Search Results (20)

Molecular dynamics (MD) simulations provide atomistic insights into nucleation and crystallization in polymers, yet interpreting their complex spatiotemporal data remains a challenge. Existing order parameters face limitations, such as failing to account for directional alignment or lacking sufficient spatial resolution, preventing them from accurately capturing the anisotropic and heterogeneous characteristics of nucleation or the surface phenomena of polymer crystallization. We introduce a novel set of local order parameters—namely, directional entropy bands— that extend scalar entropy-based descriptors by capturing first-order angular moments of the local entropy field around each particle. We compare these against conventional metrics (entropy, the crystallinity index, and smooth overlap of atomic positions (SOAP) descriptors) in equilibrium MD simulations of polymer crystallization. We show that (i) scalar entropy bands demonstrate advantages compared to SOAP in polymer phase separation at single-snapshot resolution and (ii) directional extensions (dipole projections and gradient estimates) robustly highlight the evolving crystal–melt interface, enabling earlier nucleation detection and quantitative surface profiling. UMAP embeddings of these 24–30D feature vectors reveal a continuous melt–surface–core manifold, as confirmed by supervised boundary classification. Our approach is efficient and directly interpretable, offering a practical framework for studying polymer crystallization kinetics and surface growth phenomena. Full article

Understanding the dynamics of gene regulatory networks (GRNs) across diverse cell types poses a challenge yet holds immense value in unraveling the molecular mechanisms governing cellular processes. Current computational methods, which rely solely on expression changes from bulk RNA-seq and/or scRNA-seq data, often result in high rates of false positives and low precision. Here, we introduce an advanced computational tool, DeepIMAGER, for inferring cell-specific GRNs through deep learning and data integration. DeepIMAGER employs a supervised approach that transforms the co-expression patterns of gene pairs into image-like representations and leverages transcription factor (TF) binding information for model training. It is trained using comprehensive datasets that encompass scRNA-seq profiles and ChIP-seq data, capturing TF-gene pair information across various cell types. Comprehensive validations on six cell lines show DeepIMAGER exhibits superior performance in ten popular GRN inference tools and has remarkable robustness against dropout-zero events. DeepIMAGER was applied to scRNA-seq datasets of multiple myeloma (MM) and detected potential GRNs for TFs of RORC, MITF, and FOXD2 in MM dendritic cells. This technical innovation, combined with its capability to accurately decode GRNs from scRNA-seq, establishes DeepIMAGER as a valuable tool for unraveling complex regulatory networks in various cell types. Full article

COVID-19 caused by SARS-CoV-2 has spread around the world. The receptor-binding domain (RBD) of the spike protein of SARS-CoV-2 is a critical component that directly interacts with host ACE2. Here, we simulate the ACE2 recognition processes of RBD of the WT, Delta, and OmicronBA.2 variants using our recently developed supervised Gaussian accelerated molecular dynamics (Su-GaMD) approach. We show that RBD recognizes ACE2 through three contact regions (regions I, II, and III), which aligns well with the anchor–locker mechanism. The higher binding free energy in State d of the RBD_OmicronBA.₂-ACE2 system correlates well with the increased infectivity of OmicronBA.2 in comparison with other variants. For RBD_Delta, the T478K mutation affects the first step of recognition, while the L452R mutation, through its nearby Y449, affects the RBD_Delta-ACE2 binding in the last step of recognition. For RBD_OmicronBA.₂, the E484A mutation affects the first step of recognition, the Q493R, N501Y, and Y505H mutations affect the binding free energy in the last step of recognition, mutations in the contact regions affect the recognition directly, and other mutations indirectly affect recognition through dynamic correlations with the contact regions. These results provide theoretical insights for RBD-ACE2 recognition and may facilitate drug design against SARS-CoV-2. Full article

The plant actin cytoskeleton is characterized by the basic properties of dynamic array, which plays a central role in numerous conserved processes that are required for diverse cellular functions. Here, we focus on how actins and actin-related proteins (ARPs), which represent two classical branches of a greatly diverse superfamily of ATPases, are involved in fundamental functions underlying signal regulation of plant growth and development. Moreover, we review the structure, assembly dynamics, and biological functions of filamentous actin (F-actin) from a molecular perspective. The various accessory proteins known as actin-binding proteins (ABPs) partner with F-actin to finely tune actin dynamics, often in response to various cell signaling pathways. Our understanding of the significance of the actin cytoskeleton in vital cellular activities has been furthered by comparison of conserved functions of actin filaments across different species combined with advanced microscopic techniques and experimental methods. We discuss the current model of the plant actin cytoskeleton, followed by examples of the signaling mechanisms under the supervision of F-actin related to cell morphogenesis, polar growth, and cytoplasmic streaming. Determination of the theoretical basis of how the cytoskeleton works is important in itself and is beneficial to future applications aimed at improving crop biomass and production efficiency. Full article

As one of the most influential industries in public health and the global economy, the pharmaceutical industry is facing multiple challenges in drug research, development and manufacturing. With recent developments in artificial intelligence and machine learning, data-driven modeling methods and techniques have enabled fast and accurate modeling for drug molecular design, retrosynthetic analysis, chemical reaction outcome prediction, manufacturing process optimization, and many other aspects in the pharmaceutical industry. This article provides a review of data-driven methods applied in pharmaceutical processes, based on the mathematical and algorithmic principles behind the modeling methods. Different statistical tools, such as multivariate tools, Bayesian inferences, and machine learning approaches, i.e., unsupervised learning, supervised learning (including deep learning) and reinforcement learning, are presented. Various applications in the pharmaceutical processes, as well as the connections from statistics and machine learning methods, are discussed in the narrative procedures of introducing different types of data-driven models. Afterwards, two case studies, including dynamic reaction data modeling and catalyst-kinetics prediction of cross-coupling reactions, are presented to illustrate the power and advantages of different data-driven models. We also discussed current challenges and future perspectives of data-driven modeling methods, emphasizing the integration of data-driven and mechanistic models, as well as multi-scale modeling. Full article

Precise binding affinity predictions are essential for structure-based drug discovery (SBDD). Focal adhesion kinase (FAK) is a member of the tyrosine kinase protein family and is overexpressed in a variety of human malignancies. Inhibition of FAK using small molecules is a promising therapeutic option for several types of cancer. Here, we conducted computational modeling of FAK-targeting inhibitors using three-dimensional structure–activity relationship (3D-QSAR), molecular dynamics (MD), and hybrid topology-based free energy perturbation (FEP) methods. The structure–activity relationship (SAR) studies between the physicochemical descriptors and inhibitory activities of the chemical compounds were performed with reasonable statistical accuracy using CoMFA and CoMSIA. These are two well-known 3D-QSAR methods based on the principle of supervised machine learning (ML). Essential information regarding residue-specific binding interactions was determined using MD and MM-PB/GBSA methods. Finally, physics-based relative binding free energy ($\Delta \Delta {\mathrm{G}}_{\mathrm{RBFE}}^{\mathrm{A}\to \mathrm{B}}$) terms of analogous ligands were estimated using alchemical FEP simulation. An acceptable agreement was observed between the experimental and computed relative binding free energies. Overall, the results suggested that using ML and physics-based hybrid approaches could be useful in synergy for the rational optimization of accessible lead compounds with similar scaffolds targeting the FAK receptor. Full article

Macromolecular self-assembly is at the basis of many phenomena in material and life sciences that find diverse applications in technology. One example is the formation of virus-like particles (VLPs) that act as stable empty capsids used for drug delivery or vaccine fabrication. Similarly to the capsid of a virus, VLPs are protein assemblies, but their structural formation, stability, and properties are not fully understood, especially as a function of the protein modifications. In this work, we present a data-driven modeling approach for capturing macromolecular self-assembly on scales beyond traditional molecular dynamics (MD), while preserving the chemical specificity. Each macromolecule is abstracted as an anisotropic object and high-dimensional models are formulated to describe interactions between molecules and with the solvent. For this, data-driven protein–protein interaction potentials are derived using a Kriging-based strategy, built on high-throughput MD simulations. Semi-automatic supervised learning is employed in a high performance computing environment and the resulting specialized force-fields enable a significant speed-up to the micrometer and millisecond scale, while maintaining high intermolecular detail. The reported generic framework is applied for the first time to capture the formation of hepatitis B VLPs from the smallest building unit, i.e., the dimer of the core protein HBcAg. Assembly pathways and kinetics are analyzed and compared to the available experimental observations. We demonstrate that VLP self-assembly phenomena and dependencies are now possible to be simulated. The method developed can be used for the parameterization of other macromolecules, enabling a molecular understanding of processes impossible to be attained with other theoretical models. Full article

After the SARS-CoV-2 Wuhan variant that gave rise to the pandemic, other variants named Delta, Omicron, and Omicron-2 sequentially became prevalent, with mutations spread around the viral genome, including on the spike (S) protein; in order to understand the resultant in gains in infectivity, we interrogated in silico both the equilibrium binding and the binding pathway of the virus’ receptor-binding domain (RBD) to the angiotensin-converting enzyme 2 (ACE2) receptor. We interrogated the molecular recognition between the RBD of different variants and ACE2 through supervised molecular dynamics (SuMD) and classic molecular dynamics (MD) simulations to address the effect of mutations on the possible S protein binding pathways. Our results indicate that compensation between binding pathway efficiency and stability of the complex exists for the Omicron BA.1 receptor binding domain, while Omicron BA.2′s mutations putatively improved the dynamic recognition of the ACE2 receptor, suggesting an evolutionary advantage over the previous strains. Full article

Serum is an important candidate in proteomics analysis as it potentially carries key markers on health status and disease progression. However, several important diagnostic markers found in the circulatory proteome and the low-molecular-weight (LMW) peptidome have become analytically challenging due to the high dynamic concentration range of the constituent protein/peptide species in serum. Herein, we propose a novel approach to improve the limit of detection (LoD) of LMW amino acids by combining mid-IR (MIR) and near-IR spectroscopic data using glycine as a model LMW analyte. This is the first example of near-IR spectroscopy applied to elucidate the detection limit of LMW components in serum; moreover, it is the first study of its kind to combine mid-infrared (25–2.5 μm) and near-infrared (2500–800 nm) to detect an analyte in serum. First, we evaluated the prediction model performance individually with MIR (ATR-FTIR) and NIR spectroscopic methods using partial least squares regression (PLS-R) analysis. The LoD was found to be 0.26 mg/mL with ATR spectroscopy and 0.22 mg/mL with NIR spectroscopy. Secondly, we examined the ability of combined spectral regions to enhance the detection limit of serum-based LMW amino acids. Supervised extended wavelength PLS-R resulted in a root mean square error of prediction (RMSEP) value of 0.303 mg/mL and R² value of 0.999 over a concentration range of 0–50 mg/mL for glycine spiked in whole serum. The LoD improved to 0.17 mg/mL from 0.26 mg/mL. Thus, the combination of NIR and mid-IR spectroscopy can improve the limit of detection for an LMW compound in a complex serum matrix. Full article

Adenosine receptors have been a promising class of targets for the development of new therapies for several diseases. In recent years, a renewed interest in this field has risen, thanks to the implementation of a novel class of agonists that lack the ribose moiety, once considered essential for the agonistic profile. Recently, an X-ray crystal structure of the A_2A adenosine receptor has been solved, providing insights about the receptor activation from this novel class of agonists. Starting from this structural information, we have performed supervised molecular dynamics (SuMD) simulations to investigate the binding pathway of a non-nucleoside adenosine receptor agonist as well as one of three classic agonists. Furthermore, we analyzed the possible role of water molecules in receptor activation. Full article

Precision oncology is grounded in the increasing understanding of genetic and molecular mechanisms that underly malignant disease and offer different treatment pathways for the individual patient. The growing complexity of medical data has led to the implementation of machine learning techniques that are vastly applied for risk assessment and outcome prediction using either supervised or unsupervised learning. Still largely overlooked is reinforcement learning (RL) that addresses sequential tasks by exploring the underlying dynamics of an environment and shaping it by taking actions in order to maximize cumulative rewards over time, thereby achieving optimal long-term outcomes. Recent breakthroughs in RL demonstrated remarkable results in gameplay and autonomous driving, often achieving human-like or even superhuman performance. While this type of machine learning holds the potential to become a helpful decision support tool, it comes with a set of distinctive challenges that need to be addressed to ensure applicability, validity and safety. In this review, we highlight recent advances of RL focusing on studies in oncology and point out current challenges and pitfalls that need to be accounted for in future studies in order to successfully develop RL-based decision support systems for precision oncology. Full article

Fragment-Based Drug Discovery (FBDD) has become, in recent years, a consolidated approach in the drug discovery process, leading to several drug candidates under investigation in clinical trials and some approved drugs. Among these successful applications of the FBDD approach, kinases represent a class of targets where this strategy has demonstrated its real potential with the approved kinase inhibitor Vemurafenib. In the Kinase family, protein kinase CK1 isoform δ (CK1δ) has become a promising target in the treatment of different neurodegenerative diseases such as Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis. In the present work, we set up and applied a computational workflow for the identification of putative fragment binders in large virtual databases. To validate the method, the selected compounds were tested in vitro to assess the CK1δ inhibition. Full article

Fragment-Based Drug Discovery (FBDD) approaches have gained popularity not only in industry but also in academic research institutes. However, the computational prediction of the binding mode adopted by fragment-like molecules within a protein binding site is still a very challenging task. One of the most crucial aspects of fragment binding is related to the large amounts of bound waters in the targeted binding pocket. The binding affinity of fragments may not be sufficient to displace the bound water molecules. In the present work, we confirmed the importance of the bound water molecules in the correct prediction of the fragment binding mode. Moreover, we investigate whether the use of methods based on explicit solvent molecular dynamics simulations can improve the accuracy of fragment posing. The protein chosen for this study is HSP-90. Full article

Adenosine receptors (ARs), like many otherGprotein-coupledreceptors (GPCRs), are targets of primary interest indrug design. However, one of the main limits for the development of drugs for this class of GPCRs is the complex selectivity profile usually displayed by ligands. Numerous efforts have been madefor clarifying the selectivity of ARs, leading to the development of many ligand-based models. The structure of the AR subtype A₁ (A₁AR) has been recently solved, providing important structural insights. In the present work, we rationalized the selectivity profile of two selective A₁AR and A_2AAR antagonists, investigating their recognition trajectories obtained by Supervised Molecular Dynamics from an unbound state and monitoring the role of the water molecules in the binding site. Full article

Most G protein-coupled receptors that bind the hydrophobic ligands (lipid receptors and steroid receptors) belong to the most populated class A (rhodopsin-like) of these receptors. Typical examples of lipid receptors are: rhodopsin, cannabinoid (CB), sphingosine-1-phosphate (S1P) and lysophosphatidic (LPA) receptors. The hydrophobic ligands access the receptor binding site from the lipid bilayer not only because of their low solubility in water but also because of a large N-terminal domain plug preventing access to the orthosteric binding site from the extracellular milieu. In order to identify the most probable ligand exit pathway from lipid receptors CB1, S1P1 and LPA1 orthosteric binding sites we performed at least three repeats of steered molecular dynamics simulations in which ligands were pulled in various directions. For specific ligands being agonists, the supervised molecular dynamics approach was used to simulate the ligand entry events to the inactive receptor structures. For all investigated receptors the ligand entry/exit pathway goes through the gate between transmembrane helices TM1 and TM7, however, in some cases it combined with a direction toward water milieu. Full article