Synchronous modification of ZIF-67 with cyclomatrix polyphosphazene coating for efficient flame retardancy and mechanical reinforcement of epoxy resin.

Cyclomatrix polyphosphazenes have attracted widespread attention in the field of polymer flame retardancy. Nevertheless, the optimal manifestation of their distinctive structural attributes and flame-retardant properties necessitates a judicious selection of condensation monomers and synergistic templates during the fabrication of polyphosphazene flame retardants. In our previous studies, it was discovered that when ZIF-67 is functionalized with polyphosphazene, the by-product HCl from phosphazene polycondensation causes etching on ZIF-67. Based on this "synchronous etching" effect, a series of hybrid materials comprising cyclomatrix polyphosphazene and ZIF-67, denoted as ZIF-67@PDS (PDS, poly-(cyclotriphosphazene-co-4,4'-diaminodiphenyl sulfone)), ZIF-67@PBS (PBS, poly-(cyclotriphosphazene-co-Bisphenol A)), and ZIF-67@PZS (PZS, poly-(cyclotriphosphazene-co-4,4'-sulfonyldiphenol)), was synthesized utilizing DDS (4,4'-diaminodiphenyl sulfone), BPA (Bisphenol A), and BPS (4,4'-sulfonyldiphenol) monomers as precursors, respectively. Upon the incorporation of 2.0 wt.% of ZIF-67@PDS, ZIF-67@PBS, and ZIF-67@PZS, the flame retardant and mechanical characteristics of EP composites exhibited marked enhancement. The unique structural characteristics of hybrid and the synergistic effects of Co-P-N contribute to the improvement of comprehensive properties. Compared with pure EP, EP/ZIF-67@PZS has the best enhancement effect, and its pHRR, THR, and TSP decreased by 34.0%, 30.0%, and 40.5%, respectively. In terms of mechanical strength, ZIF-67@PZS also increases the flexural strength of EP by 37.42%. Relying on the "synchronous etching" effect, this study explores and verifies the effective combination of ZIF-67 and different types of polyphosphazenes, and obtains a series of ZIF-67-derived cyclomatrix polyphosphazene hybrids with different morphologies and properties in one step. It provides a new idea and strategy for the simultaneous modification of polyphosphazene materials and the preparation of multifunctional flame retardants in the future.

Enhancing char formation of flame retardant epoxy composites: Onigiri-like ZIF-67 modification with carboxymethyl ß-cyclodextrin crosslinking.

To enhance char formation of flame retardant epoxy (EP) composites, carboxymethyl ß-cyclodextrin (CM-ß-CD) is employed as an etchant for or ZIF-67 derivatives. In the early stage, etching plays a dominant role. The mismatch in size between CM-ß-CD opening and ZIF-67 pore leads to the stacking of carboxyl cobalt complexes on the shell. When the reaction time is prolonged, crosslinking occurs between carboxyl and hydroxyl groups. Crosslinked CM-ß-CD weakens and eventually stops the etching process. Triethyl phosphate (TEP), an additive to improve flame retardancy, is also absorbed on the shell in this one-pot synthesis. Herin, the synthesis of metal-organic framework (MOF) derivatives can impart multiple functions to MOF. This novel nanohybrid significantly improved flame retardancy of EP composites with only 2.0 wt% loading. The peak heat release rate (pHRR) and total smoke production (TSP) were reduced by 54.8 and 46.9%, respectively. The integrated multi-element system resulted in an expanded and reinforced char layer. This study proposes a simple and precise method for controlling the structure of MOF-carbohydrate hybrids through competition between chemical reactions.

[Distribution Prediction of Soil Heavy Metals Based on Remote Sensing Temporal-Spatial-Spectral Features and Random Forest Model].

Obtaining soil heavy metal content characteristics and spatial distribution is crucial for preventing soil pollution and formulating environmental protection policies. We collected 304 surface soil samples (0-20 cm) in the Changqing district. At the same time, the spectral, temporal, and spatial features of soil heavy metals were derived from multi-remote sensing data; the temporal-spatial-spectral features closely related to soil heavy metals were selected via correlation analysis and used as input independent variables. The measured soil arsenic (As) content was used as the dependent variable to establish a spatial prediction model based on the random forest (RF) algorithm. The results showed the following:the As content in the soils exceeded the background value by 43.17% but did not exceed the risk screening values and intervention values, indicating slight heavy metal pollution in the soil. The accuracy ranking of the spatial prediction models with one feature type from high to low was spatial features (ratio of performance to inter-quartile range (RPIQ)=3.87)>temporal features (RPIQ=2.57)>spectral features (RPIQ=2.50). The spatial features were the most informative for predicting soil heavy metals. The models using temporal-spatial, temporal-spectral, and spatial-spectral features were superior to those using only one feature type, and the RPIQ values were 4.81, 4.21, and 4.70, respectively. The RF model with temporal-spatial-spectral features achieved the highest spatial prediction accuracy (R2=0.90; root mean square error (RMSE)=0.77; RPIQ=5.68). The As content decreased from the northwest to the southeast due to Yellow River erosion and industrial activities. The spatial prediction of soil heavy metals incorporating remote sensing temporal-spatial-spectral features and the random forest model provides effective support for soil pollution prevention and environmental risk control.

Efficacy and toxic action of the natural product natamycin against Sclerotinia sclerotiorum.

BACKGROUND: Sclerotinia stem rot caused by Sclerotinia sclerotiorum seriously endangers oilseed rape production worldwide, and the occurrence of fungicide-resistant mutants of S. sclerotiorum leads to control decline. Thus, it is critical to explore new green substitutes with different action mechanisms and high antifungal activity. Herein, the activity and the action mechanism of natamycin against S. sclerotiorum were evaluated. RESULTS: Natamycin showed potent inhibition on the mycelial growth of S. sclerotiorum, and half-maximal effective concentration (EC50 ) values against 103 S. sclerotiorum strains ranged from 0.53 to 4.04 µg/mL (mean 1.44 µg/mL). Natamycin also exhibited high efficacy against both carbendazim- and dimethachlone-resistant strains of S. sclerotiorum on detached oilseed rape leaves. No cross-resistance was detected between natamycin and carbendazim. Natamycin markedly disrupted hyphal form, sclerotia formation, integrity of the cell membrane, and reduced the content of oxalic acid and ergosterol, whereas it increased the reactive oxygen species (ROS) and malondialdehyde content. Interestingly, exogenous addition of ergosterol could reduce the inhibition of natamycin against S. sclerotiorum. Importantly, natamycin significantly inhibited expression of the Cyp51 gene, which is contrary to results for the triazole fungicide flusilazole, indicating a different action mechanism from triazole fungicides. CONCLUSION: Natamycin is a promising effective candidate for the resistance management of S. sclerotiorum. © 2023 Society of Chemical Industry.

Study on the Antifungal Activity and Potential Mechanism of Natamycin against Colletotrichum fructicola.

In this investigation, the antifungal activity, its influence on the quality of apples, and the molecular mechanism of natamycin against Colletotrichum fructicola were systematically explored. Our findings indicated that natamycin showed significant inhibition against C. fructicola. Moreover, it efficaciously maintained the apple quality by modulating the physicochemical index. Research on the antifungal mechanism showed that natamycin altered the mycelial microstructure, disrupted the plasma membrane integrality, and decreased the ergosterol content of C. fructicola. Interestingly, the exogenous addition of ergosterol weakened the antifungal activity of natamycin. Importantly, natamycin markedly inhibited the expression of Cyp51A and Cyp51B genes in C. fructicola, which was contrary to the results obtained after treatment with triazole fungicide flusilazole. All these results exhibited sufficient proof that natamycin had enormous potential to be conducive as a promising biopreservative against C. fructicola on apples, and these findings will advance our knowledge on the mechanism of natamycin against pathogenic fungi.

SFPFusion: An Improved Vision Transformer Combining Super Feature Attention and Wavelet-Guided Pooling for Infrared and Visible Images Fusion.

The infrared and visible image fusion task aims to generate a single image that preserves complementary features and reduces redundant information from different modalities. Although convolutional neural networks (CNNs) can effectively extract local features and obtain better fusion performance, the size of the receptive field limits its feature extraction ability. Thus, the Transformer architecture has gradually become mainstream to extract global features. However, current Transformer-based fusion methods ignore the enhancement of details, which is important to image fusion tasks and other downstream vision tasks. To this end, a new super feature attention mechanism and the wavelet-guided pooling operation are applied to the fusion network to form a novel fusion network, termed SFPFusion. Specifically, super feature attention is able to establish long-range dependencies of images and to fully extract global features. The extracted global features are processed by wavelet-guided pooling to fully extract multi-scale base information and to enhance the detail features. With the powerful representation ability, only simple fusion strategies are utilized to achieve better fusion performance. The superiority of our method compared with other state-of-the-art methods is demonstrated in qualitative and quantitative experiments on multiple image fusion benchmarks.

Vegetation structural shift tells environmental changes on the Tibetan Plateau over 40 years.

Structural information of grassland changes on the Tibetan Plateau is essential for understanding alterations in critical ecosystem functioning and their underlying drivers that may reflect environmental changes. However, such information at the regional scale is still lacking due to methodological limitations. Beyond remote sensing indicators only recognizing vegetation productivity, we utilized multivariate data fusion and deep learning to characterize formation-based plant community structure in alpine grasslands at the regional scale of the Tibetan Plateau for the first time and compared it with the earlier version of Vegetation Map of China for historical changes. Over the past 40 years, we revealed that (1) the proportion of alpine meadows in alpine grasslands increased from 50% to 69%, well-reflecting the warming and wetting trend; (2) dominances of Kobresia pygmaea and Stipa purpurea formations in alpine meadows and steppes were strengthened to 76% and 92%, respectively; (3) the climate factor mainly drove the distribution of Stipa purpurea formation, but not the recent distribution of Kobresia pygmaea formation that was likely shaped by human activities. Therefore, the underlying mechanisms of grassland changes over the past 40 years were considered to be formation dependent. Overall, the first exploration for structural information of plant community changes in this study not only provides a new perspective to understand drivers of grassland changes and their spatial heterogeneity at the regional scale of the Tibetan Plateau, but also innovates large-scale vegetation study paradigm.

Designed synthesis of prussian blue@Cu-doped zinc phosphate nanocomposites for chemo/chemodynamic/photothermal combined cancer therapy.

Designing a multifunctional nanoplatform that combines multiple treatments has emerged as an innovative cancer treatment strategy. A simple and clear route is put forward to develop Cu2+-doped zinc phosphate coated prussian blue nanoparticles (designated as PB@Cu2+/ZnP NPs) integrating tri-modal therapy (chemo, chemodynamic and photothermal therapy) for maximizing anti-tumor efficacy. The obtained PB@Cu2+/ZnP NPs possess drug loading capacity due to the mesoporous structure present in the Cu2+-doped ZnP shell. In addition, the Cu2+-doped ZnP shell can gradually degrade in response to the mildly acidic tumor microenvironment to release DOX and Cu2+, where the released drug plays the role of chemotherapy agent and the Cu2+ can react with intracellular glutathione to achieve a Cu-mediated Fenton-like reaction for chemodynamic therapy. Moreover, under laser irradiation, the heat garnered by the photothermal conversion of PB can be applied for photothermal therapy and enhance the generation of toxic ˙OH as well as the amount of DOX released, further boosting chemo- and chemodynamic therapy to realize a combined therapy. Importantly, the PB@Cu2+/ZnP NPs effectively limit the growth of tumors via the coordinated action of chemo/chemodynamic/photothermal therapy and no noticeable systematic toxicity can be found in mice. Taken together, the PB@Cu2+/ZnP NPs can act as a prospective therapeutic nanoplatform for multi-modal therapy of tumors.

Interleukin-37 ameliorates atherosclerosis by regulating autophagy-mediated endothelial cell apoptosis and inflammation.

Atherosclerosis is a lipid-driven chronic inflammatory disease. Endothelial dysfunction is the initiating factor of atherosclerosis. Although much work has been done on the antiatherosclerotic effects of interleukin-37 (IL-37), the exact mechanism is still not fully understood. The aim of this study was to investigate whether IL-37 attenuates atherosclerosis by protecting endothelial cells and to confirm whether autophagy plays a role in this effect. In apolipoprotein E knockout (ApoE-/-) mice fed with a high fat diet, IL-37 treatment significantly attenuated progression of atherosclerotic plaques, reduced endothelial cell apoptosis and inflammasome activation. Human umbilical vein endothelial cells (HUVECs) were treated with oxidized low-density lipoprotein (ox-LDL) to establish an endothelial dysfunction model. We observed that IL-37 alleviated ox-LDL-induced endothelial cell inflammation and dysfunction, as evidenced by decreased nod-like receptor pyrin domain-containing 3 (NLRP3) inflammasome activation, ROS production, apoptosis rate and secretion of inflammatory cytokines IL-1ß and TNF-α. Furthermore, IL-37 could activate autophagy in endothelial cells, which is characterized by the upregulation of LC3II/LC3I, the downregulation of p62 and an increase in autophagosomes. The autophagy inhibitor 3-Methyladenine (3-MA) dramatically reversed the promotion of autophagy and the protective effect of IL-37 against endothelial injury. Our data illustrate that IL-37 alleviated inflammation and apoptosis of atherosclerotic endothelial cells by enhancing autophagy. The current study provides new insights and promising therapeutic strategies for atherosclerosis.

MFR-DTA: a multi-functional and robust model for predicting drug-target binding affinity and region.

MOTIVATION: Recently, deep learning has become the mainstream methodology for drug-target binding affinity prediction. However, two deficiencies of the existing methods restrict their practical applications. On the one hand, most existing methods ignore the individual information of sequence elements, resulting in poor sequence feature representations. On the other hand, without prior biological knowledge, the prediction of drug-target binding regions based on attention weights of a deep neural network could be difficult to verify, which may bring adverse interference to biological researchers. RESULTS: We propose a novel Multi-Functional and Robust Drug-Target binding Affinity prediction (MFR-DTA) method to address the above issues. Specifically, we design a new biological sequence feature extraction block, namely BioMLP, that assists the model in extracting individual features of sequence elements. Then, we propose a new Elem-feature fusion block to refine the extracted features. After that, we construct a Mix-Decoder block that extracts drug-target interaction information and predicts their binding regions simultaneously. Last, we evaluate MFR-DTA on two benchmarks consistently with the existing methods and propose a new dataset, sc-PDB, to better measure the accuracy of binding region prediction. We also visualize some samples to demonstrate the locations of their binding sites and the predicted multi-scale interaction regions. The proposed method achieves excellent performance on these datasets, demonstrating its merits and superiority over the state-of-the-art methods. AVAILABILITY AND IMPLEMENTATION: https://github.com/JU-HuaY/MFR.

CPInformer for Efficient and Robust Compound-Protein Interaction Prediction.

Recently, deep learning has become the mainstream methodology for Compound-Protein Interaction (CPI) prediction. However, the existing compound-protein feature extraction methods have some issues that limit their performance. First, graph networks are widely used for structural compound feature extraction, but the chemical properties of a compound depend on functional groups rather than graphic structure. Besides, the existing methods lack capabilities in extracting rich and discriminative protein features. Last, the compound-protein features are usually simply combined for CPI prediction, without considering information redundancy and effective feature mining. To address the above issues, we propose a novel CPInformer method. Specifically, we extract heterogeneous compound features, including structural graph features and functional class fingerprints, to reduce prediction errors caused by similar structural compounds. Then, we combine local and global features using dense connections to obtain multi-scale protein features. Last, we apply ProbSparse self-attention to protein features, under the guidance of compound features, to eliminate information redundancy, and to improve the accuracy of CPInformer. More importantly, the proposed method identifies the activated local regions that link a CPI, providing a good visualisation for the CPI state. The results obtained on five benchmarks demonstrate the merits and superiority of CPInformer over the state-of-the-art approaches.

Dapagliflozin Alleviates Coxsackievirus B3-induced Acute Viral Myocarditis by Regulating the Macrophage Polarization Through Stat3-related Pathways.

Viral myocarditis (VMC), which is most prevalently caused by Coxsackievirus B3 (CVB3) infection, is a serious clinical condition characterized by cardiac inflammation. Dapagliflozin, a kind of sodium glucose co-transporters 2(SGLT-2) inhibitor, exhibited protective effects on plenty of inflammatory diseases, while its effect on viral myocarditis has not been studied. Recently, we found the protective effect of dapagliflozin on VMC. After CVB3 infection, dapagliflozin and STATTIC (a kind of stat3 inhibitor) were given to Balb/c male mice for 8 days, and then the severity of myocarditis was assessed. Our results indicated that dapagliflozin significantly alleviated the severity of viral myocarditis, elevated the survival rate, and ameliorated cardiac function. Besides, dapagliflozin can decrease the level of pro-inflammatory cytokines including IL-1ß, IL-6, and TNF-α. Furthermore, dapagliflozin can inhibit macrophages differentiate to classically activated macrophages (M1) in cardiac tissue and activate the Stat3 signal pathway which is reported to promote polarization of the alternatively activated macrophage (M2). And STATTIC can reverse these changes caused by dapagliflozin. In conclusion, we found that dapagliflozin treatment increased anti-inflammatory macrophage polarization and reduced cardiac injury following VMC via activating Stat3 signal pathway.

Value of M2BP in predicting in-stent restenosis in patients after coronary drug-eluting stent implantation.

OBJECTIVE: We evaluated the association between plasma levels of mac-2 binding protein (M2BP) with the risk of in-stent restenosis (ISR) after percutaneous coronary intervention (PCI). METHODS: Plasma M2BP levels were compared between 258 patients who experienced ISR at 12-months post-PCI and 258 patients, matched for age and sex, without angiographic evidence of ISR. RESULTS: The plasma M2BP level was significantly higher in the ISR than in the non-ISR group. On multivariate analysis, adjusted for potential clinical, biochemical, and angiography characteristics, M2BP remained as an independent significant predictor of ISR. CONCLUSIONS: M2BP may be an important predictive biomarker of ISR and may be useful in identifying at-risk patients.

Exonic Circular RNAs Are Involved in Arabidopsis Immune Response Against Bacterial and Fungal Pathogens and Function Synergistically with Corresponding Linear RNAs.

Circular RNAs (circRNAs) are a group of covalently closed RNAs, and their biological function is largely unknown. In this study, we focused on circRNAs that are generated from exon back-splicing (exonic circRNAs). The linear RNA counterparts encode functional proteins so that we can compare and investigate the relationship between circular and linear RNAs. We compared circRNA expression profiles between untreated and Pseudomonas syringae-infected Arabidopsis and identified and experimentally validated differentially expressed exonic circRNAs by multiple approaches. We found that exonic circRNAs are preferentially enriched in biological processes that associate with biotic and abiotic stress responses. We discovered that circR194 and circR4022 are involved in plant response against P. syringae infection, whereas circR11208 is involved in response against Botrytis cinerea infection. Intriguingly, our results indicate that these exonic circRNAs function synergistically with their corresponding linear RNAs. Furthermore, circR4022 and circR11208 also play substantial roles in Arabidopsis tolerance to salt stress. This study extends our understanding of the molecular functions of plant circRNAs.

2-Hydroxy-phytanoyl-CoA lyase (AtHPCL) is involved in phytol metabolism in Arabidopsis.

During chlorophyll degradation, large amounts of the isoprenoid alcohol phytol are released. The pathway of phytol catabolism has been studied in humans, because chlorophyll is part of the human diet, but little is known for plants. In humans, phytanoyl-CoA derived from phytol is degraded via α-oxidation by phytanoyl-CoA hydroxylase (PAHX) and 2-hydroxy-phytanoyl-CoA lyase (HPCL). Arabidopsis contains two sequences homologous to the human proteins AtPAHX and AtHPCL. Insertional mutants of Arabidopsis (pahx, hpcl) were grown under N deprivation to stimulate chlorophyll breakdown or supplemented with phytol to increase the endogenous amount of phytol. During N deprivation, chlorophyll, phytol, phytenal, upstream metabolites of phytol breakdown, and tocopherol and fatty acid phytyl esters, alternative phytol-derived lipids, accumulated in pahx and hpcl mutants, in line with the scenario that the mutations interfere with phytol degradation. AtHPCL was localized to the peroxisomes. Expression analysis of the AtHPCL sequence in the yeast Δpxp1 or Δmpo1 mutants followed by supplementation with 2-hydroxy-palmitic acid and enzyme assays of peroxisomal proteins from Col-0 and hpcl plants with 2-hydroxy-stearoyl-CoA revealed that AtHPCL harbors 2-hydroxy-acyl-CoA lyase activity. The α-dioxygenases αDOX1 and αDOX2 are involved in α-oxidation of fatty acids and could be involved in an alternative pathway of phytol degradation. However, phytol-related lipids in the αdox1, αdox2, or αdox1 αdox2 mutants were not altered compared with Col-0, indicating that αDOX1 and αDOX2 are not involved in phytol degradation. These results demonstrate that phytol degradation in Arabidopsis involves α-oxidation by AtPAHX and AtHPCL, but that it is independent of αDOX1/αDOX2.

Prediction of vegetation phenology with atmospheric reanalysis over semiarid grasslands in Inner Mongolia.

Vegetation phenology is a sensitive indicator of climate change and vegetation growth. In the present study, two phenological phases with respect to vegetation growth at the initial and mature stages, namely, the start of the season (SOS) and the peak of the season (POS), were estimated from a satellite-derived normalized difference vegetation index (NDVI) dataset over a long-term period of 32 years (1983 to 2014) and used to explore their responses to atmospheric variables, including air temperature, precipitation, solar radiation, wind speed and soil moisture. First, the forward feature selection method was used to determine whether each independent variable was linear or nonlinear to the SOS and POS. In addition, a generalized additive model (GAM) was used to analyze the correlation between the phenological phases and each independent variable at different temporal scales. The results show that soil moisture and precipitation are linearly correlated with the SOS, whereas the other variables are nonlinearly correlated. Meanwhile, soil moisture, wind speed and solar radiation are found to be nonlinearly correlated with the POS. However, air temperature and precipitation reveal a significant negative correlation with the POS. Furthermore, it was concluded that the aforementioned independent variables from the previous year could contribute to approximately 63%-85% of the SOS variations in the present year, whereas the atmospheric variables from April to June could contribute to approximately 70%-85% of the POS variations in the same year. Finally, the SOS and POS predicted by the GAM exhibit significant agreement with those derived from the satellite NDVI dataset, with the root mean square error of approximately 3 to 5 days.

A novel few-shot learning based multi-modality fusion model for COVID-19 rumor detection from online social media.

BACKGROUND: Rumor detection is a popular research topic in natural language processing and data mining. Since the outbreak of COVID-19, related rumors have been widely posted and spread on online social media, which have seriously affected people's daily lives, national economy, social stability, etc. It is both theoretically and practically essential to detect and refute COVID-19 rumors fast and effectively. As COVID-19 was an emergent event that was outbreaking drastically, the related rumor instances were very scarce and distinct at its early stage. This makes the detection task a typical few-shot learning problem. However, traditional rumor detection techniques focused on detecting existed events with enough training instances, so that they fail to detect emergent events such as COVID-19. Therefore, developing a new few-shot rumor detection framework has become critical and emergent to prevent outbreaking rumors at early stages. METHODS: This article focuses on few-shot rumor detection, especially for detecting COVID-19 rumors from Sina Weibo with only a minimal number of labeled instances. We contribute a Sina Weibo COVID-19 rumor dataset for few-shot rumor detection and propose a few-shot learning-based multi-modality fusion model for few-shot rumor detection. A full microblog consists of the source post and corresponding comments, which are considered as two modalities and fused with the meta-learning methods. RESULTS: Experiments of few-shot rumor detection on the collected Weibo dataset and the PHEME public dataset have shown significant improvement and generality of the proposed model.

Selectfluor-Promoted Intramolecular N-S Bond Formation of α-Carbamoyl Ketene Dithioacetals in the Presence of Water: Synthesis of Multifunctionalized Isothiazolones.

A practical and efficient protocol toward fully substituted isothiazolones through Selectfluor-mediated intramolecular oxidative annulation of α-carbamoyl ketene dithioacetals has been developed in the presence of H2O and metal-free conditions. Notably, the experimental results reveal that H2O was crucial to the formation of new N-S bonds and the elimination of alkyl group from the sulfur atom. This protocol provides readily prepared substrates and possesses good functional group tolerance, mild reaction conditions, and operational simplicity, which provides potential access to applications in the pharmaceutical chemistry.

Why can deep convolutional neural networks improve protein fold recognition? A visual explanation by interpretation.

As an essential task in protein structure and function prediction, protein fold recognition has attracted increasing attention. The majority of the existing machine learning-based protein fold recognition approaches strongly rely on handcrafted features, which depict the characteristics of different protein folds; however, effective feature extraction methods still represent the bottleneck for further performance improvement of protein fold recognition. As a powerful feature extractor, deep convolutional neural network (DCNN) can automatically extract discriminative features for fold recognition without human intervention, which has demonstrated an impressive performance on protein fold recognition. Despite the encouraging progress, DCNN often acts as a black box, and as such, it is challenging for users to understand what really happens in DCNN and why it works well for protein fold recognition. In this study, we explore the intrinsic mechanism of DCNN and explain why it works for protein fold recognition using a visual explanation technique. More specifically, we first trained a VGGNet-based DCNN model, termed VGGNet-FE, which can extract fold-specific features from the predicted protein residue-residue contact map for protein fold recognition. Subsequently, based on the trained VGGNet-FE, we implemented a new contact-assisted predictor, termed VGGfold, for protein fold recognition; we then visualized what features were extracted by each of the convolutional layers in VGGNet-FE using a deconvolution technique. Furthermore, we visualized the high-level semantic information, termed fold-discriminative region, of a predicted contact map from the localization map obtained from the last convolutional layer of VGGNet-FE. It is visually confirmed that VGGNet-FE could effectively extract distinct fold-discriminative regions for different types of protein folds, thereby accounting for the improved performance of VGGfold for protein fold recognition. In summary, this study is of great significance for both understanding the working principle of DCNNs in protein fold recognition and exploring the relationship between the predicted protein contact map and protein tertiary structure. This proposed visualization method is flexible and applicable to address other DCNN-based bioinformatics and computational biology questions. The online web server of VGGfold is freely available at http://csbio.njust.edu.cn/bioinf/vggfold/.