This article has been retracted: please see Elsevier Policy on Article Withdrawal (https://www.elsevier.com/locate/withdrawalpolicy).This article has been retracted at the request of the Author.The corresponding author requested to modify the title of the article, as the authors thought the new name of the model applied in the research would be better aligned with the research focus and innovations in the article. Title modification is not allowed after the publication of the article. The authors insisted that it is crucial to modify the title and decided to retract the article. The journal has agreed that the authors may submit a new version of the manuscript to the journal for review and publication, if accepted by the Editor-in-Chief.

With the advancement of adversarial techniques for malicious code, malevolent attackers have propagated numerous malicious code variants through shell coding and code obfuscation. Addressing the current issues of insufficient accuracy and efficiency in malicious code classification methods based on deep learning, this paper introduces a detection strategy for malicious code, uniting Convolutional Neural Networks (CNNs) and Transformers. This approach utilizes deep neural architecture, incorporating a novel fusion module to reparametrize the structure, which mitigates memory access costs by eliminating residual connections within the network. Simultaneously, overparametrization during linear training time and significant kernel convolution techniques are employed to enhance network precision. In the data preprocessing stage, a pixel-based image size normalization algorithm and data augmentation techniques are utilized to remedy the loss of texture information in the malicious code image scaling process and class imbalance in the dataset, thereby enhancing essential feature expression and alleviating model overfitting. Empirical evidence substantiates this method has improved accuracy and the most recent malicious code detection technologies.

As malicious code attacks continue to evolve, attackers leverage techniques like packing and code obfuscation to generate numerous variants, challenging traditional detection methods. Addressing the limitations of current deep learning-based malicious code classification approaches in feature extraction and accuracy, this paper introduces an innovative RGB visualization detection method based on a hybrid multi-head attention mechanism. Initially, a feature representation method utilizing RGB images is introduced. This approach focuses on semantic relationships between a malware’s binary information, assembly details, and API data, generating images with richer textural information. This technique effectively uncovers the deep dependencies between the original and variant versions of malicious code, providing stronger support for subsequent classification tasks. Furthermore, to tackle the issues of malware encryption and obfuscation, a deep neural network framework is adopted, incorporating a modular design philosophy and integrating a multi-head attention mechanism. This design not only enhances the expressiveness of critical features but also helps the model better focus on key aspects of the malicious code, thereby improving classification accuracy. Through comparative experiments and in-depth analysis, the effectiveness and superiority of the proposed RGB visualization method and MSA-ResNet model in the field of malicious code variant classification are validated. The accuracy rates achieved on the Kaggle and DataCon datasets are 99.49

As malicious code countermeasures evolve, attackers have responded by generating numerous malicious code variants through shelling, code obfuscation, and similar strategies. Addressing the shortcomings of current deep learning-based malicious code classification methods-namely, their limited extraction accuracy and efficiency-this study offers a novel approach that integrates CNNs and Transformer for malicious code detection. This method leverages deep neural networks as its foundation and incorporates a unique fusion module for structural reparameterization. This innovative mechanism reduces memory access costs by eliminating jump connections within the network. Furthermore, the implementation of large kernel convolution techniques enhances network accuracy. Experimental results demonstrate that our proposed method consistently outperforms the latest malicious code detection techniques in terms of both accuracy and operational efficiency.

With the development of automated malware toolkits, cybersecurity faces evolving threats. Although visualization-based malware analysis has proven to be an effective method, existing approaches struggle with challenging malware samples due to alterations in the texture features of binary images during the visualization preprocessing stage, resulting in poor performance. Furthermore, to enhance classification accuracy, existing methods sacrifice prediction time by designing deeper neural network architectures. This paper proposes PAFE, a lightweight and visualization-based rapid malware classification method. It addresses the issue of texture feature variations in preprocessing through pixel-filling techniques and applies data augmentation to overcome the challenges of class imbalance in small sample datasets. PAFE combines multi-scale feature fusion and a channel attention mechanism, enhancing feature expression through modular design. Extensive experimental results demonstrate that PAFE outperforms the current state-of-the-art methods in both efficiency and effectiveness for malware variant classification, achieving an accuracy rate of 99.25 % with a prediction time of 10.04 ms.

In response to escalating cybersecurity threats, this study aims to develop a lightweight deep learning model for efficient and accurate detection and classification of malicious code. To achieve this goal, the study introduces TriCh-RepNet, a novel network architecture that balances performance and resource utilization. The methodology involves innovatively transforming malicious code representations into image channels through a three-channel mapping technique, thereby enhancing information richness and discriminatory power. By integrating the strengths of Convolutional Neural Networks (CNNs) and Transformers, TriCh-RepNet creates a streamlined framework that optimizes network connections, reduces memory access overhead, and boosts overall efficiency. Additionally, the combination of linear training time over-parameterization and large kernel convolution techniques minimizes the model's parameter count and computational load. The model achieves accuracy rates of 99.47% and 97.51% on the Kaggle and DataCon datasets, respectively. Compared to existing methodologies, this approach stands out in terms of performance, resource consumption, and versatility. It offers a viable and efficient solution, especially in resource-constrained or real-time environments. A novel and effective solution for malicious code detection and categorization.

This study presents a novel deep learning method called BiLSTM-TCN, which aims to enhance the proficiency of recognizing air target operational intent. The traditional air target operational intent method makes judgments based on single moment data, and thus cannot effectively capture feature information on temporal data. To tackle this issue, we select suitable features for air target intent feature set construction and then encode them into time-series features. We mine the potential feature information in the data by TCN. Meanwhile, we captured the prolonged correlations within the dataset by using BiLSTM. Through experimental validation, we have shown that BiLSTM-TCN surpasses current superior techniques regarding the accuracy of identifying airborne targets' intentions.

为应对不断升级的恶意代码变种,针对现有恶意代码分类方法对特征提取能力不足、分类准确率下降的问题,文章提出了基于双向时域卷积网络(Bidirectional Temporal Convolution Network,BiTCN)和池化融合(Double Layer Pooling,DLP)的恶意代码分类方法(BiTCN-DLP).首先,该方法融合恶意代码操作码和字节码特征以展现不同细节;然后,构建BiTCN模型充分利用特征的前后依赖关系,引入池化融合机制进一步挖掘恶意代码数据内部深层的依赖关系;最后,文章在Kaggle数据集上对模型进行验证,实验结果表明,基于BiTCN-DLP的恶意代码分类准确率可达99.54%,且具有较快的收敛速度和较低的分类误差,同时,文章通过对比实验和消融实验证明了该模型的有效性.

A massive proliferation of malware variants has posed serious and evolving threats to cybersecurity. Developing intelligent methods to cope with the situation is highly necessary due to the inefficiency of traditional methods. In this paper, a highly efficient, intelligent vision-based malware variants detection method was proposed. Firstly, a bilinear interpolation algorithm was utilized for malware image normalization, and data augmentation was used to resolve the issue of imbalanced malware data sets. Moreover, the paper improved the convolutional neural network (CNN) model by combining multi-scale feature fusion (MFF) and channel attention mechanism for more discriminative and robust feature extraction. Finally, we proposed a hyperparameter optimization algorithm based on the bat algorithm, referred to as HDBA, in order to overcome the disadvantage of the traditional hyperparameter optimization method based on manual adjustment. Experimental results indicated that our model can effectively and efficiently identify malware variants from real and daily networks, with better performance than state-of-the-art solutions.

Gentamicin is an important aminoglycoside antibiotic used for treatment of infections caused by Gram-negative bacteria. Although most of the biosynthetic pathway of gentamicin has been elucidated, a remaining intriguing question is how the intermediates JI-20A and JI-20B undergo a dideoxygenation to form gentamicin C complex. Here we show that the dideoxygenation process starts with GenP-catalyzed phosphorylation of JI-20A and JI-20Ba. The phosphorylated products are converted to C1a and C2a by concerted actions of two PLP (pyridoxal 5’-phosphate)-dependent enzymes: elimination of water and then phosphate by GenB3 and double bond migration by GenB4. Each of these reactions liberates an imine which hydrolyses to a ketone or aldehyde and is then re-aminated by GenB3 using an amino donor. Crystal structures of GenB3 and GenB4 have guided site-directed mutagenesis to reveal crucial residues for the enzymes’ functions. We propose catalytic mechanisms for GenB3 and GenB4, which shed new light on the already unrivalled catalytic versatility of PLP-dependent enzymes.

Endowing mesophilic microorganisms with high-temperature resistance is highly desirable for industrial microbial fermentation. Here, we report a cold-shock protein (CspL) that is an RNA chaperone protein from a lactate producing thermophile strain ( Bacillus coagulans 2–6), which is able to recombinantly confer strong high-temperature resistance to other microorganisms. Transgenic cspL expression massively enhanced high-temperature growth of Escherichia coli (a 2.4-fold biomass increase at 45 °C) and eukaryote Saccharomyces cerevisiae (a 2.6-fold biomass increase at 36 °C). Importantly, we also found that CspL promotes growth rates at normal temperatures. Mechanistically, bio-layer interferometry characterized CspL’s nucleotide-binding functions in vitro, while in vivo we used RNA-Seq and RIP-Seq to reveal CspL’s global effects on mRNA accumulation and CspL’s direct RNA binding targets, respectively. Thus, beyond establishing how a cold-shock protein chaperone provides high-temperature resistance, our study introduces a strategy that may facilitate industrial thermal fermentation.

Hygromycin B is an aminoglycoside antibiotic widely used in industry and biological research. However, most of its biosynthetic pathway has not been completely identified due to the immense difficulty in genetic manipulation of the producing strain. To address this problem, we developed an efficient system that combines clustered regularly interspaced short palindromic repeats (CRISPR)-Cas9-associated base editing and site-specific recombination instead of conventional double-crossover-based homologous recombination. This strategy was successfully applied to the in vivo inactivation of five candidate genes involved in the biosynthesis of hygromycin B by generating stop codons or mutating conserved residues within the encoding region. The results revealed that HygJ, HygL, and HygD are responsible for successive dehydrogenation, transamination, and transglycosylation of nucleoside diphosphate (NDP)-heptose. Notably, HygY acts as an unusual radical S-adenosylmethionine (SAM)-dependent epimerase for hydroxyl carbons, and HygM serves as a versatile methyltransferase in multiple parallel metabolic networks. Based on in vivo and in vitro evidence, the biosynthetic pathway for hygromycin B is proposed.

Conventional CRISPR/Cas genetic manipulation has been profitably applied to the genus Streptomyces , the most prolific bacterial producers of antibiotics. However, its reliance on DNA double-strand break (DSB) formation leads to unacceptably low yields of desired recombinants. We have adapted for Streptomyces recently-introduced cytidine base editors (CBEs) and adenine base editors (ABEs) which enable targeted C-to-T or A-to-G nucleotide substitutions, respectively, bypassing DSB and the need for a repair template. We report successful genome editing in Streptomyces at frequencies of around 50% using defective Cas9-guided base editors and up to 100% by using nicked Cas9-guided base editors. Furthermore, we demonstrate the multiplex genome editing potential of the nicked Cas9-guided base editor BE3 by programmed mutation of nine target genes simultaneously. Use of the high-fidelity version of BE3 (HF-BE3) essentially improved editing specificity. Collectively, this work provides a powerful new tool for genome editing in Streptomyces .

作为曾经治疗细菌感染的一线临床药物,氨基糖苷类抗生素在人类与病源微生物的抗争中作出了不可磨灭的巨大贡献,也成就了这一类抗生素的辉煌.虽然,伴随着其耳毒性和肾毒性等毒副作用,以及日益严重的耐药性的严峻挑战,但借助现代科学技术的发展和认知水平的提高,氨基糖苷类抗生素许多不曾被了解的新的生物活性也正不断地丰富和拓展着它新的潜能,使之依然成为人类医药宝库中不可或缺的重要一员.基于此,本文从分子遗传学、生物化学及结构生物学角度对常见的天然和化学半合成氨基糖苷抗生素生物合成的研究进展进行简要的概述.

Gentamicin C complex from Micromonospora echinospora remains a globally important antibiotic, and there is revived interest in the semisynthesis of analogs that might show improved therapeutic properties. The complex consists of five components differing in their methylation pattern at one or more sites in the molecule. We show here, using specific gene deletion and chemical complementation, that the gentamicin pathway up to the branch point is defined by the selectivity of the methyltransferases GenN, GenD1, and GenK. Unexpectedly, they comprise a methylation network in which early intermediates are ectopically modified. Using whole-genome sequence, we have also discovered the terminal 6'-N-methyltransfer required to produce gentamicin C2b from C1a or gentamicin C1 from C2, an example of an essential biosynthetic enzyme being located not in the biosynthetic gene cluster but far removed on the chromosome. These findings fully account for the methylation pattern in gentamicins and open the way to production of individual gentamicins by fermentation, as starting materials for semisynthesis.

The readthrough of premature termination codons (PTCs) is a promising strategy for curing PTC-causing diseases. In cancers, the p53 anti-tumor activity is often disabled by forming premature terminated p53 protein (p53(PTC)). Currently, there are lack of p53(PTC)-rescuing drugs. Herein we designed a feasible strategy for identifying p53PTC-rescuing compounds using protein biosynthesis machinery in E. coli cells and the lung cancer H1299 cells both harboring p53(PTC)-GFP fusion protein expression cassettes. Rescued p53(PTC) protein enabled a GFP-tag for fluorescence spectroscopic measurements. Our data revealed that the aminoglycoside G418 not only efficiently rescued p53(PTC) in H1299 cells, but also synergistically enhanced the efficacy of antitumor drug doxorubicin. Our work gave an insight into the discovery of p53(PTC)-rescuing drugs for cancer therapy.

Hygromycin B is an aminoglycoside antibiotic with a structurally distinctive orthoester linkage. Despite its long history of use in industry and in the laboratory, its biosynthesis remains poorly understood. We show here, by in-frame gene deletion in vivo and detailed enzyme characterization in vitro, that formation of the unique orthoester moiety is catalyzed by the α-ketoglutarate- and non-heme iron-dependent oxygenase HygX. In addition, we identify HygF as a glycosyltransferase adding UDP-hexose to 2-deoxystreptamine, HygM as a methyltransferase responsible for N-3 methylation, and HygK as an epimerase. These experimental results and bioinformatic analyses allow a detailed pathway for hygromycin B biosynthesis to be proposed, including the key oxidative cyclization reactions.

Gentamicins are heavily methylated, clinically valuable pseudotrisaccharide antibiotics produced by Micromonospora echinospora. GenN has been characterized as an S-adenosyl-l-methionine-dependent methyltransferase with low sequence similarity to other enzymes. It is responsible for the 3″-N-methylation of 3″-dehydro-3″-amino-gentamicin A2, an essential modification of ring III in the biosynthetic pathway to the gentamicin C complex. Purified recombinant GenN also efficiently catalyzes 3″-N-methylation of related aminoglycosides kanamycin B and tobramycin, which both contain an additional hydroxymethyl group at the C5″ position in ring III. We have obtained eight cocrystal structures of GenN, at a resolution of 2.2 Å or better, including the binary complex of GenN and S-adenosyl-l-homocysteine (SAH) and the ternary complexes of GenN, SAH, and several aminoglycosides. The GenN structure reveals several features not observed in any other N-methyltransferase that fit it for its role in gentamicin biosynthesis. These include a novel N-terminal domain that might be involved in protein:protein interaction with upstream enzymes of the gentamicin X2 biosynthesis and two long loops that are involved in aminoglycoside substrate recognition. In addition, the analysis of structures of GenN in complex with different ligands, supported by the results of active site mutagenesis, has allowed us to propose a catalytic mechanism and has revealed the structural basis for the surprising ability of native GenN to act on these alternative substrates.