The widespread application of ionic liquids has been limited by uncertainties related to their ecotoxicity. In this article, we investigate the relationship between molecular descriptors of ionic liquids and acetylcholinesterase (AChE) inhibition using a QSAR approach coupled with machine learning models. An initial set of descriptors underwent a preprocessing and selection process based on multiple criteria, including importance by models, correlation with the target variable, statistical relevance, stability, and non-redundancy, followed by iterative sequential selection. Predictive models were built and evaluated from multiple random partitions of the data, showing high performance (R[2]>0.8700). The analysis of the descriptors indicated that inhibitory activity is strongly associated with the structural characteristics of the cations, especially properties related to hydrophobicity, electronic distribution, and molecular topology. Cations with aromatic systems showed greater affinity for the enzyme, while very long alkyl chains reduced activity due to steric effects. On the other hand, anions played a secondary role. The results contribute to the understanding of the structure-activity relationship and aid in the development of ionic liquids with lower toxic potential.

The KRAS oncogene, which plays a crucial role in numerous human malignancies, remains a significant challenge. It is amongst the most mutated oncogenes within human cancers, specifically in pancreatic, colorectal, and non-small cell lung cancers, where it is a key driver of cancer initiation and growth. Despite its importance, KRAS is difficult to target directly due to the absence of deep binding pockets, its high affinity for GTP/GDP, and rapid cycling between active and inactive states. Thus, an effective strategy will be to target downstream effectors of KRAS, such as MEK, PI3K, and AKT. Mitogen-activated protein kinase 1 (MEK1) emerged as a promising target. In the current study, pharmacophore models were developed via a ligand-based pharmacophore approach and validated by using the Ligand Pharmacophore Mapping tool by calculating the EF1%, EF5%, EF10%, BEDROC and GH scores. Then, pharmacophore-based virtual screening was performed for the identification of novel hit molecules. A total of 22042 novel hit molecules were obtained, further narrowed to 2430 hit molecules through ADMET and Toxicity screening. Molecular Docking studies identified three top ranked compounds, PubChem-125589749 (Rank 1), PubChem-122594448 (Rank 2) and PubChem-49842405 (Rank 3), with score -76.91, -75.54 and -75.3184 kcal/mol, respectively, whereas Avutometinib showed a score of -61.78 kcal/mol. MD simulations over 100 ns and MM-GBSA calculations show that PubChem-125589749 (Rank 1) exhibited the most favourable interaction profile and stability, followed by PubChem-49842405 (Rank 3). MM-GBSA analysis of PubChem-125589749 (Rank 1) and Rank 3 showed average ΔGbinding energy -126.19 ± 15.29 and -114.81 ± 13.33. Further extended MD simulations for 200 ns and MM-GBSA calculations confirmed the binding stability and binding energy over a longer time. The reproducibility of the MD simulations was assessed by performing an independent MD simulation of the complexes and analysing the results.

This article investigates the rare-earth-doped spinel chalcogenides Sr(Pr/Nd)2S4 for spintronics and thermoelectric applications. Spin-dependent structure optimization reveals their stability in the ferromagnetic state. Tolerance factors, formation energies, ab initio molecular dynamics simulations, and phonon band structures are calculated to evaluate their structural and thermodynamic stability. The spin-polarization density calculations and band structure analysis show 100% spin polarization and a ferromagnetic semiconducting nature. The spin-↑ channel has band gaps of ∼1.75 and 0.70 eV, and the spin-↓ channel is insulating. The exchange energies (Δx(f), Δx(pf)) and exchange constants (Noα, Noβ) are measured based on the strong coupling between the 4f-state of Pr/Nd and 3p-orbitals of S, leading to the creation of localized magnetic moments and the emergence of intriguing magnetic phenomena. Additionally, analyzing the electrical conductivity, power factor, thermal conductivity, and Seebeck coefficient by means of the BoltzTrap code helps to assess their suitability for transport applications. Thermoelectric performance has been measured by the figure of merit values of 0.84 and 0.58 for SrPr2S4 and SrNd2S4, respectively. The substantial values of transport parameters make these materials an evolving class for Spin electronics and thermal energy harvesting systems.

Hybrid membranes composed of graphene oxide (GO) and molybdenum disulfide (MoS2) offer promising opportunities for overcoming the permeability-selectivity trade-off in membrane-based desalination. However, the influence of layer composition and stacking sequence on water and ion transport remains insufficiently understood at the molecular level. In this work, non-equilibrium molecular dynamics simulations were employed to systematically investigate pressure-driven water and ion transport through six GO-MoS2 membrane architectures, including pristine GO and MoS2 membranes and four hybrid stacking configurations. The results demonstrate that membrane architecture strongly governs the balance between water transport and ion exclusion. Pristine MoS2 exhibits the highest water transport rate, whereas GO-containing configurations generally show lower ion accessibility to the confined membrane region. Molecular-level analyses reveal that water transport is governed by the coupled effects of channel confinement, interfacial interactions, molecular packing, hydrogen bonding, and surface chemistry. Hydration-shell analysis further shows distinct responses of Na[+] and Cl[-] under confinement, contributing to their different penetration behavior. The results highlight that the stacking sequence and surface composition of GO-MoS2 membranes provide an effective means of tuning water transport and ion exclusion, offering molecular-level insights for the design of 2D hybrid membranes for desalination.

The equiatomic quaternary Heusler alloy AlTiOsMn exhibits a pronounced correlation-induced transition from a metallic ferromagnet to a half-metallic state, accompanied by thermally stable spin-selective transport. The ferromagnetic Type 3 configuration is the lowest-energy arrangement, with a lattice parameter of 6.06 Å and an energy above the convex hull of 0.037 eV/atom. Phonon calculations and ab-initio molecular dynamics support dynamical and thermal stability, while the lattice thermal conductivity decreases from 16.75 W/mK at 300 K to 5.39 W/mK at 900 K. Most importantly, the self-consistent Hubbard correction transforms the GGA metallic state into a half-metallic ferromagnet with a 0.19 eV half-metallic gap, 100% spin polarization, and a total magnetic moment of 4 μB. This electronic reconstruction is accompanied by a distinct spin-dependent thermoelectric response, with Ssp reaching about 78 μV/K and ZTsp attaining 0.27 at 900 K in GGA and 0.23 in GGA + U. The results establish a direct link between correlation-driven half-metallicity and spin-selective thermoelectric transport, highlighting AlTiOsMn as a promising candidate for spintronic and spin-caloritronic applications.

Sarsasapogenin, a spirostanol sapogenin with reported activity against ERα-positive breast cancer cells, has no defined molecular target, and its metabolic fate has not been considered in computational studies of this compound class. Network pharmacology, biotransformation profiling, and multi-level molecular modeling were combined to address both questions. Of 108 targets shared between Sarsasapogenin and breast cancer, ERα gave the most favorable docking energy among ten hub proteins (-10.46 kcal/mol), against -8.73 kcal/mol for the reference modulator Bazedoxifene. BioTransformer predicted 16 metabolites, of which five phase-I derivatives retaining the spirostanol scaffold bound ERα within a narrow window (-10.01 to -10.30 kcal/mol). The representative metabolite BTM00010, a 6-hydroxylated derivative, retained affinity at -10.30 kcal/mol and formed a hydrogen bond to Val533 that is absent from the parent pose. Over 100 ns of simulation, all three complexes reached comparable plateaus in RMSD (0.20-0.30 nm) and radius of gyration (1.68-1.80 nm), and MMGBSA ranked them in the same order as docking (-47.30, -21.48, and -11.83 kcal/mol). Post-dynamics analysis showed lower collective-motion amplitude for the metabolite complex than for the parent, and a correlated-motion network closer to the parent than to the reference modulator. Predicted phase-I hydroxylation, therefore, does not abolish ERα engagement, which argues for evaluating biotransformation products alongside the parent compound in computational screening of plant sapogenins.

Molecular docking is widely used to examine ligand interactions with antimicrobial resistance (AMR)-associated proteins, but favorable docking scores do not necessarily indicate mechanistically meaningful binding poses. This study developed a target-level diagnostic framework to determine when structure-based docking outputs are interpretable, conditionally useable, or unsuitable for early-stage AMR research. Eight AMR-relevant targets were evaluated using 30 predefined probe ligands spanning controlled differences in size, polarity, flexibility, and interaction density under a fixed AutoDock Vina protocol. Docking behavior was assessed through crystallographic redocking, reference-anchored pocket localization, PLIP-supported residue-contact analysis, pose classification, and a four-component Composite Interpretability Index applied only within individual targets. Crystallographic redocking showed consistent rank-1 pose reproduction for ErmC, DHFR, and GyrB, whereas AmpC, NDM-1, and VanA showed less stable rank-1 selection despite evidence of recoverable crystallographic poses. Large polar polyphenols and some stress-test ligands frequently produced favorable scores, but score trends were nonmonotonic and target-dependent; favorable scores also occurred in peripheral or out-of-pocket poses, demonstrating partial decoupling between score magnitude and mechanistic localization. At the target level, DHFR and GyrB each produced 29 of 30 highly interpretable probe cases with no gate flags and were classified as docking reliable. AmpC and VanA were classified as docking acceptable, PBP2a, KPC-2, and NDM-1 as docking constrained, and ErmC as unsuitable for generalized heterogeneous-ligand prioritization, with 11 gate flags and the most heterogeneous interpretability profile. The framework therefore provides a conservative pre-deployment gate for determining whether docking should be used routinely, conditionally, or excluded for a given AMR target.

Classical sulfonamides share a formal SO2NH motif, yet their molecular environments impose distinct frontier-orbital responses. We examined thirteen classical sulfonamides and dapsone using a staged, quality-controlled workflow that separated exploratory GFN2-xTB/Composite-SEPI (the study-specific exploratory electronic-prioritization score) screening from full-panel B3LYP-D3BJ/def2-SVP optimization/frequency analysis, canonical PBE0-D4/def2-TZVP refinement with the conductor-like polarizable continuum model for water (CPCM(water)), and an orthogonal ωB97X-D4 sensitivity layer. Revision-stage robustness tests additionally challenged geometry, additive D4 dispersion, population partitioning, and the basis-set dependence of charged-state Fukui descriptors. Corrected B3LYP and fixed-geometry PBE0 gaps preserved the same lower-gap regime (Pearson r = 0.935; Spearman ρ = 0.934). Full-panel PBE0 reoptimization showed modest but measurable gap sensitivity (mean absolute error (MAE) = 0.0595 eV; root-mean-square error (RMSE) = 0.0696 eV; 7/14 with |Δgap| ≥ 0.050 eV) while preserving the lower-gap top-three identity. PBE0 and ωB97X-D4 retained the relative gap landscape (r = 0.959; ρ = 0.978), and Mulliken-Löwdin cross-checks preserved the dominant HOMO and LUMO fragments for 14/14 molecules. In sulfadiazine, sulfamerazine, and sulfamethazine, the occupied and virtual frontier orbitals remained spatially separated between the aryl/aniline and diazine-containing regions. Diffuse-basis def2-TZVPD tests yielded negative vertical electron affinities and positive anion α-spin singly occupied molecular orbital (SOMO) energies for 14/14 molecules, making f+ and the dual descriptor basis-sensitive while f- remained comparatively robust. Formal SO2NH identity therefore masks reproducible electronic heterogeneity, but the reported quantities remain method-dependent isolated-molecule descriptors and do not establish charge-transfer states or biological activity predictions.

Histone deacetylase 8 (HDAC8) is an emerging epigenetic target implicated in cancer, neurodegenerative disorders, and other human diseases, driving the urgent need for potent and selective inhibitors. In this study, we developed and validated a deep learning-based computational pipeline for the accurate prediction of HDAC8 inhibitor potency, integrating cheminformatics, neural networks, molecular docking, and molecular dynamics simulations. A large, curated dataset of 4583 HDAC8 inhibitors with experimental pIC50 values was retrieved from the ChEMBL database (target CHEMBL3192) and preprocessed using RDKit. Each molecule was represented as a 1033-dimensional feature vector combining nine physicochemical descriptors with a 1024-bit Morgan fingerprint. Through systematic hyperparameter ablation comparing eight neural network architectures, the optimal configuration comprising three hidden layers (512 → 256 → 64 units) with batch normalization and dropout regularization achieved a moderate but useful test set R[2] of 0.617 and RMSE of 0.642 log units. Residual analysis confirmed model reliability across the central tendency of chemical space, though systematic regression-to-the-mean effects were observed for extreme potency values. Leveraging an extrapolation-trained model with SELFIES-based evolutionary generation, we identified ten novel HDAC8 inhibitor candidates with predicted pIC50 values ranging from 10.49 to 11.07, all satisfying Lipinski's Rule of Five. The highest-priority candidate (Compound 1, predicted pIC50 = 11.07) was subjected to molecular docking and 200 ns molecular dynamics simulations, revealing a stable binding mode (protein RMSD = 1.8-2.2 Å, ligand RMSD = 1.5-1.8 Å) with persistent hydrogen bonding interactions (GLY 131 occupancy = 27%, GLY 120 occupancy = 22%). This integrated computational pipeline demonstrates the power of deep learning for accelerated HDAC8 inhibitor discovery, providing a robust framework for identifying and optimizing potent, drug-like candidates. These computational predictions, while promising, should be interpreted as hypotheses that require rigorous experimental validation.

Accurate protein structure prediction is fundamental to structure-based drug discovery. However, the practical performance of deep learning-based models such as AlphaFold compared with traditional homology modeling approaches like SWISS-MODEL remains incompletely evaluated in realistic drug-design workflows. Here, we systematically benchmark AlphaFold2 and SWISS-MODEL using a curated dataset of 20 kinase structures with co-crystallized ligands. Predicted models were evaluated under three conditions: the raw predicted structures, structures subjected to energy minimization, and structures subjected to triplicate 100-ns unbiased molecular dynamics (MD) simulations for structural refinement. Model performance was assessed using structural accuracy metrics, multi-software docking (AutoDock Vina, AutoDock, and MOE), and post-docking MD trajectory stability analysis, together with MM/GBSA binding energy estimation. As expected, experimental structures consistently showed the best docking performance. Among predicted models, SWISS-MODEL produced slightly lower RMSD values and better interaction similarity than AlphaFold, although overall docking scores were statistically comparable. MD refinement prior to docking reduced structural suitability by increasing binding-pocket deviations and was associated with misoriented ligand poses in subsequent docking calculations, while energy minimization provided little to no improvement. Notably, the reduction in docking performance was not limited to computationally predicted structures but was also observed for experimental structures solved in complex with different ligands. Thus, rigid docking alone tended to generate a substantial number of false-positive poses. However, MD simulations applied after docking effectively identified unstable ligand poses and reduced false-positive predictions by detecting ligand dissociation. Overall true-positive rates were 30% for SWISS-MODEL and 35% for AlphaFold2. These findings highlight the importance of dynamic validation in structure-based drug discovery workflows.

Research studies on doped titanium clusters have not received significant attention in both theoretical and experimental fields. This is attributed to the complexity of almost empty d orbital for titanium metal, which results in its unique bonding properties. Thus, bimetallic systems offer a new degree of freedom to extend the titanium nanoclusters' properties by doping with other transition metal elements. In this study, the density functional theory (DFT) approach was employed with the PBEsol exchange-correlation functional to investigate the structural and electronic properties of TiNM (N = 1 - 19, M = Ru, Rh) nanoclusters. The investigation revealed that Rh mostly favours to be placed on the edge and surface, whereas the Ru impurity mostly favours the apex of the nanoclusters. Rhodium and ruthenium dopants are observed to enhance the binding energy of pure Ti nanoclusters, with Ru dopant revealing the least binding energies. The relative stability showed that the doping of Rh and Ru converted the N = 13 cluster into the magic cluster, whereas the dissociation energy showed reduction in stability strength for N = 7 and enhanced stability for N = 13. Notably, the local maxima of the HOMO-LUMO gap for Rh, Ru doped Ti clusters occur at N = 13, suggesting relatively stable cluster. VIP, VEA, η, S, electronegativity, chemical potential, electrophilicity, and wavelength suggested that the results have strong quantum size effects at small N and progressive quasi-metallic stabilisation at larger sizes, with Rh inducing stronger oscillatory and localised reactivity while Ru promotes smoother electronic evolution. The density of states (DOS) for both Rh and Ru impurities stabilise the Ti13 cluster electronically, with rhodium emerging as the more favourable dopant.

Enzyme immobilization depends not only on support chemistry, but also on the spatial orientation adopted by the protein after attachment. This study presents a computational framework guided by molecular geometry to screen favorable immobilization orientations of plant peroxidases on modified surfaces. Horseradish peroxidase, soybean peroxidase, and royal palm tree peroxidase were analyzed as three distinct case studies. Surface exposed anchoring residues were identified from experimental structures, and rigid body orientation sampling generated enzyme and surface configurations. Each configuration was evaluated using a functional orientation score that combined heme accessibility, catalytic channel exposure, anchoring site proximity, protein and surface separation, and steric feasibility. Three support chemistries, carboxylated, hydroxylated, and aminated, were assessed, producing nine enzyme and surface combinations. Among the three simplified surface chemistries evaluated, the hydroxylated neutral model produced the highest composite compatibility score for each of the three peroxidases. The preferred surface therefore represented three of the nine evaluated combinations, while the remaining six showed lower overall compatibility. Comparison of mean and maximum scores indicated that the best orientation did not always represent the most stable orientation ensemble, highlighting the need to consider score dispersion and sensitivity. Reactive residue mapping also identified enzyme specific anchoring regions capable of preserving catalytic channel exposure and limiting heme occlusion. Overall, the workflow integrates structural accessibility, orientation quality, and surface compatibility within an interpretable screening strategy. It provides a low cost route for prioritizing immobilization conditions before testing and can be extended to other enzymes whose catalytic performance depends on maintaining access to buried active sites after attachment.

Lysine crotonylation (Kcr) is an important post-translational modification (PTM) involved in diverse biological processes, including chromatin regulation, protein function modulation, and cellular signaling. Although mass spectrometry-based proteomics has substantially expanded the identification of Kcr sites, experimental screening remains labor-intensive, costly, and difficult to apply at proteome scale. Computational methods provide an efficient strategy for prioritizing candidate Kcr sites. However, most existing predictors mainly rely on sequence-derived representations and insufficiently exploit protein structural context. In this study, we propose BLOSSOM-Kcr, a structure-informed deep learning framework for Kcr site prediction. BLOSSOM-Kcr integrates BLOSUM62-based sequence substitution features with residue-level structural descriptors, including secondary structure, solvent accessibility, backbone geometry, and spatial neighborhood information. The fused residue-level representation is further processed by residual convolutional blocks, channel attention, bidirectional long short-term memory (BiLSTM) layers, and attention pooling to capture local motif patterns, informative feature dimensions, and contextual dependencies surrounding candidate lysine residues. Fivefold cross-validation was performed for model optimization and comparative analysis, while an independent test set was used for final evaluation against existing Kcr site predictors. On the independent test set, BLOSSOM-Kcr achieved an AUC of 0.9023, an MCC of 0.6479, and an F1-score of 0.8336, outperforming representative Kcr site predictors. These results suggest that BLOSSOM-Kcr provides an effective structure-aware framework for Kcr site prediction.

Non-small-cell lung cancer (NSCLC), the predominant histological subtype accounting for more than 85% of lung cancer cases, remains the leading cause of cancer-related mortality worldwide. Despite the clinical advances afforded by targeted therapies, the frequent emergence of oncogenic driver mutations drives drug resistance and therapeutic failure, underscoring the critical need for novel molecular targets and alternative therapeutic strategies. AKT1, a serine/threonine kinase and central effector of the PI3K/AKT/mTOR signaling cascade, regulates key oncogenic processes including cell growth, survival, proliferation, and apoptotic evasion, and dysregulated hyperactivation is observed in 30-75% of NSCLC cases. Selective allosteric inhibition of AKT1, which targets the unique PH-kinase domain interface rather than the highly conserved ATP-binding pocket, has emerged as a more isoform-specific and therapeutically advantageous strategy compared with pan-AKT or ATP-competitive inhibition. In the present study, a ChemDiv AKT1-targeted compound library comprising 14,043 small molecules was subjected to hierarchical structure-based virtual screening using sequential high-throughput virtual screening (HTVS), standard precision (SP), and extra precision (XP) docking protocols, followed by binding free energy calculations using the MM-GBSA method. The top XP-ranked compounds, compound 1 (docking score: -11.771 kcal/mol), compound 2 (-11.428 kcal/mol), and compound 3 (-11.167 kcal/mol), achieved more favourable XP docking scores at the AKT1 allosteric site (PDB ID: 3O96) than established pan-AKT inhibitors, including MK-2206, miransertib, Bay1125976, and vevorisertib, although MM-GBSA rescoring did not reproduce this ordering. Comparative docking against the allosteric sites of AKT2 and AKT3 revealed a consistent preference for AKT1 across all three compounds, with a magnitude closely comparable to that of the reference allosteric ligand Inhibitor VIII, whose experimentally determined isoform rank order the protocol reproduced. Comprehensive 300-ns all-atom molecular dynamics (MD) analysis, encompassing RMSD and RMSF profiling, principal component analysis, dynamic cross-correlation matrix analysis, free energy landscape evaluation, and protein-ligand contact profiling, consistently demonstrated that compounds 1 and 2 maintain stable, high-affinity interactions with AKT1, exhibiting persistent engagement with the critical allosteric residues Trp80 and Tyr272 throughout the simulation trajectory. Furthermore, both compounds exhibited favourable pharmacokinetic profiles with predicted human oral absorption values over 80%, supporting their potential for preclinical development.

We present a first-principles investigation of pristine and transition-metal (Cu, Ag, Au) decorated TiO2 nanotubes (TiO2-NTs) for CO and CO2 sensing. Density functional theory calculations show that metal substitution at the O1 site significantly tailors the electronic structure through band-gap modulation, localized states, and spin polarization while preserving structural stability. Natural bond orbital and density-of-states analyses reveal strong charge polarization induced by Cu, moderate electron donation from Ag, and partial back-donation effects associated with Au. Adsorption results indicate that CO2 interacts most strongly with pristine TiO2-NTs, whereas CO exhibits pronounced chemisorption on Au-decorated nanotubes via C-Ti bond formation, supported by orbital hybridization and noncovalent interaction analyses. Recovery-time calculations demonstrate that Ag- and Au-decorated TiO2-NTs enable fast and reversible desorption, making them suitable for reusable sensing applications. Overall, transition-metal decoration emerges as an effective strategy to enhance the sensitivity, selectivity, and reversibility of TiO2-NT-based gas sensors for CO and CO2 detection.