SecureNet: Proactive intellectual property protection and model security defense for DNNs based on backdoor learning.

With the widespread application of deep neural networks (DNNs), the risk of privacy breaches against DNN models is constantly on the rise, resulting in an increasing need for intellectual property (IP) protection for such models. Although neural network watermarking techniques are widely used to safeguard the IP of DNNs, they can only achieve passive protection and cannot actively prevent unauthorized users from illicit use or embezzlement of the trained DNN models. Therefore, the development of proactive protection techniques to prevent IP infringement is imperative. To this end, we propose SecureNet, a key-based access license framework for DNN models. The proposed approach involves injecting license keys into the model through backdoor learning, enabling correct model functionality only when the appropriate license key is included in the input. To ensure the reusability of DNN models, we also propose a license key replacement algorithm. In addition, based on SecureNet, we designed defense mechanisms against adversarial attacks and backdoor attacks, respectively. Furthermore, we introduce a fine-grained authorization method that enables flexible granting of model permissions to different users. We have designed four license-key schemes with different privileges, tailored to various scenarios. We evaluated SecureNet on five benchmark datasets including MNIST, Cifar10, Cifar100, FaceScrub, and CelebA, and assessed its performance on six classic DNN models: LeNet-5, VGG16, ResNet18, ResNet101, NFNet-F5, and MobileNetV3. The results demonstrate that our approach outperforms the state-of-the-art model parameter encryption methods by at least 95% in terms of computational efficiency. Additionally, it provides effective defense against adversarial attacks and backdoor attacks without compromising the model's overall performance.

The causal relationship between gut microbiota and type 2 diabetes: a two-sample Mendelian randomized study.

Background: Type 2 diabetes mellitus (T2DM) is a commonly observed metabolic anomaly globally, and as of the present time, there's no recognized solution. There is an increasing body of evidence from numerous observational studies indicating a significant correlation between gut flora and metabolic disease progression, particularly in relation to T2DM. Despite this, the direct impact of gut microbiota on T2DM isn't fully understood yet. Methods: The summary statistical figures for intestinal microbiota were sourced from the MiBioGen consortium, while the summary statistical data for T2DM were gathered from the Genome-Wide Association Studies (GWAS) database. These datasets were used to execute a two-sample Mendelian randomization (MR) investigation. The Inverse Variance Weighted (IVW), Maximum Likelihood, MR-Egger, Weighted Median, and Weighted Models strategies were employed to assess the impact of gut microbiota on T2DM. Findings were primarily obtained using the IVW technique. Techniques like MR-Egger were employed to identify the occurrence of horizontal pleiotropy among instrumental variables. Meanwhile, Cochran's Q statistical measures were utilized to assess the variability or heterogeneity within these instrumental variables. Results: The outcomes from the IVW analysis demonstrated that the genus Alistipes (OR = 0.998, 95% confidence interval: 0.996-1.000, and P = 0.038), genus Allisonella (OR = 0.998, 95% confidence interval: 0.997-0.999, P = 0.033), genus Flavonifractor (OR = 0.995, 95% confidence interval: 0.993-0.998, P = 3.78 × 10-3), and genus Haemophilus (OR = 0.995, 95% confidence interval: 0.993-0.998, P = 8.08 × 10-3) all acted as defense elements against type 2 diabetes. Family Clostridiaceae1 (OR = 1.003, 95% confidence interval: 1.001-1.005, P = 0.012), family Coriobacteriaceae (OR = 1.0025, 95% confidence interval: 1.000-1.005, P = 0.043), genus Actinomyces (OR = 1.003,95% confidence interval: 1.001-1.005, P = 4.38 × 10-3), genus Candidatus Soleaferrea (OR = 1.001,95% confidence interval: 1.000-1.002 P = 0.012) were risk factors for type 2 diabetes. False Discovery Rate correction was performed with finding that genus.Allisonella, genus.Alistipes, family Coriobacteriaceaeand T2DM no longer displayed a significant causal association. In addition, no significant heterogeneity or horizontal pleiotropy was found for instrumental variable. Conclusion: This MR study relies on genetic variation tools to confirm the causal effect of genus Flavonifractor, genus Haemophilus, family Clostridiaceae1, genus Actinomyces and genus Candidatus Soleaferrea on T2DM in the gut microbiome, providing new directions and strategies for the treatment and early screening of T2DM, which carries significant clinical relevance. To develop new biomarkers and better understand targeted prevention strategies for T2DM, further comprehensive investigations are required into the protective and detrimental mechanisms exerted by these five genera against T2DM.

Water footprint and virtual water trade analysis in water-rich basins: Case of the Chaohu Lake Basin in China.

Water footprints and virtual water are widely used as essential tools for water use and conservation analysis of basins worldwide. Despite the importance of water-rich basins as the main force for water-saving, water use analysis has been mainly for water-scarce basins rather than water-rich basins in the existing literature. To fill the gap, in this paper, we investigate the water footprint and virtual water trade in a water-rich basin, namely the Chaohu Lake Basin in China, from 2007 to 2017 using input-output analysis. The results show that: (1) Water use efficiency in the Chaohu Lake Basin was significantly improved. The overall trend of the water intensity was declining, decreasing by 10.21 % in 2017 versus 2012; (2) The internal and external water footprints showed an upward trend, and the growth rate of total water footprint was 36.66 %; (3) The basin was a net virtual water exporter, but the net export flows of virtual water has decreased significantly. The virtual water net export flow decreased by 0.12 billion m3 in 2017 versus 2012; (4) Water resources in the basin were mainly used locally, and its supply to other provinces was minimal. Compared with some water-scarce basins such as the Heihe River Basin and Haihe River Basin, the Chaohu Lake Basin shows significant gaps in the virtual water export flow per capita and behaves differently in the proportion of virtual water transfer. Based on the above findings, we conclude with some guidance and implications for local governments and policymakers.

Effect of Land Expropriation on Land-Lost Farmers' Health: Empirical Evidence from Rural China.

With rapid urbanization and industry development, China has witnessed substantial land acquisition. Using the rural household survey data, this paper examines the impact of land expropriation on land-lost farmers' self-reported health with the ordered probit model and investigates the possible mechanisms. The results show that the land expropriation puts higher health risks over those land-lost farmers and the health status of land-lost farmers is significantly worse than that of those with land. Land expropriation has a negative impact on the land-lost farmer's health through income effects and psychological effects. The health status of land-lost farmers can be enhanced through amending current land requisition policies, increasing the amount of compensation, improving the earning capacity of land-lost farmers and strengthening mental health education.

Collagen Features of Dermatofibrosarcoma Protuberans Skin Base on Multiphoton Microscopy.

Dermatofibrosarcoma protuberans is a rare, low-grade skin fibroblastic tumor which tends to recur locally due to its high misdiagnosis. Dermatofibrosarcoma protuberans usually spreads through the intracutaneous and subcutaneous layers into the deep dermis layer in which the main component is collagen. Therefore, alterations in collagen shape and content are important for accurate diagnosis of dermatofibrosarcoma protuberans. In this study, multiphoton microscopy was employed to observe normal human skin and dermatofibrosarcoma protuberans skin. Then, a centerline based on an algorithm that skeletonizes a binary image of fibers was applied to analyze collagen shapes in 2 types of skin. Then, collagen content, including intensity and density, was quantitatively obtained to demonstrate differences between the 2 skin types. Results indicate that collagen shape and density can be considered as auxiliary diagnostic parameters to improve the accuracy of dermatofibrosarcoma protuberans diagnosis.

Flux balance analysis predicts Warburg-like effects of mouse hepatocyte deficient in miR-122a.

The liver is a vital organ involving in various major metabolic functions in human body. MicroRNA-122 (miR-122) plays an important role in the regulation of liver metabolism, but its intrinsic physiological functions require further clarification. This study integrated the genome-scale metabolic model of hepatocytes and mouse experimental data with germline deletion of Mir122a (Mir122a-/-) to infer Warburg-like effects. Elevated expression of MiR-122a target genes in Mir122a-/-mice, especially those encoding for metabolic enzymes, was applied to analyze the flux distributions of the genome-scale metabolic model in normal and deficient states. By definition of the similarity ratio, we compared the flux fold change of the genome-scale metabolic model computational results and metabolomic profiling data measured through a liquid-chromatography with mass spectrometer, respectively, for hepatocytes of 2-month-old mice in normal and deficient states. The Ddc gene demonstrated the highest similarity ratio of 95% to the biological hypothesis of the Warburg effect, and similarity of 75% to the experimental observation. We also used 2, 6, and 11 months of mir-122 knockout mice liver cell to examined the expression pattern of DDC in the knockout mice livers to show upregulated profiles of DDC from the data. Furthermore, through a bioinformatics (LINCS program) prediction, BTK inhibitors and withaferin A could downregulate DDC expression, suggesting that such drugs could potentially alter the early events of metabolomics of liver cancer cells.

Self-compatibility of 'Zaoguan' (Pyrus bretschneideri Rehd.) is associated with style-part mutations.

The pear cultivar 'Zaoguan' (S(4)S(34)) is the a self-compatible descendant of 'Yali' (S(21)S(34)) × 'Qingyun'(S(4)S(9)). Two self-incompatible cultivars 'Xinya' and 'Yaqing', also S-genotyped as S(4)S(34) for the S-RNase gene, were used as controls. Field pollination data revealed that 'Zaoguan' displayed SC, whereas 'Xinya' and 'Yaqing' showed self-incompatibility (SI) upon self-pollination. Reciprocal pollinations between the varieties showed that most of the 'Zaoguan' flowers pollinated with 'Xinya' or 'Yaqing' pollen set fruits but that few of the 'Xinya' or 'Yaqing' flowers set fruit when pollinated with 'Zaoguan' pollen. The pollen performance was monitored with fluorescence microscopy, and we observed that 'Zaoguan' accepted self-pollen as well as 'Xinya' or 'Yaqing' pollen, whereas 'Xinya' or 'Yaqing' rejected self-pollen and 'Zaoguan' pollen. The S(34)-RNase but not the S(4)-RNase could be detected in all selfed progeny of 'Zaoguan'. Comparisons of the 2D-PAGE profiles of the stylar extracts from the three cultivars showed that the S(4)-RNase protein expressed normally, but the S(34)-RNase of 'Zaoguan' was not found. Thus, we concluded that the stylar S(34) products were defective in 'Zaoguan' and that the S (4)-allele functioned normally. The nucleotide sequences of the S(4)- and S(34)-RNase of 'Zaoguan' showed no differences from those of 'Xinya' or 'Yaqing', and they transcribed normally. These results indicate that SC in 'Zaoguan' was due to the loss of the S(34)-RNase caused by unknown post-transcriptional factors.

Reciprocal regulation of Ca²+-activated outward K+ channels of Pyrus pyrifolia pollen by heme and carbon monoxide.

• The regulation of plant potassium (K+) channels has been extensively studied in various systems. However, the mechanism of their regulation in the pollen tube is unclear. • In this study, the effects of heme and carbon monoxide (CO) on the outward K+ (K+(out)) channel in pear (Pyrus pyrifolia) pollen tube protoplasts were characterized using a patch-clamp technique. • Heme (1 µM) decreased the probability of K+(out) channel opening without affecting the unitary conductance, but this inhibition disappeared when heme was co-applied with 10 µM intracellular free Ca²+. Conversely, exposure to heme in the presence of NADPH increased channel activity. However, with tin protoporphyrin IX treatment, which inhibits hemeoxygenase activity, the inhibition of the K+(out) channel by heme occurred even in the presence of NADPH. CO, a product of heme catabolism by hemeoxygenase, activates the K+(out) channel in pollen tube protoplasts in a dose-dependent manner. The current induced by CO was inhibited by the K+ channel inhibitor tetraethylammonium. • These data indicate a role of heme and CO in reciprocal regulation of the K+(out) channel in pear pollen tubes.

Self-compatibility of 'Katy' apricot (Prunus armeniaca L.) is associated with pollen-part mutations.

Apricot (Prunus armeniaca L.) cultivars originated in China display a typical S-RNase-based gametophytic self-incompatibility (GSI). 'Katy', a natural self-compatible cultivar belonging to the European ecotype group, was used as a useful material for breeding new cultivars with high frequency of self-compatibility by hybridizing with Chinese native cultivars. In this work, the pollen-S genes (S-haplotype-specific F-box gene, or SFB gene) of 'Katy' were first identified as SFB1 and SFB (8), and the S-genotype was determined as S1 S8. Genetic analysis of 'Katy' progenies under controlled pollination revealed that the stylar S1-RNase and S8-RNase have a normal function in rejecting wild-type pollen with the same S-haplotype, while the pollen grains carrying either the SFB1 or the SFB8 gene are both able to overcome the incompatibility barrier. However, the observed segregation ratios of the S-genotype did not fit the expected ratios under the assumption that the pollen-part mutations are linked to the S-locus. Moreover, alterations in the SFB1 and SFB8 genes and pollen-S duplications were not detected. These results indicated that the breakdown of SI in 'Katy' occurred in pollen, and other factors not linked to the S-locus, which caused a loss of pollen S-activity. These findings support a hypothesis that modifying factors other than the S-locus are required for GSI in apricot.

S-RNase disrupts tip-localized reactive oxygen species and induces nuclear DNA degradation in incompatible pollen tubes of Pyrus pyrifolia.

Pear (Pyrus pyrifolia L.) has an S-RNase-based gametophytic self-incompatibility (SI) mechanism, and S-RNase has also been implicated in the rejection of self-pollen and genetically identical pollen. However, RNA degradation might be only the beginning of the SI response, not the end. Recent in vitro studies suggest that S-RNase triggers mitochondrial alteration and DNA degradation in the incompatible pollen tube of Pyrus pyrifolia, and it seems that a relationship exists between self S-RNase, actin depolymerization and DNA degradation. To further uncover the SI response in pear, the relationship between self S-RNase and tip-localized reactive oxygen species (ROS) was evaluated. Our results show that S-RNase specifically disrupted tip-localized ROS of incompatible pollen tubes via arrest of ROS formation in mitochondria and cell walls. The mitochondrial ROS disruption was related to mitochondrial alteration, whereas cell wall ROS disruption was related to a decrease in NADPH. Tip-localized ROS disruption not only decreased the Ca(2+) current and depolymerized the actin cytoskeleton, but it also induced nuclear DNA degradation. These results indicate that tip-localized ROS disruption occurs in Pyrus pyrifolia SI. Importantly, we demonstrated nuclear DNA degradation in the incompatible pollen tube after pollination in vivo. This result validates our in vitro system in vivo.

S-RNase triggers mitochondrial alteration and DNA degradation in the incompatible pollen tube of Pyrus pyrifolia in vitro.

Pear (Pyrus pyrifolia L.) has a S-RNase-based gametophytic self-incompatibility (SI) mechanism, and S-RNase has also been implicated in the rejection of self-pollen and genetically identical pollen. No studies, however, have examined the extent of organelle alterations during the SI response in Pyrus pyrifolia. Consequently, this study focused on the alterations to mitochondria and nuclear DNA in incompatible pollen tubes of the pear. Methylthiazolyldiphenyl-tetrazolium bromide was used to evaluate the viability of pollen tubes under S-RNase challenge. The results showed that the viability of the control and compatible pollen tubes decreased slightly, but that of the incompatible pollen and pollen tubes began to decline at 30 min. The mitochondrial membrane potential (Delta psi(mit)) was also tested with rhodamine 123 30 min after SI challenge, and was shown to have collapsed in the incompatible pollen tubes after exposure to S-RNase. Western blotting 2 h after SI challenge confirmed that the Delta psi(mit) collapse induced leakage of cytochrome c into the cytosol. Swollen mitochondria were detected by transmission electron microscopy as early as 1 h after SI challenge and the degradation of nuclear DNA was observed by both 4,6-diamidino-2-phenylindole and transferase-mediated dUTP nick-end labeling. These diagnostic features of programmed cell death (PCD) suggested that PCD may specifically occur in incompatible pollen tubes.

Ubiquitination and degradation of Tal1/SCL are induced by notch signaling and depend on Skp2 and CHIP.

Notch signaling controls diverse eukaryotic differentiation processes in multiple cell types, thus demanding versatile tools with which Notch triggers downstream events. Ubiquitin-mediated proteolysis has previously been shown to be one such tool with which Notch regulates the turnover of the basic helix-loophelix transcription factor, E47. Here, we show that Notch signaling also accelerated the degradation of Tal1/SCL (T cell acute leukemia 1/stem cell leukemia) protein, a basic helix-loop-helix protein involved in the development of hematopoietic, vascular, and neuronal tissues. Notch-induced Tal1/SCL degradation was mediated by ubiquitination and proteasomes. The sequence responsible for Tal1 degradation was localized to a region in the C terminus of Tal1, which is evolutionarily conserved, thus suggesting a functional significance. Analogous to the situation for E47, Notch-induced Tal1/SCL degradation not only required Skp2, a substrate-binding subunit of SCF ubiquitin ligase complexes, but also relied on CHIP, a chaperone-binding protein with a ubiquitin ligase activity. In contrast to the fact that the N-terminal tetratricopeptide region (TPR) domain of CHIP is necessary and sufficient for E47 ubiquitination and degradation, CHIP promoted Tal1 degradation with both chaperone binding and ubiquitin ligase activities, which are mediated by its TPR domain and U box, respectively. Although the TPR domain was not involved in Tal1/SCL binding, it was required for enhancing its degradation. Likewise, the ubiquitin ligase activity of CHIP was dispensable for Tal1/SCL binding but essential for degradation. These findings provide both novel mechanistic insights into the operation of cullin-based ubiquitin ligase complexes and potential means by which Notch and Tal1/SCL regulate eukaryotic development.