Search

The adversary makes malicious samples capable of triggering erroneous judgments in deep learning models by introducing imperceptible perturbations to the original benign texts. These malicious samples are referred to as adversarial texts. The exploration of adversarial text generation methods not only facilitates our understanding of the robustness of mainstream deep neural networks against such adversarial attacks but also aids in developing appropriate defensive strategies. Nevertheless, the mainstream research on textual adversarial attacks has mainly focused on attack effectiveness, overlooking the associated attack cost. For real-world attacks, considerations such as the time cost, material cost, manpower cost, and various constraints are also crucial. In this paper, we propose a low-cost adversarial text generation method based on the universal strategy in the black-box attack scenario, Universal Chinese Text Tricker (UCTT), for tendency classification on Chinese texts. UCTT is both text-independent and model-independent, which markedly reduces its attack cost. Instead of crafting adversarial texts for a specific text, UCTT generates universal perturbations based on a universal word substitution list, which is applicable to any data in tendency classification datasets. Given a perturbation rate, we can use the word list to craft adversarial texts by simple substitutions without accessing the target model. In the framework of adversarial text generation based on word importance, UCTT utilizes count, arithmetic progression, linear normalization, and nonlinear normalization to calculate the scores of the important words in the dataset and then computes the candidate word frequencies, which in turn constructs the universal word substitution list. Compared with other black-box methods, the experimental results on real-world tendency classification datasets show that UCTT exhibits an effective attack capability while significantly reducing the attack cost. Compared to the powerful baseline we designed that exceeds the SOTA, UCTT improves the efficiency of adversarial text generation by up to a factor of 7 without accessing the target model. In addition to demonstrating excellent attack performance on mainstream models, UCTT is also capable of attacking the powerful ChatGPT in the physical world, which cannot be directly attacked by traditional adversarial text generation methods due to the hard labels produced by the target model. [ABSTRACT FROM AUTHOR]

Children now are facing an increasing risk of early life stress (ELS), which leads to detrimental psychosocial outcomes. Behavior studies suggested that positive parental interactions might moderate the negative impact of ELS, but the related biological alteration remains unclear. This study aims to investigate whether positive parent–child interactions moderate the association between maltreatment (as a severe form of ELS) and hair cortisol concentration (HCC), as well as between HCC and psychosocial outcomes in young children. Participants were 6-year-old Chinese children (N = 257, Mage = 6.2, 121 were male) selected by stratified cluster random sampling from a Shanghai population representative cohort. Proximal 3 cm hair strands were analyzed using liquid chromatography coupled with tandem mass spectrometry for HCC. Children’s psychosocial outcome was evaluated using the parental report Strengths and Difficulties Questionnaire (SDQ). Parents also reported the frequency of positive parent–child interactions using the Chinese Parent–Child Interaction Scale (CPCIS) as well as the history of maltreatment. Multi-level logistic regression models adjusting for individual, kindergarten, and district confounders were used to evaluate the associations between maltreatment, HCC, and psychosocial outcomes. Interactions terms tested whether more frequent positive parent–child interactions moderates the association between maltreatment and HCC, as well as between HCC and psychosocial outcomes. Maltreated children exhibited higher levels of HCC (B = 1.20, 95% CI: 0.38,2.02; p = 0.004), and children with higher HCC exhibited poorer psychosocial outcomes (B = 0.34, 95% CI: 0.18,0.51; p < 0.001). Positive parent–child interactions did not have a moderating effect on the association between maltreatment and HCC, but they demonstrated a moderating effect on the association between increased HCC and psychosocial outcomes (interaction term: B = −0.42, 95% CI: −0.75,−0.10; p = 0.01). These findings provide evidence that positive parental interaction may serve as a moderator between chronic cortisol exposure and psychosocial problems. It highlights the importance of frequent parent–child interactions, especially among children under a high risk of ELS. [ABSTRACT FROM AUTHOR]

The dysfunction of airway smooth muscle cells (ASMCs) is one of the key factors in the pathogenesis of asthma. How miR-98-5p works in asthma has not been completely elucidated. This work focused on how miR-98-5p functions in the proliferation and migration of human ASMCs treated with interleukin-13 (IL-13). MiR-98-5p expression in plasma of asthmatic patients and IL-13-stimulated ASMCs was probed by quantitative real-time polymerase chain reaction (qRT-PCR). RAS-relevant C3 botulinum toxin substrate 1 (RAC1) protein expression in ASMCs was assessed by Western blot. The growth of ASMCs was measured by cell counting kit-8 (CCK-8) assay and 5-ethynyl-2′-deoxyuridine (EdU) assay. The migration of ASMCs was examined by Transwell assay. Besides, the apoptosis of ASMCs was analyzed by flow cytometry. The targeting relationship between miR-98-5p and RAC1 3′-UTR was verified by a dual-luciferase reporter gene assay. MiR-98-5p expression was reduced in patients' plasma and IL-13-stimulated ASMCs, and RAC1 expression was upregulated in ASMCs treated with IL-13. MiR-98-5p overexpression inhibited IL-13-induced proliferation and migration of ASMCs while promoting the apoptosis. The opposite result was observed after inhibiting miR-98-5p expression. Besides, RAC1 was identified as a direct downstream target of miR-98-5p in ASMCs. The restoration of RAC1 expression counteracted the impacts of miR-98-5p overexpression on IL-13-stimulated proliferation, migration, and apoptosis of ASMCs. MiR-98-5p inhibits IL-13-induced proliferation and migration and accelerates the apoptosis of ASMCs by downregulating RAC1 expression. [ABSTRACT FROM AUTHOR]

Key Message: PbMYB1L enhances the cold tolerance and anthocyanin accumulation of transgenic Arabidopsis by regulating the expression of genes related to the cold-responsive genes pathway and anthocyanin synthesis pathway. MYB transcription factors (TFs) have been demonstrated to play diverse roles in plant growth and development. In the present study, we identified a novel R2R3-MYB transcription factor, PbMYB1L, from the peel of 'Red Zaosu' pear (Pyrus bretschneideri), which was induced by cold stress and acted as a positive regulator in anthocyanin biosynthesis. Notably, the transgenic Arabidopsis lines exhibited enhanced tolerance to cold stress. Compared to the Arabidopsis wild-type plants, the transgenic lines displayed longer primary roots and reduced reactive oxygen species (ROS) levels including O2−, hydrogen peroxide (H2O2), and malondialdehyde (MDA). Furthermore, significant upregulation of key cold-responsive genes AtCBF1, AtCBF2, AtCBF3, AtCBF4, and AtKIN1 was observed in the transgenic plants under cold stress conditions compared to wild type. Arabidopsis plants overexpressing PbMYB1L had significant anthocyanin accumulation in leaves after cold treatment with quantitative results indicating higher expression of anthocyanin structural genes compared to wild type. These findings suggest that PbMYB1L not only plays a vital role in conferring cold tolerance but also acts as a crucial regulator of anthocyanin biosynthesis. [ABSTRACT FROM AUTHOR]

Citrus rootstock can modify plant growth and enhance stress resistance. There are many genotypes and species used as citrus rootstocks. Although multiple citrus rootstocks whole genome sequence and transcriptome databases have previously been published, no suitable internal reference genes have been investigated for standardization of gene expression via quantitative real-time PCR. Here we reported the first systematic and comprehensive analysis of reference genes for quantitative real-time PCR standardization in various citrus rootstocks. The expression stability of ten candidate reference genes in diverse sample subsets of flooding, drought, alkaline and cold treatments was evaluated using four statistical algorithms, geNorm, NormFinder, BestKeeper and ΔCt method. The results demonstrated that the expression stability of reference genes varied under different experimental conditions. In addition, normalization of gene expression of 9-cis-epoxycarotenoid dioxygenase 3 (NCED 3), involved in abscisic acid biosynthesis, was conducted to further confirm the reliability of the reference genes. Overall, EF-1α, DIM 1, GAPC and UBC expressed much more stably. ACTIN and GAPDH were not recommended for normalization in given experimental conditions due to low stability. Our main contribution was to identify reference genes with suitable and stable expression in citrus rootstocks varied across selected experimental conditions. Furthermore, these results will provide valuable information for future research on citrus rootstocks. [ABSTRACT FROM AUTHOR]