Effects of body posture on aortic valve hemodynamics and biomechanics using the fluid-structure interaction method.

Bioprosthetic heart valve (BHV), the most widely and commonly used valves in clinical practice, are susceptible to fatigue damage. Biological valves are always in one or fewer body postures before sampling in pigs and bovines. Nevertheless, human body positions are far more than them. Variations in body position significantly affect the intrinsic environment of blood pressure (BP), heart rate (HR), and peripheral resistance (PR). Such boundary condition changes will inevitably affect the implanted biological valve. In this paper, the immersed boundary method was used to simulate the motion of the aortic valve during the entire cardiac cycle in five postural blood flow environments: upright, sitting, prone, supine and orthostatic hypotension (OH). Several hemodynamic and biomechanical parameters, including the transvalvular pressure gradient and valve displacement, were evaluated. The results showed that the OH group exhibited the worst performance of the valves, accompanied by the greatest regurgitation and high-frequency flutter, predisposing patients to thrombosis and fatigue calcification. For BHVs to serve longer, patients implanted with BHV should avoid OH in their daily routine.

Enabling High Strength and Toughness Polyurethane through Disordered-Hydrogen Bonds for Printable, Recyclable, Ultra-Fast Responsive Capacitive Sensors.

The rapid advancement of smart, flexible electronic devices has paralleled a surge in electronic waste (e-waste), exacerbating massive resource depletion and serious environmental pollution. Recyclable materials are extensively investigated to address these challenges. Herein, this study designs a unique polyurethane (SPPUs) with ultra-high strength up to 60 MPa and toughness of 360 MJ m-3. This synthetic SPPUs can be fully recycled at room temperature by using green solvents of ethanol. Accordingly, the resultant SPPU-Ni composites, created by mixing the ethanol-dissolved SPPUs solution with nickel (Ni) powder, effectively combine the flexibility and recyclability of SPPUs with the electrical conductivity of the nickel filler. Notably, this work develops the printable capacitive sensors (PCBS) through transcribing the paste of SPPUs-Ni slurry onto PET film and paper using screen-printing technology. The devised PCBS have fast response time ≈50 ms, high resolution, and multiple signal recognition capabilities. Remarkably, SPPUs and Ni powder can be fully recycled by only dissolving the waste PCBS in ethanol. This work offers a sustainable solution to the growing e-waste problem in recyclable flexible electronics.

GIHP: Graph convolutional neural network based interpretable pan-specific HLA-peptide binding affinity prediction.

Accurately predicting the binding affinities between Human Leukocyte Antigen (HLA) molecules and peptides is a crucial step in understanding the adaptive immune response. This knowledge can have important implications for the development of effective vaccines and the design of targeted immunotherapies. Existing sequence-based methods are insufficient to capture the structure information. Besides, the current methods lack model interpretability, which hinder revealing the key binding amino acids between the two molecules. To address these limitations, we proposed an interpretable graph convolutional neural network (GCNN) based prediction method named GIHP. Considering the size differences between HLA and short peptides, GIHP represent HLA structure as amino acid-level graph while represent peptide SMILE string as atom-level graph. For interpretation, we design a novel visual explanation method, gradient weighted activation mapping (Grad-WAM), for identifying key binding residues. GIHP achieved better prediction accuracy than state-of-the-art methods across various datasets. According to current research findings, key HLA-peptide binding residues mutations directly impact immunotherapy efficacy. Therefore, we verified those highlighted key residues to see whether they can significantly distinguish immunotherapy patient groups. We have verified that the identified functional residues can successfully separate patient survival groups across breast, bladder, and pan-cancer datasets. Results demonstrate that GIHP improves the accuracy and interpretation capabilities of HLA-peptide prediction, and the findings of this study can be used to guide personalized cancer immunotherapy treatment. Codes and datasets are publicly accessible at: https://github.com/sdustSu/GIHP.

MGPPI: multiscale graph neural networks for explainable protein-protein interaction prediction.

Protein-Protein Interactions (PPIs) involves in various biological processes, which are of significant importance in cancer diagnosis and drug development. Computational based PPI prediction methods are more preferred due to their low cost and high accuracy. However, existing protein structure based methods are insufficient in the extraction of protein structural information. Furthermore, most methods are less interpretable, which hinder their practical application in the biomedical field. In this paper, we propose MGPPI, which is a Multiscale graph convolutional neural network model for PPI prediction. By incorporating multiscale module into the Graph Neural Network (GNN) and constructing multi convolutional layers, MGPPI can effectively capture both local and global protein structure information. For model interpretability, we introduce a novel visual explanation method named Gradient Weighted interaction Activation Mapping (Grad-WAM), which can highlight key binding residue sites. We evaluate the performance of MGPPI by comparing with state-of-the-arts methods on various datasets. Results shows that MGPPI outperforms other methods significantly and exhibits strong generalization capabilities on the multi-species dataset. As a practical case study, we predicted the binding affinity between the spike (S) protein of SARS-COV-2 and the human ACE2 receptor protein, and successfully identified key binding sites with known binding functions. Key binding sites mutation in PPIs can affect cancer patient survival statues. Therefore, we further verified Grad-WAM highlighted residue sites in separating patients survival groups in several different cancer type datasets. According to our results, some of the highlighted residues can be used as biomarkers in predicting patients survival probability. All these results together demonstrate the high accuracy and practical application value of MGPPI. Our method not only addresses the limitations of existing approaches but also can assists researchers in identifying crucial drug targets and help guide personalized cancer treatment.

DNA-functionalized MOF fluorescent probes for the enzyme-free and pretreatment-free detection of MicroRNA in serum.

MicroRNA (miRNA) is a promising biomarker that plays an important role in various biomedical applications, especially in cancer diagnosis. However, the current miRNA detection technology has inherent limitations such as complex operation, expensive testing cost and excessive detection time. In this study, a dual signal amplification biosensor based on DNA-functionalized metal-organic frameworks (MOFs) fluorescent probes, MFPBiosensor, was established for the enzyme-free and pretreatment-free detection of the colon cancer (CC) marker miR-23a. DNA-functionalized MOFs NH2-MIL-53(Al) (DNA@MOFs) were synthesized as fluorescent probes with specific recognition functions. A single DNA@MOF carries a large number of fluorescent ligands 2-aminoterephthalic acid (NH2-H2BDC), which can generate strong fluorescence signals after alkaline hydrolysis. Combined with catalyzed hairpin assembly (CHA), an efficient isothermal amplification technique, the dual signal enhancement strategy reduced matrix interference and sensitized the signal response. The established MFPBiosensor successfully detected extremely low levels of miRNA in complex biological samples with acceptable sensitivity and specificity. With a single detection cost of $0.583 and a test time of 50 min, the excellent inexpensive and rapid advantage of the MFPBiosensor is highlighted. More importantly, the subtle design enables the MFPBiosensor to achieve convenient batch detection, where miRNA in serum can be directly detected without any pretreatment process or enzyme. In conclusion, MFPBiosensor is a promising biosensor with substantial potential for commercial miRNA detection and clinical diagnostic applications of CC.

Multiregion WES of metastatic pancreatic neuroendocrine tumors revealed heterogeneity in genomic alterations, immune microenvironment and evolutionary patterns.

Pancreatic neuroendocrine tumors (PanNETs), though uncommon, have a high likelihood of spreading to other body parts. Previously, the genetic diversity and evolutionary patterns in metastatic PanNETs were not well understood. To investigate this, we performed multiregion sampling whole-exome sequencing (MRS-WES) on samples from 10 patients who had not received prior treatment for metastatic PanNETs. This included 29 primary tumor samples, 31 lymph node metastases, and 15 liver metastases. We used the MSK-MET dataset for survival analysis and validation of our findings. Our research indicates that mutations in the MEN1/DAXX genes might trigger the early stages of PanNET development. We categorized the patients based on the presence (MEN1/DAXXmut, n = 7) or absence (MEN1/DAXXwild, n = 3) of these mutations. Notable differences were observed between the two groups in terms of genetic alterations and clinically relevant mutations, confirmed using the MSK-MET dataset. Notably, patients with mutations in MEN1/DAXX/ATRX genes had a significantly longer median overall survival compared to those without these mutations (median not reached vs. 43.63 months, p = 0.047). Multiplex immunohistochemistry (mIHC) analysis showed a more prominent immunosuppressive environment in metastatic tumors, especially in patients with MEN1/DAXX mutations. These findings imply that MEN1/DAXX mutations lead PanNETs through a unique evolutionary path. The disease's progression pattern indicates that PanNETs can spread early, even before clinical detection, highlighting the importance of identifying biomarkers related to metastasis to guide personalized treatment strategies.

High-quality genome assembly enables prediction of allele-specific gene expression in hybrid poplar.

Poplar (Populus) is a well-established model system for tree genomics and molecular breeding, and hybrid poplar is widely used in forest plantations. However, distinguishing its diploid homologous chromosomes is difficult, complicating advanced functional studies on specific alleles. In this study, we applied a trio-binning design and PacBio high-fidelity long-read sequencing to obtain haplotype-phased telomere-to-telomere genome assemblies for the 2 parents of the well-studied F1 hybrid "84K" (Populus alba × Populus tremula var. glandulosa). Almost all chromosomes, including the telomeres and centromeres, were completely assembled for each haplotype subgenome apart from 2 small gaps on one chromosome. By incorporating information from these haplotype assemblies and extensive RNA-seq data, we analyzed gene expression patterns between the 2 subgenomes and alleles. Transcription bias at the subgenome level was not uncovered, but extensive-expression differences were detected between alleles. We developed machine-learning (ML) models to predict allele-specific expression (ASE) with high accuracy and identified underlying genome features most highly influencing ASE. One of our models with 15 predictor variables achieved 77% accuracy on the training set and 74% accuracy on the testing set. ML models identified gene body CHG methylation, sequence divergence, and transposon occupancy both upstream and downstream of alleles as important factors for ASE. Our haplotype-phased genome assemblies and ML strategy highlight an avenue for functional studies in Populus and provide additional tools for studying ASE and heterosis in hybrids.

Short-term and Long-term Outcomes of Robotic Enucleation of Tumors Located in the Pancreatic Head and Uncinate Process.

OBJECTIVE: To assess short-term and long-term outcomes following robotic enucleation (REn) of tumors in the proximal pancreas. BACKGROUND: Despite the advantages of preserving function via pancreatic enucleation, controversies persist, since this can be associated with severe complications, such as clinically relevant postoperative pancreatic fistula, especially when performed near the main pancreatic duct. The safety and efficacy of REn in this context remain largely unknown. METHODS: A retrospective analysis was performed of all patients who underwent REn for benign and low-grade malignant neoplasms in the pancreatic head and uncinate process between January 2005 and December 2021. Clinicopathologic, perioperative, and long-term outcomes were compared with a similar open enucleation (OEn) group. RESULTS: Of 146 patients, 92 underwent REn with a zero conversion-to-open rate. REn was superior to OEn in terms of shorter operative time (90.0 minutes vs 120.0 minutes, P<0.001), decreased blood loss (20.0 mL vs 100.0 min, P=0.001), and lower clinically relevant postoperative pancreatic fistula rate (43.5% vs 61.1%, P=0.040). Bile leakage rate, major morbidity, 90-day mortality, and length of hospital stay were comparable between groups. No post-REn grade C POPF or grade IV/V complication was identified. Subgroup analyses for uncinate process tumors and proximity to the main pancreatic duct did not demonstrate inferior postoperative outcomes. In a median follow-up period of 50 months, REn outcomes were comparable to OEn regarding recurrence rate and pancreatic endocrine or exocrine function. CONCLUSIONS: REn for pancreatic head and uncinate process tumors improved clinically relevant outcomes without increased major complications compared to OEn, while demonstrating comparable long-term oncological and functional outcomes.

High-quality maple genome reveals duplication-facilitated leaf color diversity.

Acer truncatum is a horticultural tree species with individuals that display either yellow or red leaves in autumn, giving it high ornamental and economic value. 'Lihong' of A. truncatum is an excellent cultivar due to its characteristic of having autumn leaves that turn a bright and beautiful shade of red, while its closely related cultivar 'Bunge' does not. However, the molecular mechanism underlying the color change in the cultivar 'Lihong' is still unclear. Here, we assembled a high-quality genome sequence of Acer truncatum 'Lihong' (genome size = 688 Mb, scaffold N50 = 9.14 Mb) with 28,438 protein-coding genes predicted. Through comparative genomic analysis, we found that 'Lihong' had experienced more tandem duplication events although it's a high degree of collinearity with 'Bunge'. Especially, the expansion of key enzymes in the anthocyanin synthesis pathway was significantly uneven between the two varieties, with 'Lihong' genome containing a significantly higher number of tandem/dispersed duplication key genes. Further transcriptomic, metabolomic, and molecular functional analyses demonstrated that several UFGT genes, mainly resulting from tandem/dispersed duplication, followed by the promoter sequence variation, may contribute greatly to the leaf color phenotype, which provides new insights into the mechanism of divergent anthocyanin accumulation process in the 'Lihong' and 'Bunge' with yellow leaves in autumn. Further, constitutive expression of two UFGT genes, which showed higher expression in 'Lihong', elevated the anthocyanin content. We proposed that the small-scale duplication events could contribute to phenotype innovation.

Telomere-to-telomere and haplotype-resolved genome assembly of the Chinese cork oak (Quercus variabilis).

The Quercus variabilis, a deciduous broadleaved tree species, holds significant ecological and economical value. While a chromosome-level genome for this species has been made available, it remains riddled with unanchored sequences and gaps. In this study, we present a nearly complete comprehensive telomere-to-telomere (T2T) and haplotype-resolved reference genome for Q. variabilis. This was achieved through the integration of ONT ultra-long reads, PacBio HiFi long reads, and Hi-C data. The resultant two haplotype genomes measure 789 Mb and 768 Mb in length, with a contig N50 of 65 Mb and 56 Mb, and were anchored to 12 allelic chromosomes. Within this T2T haplotype-resolved assembly, we predicted 36,830 and 36,370 protein-coding genes, with 95.9% and 96.0% functional annotation for each haplotype genome. The availability of the T2T and haplotype-resolved reference genome lays a solid foundation, not only for illustrating genome structure and functional genomics studies but also to inform and facilitate genetic breeding and improvement of cultivated Quercus species.

Berberine protects mice against type 2 diabetes by promoting PPARγ-FGF21-GLUT2-regulated insulin sensitivity and glucose/lipid homeostasis.

Type 2 diabetes (T2D) is a chronic, burdensome disease that is characterized by disordered insulin sensitivity and disturbed glucose/lipid homeostasis. Berberine (BBR) has multiple therapeutic actions on T2D, including regulation of glucose and lipid metabolism, improvement of insulin sensitivity and energy expenditure. Recently, the function of BBR on fibroblast growth factor 21 (FGF21) has been identified. However, if BBR ameliorates T2D through FGF21, the underlying mechanisms remain unknown. Herein, we used T2D wild type (WT) and FGF21 global knockout (FKO) mice [mouse T2D model: established by high-fat diet (HFD) feeding plus streptozotocin (STZ) injection], and hepatocyte-specific peroxisome proliferator activated receptor γ (PPARγ) deficient (PPARγHepKO) mice, and cultured human liver carcinoma cells line, HepG2 cells, to characterize the role of BBR in glucose/lipid metabolism and insulin sensitivity. We found that BBR activated FGF21 expression by up-regulating PPARγ expression at the cellular level. Meanwhile, BBR ameliorated glucosamine hydrochloride (Glcn)-induced insulin resistance and increased glucose transporter 2 (GLUT2) expression in a PPARγ/FGF21-dependent manner. In T2D mice, BBR up-regulated the expression of PPARγ, FGF21 and GLUT2 in the liver, and GLUT2 in the pancreas. BBR also reversed T2D-induced insulin resistance, liver lipid accumulation, and damage in liver and pancreas. However, FGF21 deficiency diminished these effects of BBR on diabetic mice. Altogether, our study demonstrates that the therapeutic effects of BBR on T2D were partly accomplished by activating PPARγ-FGF21-GLUT2 signaling pathway. The discovery of this new pathway provides a deeper understanding of the mechanism of BBR for T2D treatment.

Study on soil hydraulic properties of slope farmlands with different degrees of erosion degradation in a typical black soil region.

In order to explore the impact of soil erosion degradation on soil hydraulic properties of slope farmland in a typical black soil region, typical black soils with three degrees of erosion degradation (light, moderate and heavy) were selected as the research objects. The saturated hydraulic conductivity, water holding capacity and water supply capacity of the soils were analyzed, as well as their correlations with soil physicochemical properties. The results showed that the saturated hydraulic conductivity of black soils in slope farmlands decreased with erosion degradation degree, which was higher in 0-10 cm soil layer than in 10-20 cm soil layer. The water holding capacity and water supplying capacity of typical black soils also decreased with the increase of erosion degradation degree, and both of them were stronger in the upper soil than in the lower soil. With the aggravation of erosion degradation of black soils, soil organic matter content decreased while soil bulk density increased, leading to the decline of soil hydraulic conductivity. The increase of soil bulk density and the decrease of contents of organic matter and >0.25 mm water stable aggregates were the main factors leading to the decrease of soil water holding capacity. These findings provide scientific basis and basic data for rational utilization of soil water, improvement of land productivity and prevention of soil erosion.

Multi-omics analysis of oxidative stress and apoptosis in hepatopancreas cells induced by Polyascus gregaria parasitizing the Eriocheir sinensis.

Polyascus gregaria, a parasitic barnacle, poses a significant threat to Eriocheir sinensis farms by inhibiting crab growth. However, the molecular and pathological mechanisms behind P. gregaria infection in the hepatopancreas of E. sinensis remain unclear. In this study, we investigated the impact and underlying mechanisms of P. gregaria infection on E. sinensis through analyzing the infected hepatopancreatic tissues by tandem mass tag technology and RNA-Seq high-throughput sequencing. Among the identified 10,693 differentially expressed genes, 294 genes were significantly altered following P. gregaria infection, including 92 upregulated and 202 downregulated genes. Gene Ontology and Kyoto Encyclopedia of Genes and Genomes analyses further revealed the involvement of these genes in oxidative decomposition, lipid metabolism, inflammation, and hepatopancreas metabolism. Meanwhile, the identified 253 differentially expressed proteins, including 143 upregulated and 110 downregulated proteins, are mainly related to cellular and metabolic processes, catalytic activity, and cell components. The pathway analysis indicated their enrichment in glycolysis/gluconeogenesis, oxidative phosphorylation, endoplasmic reticulum protein processing, and actin cytoskeleton regulation. The involvement of these differentially expressed genes and proteins in the peroxisome proliferator-activated receptors pathway during host immune responses against P. gregaria infection has been highlighted. Furthermore, pathological examinations and biochemical indicators jointly demonstrated the hepatopancreatic damage and increased oxidative stress and apoptosis in the infected E. sinensis. Collectively, our study provides crucial insights into the mechanisms underlying the E. sinensis-P. gregaria interactions, and may contribute to the development of novel strategies for parasite control and reducing economic losses in aquaculture.

Unraveling the evolutionary dynamics of the TPS gene family in land plants.

Terpenes and terpenoids are key natural compounds for plant defense, development, and composition of plant oil. The synthesis and accumulation of a myriad of volatile terpenoid compounds in these plants may dramatically alter the quality and flavor of the oils, which provide great commercial utilization value for oil-producing plants. Terpene synthases (TPSs) are important enzymes responsible for terpenic diversity. Investigating the differentiation of the TPS gene family could provide valuable theoretical support for the genetic improvement of oil-producing plants. While the origin and function of TPS genes have been extensively studied, the exact origin of the initial gene fusion event - it occurred in plants or microbes - remains uncertain. Furthermore, a comprehensive exploration of the TPS gene differentiation is still pending. Here, phylogenetic analysis revealed that the fusion of the TPS gene likely occurred in the ancestor of land plants, following the acquisition of individual C- and N- terminal domains. Potential mutual transfer of TPS genes was observed among microbes and plants. Gene synteny analysis disclosed a differential divergence pattern between TPS-c and TPS-e/f subfamilies involved in primary metabolism and those (TPS-a/b/d/g/h subfamilies) crucial for secondary metabolites. Biosynthetic gene clusters (BGCs) analysis suggested a correlation between lineage divergence and potential natural selection in structuring terpene diversities. This study provides fresh perspectives on the origin and evolution of the TPS gene family.

Retinoic acid delays murine palatal shelf elevation by inhibiting Wnt5a-mediated noncanonical Wnt signaling and downstream Cdc-42/F-actin remodeling in mesenchymal cells.

BACKGROUND: Mammalian palatal shelves erupted from maxillary prominences undergo vertical extention, transient elevation, and horizontal growth to fuse. Previous studies in mice reported that the retinoic acid (RA) contributed to cleft palate in high incidence by delaying the elevating procedure, but little was known about the underlying biological mechanisms. METHODS: In this study, hematoxylin-eosin and immunofluorescence staining were employed to evaluate the phenotypes and the expression of related markers in the RA-treated mice model. In situ hybridization and RT-qPCR were used to detect the expression of genes involved in Wnt signaling pathway. The palatal mesenchymal cells were cultured in vitro, and stimulated with RA or CASIN, and co-treated with Foxy5. Wnt5a and Ccd42 expression were evaluated by immunofluorescence staining. Phalloidin was used to label the microfilament cytoskeleton (F-actin) in cultured cells. RESULTS: We revealed that RA resulted in 100% incidence of cleft palate in mouse embryos, and the expression of genes responsible for Wnt5a-mediated noncanonical Wnt signal transduction were specifically downregulated in mesenchymal palatal shelves. The in vitro study of palatal mesenchymal cells indicated that RA treatment disrupted the organized remodeling of cytoskeleton, an indicative structure of cell migration regulated by the small Rho GTPase Cdc42. Moreover, we showed that the suppression of cytoskeleton and cell migration induced by RA was partially restored using the small molecule Foxy-5-mediated activation of Wnt5A, and this restoration was attenuated by CASIN (a selective GTPase Cdc42 inhibitor) again. CONCLUSIONS: These data identified a crucial mechanism for Wnt5a-mediated noncanonical Wnt signaling in acting downstream of Rho GTPase Cdc42 to regulate cytoskeletal remodeling and cell migration during the process of palate elevation. Our study provided a new explanation for the cause of cleft palate induced by RA.

The Synergistic Effects between Liquid Crystal and Crystalline Phase on Photo-Responsive Elastomers toward Quick Photo-Responsive Performance.

Adopting only a small amount of azobenzene molecular to design liquid crystal photo-responsive materials capable of quick response and flexible adjustability is in high demand but is challenging. Herein, azobenzenemolecules into polyurethane elastomer containing crystalline structure for preparing azobenzene liquid-crystal elastomers (ALCEs) are demonstrated and this phenomenon of the synergistic effects between liquid crystal and crystalline phase is discovered. The key point of the work is that the synthetic ALCEs can utilize the reversible isomerism capability of azobenzene molecules under light irradiation, which can pry the motion of the macromolecular crystalline region in system to realize the large macroscopic deformation of the photo-responsive behavior. Obviously, the ALCEs sample containing azobenzene molecule and polyethylene glycol crystallization can quickly bend, illuminated by ultraviolet light and rapidly straighten under green light. Under the same ultraviolet irradiation, the bending speed, final bending angle, recovery rate and recovery ratio of ALCEs are larger than that of ALCEs without any crystalline structure. This ALCEs based on the synergistic effects between liquid crystal and crystalline phase can break through the current dilemma that the application of traditional azobenzene photo-responsive materials is limited by their concentration, greatly expanding the design thought and their scope of application.

Omni-Frequency Channel-Selection Representations for Unsupervised Anomaly Detection.

Density-based and classification-based methods have ruled unsupervised anomaly detection in recent years, while reconstruction-based methods are rarely mentioned for the poor reconstruction ability and low performance. However, the latter requires no costly extra training samples for the unsupervised training that is more practical, so this paper focuses on improving reconstruction-based method and proposes a novel O mni-frequency C hannel-selection R econstruction (OCR-GAN) network to handle sensory anomaly detection task in a perspective of frequency. Concretely, we propose a Frequency Decoupling (FD) module to decouple the input image into different frequency components and model the reconstruction process as a combination of parallel omni-frequency image restorations, as we observe a significant difference in the frequency distribution of normal and abnormal images. Given the correlation among multiple frequencies, we further propose a Channel Selection (CS) module that performs frequency interaction among different encoders by adaptively selecting different channels. Abundant experiments demonstrate the effectiveness and superiority of our approach over different kinds of methods, e.g., achieving a new state-of-the-art 98.3 detection AUC on the MVTec AD dataset without extra training data that markedly surpasses the reconstruction-based baseline by +38.1 ↑ and the current SOTA method by +0.3 ↑ . The source code is available in the additional materials.

Face-based age estimation using improved Swin Transformer with attention-based convolution.

Recently Transformer models is new direction in the computer vision field, which is based on self multihead attention mechanism. Compared with the convolutional neural network, this Transformer uses the self-attention mechanism to capture global contextual information and extract more strong features by learning the association relationship between different features, which has achieved good results in many vision tasks. In face-based age estimation, some facial patches that contain rich age-specific information are critical in the age estimation task. The present study proposed an attention-based convolution (ABC) age estimation framework, called improved Swin Transformer with ABC, in which two separate regions were implemented, namely ABC and Swin Transformer. ABC extracted facial patches containing rich age-specific information using a shallow convolutional network and a multiheaded attention mechanism. Subsequently, the features obtained by ABC were spliced with the flattened image in the Swin Transformer, which were then input to the Swin Transformer to predict the age of the image. The ABC framework spliced the important regions that contained rich age-specific information into the original image, which could fully mobilize the long-dependency of the Swin Transformer, that is, extracting stronger features by learning the dependency relationship between different features. ABC also introduced loss of diversity to guide the training of self-attention mechanism, reducing overlap between patches so that the diverse and important patches were discovered. Through extensive experiments, this study showed that the proposed framework outperformed several state-of-the-art methods on age estimation benchmark datasets.