Ectopic Expression of PsnNAC090 Enhances Salt and Osmotic Tolerance in Transgenic Tobacco.

The NAC transcription factor family is well known to play vital roles in plant development and stress responses. For this research, a salt-inducible NAC gene, PsnNAC090 (Po-tri.016G076100.1), was successfully isolated from Populus simonii × Populus nigra. PsnNAC090 contains the same motifs at the N-terminal end of the highly conserved NAM structural domain. The promoter region of this gene is rich in phytohormone-related and stress response elements. Transient transformation of the gene in the epidermal cells of both tobacco and onion showed that the protein was targeted to the whole cell including the cell membrane, cytoplasm and nucleus. A yeast two-hybrid assay demonstrated that PsnNAC090 has transcriptional activation activity with the activation structural domain located at 167-256aa. A yeast one-hybrid experiment showed that PsnNAC090 protein can bind to ABA-responsive elements (ABREs). The spatial and temporal expression patterns of PsnNAC090 under salt and osmotic stresses indicated that the gene was tissue-specific, with the highest expression level in the roots of Populus simonii × Populus nigra. We successfully obtained a total of six transgenic tobacco lines overexpressing PsnNAC090. The physiological indicators including peroxidase (POD) activity, superoxide dismutase (SOD) activity, chlorophyll content, proline content, malondialdehyde (MDA) content and hydrogen peroxide (H2O2) content were measured in three transgenic tobacco lines under NaCl and polyethylene glycol (PEG) 6000 stresses. The findings reveal that PsnNAC090 improves salt and osmotic tolerance by enhancing reactive oxygen species (ROS) scavenging and reducing membrane lipid peroxide content in transgenic tobacco. All the results suggest that the PsnNAC090 gene is a potential candidate gene playing an important role in stress response.

Label-Aware Distribution Calibration for Long-Tailed Classification.

Real-world data usually present long-tailed distributions. Training on imbalanced data tends to render neural networks perform well on head classes while much worse on tail classes. The severe sparseness of training instances for the tail classes is the main challenge, which results in biased distribution estimation during training. Plenty of efforts have been devoted to ameliorating the challenge, including data resampling and synthesizing new training instances for tail classes. However, no prior research has exploited the transferable knowledge from head classes to tail classes for calibrating the distribution of tail classes. In this article, we suppose that tail classes can be enriched by similar head classes and propose a novel distribution calibration (DC) approach named as label-aware DC (). transfers the statistics from relevant head classes to infer the distribution of tail classes. Sampling from calibrated distribution further facilitates rebalancing the classifier. Experiments on both image and text long-tailed datasets demonstrate that significantly outperforms existing methods. The visualization also shows that provides a more accurate distribution estimation.

Enriching query semantics for code search with reinforcement learning.

Code search is a common practice for developers during software implementation. The challenges of accurate code search mainly lie in the knowledge gap between source code and natural language (i.e., queries). Due to the limited code-query pairs and large code-description pairs available, the prior studies based on deep learning techniques focus on learning the semantic matching relation between source code and corresponding description texts for the task, and hypothesize that the semantic gap between descriptions and user queries is marginal. In this work, we found that the code search models trained on code-description pairs may not perform well on user queries, which indicates the semantic distance between queries and code descriptions. To mitigate the semantic distance for more effective code search, we propose QueCos, a Query-enriched Code search model. QueCos learns to generate semantic enriched queries to capture the key semantics of given queries with reinforcement learning (RL). With RL, the code search performance is considered as a reward for producing accurate semantic enriched queries. The enriched queries are finally employed for code search. Experiments on the benchmark datasets show that QueCos can significantly outperform the state-of-the-art code search models.

CRaDLe: Deep code retrieval based on semantic Dependency Learning.

Code retrieval is a common practice for programmers to reuse existing code snippets in the open-source repositories. Given a user query (i.e., a natural language description), code retrieval aims at searching the most relevant ones from a set of code snippets. The main challenge of effective code retrieval lies in mitigating the semantic gap between natural language descriptions and code snippets. With the ever-increasing amount of available open-source code, recent studies resort to neural networks to learn the semantic matching relationships between the two sources. The statement-level dependency information, which highlights the dependency relations among the program statements during the execution, reflects the structural importance of one statement in the code, which is favorable for accurately capturing the code semantics but has never been explored for the code retrieval task. In this paper, we propose CRaDLe, a novel approach for Code Retrieval based on statement-level semantic Dependency Learning. Specifically, CRaDLe distills code representations through fusing both the dependency and semantic information at the statement level, and then learns a unified vector representation for each code and description pair for modeling the matching relationship. Comprehensive experiments and analysis on real-world datasets show that the proposed approach can accurately retrieve code snippets for a given query and significantly outperform the state-of-the-art approaches on the task.

Acute and sub-chronic toxicity of glucose-cysteine Maillard reaction products in Sprague-Dawley rats.

Maillard reaction products (MRPs) derived from glucose-cysteine reactions have excellent anti-browning ability. However, there is a lack of information about their acute and sub-chronic toxicities. To our knowledge, the present study is the first to evaluate the acute and sub-chronic toxicities of MRPs in experimental animals. Acute toxicity testing and analysis by Horn's method showed that the median lethal oral dose (LD50) of MRPs in rats was 6.81 g/kg body weight. The sub-chronic toxicity test involved feeding rats with diet containing 0, 0.43, 0.85, or 1.70% (w/w) MRPs for 90 days. These treatments did not affect mortality, gross pathology, histology, hematology, or blood chemistry, and there were no dose-dependent changes in feed consumption. Based on these results, the dietary no-observed-adverse-effect level (NOAEL) for 90-day exposure was 1.29 and 1.51 g MRPs/kg body weight/day for male and female rats, respectively.