Multimodal Rebalanced Network with bidirectional interaction for tumor segmentation in PET/CT images

Harnessing GhGRDP1 natural variation for enhanced cotton seed yield: a rhamnose-dependent strategy in breeding.

TD2-LTS: Transformer based group-channel interaction for medical image segmentation

JBD-Net: Joint bidirectional inter-layer interaction and dynamic weighting network for polyp image segmentation

The complex anatomy of the ovine cervix presents a significant barrier to transcervical artificial insemination, necessitating sperm deposition at the external cervical Os. This study aimed to evaluate the cervical microarchitecture to elucidate the mechanisms orchestrating sperm transport across the cervix. Cervices from crossbred ewes (n=13) were harvested during the follicular (n=6) and luteal (n=7) phase, sectioned into 10 transverse segments from external to internal Os (1=External cervical Os; 10=Internal cervical Os), and analysed for epithelial morphology and microgroove architecture. Columnar epithelium predominated throughout the cervix, with a squamocolumnar junction localised at the external Os. Columnar cell height was significantly greater during the follicular phase (33.3±1.31μm) than the luteal phase (26.7±1.21μm; p<0.001). Cervical microgroove surface area increased during the follicular phase (p<0.05). Microgroove surface area, depth, and tortuosity (deviation from a straight line) increased progressively toward the internal Os and were enhanced during the follicular phase (p<0.05). Sialic acid expression was elevated in secondary and tertiary microgrooves during the follicular phase (p<0.001), with mucin-bound sialic acids showing spatial variation along the canal. Segment 4 exhibited the lowest sialic acid expression (p<0.01). Reproductive phase by cervical segment interactions affected both the secondary, tertiary microgroove epithelium, and surrounding luminal region adjacent to the tertiary microgrooves (p<0.001). This study provides a detailed characterisation of the microarchitecture of the ovine cervical canal which is critical to understanding the mechanisms orchestrating sperm transport.

We present computer simulations of a minimal polymer model in water-methanol mixed solvents that captures key features of the cononsolvency behavior of amphiphilic polymers in alcohol-water mixtures. Our results indicate that the effective interactions between polymer segments vary nonmonotonically with the water-methanol composition. Their temperature dependence, however, shows a transition from good-to-poor solvent conditions only in the aqueous, low-methanol regime, up to the solvent composition X* corresponding to the minimum of the lower critical solution temperature. Consistent with this, the polymer radius of gyration decreases sharply with increasing temperature for X < X*. This temperature sensitivity, however, disappears at X = X*, even though the temperature dependence of effective segmental interactions still indicates a good-to-poor solvent transition. This loss of thermoresponsivity, in agreement with previous experimental observations, arises from preferential methanol adsorption, which reduces the role of hydrophobic hydration and leads to a vanishing heat of polymer collapse. Finally, we highlight the role of mixing entropy, arising from methanol-water mixing during polymer collapse, in driving the cononsolvency effect at a fixed temperature.

CIA-net: Cross-modality interaction and aggregation network for ovarian tumor segmentation from multi-modal MRI.

Global-to-Local Deep Interaction and Boundary-Aware Transformer for accurate polyp segmentation

As the core of modern industrial automation, Industrial Internet of Things (IIoT) can achieve data collection, control, and remote operations. However, the complex continuous connectivity exposes IIoT to increasing cyber intrusions. Existing intrusion detection methods often fail to adapt to diverse temporal attack patterns, lack effective feature modeling, and perform poorly under class imbalance. To this end, this article proposes a novel transformer-based IIoT intrusion detection method. First, an adaptive multi-scale window segmentation (AMSWS) strategy is proposed to dynamically segment traffic sequences via traffic variation. This can capture short-term and long-term intrusion behaviors. Second, an attention interaction transformer (AITrans) is proposed to enhance feature modeling by enabling inter-head and inter-layer attention interaction. Finally, an adaptive focal loss (AFL) is proposed to mitigate class imbalance by incorporating adaptive class weight and hardness-aware adjustment. This can improve detection performance on minority-class intrusions. Extensive experiments validate the effectiveness of the proposed method for IIoT intrusion detection.

The behavior of shield tunnel lining structures is known to be influenced by segmental joints. Most studies conducted in this area use simplified models, which may not properly simulate the behavior of the segmental joints. This study utilizes a full-reinforced concrete segment model to rigorously investigate the seismic behavior of joints in a segmental tunnel lining, explicitly accounting for segment–segment contact, interaction, and joint bolts. Specifically, a comprehensive full dynamic analysis of a two-dimensional (2D) lining–soil model, incorporating nonlinear constitutive models for both concrete (CDPM) and soil (Mohr–Coulomb), was conducted to investigate the effects of joint bolt type, seismic intensity, and vertical excitation component on the seismic response. The lining–soil model was excited using three ground motions. The results indicate that the joint rotation is significantly influenced by the amplitude and frequency content of ground motions, which has implications for the watertightness of the gasketed joint. In particular, including the vertical component of the excitations was found to increase the diametral deformation by at least 150% and tended to increase other structural responses. Moreover, the bolt tension increased significantly by over 400% with only a 150% increase in seismic intensity, highlighting the strong nonlinear sensitivity. However, due to the inherent constraints of the 2D plane-strain assumption, the influence of the bolt type remains inconclusive.

MMD U-Net: A multi-scale interaction and dynamic upsampling network for pancreas and pancreatic tumor segmentation in CT images

Strong adhesives have tremendous application potential across various fields. However, those relying on chemical bonding tend to become brittle after curing and are difficult to remove or recycle upon adhesion failure. Currently, adhesives based on dynamic interactions often suffer from poor adhesion reversibility and solvent resistance. Herein, a superstrong adhesive was fabricated via block copolymerization of acrylic acid (AA), methyl methacrylate, styrene, polydimethylsiloxane (PDMS), and quaternary ammonium salt (QAS). Owing to strong interactions of rigid ionic segments and enhanced toughness by soft PDMS segments, the copolymer achieves adhesion strengths exceeding 10.0 MPa on various substrates, with a maximum of ∼20.9 MPa on glass, exceeding classic strong adhesives such as epoxy and ethyl α-cyanoacrylate. Especially, the zwitterionic adhesives present high adhesion reversibility, with 57-94% adhesion recovery on polar and nonpolar substrates via solvent-triggered dynamic bond reorganization. Moreover, the integration of diverse molecular interactions, including electrostatic interactions, hydrogen bonding, and hydrophobic interactions, endows the adhesive with high resistance to organic and aqueous media. The copolymer exhibits exceptional antibacterial activity against Gram-positive and Gram-negative bacteria, predominantly from the QAS and AA components. Such a strong, reversible, robust, versatile, and antibacterial adhesive holds great promise for achieving durable adhesion under practical application conditions.

This study explores the influence of propylene carbonate (PC) on the properties of casting polyurethane elastomers. Five formulations were developed with varying PC content: 0% (PU0), 1% (PU1), 5% (PU5), 10% (PU10), and 20% (PU20) by weight in the prepolymer, while maintaining a consistent curing ratio. The effect of PC incorporation on hardness, resilience, segmental interaction, stress-strain behavior, thermal stability, and abrasion resistance was systematically evaluated to understand the impact of PC on the phase structure and performance of the elastomers. The successful synthesizing of polyurethanes was confirmed by spectroscopy analysis. The hardness values showed a progressive decline with increasing PC content, indicating a reduction in cross-link density and increased polymer chain mobility. The resilience of samples increases progressively from PU0 (28%) to PU20 (47%), indicating that PC enhances the elastomer’s ability to recover energy upon deformation. At higher PC levels, the system transitions toward rubber-like elasticity. By modifying the cross-link density and enhancing chain mobility, PC improves the elongation of the polymer while reducing tensile strength. In Addition, abrasion loss decreased with increasing PC concentration. This study provides insights into optimizing the composition of polyurethane elastomers for diverse applications where tailored physical and mechanical properties are required. Adding PC offers a simple and cost-effective method for tailoring the properties of ester-based polyurethane elastomers.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-025-32349-4.

An efficient self-adaptive planning algorithm for shield machine attitude and segment assembly considering machine–segment interaction

Due to the promotion of the digital wave, the traditional marketing model in the home industry has fallen into the dilemma of single experience and high decision-making costs. Users' demand for personalized and immersive experiences is becoming increasingly prominent, and experiential marketing has become a key way for brands to overcome bottlenecks. The application of AR and VR real-life technology in marketing continues to deepen. However, the integration of the two in home experience marketing faces challenges such as insufficient scene adaptation and homogeneous user interaction. Research on strategy optimization is relatively scattered. This study takes user experience as a guide and combines it with the CAS paper structure framework to explore the upgrade path in home experience marketing. By sorting out the current status of technological development and application, analyzing relevant cases of brands such as IKEA and Easyhome, the authors can understand the strengths, weaknesses, opportunities, and crises of current strategies. Through the integration of technology, a marketing closed-loop system can be formed, which includes "precise insight into needs - immersive scene experience - personalized solution output - efficient conversion and retention". Different consumer scenarios (partial revitalization/whole house adaptation) require differentiated design of technology application focus, providing optimization methods in dimensions such as technology integration, scene segmentation, and user interaction, providing practical reference for home brands and filling the gap in system strategy research in this field.

Influenza A, a negative-sense RNA virus, has a genome that consists of eight single-stranded RNA segments. Influenza co-infections can result in reassortant viruses that contain gene segments from multiple strains, causing pandemic outbreaks with severe consequences for human health. The outcome of reassortment is likely influenced by a selective sequence-specific genome packaging mechanism. To uncover the contributions of individual segment pairings to selective packaging, we set out to statistically analyse packaging defects and inter-segment distances in individual A/Puerto Rico/8/34 (H1N1) (PR8) virus particles. To enable such analysis, we developed a multiplexed DNA-PAINT approach capable of assessing the segment stoichiometry of >10 000 individual virus particles in one experiment; our approach can also spatially resolve the individual segments inside complete virus particles with a localization precision of ∼10 nm. Our results show the influenza genome can be assembled through multiple pathways in a redundant and cooperative process guided by preferentially interacting segment pairs and aided by synergistic effects that enhance genome assembly, driving it to completion. Our structural evidence indicates that the interaction strength of segment pairs affects the spatial configuration of the gene segments, which appears to be preserved in mature virions. As our method quantified the interactions of whole influenza segments instead of identifying individual sequence-based interactions, our results can serve as a template to quantify the contributions of individual sequence motifs to selective packaging.

Integrins are transmembrane receptors that mediate cell adhesion and signaling. They perform allosteric rearrangements to transmit external signals to intracellular regions, while intracellular signaling events can also influence their extracellular behavior. The αIIbβ3 integrin, found in human platelets, is involved in the thrombosis and hemostatic processes. Protein-lipid interactions can regulate membrane organization, influencing integrin conformational states and downstream signaling. In this study, we expressed and purified peptides comprising the transmembrane (TM), the intracellular (IC) segments and a small portion of the extracellular segments (EC) of αIIb and β3 integrin and studied the surface behavior of the pure peptides and their mixtures with 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), with a mixture of the zwitterionic lipid POPC with the anionic 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol (POPG), and with 1,2-dipalmitoyl-sn-3-glycerophosphocholine (DPPC) using Langmuir monolayers and Brewster angle microscopy (BAM). The peptides formed stable monolayers at the air/water interface. Mixtures with the unsaturated lipid POPC and a mixture of POPC-POPG showed a negative deviation from ideal mixing, while the mixtures with the saturated DPPC deviated positively from ideal mixing. BAM imaging revealed that pure peptides formed a continuous network of reflective threads surrounding dark patches. In the mixtures with DPPC, the threads were fragmented, while POPC mixtures maintained the connectivity. These findings showed that these peptides from integrin αIIbβ3 have different interactions with different lipids, which could have implications for integrin activation and function in platelet membranes.

Thermoplastic polyurethane (TPU) is a versatile, melt-processable elastomer valued for its flexibility and durability. However, expanding its application to advanced industries requires enhanced multifunctionality, such as improved flame retardancy and thermal conductivity, without compromising its intrinsic mechanical properties. In this study, we present a fully dry, solvent-free strategy that integrates mechanochemistry (MC) with reactive extrusion (REX) to fabricate multifunctional TPU composites. Unlike conventional methods that rely on solution-based processing or hard segment interactions, this approach leverages MC to pre-establish a bonded filler network, which is then grafted onto the TPU soft segments via reactive coupling during REX. This sequential MC–REX process enables uniform filler dispersion and enhanced interfacial compatibility, while eliminating the need for solvents or additional compatibilizers. Characterization results demonstrate that composites prepared via MC–REX exhibit improved flame retardancy, enhanced thermal conductivity, and minimized deterioration in mechanical properties compared to conventional composites. Moreover, the composites retain their mechanical performance after multiple reprocessing cycles, indicating improved recyclability. This work offers a scalable, eco-friendly processing platform for designing next-generation TPU composites with tunable multifunctionality through controlled soft segment–filler interactions.

Objectives: To comprehensively examine the association between spinopelvic alignment and muscle shortening in healthy young men, focusing on the individual and interactive effects of thoracic kyphosis, lumbar lordosis, and anterior pelvic tilt using SHapley Additive exPlanation (SHAP) analysis. Methods: Forty-one healthy young adult men participated in this cross-sectional study. Thoracic kyphosis, lumbar lordosis, and anterior pelvic tilt were measured using a flexible curve ruler and inclinometer. Muscle length indices for six muscles (iliopsoas, rectus femoris, gluteus maximus, hamstrings, back extensors, and abdominals) were assessed via standardized physical examinations and image analysis. A machine learning model was developed, and SHAP analysis applied to determine individual and interactive contributions of spinopelvic angles to each muscle length index. Results: SHAP analysis showed that hip-related muscle shortening (iliopsoas, rectus femoris, hamstrings, gluteus maximus) was influenced by both individual alignments and interactions, especially between thoracic kyphosis and lumbar lordosis. Lumbar lordosis was most associated with iliopsoas shortening (SHAP = −0.09), while anterior pelvic tilt was linked to hamstring shortening (SHAP = −0.30). Thoracic kyphosis was the key factor for rectus femoris shortening (SHAP = −0.05). Interactive effects exceeded individual contributions for the rectus femoris, gluteus maximus, and hamstrings. In contrast, spinal alignment had minimal influence on the back extensors and abdominals. Conclusions: Both individual and intersegmental spinal alignments are associated with muscle shortening, particularly in hip-related muscles. The interaction between thoracic kyphosis and lumbar lordosis plays a pivotal role. These findings underscore the importance of evaluating segmental spinal interactions when assessing muscle flexibility and posture.

Poly(ε-caprolactone) (PCL) has been widely studied due to its excellent biodegradability and biocompatibility. However, its low melting temperature (∼60 °C) and limited tensile strength (∼10 MPa) restrict its application in high-performance or thermally demanding environments. To overcome these limitations, we developed a novel two-step melt polycondensation strategy to synthesize quasi-alternating poly(ester amide)s (PEAs) by incorporating diamide diols derived from ε-caprolactone. This synthetic approach enables precise control over the polymer microstructure, leading to PEAs with tunable crystallinity, enhanced chain regularity, and improved segmental interactions. The successful synthesis of PEAs was confirmed using nuclear magnetic resonance and infrared spectroscopy. The results demonstrate that PEAs possess adjustable melting temperatures (Tm = 134-139 °C), moderate crystallinity (Xc = 17-35%), excellent transparency (T > 85%), tensile strength (15-18 MPa), and high elongation at break (ε = 68-800%). This work offers a promising route toward high-performance, sustainable materials for packaging and biomedical applications.