On the other hand, Ivermectin interacted with THR237, THR239, LEU271, LEU272, and LEU287, displaying an RMSD of 1.87 Å, indicating enduring interactions. Both ligands stabilized Mpro(A), with Ivermectin showing stability and persistent interactions despite forming fewer hydrogen bonds. These findings offer detailed insights into how Nirmatrelvir and Ivermectin bind to the SARS-CoV-2 main protease, providing valuable information for potential therapeutic strategies against COVID-19.

Current trends on the early diagnosis of Alzheimer’s Disease by means of neural computation methods

Abstract—The prevalence of dementia is expected to increment in the next decades as the elderly population grows and ages. Hence, Alzheimer’s Disease (AD), as the most frequent dementia, will be more problematic from a socioeconomic point of view. Different diagnostic criteria have been proposed by clinicians for the early diagnosis of AD. After discarding the longitudinal and prognosis articles, a selection of articles from the last decade and based on Artificial Neural Networks (ANNs) was collated from the PubMed database, and complemented with researches extracted from others. The latest trends on this field were discovered in these selected articles, which were later discussed. Only articles based whether on shallow ANNs, Deep Learning (DL) or a mix of both were included. The total number of cross-sectional articles that complied with our selection criteria was 154. Convolutional Neural Networks (CNNs) combined with neuroimaging has been the most popular approach, yielding very good performance results. Approaches based on nonneuroimaging techniques, such as gait, genetics, speech and neuropsychological tests, were less common but have their own advantages. Multimodality solutions may become even more prevalent in the near future. Similarly, novel diagnostic criteria will appear and the popularity of currently not-so-common ones will expand. A new proposal emerged from these trends, which is based on ontogenetic ANNs.

Unlocking the Potential of VCI Films: A Comprehensive Study on Protective Efficiency

ABSTRACT Volatile Corrosion Inhibitor films have become an essential tool in safeguarding metal surfaces from corrosion. To maximize the effectiveness of VCI films, it is crucial to understand the relationship between their thermal properties and vapor phase inhibition ability. In this study, the optimal and modified standard test was performed for evaluation of VCI films ability and also investigate how their thermal characteristics influence their protective efficiency via the weight-loss method, and differential scanning calorimetry (DSC) analysis. The results showed that under the simulated and standard test conditions, the steel surface exposed to VCI films shows a model of a “preservative memory for VCI material”, at higher temperature/humidity. This memory is affected by different adsorption power due to the roughness of the surface, humidity and temperature as well as the rate of release of inhibitors. The results of the computational method confirmed that the orientation of the inhibitor molecule when interact with the polymer chain is the determining factor in the inhibition efficiency, consistent with the FTIR results. This behavior caused by chemical or electrochemical reaction with the temperature-dependent release rate of inhibitors has a greater effect on inhibition.

A Method Utilizing an Image Visibility Graph to Portray the Arrangement of Genomic Data Sequencing, Gene Frequencies for The Peptidoglycan-Associated Lipoprotein (Pal) Gene in Brucella Spp., and Prevalence of Brucellosis in Nineveh

Brucellosis is a disease triggered by various Gram-negative bacteria (Brucella spp.), impacting animals through miscarriages and causing fever in humans. The verified reality that peptidoglycan-associated lipoprotein (Pal) proteins eist in multiple bacteria and display remarkable vaccine potential shifts the spotlight to this protein and its importance within bacterial systems. Furthermore, gaining insight into the implications of mutations in this protein becomes of utmost importance. Biological informatics software was utilized to carry out a comparative study of the Pal gene across different species of Brucella bacteria. Computational methods were employed to represent sequences using both CGR and FCGR techniques. The Pal protein’s secondary and tertiary structures were built, and comparisons were drawn regarding mutations present among various Brucella species. An assessment of instances of infection was conducted, and statistical analyses were eecuted concerning brucellosis in the Nineveh region. While no significant differences were observed in the Pal gene among Brucella species, noticeable variations became evident when employing CGR and FCGR methodologies. FCGR proves to be more precise for visualization compared to CGR. Particularly, B. endophytica ehibited the most notable changes in the Pal protein when compared to the standard type within the species. The frequency of brucellosis infections increases in correlation with rising population numbers. The initial CGR images might not ehibit noticeable differences, but utilizing the FCGR method makes it easier to distinguish these variations. This affirms that FCGR is more effective for assessing the degree of dissimilarities. The presence of brucellosis poses a concern, especially in regions characterized by dense populations. It is important to emphasize public education regarding the adoption of suitable health precautions to thwart the disease, coupled with the importance of vaccination.

Zintl Phase versus Covalent Metal: Chemical Bonding in Silicon Dumbbells of Ca5Si3 and CaSi3.

Silicon dumbbells constitute identifiable anionic molecular species in Zintl phases and so-called covalent metals holding units with homopolar bonding inside a metallic framework. Based on electron-precise Ca5Si3 and metallic CaSi3, the chemical bonding in Si2 units is investigated by computational quantum chemical methods considering the dual nature of the wave function. This concerted wave-vector and real space study substantiates that the Si2 dumbbells in Ca5Si3 can be referred to as molecular building units Si26- with additional metallic and ionic contributions in the solid. In the covalent metal CaSi3, however, the bonding within the dumbbells falls short of fulfilling the octet rule. As a result, antibonding states of the Si2 building units are depopulated and attend metallic interactions, simultaneously giving rise to stronger covalent Si-Si bonds.

Growing Glycans in Rosetta: Accurate de novo glycan modeling, density fitting, and rational sequon design.

Carbohydrates and glycoproteins modulate key biological functions. However, experimental structure determination of sugar polymers is notoriously difficult. Computational approaches can aid in carbohydrate structure prediction, structure determination, and design. In this work, we developed a glycan-modeling algorithm, GlycanTreeModeler, that computationally builds glycans layer-by-layer, using adaptive kernel density estimates (KDE) of common glycan conformations derived from data in the Protein Data Bank (PDB) and from quantum mechanics (QM) calculations. GlycanTreeModeler was benchmarked on a test set of glycan structures of varying lengths, or "trees". Structures predicted by GlycanTreeModeler agreed with native structures at high accuracy for both de novo modeling and experimental density-guided building. We employed these tools to design de novo glycan trees into a protein nanoparticle vaccine to shield regions of the scaffold from antibody recognition, and experimentally verified shielding. This work will inform glycoprotein model prediction, glycan masking, and further aid computational methods in experimental structure determination and refinement.

Numerical investigation on mixing behaviour in the contact zone of an industrial‐scale dissolved air flotation tank

AbstractThe mixing behaviour of the wastewater and the recycle bubbly water in the contact zone could affect the separation performance of dissolved air flotation systems. In this study, the mixing characteristics in the contact zone of an industrial‐scale dissolved air flotation tank were investigated using the computational fluid dynamics method. At first, the influence of the turbulence model on the simulation results was clarified. It is demonstrated that the realizable k‐ε model could predict the gas holdup distribution with better accuracy, and the maximum relative deviation between the predicted gas holdup and the experimental results is only 10%. The influence of the geometry and operating parameters on the mixing process was then studied numerically. It is found that the mixing performance in the contact zone could be improved by decreasing the height position of the nozzle inlet in a certain range. The reason is that microbubbles disperse better in the bottom region and a double circulation flow state forms in the contact zone when the nozzle inlet is installed at a low height. Additionally, the mixing process could be further enhanced by decreasing the diameter of the nozzle inlet. As to the operating conditions, increasing the recycle ratio could not only increase the gas holdup but also improve the mixing performance in the contact zone. When the recycle flow rate remains constant, increasing the wastewater flow rate will lead to a decrease in gas holdup and deterioration of the mixing effect.

Fabrication, structural elucidation, and DFT calculation of some new hydrophilic metal chelates based on N N′‐(1‐methyl‐2‐oxoindolin‐3‐ylidene)benzohydrazide ligand: Pharmaceutical studies and molecular docking approach

Some novel FeIII, CuII, and PdII chelates incorporating N′‐(1‐methyl‐2‐oxoindolin‐3‐ylidene)benzohydrazide (MIBA) were fabricated. The tested compounds were investigated using thermogravimetric analysis (TGA), CHN, spectra analysis (IR, mass spectra, and NMR), melting point, magnetic moments, molar conductance, ultraviolet–visible spectroscopy, powder X‐ray diffraction, and computational studies. The conductance results showed that the tested FeIII, CuII, and PdII chelates are electrolytes. Magnetic and electronic spectra are applied to deduce the coordinating ability of the tested ligand, and the geometric structure of the studied chelates is found to be octahedral, distorted octahedral, and square planar for FeIII, CuII, and PdII chelates, respectively. The TGA study of these studied complexes displays that the hydrated H2O molecules, acetate, and nitrate are removed in the first and second degradation steps followed directly by degradation of the studied ligand leaving metal oxide as residue. The thermodynamic factors, like ΔS*, ΔH*, E*, A, and ΔG* are evaluated from the TGA curves and explained. The density functional theory (DFT)/B3LYP computation method was applied for the estimation of the molecular electrostatic potential (MEP; highest occupied molecular orbital [HOMO] and lowest unoccupied molecular orbital [LUMO]) energy for the studied compounds. In an in vitro study, the antimicrobial effects of the prepared compounds were screened on various strains of bacteria and fungi. It was found that tested compounds exposed a good biological efficacy through IC50 results close to reference drugs and antitumor potential against (MCF‐7, Hep‐G2, and HC‐T116) cell lines. The data obtained displayed that the studied chelates showed promising antitumor activity. The studied metal chelates were screened for in vitro antioxidant efficacy using DPPH assay. The studied compounds explained dynamic satisfying performance. Also, the crystal structures of breast cancer protein (PDB ID: 3HB5) and Escherichia coli (PDB ID: 2VF5) were performed by molecular docking simulation. Data of docking simulation suggestions are which tested compounds have biological behavior as well as have obvious benefit in the pharmaceutical business.

Computational discovery of dual potential inhibitors of SARS-CoV-2 spike/ACE2 and Mpro: 3D-pharmacophore, docking-based virtual screening, quantum mechanics and molecular dynamics.

To find drugs against COVID-19, caused by the SARS-CoV-2, promising targets include the fusion of the viral spike with the human angiotensin-converting enzyme 2 (ACE2) as well as the main protease (Mpro). These proteins are responsible for viral entry and replication, respectively. We combined several state-of-the-art computational methods, including, protein-ligand interaction fingerprint, 3D-pharmacophores, molecular-docking, MM-GBSA, DFT, and MD simulations to explore two databases: ChEMBL and NANPDB to identify molecules that could both block spike/ACE2 fusion and inhibit Mpro. A total of 1,690,649 compounds from the two databases were screened using the pharmacophore model obtained from PLIF analysis. Five recent complexes of Mpro co-crystallized with different ligands were used to generate the pharmacophore model, allowing 4,829 compounds that passed this prefilter. These were then submitted to molecular docking against Mpro. The 5% top-ranked docking hits from docking result having scores -8.32kcalmol-1 were selected and then docked against spike/ACE2. Only four compounds: ChEMBL244958, ChEMBL266531, ChEMBL3680003, and 1-methoxy-3-indolymethyl glucosinolate (4) displayed binding energies 8.21kcalmol-1 (for the native ligand) were considered as putative dual-target inhibitors. Furthermore, predictive ADMET, MM-GBSA and DFT/6-311G(d,p) were performed on these compounds and compared with those of well-known antivirals. DFT calculations showed that ChEMBL244958 and compound 4 had significant predicted reactivity values. Molecular dynamics simulations of the docked complexes were run for 100ns and used to validate the stability docked poses and to confirm that these hits are putative dual binders of the spike/ACE2 and the Mpro.

DDI Prediction with Heterogeneous Information Network - Meta-Path Based Approach.

Drug-drug interaction (DDI) indicates where a particular drug's desired course of action is modified when taken with other drug (s). DDIs may hamper, enhance, or reduce the expected effect of either drug or, in the worst possible scenario, cause an adverse side effect. While it is crucial to identify drug-drug interactions, it is quite impossible to detect all possible DDIs for a new drug during the clinical trial. Therefore, many computational methods are proposed for this task. This paper presents a novel method based on a heterogeneous information network (HIN), which consists of drugs and other biomedical entities like proteins, pathways, and side effects. Afterward, we extract the rich semantic relationships among these entities using different meta-path-based topological features and facilitate DDI prediction. In addition, we present a heterogeneous graph attention network-based end-to-end model for DDI prediction in the heterogeneous graph. Experimental results show that our proposed method accurately predicts DDIs and outperforms the baselines significantly.

КОМПЬЮТЕРНОЕ МОДЕЛИРОВАНИЕ УСЛОВИЙ РАБОТЫ ДИСКОВ БОРОН

The working parts of tillage machines, including harrow discs, wear out and fail during tillage. Basically, they are hardened at the manufacturing stage to increase wear resistance. Conducting tests to verify their operability in order to select the most effective hardening technologies requires a lot of time and material costs. A modern alternative to these tests is the use of computer simulation at the initial stage of the selection and development of hardening technologies. (Research purpose) The research purpose is investigating and evaluating the influence of technological factors on the wear resistance of the working parts of tillage machines, in particular harrow discs, using computer modeling tools. (Materials and methods) Carried out computer simulation of their working conditions using the computational fluid dynamics method by using the SOLIDWORKS Flow Simulation program to evaluate the effectiveness and selection of various hardening technologies. We created solid-state disk models that were used to simulate the operating conditions of harrow disks with various parameters. (Results and discussion) As a result of modeling, pressure distribution patterns were obtained, which are a color gradient on the selected surface of the model, showing the values of soil pressure on the disk. The pressure distribution on the disc of the disc harrow in the factory version and with reinforcing rollers at various angles of attack of 10, 15 and 20 degrees and speeds of 6, 12 and 24 kilometers per hour was shown. The patterns of pressure distribution on the disk with various geometric parameters of the applied reinforcing rollers were revealed. (Conclusions) With all the various technological factors presented, the harrow discs with reinforcing rollers showed a decrease in soil pressure on their surface. It was found that the pressure of the soil on the disk is most strongly reduced when applying rollers with a width of 6 and a height of no more than 2 millimeters.

Manifold learning uncovers nonlinear interactions between the adolescent brain and environment that predict emotional and behavioral problems.

Background To progress adolescent mental health research beyond our present achievements - a complex account of brain and environmental risk factors without understanding neurobiological embedding in the environment - we need methods to unveil relationships between the developing brain and real-world environmental experiences. Methods We investigated associations among brain function, environments, and emotional and behavioral problems using participants from the Adolescent Brain and Cognitive Development Study (N=2,401 female). We applied manifold learning, a promising technique for uncovering latent structure from high-dimensional biomedical data like functional magnetic resonance imaging (fMRI). Specifically, we developed exogenous PHATE (E-PHATE) to model brain-environment interactions. We used E-PHATE embeddings of participants' brain activation during emotional and cognitive processing to predict individual differences in cognition and emotional and behavioral problems, both cross-sectionally and longitudinally. Results E-PHATE embeddings of participants' brain activation and environments at baseline show moderate-to-large associations with total, externalizing, and internalizing problems at baseline, across several subcortical regions and large-scale cortical networks, relative to the zero-to-small effects achieved by voxel or PHATE methods. E-PHATE embeddings of the brain and environment at baseline also relate to emotional and behavioral problems two years later. These longitudinal predictions show a consistent, moderate effect in the frontoparietal and attention networks. Conclusions Adolescent brain's embedding in the environment yields enriched insight into emotional and behavioral problems. Using E-PHATE, we demonstrate how the harmonization of cutting-edge computational methods with longstanding developmental theories advances detection and prediction of adolescent emotional and behavioral problems.

Guest Editorial: Multi-Fluid and Fluid–Structure Interactions in Biological Systems: Computational Methods and Applications

Guest Editorial: Multi-Fluid and Fluid–Structure Interactions in Biological Systems: Computational Methods and Applications

A Systematic Review of Computational Approaches to Deciphering Bronze Age Aegean and Cypriot Scripts

Abstract This article provides a detailed insight into computational approaches for deciphering Bronze Age Aegean and Cypriot scripts, namely, the Archanes script and the Archanes formula, Phaistos Disk, Cretan hieroglyphic (including the Malia Altar Stone and Arkalochori Axe), Linear A, Linear B, Cypro-Minoan, and Cypriot scripts. The unique contributions of this article are threefold: (1) a thorough review of major Bronze Age Aegean and Cypriot scripts and inscriptions, digital data and corpora associated with them, existing computational decipherment methods developed in order to decipher them, and possible links to other scripts and languages; (2) the definition of 15 major challenges that can be encountered in computational decipherments of ancient scripts; and (3) an outline of a computational model that could possibly be used to simulate traditional decipherment processes of ancient scripts based on palaeography and epigraphy. In the context of this article the term decipherment denotes the process of discovery of the language and/or the set of symbols behind an unknown script, and the meaning behind it.

The syntactic position of the “people” in the political discourse of left and right populism (NLP-based syntactic analysis experience)

This paper attempts to measure populism in English-language speeches of politicians using computational linguistics methods. The relevance of this study is related not only to the rise of populism in the world and the importance of understanding the mechanisms of political discourse, but also to the lack of linguistic research in the context of corpus studies. Most of the methods proposed to date require significant resources or suffer from structural limitations, especially when they rely heavily on the human annotation process to analyze extensive corpus data. The material for the study was public speeches and interviews of right-wing populist politician and 45th President of the United States Donald Trump and political rhetoric of left-wing populist politician Senator Bernie Sanders. The corpus analysis was based on transcripts downloaded from The American Presidency Project and Miller Center websites. The syntactic position of key lexical units is determined using a comprehensive method based on dependency grammar combined with natural language processing (NLP) and van Leeuwen’s linguistic theory. The application of natural language processing methods confirms the assumption that the frequency of politicians’ reference to the people is not the only relevant indicator for measuring people-centrism in (populist) political discourse, as usually proposed in the political science literature. The results of the study indicate that there is no predominant role of the people in the discourse of Trump and Sanders. In most cases, the people appear in a complementary or adjunct position, indicating their secondary importance in their rhetoric. Given that populism implies proximity to the people and is positioned as “vox populi” – “the voice of the people”, it can be assumed that for both Trump and Sanders, addressing the people is only a way to achieve their political goals.

Ion-Specific Effects on Ion and Polyelectrolyte Solvation.

Ion-specific effects on aqueous solvation of monovalent counter ions, Na , K , Cl , and Br , and two model polyelectrolytes (PEs), poly(styrene sulfonate) (PSS) and poly(diallyldimethylammonium) (PDADMA) were here studied with ab initio molecular dynamics (AIMD) and classical molecular dynamics (MD) simulations based on the OPLS-aa force-field which is an empirical fixed point-charge force-field. Ion-specific binding to the PE charge groups was also characterized. Both computational methods predict similar response for the solvation of the PEs but differ notably in description of ion solvation. Notably, AIMD captures the experimentally observed differences in Cl and Br anion solvation and binding with the PEs, while the classical MD simulations fail to differentiate the ion species response. Furthermore, the findings show that combining AIMD with the computationally less costly classical MD simulations allows benefiting from both the increased accuracy and statistics reach.