Search

Deep neural networks (DNNs) have been widely applied in various domains in artificial intelligence including computer vision and natural language processing. A DNN is typically trained for many epochs and then a validation dataset is used to select the DNN in an epoch (we simply call this epoch ``the last epoch") as the final model for making predictions on unseen samples, while it usually cannot achieve a perfect accuracy on unseen samples. An interesting question is ``how many test (unseen) samples that a DNN misclassifies in the last epoch were ever correctly classified by the DNN before the last epoch?". In this paper, we empirically study this question and find on several benchmark datasets that the vast majority of the misclassified samples in the last epoch were ever classified correctly before the last epoch, which means that the predictions for these samples were flipped from ``correct" to ``wrong". Motivated by this observation, we propose to restrict the behavior changes of a DNN on the correctly-classified samples so that the correct local boundaries can be maintained and the flipping error on unseen samples can be largely reduced. Extensive experiments on different benchmark datasets with different modern network architectures demonstrate that the proposed flipping error reduction (FER) approach can substantially improve the generalization, the robustness, and the transferability of DNNs without introducing any additional network parameters or inference cost, only with a negligible training overhead.

IntroductionIt is believed that ovarian cancer (OC) is the most deadly form of gynecological cancer despite its infrequent occurrence, which makes it one of the most salient public health concerns. Clinical and preclinical studies have revealed that intratumoral CD4+ T cells possess cytotoxic capabilities and were capable of directly killing cancer cells. This study aimed to identify the CD4+ conventional T cells-related genes (CD4TGs) with respect to the prognosis in OC.MethodsWe obtained the transcriptome and clinical data from the Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases. CD4TGs were first identified from single-cell datasets, then univariate Cox regression was used to screen prognosis-related genes, LASSO was conducted to remove genes with coefficient zero, and multivariate Cox regression was used to calculate riskscore and to construct the CD4TGs risk signature. Kaplan-Meier analysis, univariate Cox regression, multivariate Cox regression, time-dependent receiver operating characteristics (ROC), decision curve analysis (DCA), nomogram, and calibration were made to verify and evaluate the risk signature. Gene set enrichment analyses (GSEA) in risk groups were conducted to explore the tightly correlated pathways with the risk group. The role of riskscore has been further explored in the tumor microenvironment (TME), immunotherapy, and chemotherapy. A risk signature with 11 CD4TGs in OC was finally established in the TCGA database and furtherly validated in several GEO cohorts.ResultsHigh riskscore was significantly associated with a poorer prognosis and proven to be an independent prognostic biomarker by multivariate Cox regression. The 1-, 3-, and 5-year ROC values, DCA curve, nomogram, and calibration results confirmed the excellent prediction power of this model. Compared with the reported risk models, our model showed better performance. The patients were grouped into high-risk and low-risk subgroups according to the riskscore by the median value. The low-risk group patients tended to exhibit a higher immune infiltration, immune-related gene expression and were more sensitive to immunotherapy and chemotherapy.DiscussionCollectively, our findings of the prognostic value of CD4TGs in prognosis and immune response, provided valuable insights into the molecular mechanisms and clinical management of OC.

Two actinide isotopes, $^{241}$Am and $^{244}$Cm, produced and chemically purified by the HFIR/REDC complex at ORNL are candidates for target materials of heavy-ion fusion reaction experiments for the synthesis of new superheavy elements (SHEs) with $Z>118$. In the framework of the dinuclear system model with a dynamical potential energy surface (DNS-DyPES model), we systematically study the $^{48}$Ca-induced reactions that have been applied to synthesize SHEs with $Z=112$--118, as well as the hot-fusion reactions with $^{241}$Am and $^{244}$Cm as targets which are promising for synthesizing new SHEs with $Z=119$--122. Detailed results including the maximal evaporation residue cross section and the optimal incident energy for each reaction are presented and discussed.

BackgroundHyperuricemia and right ventricular dysfunction (RVD) are both widespread in heart failure with preserved ejection fraction (HFpEF) patients. RVD is associated with a poor prognosis in HFpEF. The correlation between serum uric acid (UA) levels and right ventricular function is unclear. The prognostic performance of UA in patients with HFpEF needs further validation.Methods and resultsA total of 210 patients with HFpEF were included in the study and divided into two groups according to UA level: the normal UA group (≤7 mg/dl) and the high UA group (>7 mg/dl). The variables examined included clinical characteristics, echocardiography, and serum biochemical parameters. Right ventricular function was assessed by tricuspid annular plane systolic excursion (TAPSE) and tricuspid annular peak systolic velocity (TAPSV). Baseline characteristics were compared between the two groups, and the correlation between baseline UA and RVD was assessed using multifactorial binary logistic regression. Kaplan–Meier curves were used to describe all-cause mortality and heart failure readmission. Results showed that right ventricular function parameters were worse in the high UA group. After adjusting for UA, left ventricular posterior wall thickness (LVPWT), N-terminal B-type natriuretic peptide (NT-proBNP), atrial fibrillation (AF), and low-density lipoprotein cholesterol (LDL-C), UA (odds ratio = 2.028; p < 0.001) was independently associated with RVD, and UA >7 mg/dl (HR = 2.98; p < 0.001) was associated with heart failure readmission in patients with HFpEF.ConclusionElevated serum UA is closely associated with RVD and significantly associated with the heart failure readmission rate in patients with HFpEF.

Hepatoblastoma (HB) is a common primary malignant liver tumour in children, mainly treated by means of traditional chemotherapy using platinum and doxorubicin (ADM). There has been limited progress in the research and development of new drugs for treating HB.A tumour biopsy from a child with HB was implanted into immunodeficient mice. The primary tumour and patient-derived xenograft (PDX) tumour were extensively characterised by histology, immunohistochemistry (IHC), and humanisation identification. We used the PDX model to evaluate the anti-tumour effects of anlotinib oxaliplatin (L-OHP) and sorafenib on childhood HB.The established PDX model maintained the histological characteristics of the primary tumour. Anlotinib, L-OHP, and sorafenib can significantly inhibit the tumour growth in the PDX model. There was no obvious damage of the drugs to the heart, liver and kidney of the mice, and the side effects observed were light.We have successfully established a PDX model of childhood HB. The model retains important molecular characteristics of human primary tumours. Using the model, it was found that anlotinib, L-OHP, and sorafenib have a good inhibitory effect on the growth of childhood HB. This provides a preliminary research basis for the clinical application of the drugs.

Anoplocaryum (Boraginaceae), a small genus previously known from Russia and Mongolia, is newly recorded from China. The specimen was discovered in the rock crevices of Zhangjiakou City, Hebei province, China, and was identified as A. helenae Volot. based on both morphological and molecular data. This study updates our knowledge on the geographic distribution of the genus Anoplocaryum. Two taxonomic keys were provided to distinguish Anoplocaryum and Chinese related genera, also to index taxa in Anoplocaryum.

We introduce Mind2Web, the first dataset for developing and evaluating generalist agents for the web that can follow language instructions to complete complex tasks on any website. Existing datasets for web agents either use simulated websites or only cover a limited set of websites and tasks, thus not suitable for generalist web agents. With over 2,000 open-ended tasks collected from 137 websites spanning 31 domains and crowdsourced action sequences for the tasks, Mind2Web provides three necessary ingredients for building generalist web agents: 1) diverse domains, websites, and tasks, 2) use of real-world websites instead of simulated and simplified ones, and 3) a broad spectrum of user interaction patterns. Based on Mind2Web, we conduct an initial exploration of using large language models (LLMs) for building generalist web agents. While the raw HTML of real-world websites are often too large to be fed to LLMs, we show that first filtering it with a small LM significantly improves the effectiveness and efficiency of LLMs. Our solution demonstrates a decent level of performance, even on websites or entire domains the model has never seen before, but there is still a substantial room to improve towards truly generalizable agents. We open-source our dataset, model implementation, and trained models (https://osu-nlp-group.github.io/Mind2Web) to facilitate further research on building a generalist agent for the web., Comment: Website: https://osu-nlp-group.github.io/Mind2Web Updated with supplementary material

Chain-of-Thought (CoT) prompting can dramatically improve the multi-step reasoning abilities of large language models (LLMs). CoT explicitly encourages the LLM to generate intermediate rationales for solving a problem, by providing a series of reasoning steps in the demonstrations. Despite its success, there is still little understanding of what makes CoT prompting effective and which aspects of the demonstrated reasoning steps contribute to its performance. In this paper, we show that CoT reasoning is possible even with invalid demonstrations - prompting with invalid reasoning steps can achieve over 80-90% of the performance obtained using CoT under various metrics, while still generating coherent lines of reasoning during inference. Further experiments show that other aspects of the rationales, such as being relevant to the query and correctly ordering the reasoning steps, are much more important for effective CoT reasoning. Overall, these findings both deepen our understanding of CoT prompting, and open up new questions regarding LLMs' capability to learn to reason in context., ACL-23 Camera Ready. Code and model input/output are available at https://github.com/sunlab-osu/Understanding-CoT

A key missing capacity of current language models (LMs) is grounding to real-world environments. Most existing work for grounded language understanding uses LMs to directly generate plans that can be executed in the environment to achieve the desired effects. It thereby casts the burden of ensuring grammaticality, faithfulness, and controllability all on the LMs. We propose Pangu, a generic framework for grounded language understanding that capitalizes on the discriminative ability of LMs instead of their generative ability. Pangu consists of a symbolic agent and a neural LM working in a concerted fashion: The agent explores the environment to incrementally construct valid plans, and the LM evaluates the plausibility of the candidate plans to guide the search process. A case study on the challenging problem of knowledge base question answering (KBQA), which features a massive environment, demonstrates the remarkable effectiveness and flexibility of Pangu: A BERT-base LM is sufficient for setting a new record on standard KBQA datasets, and larger LMs further bring substantial gains. Pangu also enables, for the first time, effective few-shot in-context learning for KBQA with large LMs such as Codex., 18 pages, 6 figures, 6 tables; accepted to ACL'2023

At present, the local ventilator control mainly uses traditional PID controllers. The parameters of the controller are fixed and selected based on experience. It is difficult to achieve the optimal combination of membership functions and fuzzy rules, which makes it difficult for the controller to meet the adaptive control requirements of local ventilator. In order to solve this problem, a fuzzy PID controller for mine local ventilator optimized by improved genetic algorithm is designed. In the improved genetic algorithm, the Euclidean distance is introduced to increase the diversity of the population, and the adaptive crossover and mutation probability are introduced to improve the convergence of the algorithm. In the encoding process, the proportional fraction is used to indirectly optimize the membership functions, which enables the improved genetic algorithm to optimize both the membership functions and fuzzy rules. The optimized membership functions and fuzzy rules are imported into the fuzzy PID controller. The controller can adaptively adjust the air volume of the local ventilator through the inverter according to the different working status of the local ventilator so as to realize the dynamic adjustment of the local ventilator. The simulation results show that compared with the traditional fuzzy PID controller, the improved fuzzy PID controller can basically achieve no overshoot, the rise time is shortened by 56.25%, and the stabilization time is shortened by 47.06%, which can better meet the control requirements of mine local ventilator.

Cholestasis is a main clinical feature of biliary atresia (BA), which leads to liver fibrosis (LF). The focus of BA treatment is preventing and slowing the progress of LF. This study reports the improvement effect of anlotinib on common bile duct ligature (BDL)-induced LF in young rats. The BDL young rats were treated with anlotinib and the serum levels of aspartate aminotransferase, alanine aminotransferase, albumin, and total bilirubin were determined. Histological staining was performed and pathological changes in liver tissue were observed. The expression levels of α-SMA, collagen I, CD31, TGF-β1, phospho-VEGFR2, phospho-4E/BP1, and phospho-S6K1 were determined. The results showed that anlotinib significantly improved the liver function and histopathological injury of BDL rats, inhibited the deposition of collagen and hepatocyte apoptosis, and downregulated the protein expression of α-SMA and collagen I. Furthermore, anlotinib treatment significantly inhibited microvascular formation in the liver and downregulated the expression level of phospho-VEGFR2, thereby suggesting that the antifibrosis effect of anlotinib may be achieved by antiangiogenesis. In addition, anlotinib downregulated the expression of phospho-S6K1 and upregulated the expression of phospho-4E/BP1, two downstream proteins of the mammalian target of rapamycin (mTOR) pathway. MHY1485, an agonist of mTOR, significantly reversed the inhibitory effect of anlotinib on angiogenesis and LF but did not influence the effect of anlotinib on the downregulation of phospho-VEGFR2 expression. Together, the above-mentioned results suggest that the effect of anlotinib on BDL-induced LF involves at least antiangiogenesis regulated by the VEGFR2/mTOR signaling pathway.

The Confinement effect is an effective strategy for enhancing the performance of catalysts during various reactions. We developed a strategy to encapsulating Palladium nanoparticles with a dopamine-derived N-doped multilayer carbon shell to catalyze the Heck reaction between iodobenzene and carbon–carbon double-bond compounds (styrene and methyl acrylate). The catalytic performances of Pd nanoparticles encapsulated within a N-doped carbon layer (Pd@N-C) and those attached to the external surface of such a layer (Pd/N-C) were evaluated and compared. Pd@N-C exhibited a lower reaction activation energy (78 kJ/mol) than that of Pd/N-C (106 kJ/mol). Thus, the activity of Pd@N-C during the Heck reaction between iodobenzene and methyl acrylate was approximately 20 times higher than that of Pd/N-C-200. The results of thermal filtration and recycling tests indicated that Pd@N-C showed a lower degree of leaching than that of Pd/N-C-200 owing to the spatial restriction effect of the cavity in the case of the former. Hence, the superior catalytic performance of Pd@N-C can be ascribed to the confinement of the Pd nanoparticles within the N-doped carbon layer. Finally, we used these catalysts to synthesize a range of para-substituted iodobenzene derivatives, thus demonstrating the potential of Pd-encapsulating catalysts for use in the Heck reaction on the technical scale.

Background Colon cancer is one of the most common malignancies worldwide, and chemotherapy is a widely used strategy in colon cancer clinical therapy. However, chemotherapy resistance is a major cause of disease recurrence and progression in colon cancer, and thus novel drugs for treatment are urgently needed. Tetramethylpyrazine (TMP), a component of the traditional Chinese medicine Chuanxiong Hort, has been proven to exhibit a beneficial effect in tumors. Aim To investigate the potential anticancer activity of TMP in colon cancer and its underlying mechanisms. Methods Colon cancer cells were incubated with different concentrations of TMP. Cell viability was evaluated by crystal violet staining assay and cell counting kit-8 assay, and cell apoptosis and cell cycle were assessed by flow cytometry. Results TMP significantly inhibited the proliferation of colon cancer cells in a dose- and time-dependent manner. In addition, flow cytometry revealed that TMP induced cell cycle arrest at the G0/G1 phase. TMP treatment caused early stage apoptosis in SW480 cells, whereas it caused late stage apoptosis in HCT116 cells. Conclusion Our studies demonstrated that TMP inhibits the proliferation of colon cancer cells in a dose- and time-dependent manner by inducing apoptosis and arresting the cell cycle at the G0/G1 phase. Our findings suggest that TMP might serve as a potential novel therapeutic drug in the treatment of human colon cancer.

As one of the most concerned technologies in the field of artificial intelligence in recent ten years, deep learning has achieved excellent results in many applications, but the current learning strategies rely heavily on a large number of labeled data. In many practical problems, it is not feasible to obtain a large number of labeled training data, so it increases the training difficulty of the model. But it is easy to obtain a large number of unlabeled data. Semi-supervised learning makes full use of unlabeled data, provides solutions and effective methods to improve the performance of the model under the condition of limited labeled data, and achieves high recognition accuracy in the task of image classification. This paper first gives an overview of semi-supervised learning, and then introduces the basic ideas commonly used in classification algorithms. It focuses on the comprehensive review of image classification methods based on semi-supervised deep learning framework in recent years, including multi- view training, consistency regularization, diversity mixing and semi-supervised generative adversarial networks. It summarizes the common technologies of various methods, analyzes and compares the differences of experimental results of different methods. Finally, this paper thinks about the existing problems and looks forward to the feasible research direction in the future.

We present TacoBot, a task-oriented dialogue system built for the inaugural Alexa Prize TaskBot Challenge, which assists users in completing multi-step cooking and home improvement tasks. TacoBot is designed with a user-centered principle and aspires to deliver a collaborative and accessible dialogue experience. Towards that end, it is equipped with accurate language understanding, flexible dialogue management, and engaging response generation. Furthermore, TacoBot is backed by a strong search engine and an automated end-to-end test suite. In bootstrapping the development of TacoBot, we explore a series of data augmentation strategies to train advanced neural language processing models and continuously improve the dialogue experience with collected real conversations. At the end of the semifinals, TacoBot achieved an average rating of 3.55/5.0., Published in 1st Proceedings of Alexa Prize TaskBot (Alexa Prize 2021). TacoBot won 3rd place in the challenge. See project website https://sunlab-osu.github.io/tacobot/ for details

The golden snub-nosed monkey is one of the most endangered animal species endemic to China. In order to explore the characteristics and health risks of golden snub-nosed monkeys exposed to heavy metals, we collected the plant food sources, soil, and water samples from the golden snub-nosed monkey habitat in the Shennongjia Mountains; examined the contents of seven heavy metals (Cu, Zn, Pb, Cd, Cr, Ni, and As); and adopted the comprehensive pollution index, potential ecological risk index, and Nemerow index to evaluate pollutants in the water, soil, and food plants. At the same time, the Target Hazard Quotient method was used to assess heavy metals in the food plants. The results showed that the heavy metal concentration of the habitat water was 0.004-1.220 μg·L

Thallium (Tl) is a cumulative high toxicant in the environment, but few studies have investigated the comprehensive health effects underlying chronic Tl exposure at trace levels. This study aims to evaluate the liver, kidney, lung and other potential health effects associated with chronic Tl exposure at trace levels in rural areas of China. Urinary Tl concentrations of 2883 adults from rural areas of 12 provinces in China were measured and 2363 participants were involved in the final analysis. Indicators of liver and kidney functions in the serum, as well as the lung function indicators, were determined in the participants. General linear regression and restricted cubic spline regression were combined to study the associations between urinary Tl and health indicators or outcomes. In this study, the detected rate of Tl in the urine of the participants was 97.28 %. When the urinary Tl concentration was ranged at the fourth quintile, the risk of having liver function disorder was 70 % higher [Odds ratio (OR) = 1.70 (95 % confidence intervals (CI): 1.30, 2.22)] in all the participants, whereas the farmers were more likely to have the disorder [OR = 2.08 (95 % CI: 1.49, 2.92)] than the non-farmers [OR = 1.20 (95 % CI: 0.77, 1.88)]. Nonlinear associations between most of the liver health indicators and urinary Tl were identified, of which serum bilirubin was strongly associated with the elevation of urinary Tl when its concentration was0.40 μg/g creatinine. Besides, urinary Tl was negatively associated with lung health indicators. Our study proposes the safety re-assessment of the current exposure level of Tl in the environment, especially in rural areas of China.

Identifying informal e-waste recycling activity is crucial for preventing health hazards caused by e-waste pollution. This study attempted to build a prediction model for e-waste recycling activity based on the differential exposure biomarkers of the populations between the e-waste recycling area (ER) and non-ER. This study recruited children in ER and non-ER and conducted a quasi-experiment among the adult investigators to screen differential exposure or effect biomarkers by measuring urinary 25 volatile organic compound (VOC) metabolites, 18 metals/metalloids, and 8-hydroxy-2'-deoxyguanosine (8-OHdG). Compared with children of the non-ER, the ER children had higher metal/metalloid (e.g., manganese [Mn], lead [Pb], antimony [Sb], tin [Sn], and copper [Cu]) and VOC exposure (e.g., carbon-disulfide, acrolein, and 1-bromopropane) levels, oxidative DNA damage, and non-carcinogenic risks. Individually added 8-OHdG, VOC metabolites, and metals/metalloids to the support vector machine (SVM) classifier could obtain similar classification effects, with the area under curve (AUC) ranging from 0.741 to 0.819. The combined inclusion of 8-OHdG and differential VOC metabolites, metals/metalloids, and mixed indexes (e.g., product items or ratios of different metals/metalloids) in the SVM classifier showed the highest performance in predicting e-waste recycling activity, with an AUC of 0.914 and prediction accuracy of 83.3 %. "Sb × Mn", followed by "Sn × Pb/Cu", "Sb × Mn/Cu", and "Sn × Pb", were the top four important features in the models. Compared with non-ER children, the levels of urinary Mn, Pb, Sb, Sn, and Cu in ER children were 1.2 to 2.4 times higher, while the levels of "Sb × Mn", "Sn × Pb/Cu", "Sb × Mn/Cu", and "Sn × Pb" were 3.5 to 4.7 times higher, suggesting that these mixed indexes could amplify the differences between e-waste exposed and non-e-waste exposed populations. With the continued inclusion of new biomarkers of e-waste pollution in the future, our prediction model is promising for screening informal e-waste recycling sites.

While Pre-trained Language Models (PLMs) internalize a great amount of world knowledge, they have been shown incapable of recalling these knowledge to solve tasks requiring complex & multi-step reasoning. Similar to how humans develop a "chain of thought" for these tasks, how can we equip PLMs with such abilities? In this work, we explore an iterative prompting framework, a new prompting paradigm which progressively elicits relevant knowledge from PLMs for multi-step inference. We identify key limitations of existing prompting methods, namely they are either restricted to queries with a single identifiable relation/predicate, or being agnostic to input contexts, which makes it difficult to capture variabilities across different inference steps. We propose an iterative context-aware prompter, which addresses these limitations by learning to dynamically synthesize prompts conditioned on the current step's contexts. Experiments on three datasets involving multi-step reasoning show the effectiveness of the iterative scheme and the context-aware prompter design., EMNLP-2022 Camera-Ready. Code will be available at https://github.com/sunlab-osu/IterPrompt

BACKGROUND Recent evidence showed that combining endoscopic submucosal dissection (ESD) and laparoscopic sentinel lymph node dissection may avoid unnecessary gastrectomy in treating early mucinous gastric cancer (EMGC) patients with risks of positive lymph node metastasis (pLNM). AIM To explore the predictive factors for pLNM in EMGC, and to optimize the clinical application of combing ESD and sentinel lymph node dissection in a proper subgroup of patients with EMGC. METHODS Thirty-one patients with EMGC who had undergone gastrectomy with lymph node dissection were consecutively enrolled from January 1988 to December 2016. Univariate and multivariate logistic regression analyses were used to estimate the association between the rates of pLNM and clinicopathological factors, providing odds ratio (OR) with 95% confidence interval. And the association between the number of predictors and the pLNM rate was also investigated. RESULTS Depth of invasion (OR = 7.342, 1.127-33.256, P = 0.039), tumor diameter (OR = 9.158, 1.348-29.133, P = 0.044), and lymphatic vessel involvement (OR = 27.749, 1.821-33.143, P = 0.019) turned out to be significant and might be the independent risk factors for predicating pLNM in the multivariate analysis. For patients with 1, 2, and 3 risk factors, the pLNM rates were 9.1%, 33.3%, and 75.0%, respectively. pLNM was not detected in seven patients without any of these risk factors. CONCLUSION ESD might serve as a safe and sufficient treatment for intramucosal EMGC if tumor size ≤ 2 cm, and when lymphatic vessel involvement is absent by postoperative histological examination. Combining ESD and sentinel lymph node dissection could be recommended as a safe and effective treatment for EMGC patients with a potential risk of pLNM.

Key to the Anoplocaryum and Chinese related genera 1. Calyx lobes unequal, strongly enlarged in fruit, bilaterally compressed, becoming somewhat clamshell-shaped.............. Asperugo - Calyx lobes subequal, slightly enlarged in fruit, not clamshell-shaped.............................................................................................2 2. Nutlets tetrahedral..............................................................................................................................................................................3 - Nutlets dihedral, dorsiventrally compressed......................................................................................................................................4 3. Stamens included................................................................................................................................................................ Trigonotis - Stamens exserted................................................................................................................................................................. Mertensia 4. Nutlets lenticular, smooth...................................................................................................................................................... Myosotis - Nutlets grooved, wrinkled............................................................................................................................................ Anoplocaryum, Published as part of Hao, Jia-Chen, Fan, Ying-Xin, Zhu, Deng-Xiang & Lin, Qin-Wen, 2022, Anoplocaryum: A newly recorded genus of Boraginaceae from China, pp. 205-212 in Phytotaxa 530 (2) on page 210, DOI: 10.11646/phytotaxa.530.2.7, http://zenodo.org/record/5832746

Anoplocaryum helenae Volot. (1996: 147) (Fig. 2, 3, 4) Type:��� Russia. [The Sakha Republic], South Yakutia, river Big Ylymakh (basin of the Aldan river), the mouth of the river Cold (Agdy), in shady damp rock crevices, 24 July 1988, 420 m, K.A. Volotovsky s.n. (holotype LE; isotypes SASY, NSK0000079!). Specimens examined:��� CHINA. Hebei Province, Zhangjiakou City, N40��49���57���, E115��26���9���, ca. 2034 m, 4 September 2020, Lin et al. 2020-17 (PE!). Habitat:���moist crevices of granitic rocks. Known distribution:��� Russia, China. Phenology:���flowering in May���July; fruiting in July���August., Published as part of Hao, Jia-Chen, Fan, Ying-Xin, Zhu, Deng-Xiang & Lin, Qin-Wen, 2022, Anoplocaryum: A newly recorded genus of Boraginaceae from China, pp. 205-212 in Phytotaxa 530 (2) on page 210, DOI: 10.11646/phytotaxa.530.2.7, http://zenodo.org/record/5832746

HTML documents are an important medium for disseminating information on the Web for human consumption. An HTML document presents information in multiple text formats including unstructured text, structured key-value pairs, and tables. Effective representation of these documents is essential for machine understanding to enable a wide range of applications, such as Question Answering, Web Search, and Personalization. Existing work has either represented these documents using visual features extracted by rendering them in a browser, which is typically computationally expensive, or has simply treated them as plain text documents, thereby failing to capture useful information presented in their HTML structure. We argue that the text and HTML structure together convey important semantics of the content and therefore warrant a special treatment for their representation learning. In this paper, we introduce a novel representation learning approach for web pages, dubbed DOM-LM, which addresses the limitations of existing approaches by encoding both text and DOM tree structure with a transformer-based encoder and learning generalizable representations for HTML documents via self-supervised pre-training. We evaluate DOM-LM on a variety of webpage understanding tasks, including Attribute Extraction, Open Information Extraction, and Question Answering. Our extensive experiments show that DOM-LM consistently outperforms all baselines designed for these tasks. In particular, DOM-LM demonstrates better generalization performance both in few-shot and zero-shot settings, making it attractive for making it suitable for real-world application settings with limited labeled data.

Eupatilin (5,7-dihydroxy-3 ′ ,4 ′ ,6-trimethoxyflavone) is a pharmacologically active flavone that has been isolated from a variety of medicinal plants and possesses a number of pharmacological properties. This study evaluates the antioxidant and antiapoptotic effects of eupatilin on cisplatin-induced ototoxicity using in vitro and in vivo models including HEI-OC1 cells, cochlear hair cells, and zebrafish. Employing a CCK8 assay and Annexin V-FITC/PI double staining, we found that eupatilin significantly alleviated cisplatin-induced apoptosis and increased hair cell viability. The level of reactive oxygen species (ROS) was evaluated by CellROX green and MitoSOX Red staining. The results showed that eupatilin possesses antioxidant activity. MitoTracker Red staining indicated that eupatilin remarkably decreased mitochondrial damage. Furthermore, we demonstrated that eupatilin protects hair cells from cisplatin-induced damage. Mechanistic studies in cisplatin-induced HEI-OC1 cells revealed that eupatilin promoted Bcl-2 expression, downregulated Bax expression, reversed the increase in caspase-3 and PARP activity, and reduced the expression of phosphorylated p38 and JNK. Our data suggest a novel role for eupatilin as a protective agent against ototoxic drug-induced hair cell apoptosis by inhibiting ROS generation and modulating mitochondrial-related apoptosis.

Ovarian cancer (OV) is the most lethal gynecologic malignancy. One major reason of the high mortality of the disease is due to platinum-based chemotherapy resistance. Increasing evidence reveal the important biological functions and clinical significance of zinc finger proteins (ZNFs) in OV. In the present study, the relationship between the zinc finger protein 76 (ZNF76) and clinical outcome and platinum resistance in patients with OV was explored. We further analyzed ZNF76 expression via multiple gene expression databases and identified its functional networks using cBioPortal. RT-qPCR and IHC assay shown that the ZNF76 mRNA and protein expression were significantly lower in OV tumor than that in normal ovary tissues. A strong relationship between ZNF76 expression and platinum resistance was determined in patients with OV. The low expression of ZNF76 was associated with worse survival in OV. Multivariable analysis showed that the low expression of ZNF76 was an independent factor predicting poor outcome in OV. The prognosis value of ZNF76 in pan-cancer was validated from multiple cohorts using the PrognoScan database and GEPIA 2. A gene-clinical nomogram was constructed by multivariate cox regression analysis, combined with clinical characterization and ZNF76 expression in TCGA. Functional network analysis suggested that ZNF76 was involved in several biology progressions which associated with OV. Ten hub genes (CDC5L, DHX16, SNRPC, LSM2, CUL7, PFDN6, VARS, HSD17B8, PPIL1, and RGL2) were identified as positively associated with the expression of ZNF76 in OV. In conclusion, ZNF76 may serve as a promising prognostic-related biomarker and predict the response to platinum in OV patients.

Background Chlorinated polyfluorinated ether sulfonic acids (Cl-PFESA 6:2 and 8:2), used as perfluorooctane sulfonate (PFOS) alternatives, affect lipid metabolism in vivo and in vitro studies. The association between Cl-PFESAs exposure and the prevalence of overweight /obesity in human is unknown. Objectives We investigated associations of serum 6:2 Cl-PFESA and 8:2 Cl-PFESA with overweight/obesity status in adults. Methods We quantified four perfluoroalkyl substances (PFAS), including 6:2 Cl-PFESA, 8:2 Cl-PFESA, PFOS, and perfluorooctanoic acid (PFOA) in 1275 Chinese adults from the Isomers of C8 Health Project in China study. Participants were categorized into normal weight group [body mass index (BMI) < 25 kg/m2] and overweight/obesity group (BMI ≥ 25 kg/m2). Results Adjusted for potential confounders, BMI in the second quartile of each ln-ng/mL greater concentration of 6:2 Cl-PFESA and 8:2 Cl-PFESA were 0.45 [95% confidence interval (CI): 0.08, 0.82], and 0.39 (95% CI:0.03, 0.76) significantly higher than the lowest quartile, respectively. Cl-PFESAs displayed inverted U-shaped associations with the risk of overweight/obesity, and the inflection point of 6:2 Cl-PFESA and 8:2 Cl-PFESA were 1.80 ng/mL, 0.01 ng/mL, respectively. PFOS was associated with waist circumference (WC) but not BMI in each quartile. For PFOA, the associations with outcomes were linearly positive (P for trend < 0.05). Conclusions This study reports the first observations on non-monotonic associations between serum 6:2 Cl-PFESA and 8:2 Cl-PFESA concentrations and the prevalence of overweight/obesity in adults. More epidemiological investigations are required to confirm the observed associations.

The rural land right has paved the way for the deepening of China’s agricultural land system, which is critical to the successful implementation of the rural revitalization plan in the new era. Based on the micro-survey data of farmers in Yunnan Province, we use an OLS model to empirically test the impact of rural land rights on farmers’ income in underdeveloped areas, and we use a stepwise regression method to test the mechanism role of land transfer willingness and behavior in the rural land right on farmers’ income. Benchmark regression results show that the implementation of rural land rights policy significantly increased the total income of farmers, which mainly comes from farm income. Although the rural land right policy has no significant impact on farmers’ non-farm income, the sign of right coefficient is positive. Mechanism analysis shows that the rural land right can increase the income of farmers by encouraging them to transfer to land. It shows that the property right of land is clear, which makes it easier for land to transfer and trade, and farmers can achieve a certain degree of scale operation, thereby increasing income. Finally, it is suggested that future fiscal funding in impoverished regions be directed toward land transfer links and that accompanying rules and regulations adequately protect farmers’ land contract management rights.

BACKGROUND There are several surgical options for treating early gastric cancers (EGCs), such as endoscopic resection, laparoscopic or open gastrectomy with D1 or D2 lymphadenectomy. Endoscopic resection for EGC with low risk of lymph node metastasis has been widely accepted as a therapeutic alternative. The role of endoscopic submucosal dissection (ESD) in treating EGC is not well established, especially when compared with resection surgery cases in a long-term follow-up scope. AIM To compare the safety and efficacy of the short- and long-term outcomes between ESD and resection surgery. METHODS We searched the databases of PubMed, EMBASE, Web of Science, and the Cochrane Library from January 1990 to June 2018, enrolling studies reporting short- or long-term outcomes of ESD in comparison with resection surgery for EGC. The quality of the studies was assessed by the Newcastle-Ottawa Quality Assessment Scale. Stata software (version 12.0) was used for the analysis. Pooling analysis was conducted using either fixed- or random-effects models depending on heterogeneity across studies. RESULTS Fourteen studies comprising 5112 patients were eligible for analysis (2402 for EGC and 2710 for radical surgery). Our meta-analysis demonstrated that the ESD approach showed advantages through decreased operation time [weighted mean difference (WMD): -140.02 min, 95%CI: -254.23 to -34.82 min, P = 0.009], shorter hospital stay (WMD: -5.41 d, 95% CI: -5.93 to -4.89 d, P < 0.001), and lower postoperative complication rate [Odds ratio (OR) = 0.39, 95%CI: 0.28-0.55, P < 0.001). Meanwhile, EGC patients who underwent ESD had higher recurrence rate (OR = 9.24, 95%CI: 5.94-14.36, P < 0.001) than resection surgery patients. However, the long-term survival including overall survival [Hazard ratio (HR) = 0.51, 95%CI: 0.26-1.00, P = 0.05] and event-free survival (HR = 1.59, 95%CI: 0.66-9.81, P = 0.300) showed no significant differences between these two groups. CONCLUSION In the treatment of EGC, ESD was safe and feasible in comparison with resection surgery, with advantages in several surgical and post-operative recovery parameters. Although the recurrence rate was higher in ESD group, the long-term survival was still comparable in these two groups, suggesting ESD could be recommended as standard treatment for EGC with indications.

BACKGROUND Colon cancer is one of the most common malignancies worldwide, and chemotherapy is a widely used strategy in colon cancer clinical therapy. However, chemotherapy resistance is a major cause of disease recurrence and progression in colon cancer, and thus novel drugs for treatment are urgently needed. Tetramethylpyrazine (TMP), a component of the traditional Chinese medicine Chuanxiong Hort, has been proven to exhibit a beneficial effect in tumors. AIM To investigate the potential anticancer activity of TMP in colon cancer and its underlying mechanisms. METHODS Colon cancer cells were incubated with different concentrations of TMP. Cell viability was evaluated by crystal violet staining assay and cell counting kit-8 assay, and cell apoptosis and cell cycle were assessed by flow cytometry. RESULTS TMP significantly inhibited the proliferation of colon cancer cells in a dose- and time-dependent manner. In addition, flow cytometry revealed that TMP induced cell cycle arrest at the G0/G1 phase. TMP treatment caused early stage apoptosis in SW480 cells, whereas it caused late stage apoptosis in HCT116 cells. CONCLUSION Our studies demonstrated that TMP inhibits the proliferation of colon cancer cells in a dose- and time-dependent manner by inducing apoptosis and arresting the cell cycle at the G0/G1 phase. Our findings suggest that TMP might serve as a potential novel therapeutic drug in the treatment of human colon cancer.

We present ReasonBert, a pre-training method that augments language models with the ability to reason over long-range relations and multiple, possibly hybrid contexts. Unlike existing pre-training methods that only harvest learning signals from local contexts of naturally occurring texts, we propose a generalized notion of distant supervision to automatically connect multiple pieces of text and tables to create pre-training examples that require long-range reasoning. Different types of reasoning are simulated, including intersecting multiple pieces of evidence, bridging from one piece of evidence to another, and detecting unanswerable cases. We conduct a comprehensive evaluation on a variety of extractive question answering datasets ranging from single-hop to multi-hop and from text-only to table-only to hybrid that require various reasoning capabilities and show that ReasonBert achieves remarkable improvement over an array of strong baselines. Few-shot experiments further demonstrate that our pre-training method substantially improves sample efficiency., Accepted to EMNLP'2021. Our code and pre-trained models are available at https://github.com/sunlab-osu/ReasonBERT

Although the potential serious threat of anthropogenic heat on human health was receiving considerable attention worldwide, its long-term health effect on blood pressure (BP) remained unknown. We aimed to evaluate the associations of long-term anthropogenic heat exposure with different components of BP and hypertension.In this cross-sectional study (Liaoning province, China) conducted in 2009, we included a total of 24,845 Chinese adults (18-74 years). We estimated the anthropogenic heat exposure in 2008 using multisource remote sensing images and ancillary data. We measured systolic blood pressure (SBP), diastolic blood pressure (DBP), mean arterial pressure (MAP) and pulse pressure (PP), and defined hypertension. We used generalized linear mixed model to examine the associations.In the adjusted model, the estimates indicated that the difference in SBP, MAP and PP for those in highest quartiles of total anthropogenic heat exposure was greater compared with the lowest quartile (highest quartile: β = 1.11 [95% CI: 0.28-1.94], 0.60 [95% CI: 0.04-1.17], 0.76 [95% CI: 0.17-1.35]). Compared with the lowest quartile, the odds of hypertension were higher among those in higher quartiles (second quartile: OR = 1.17 [95% CI: 1.05-1.30]; third quartile:1.10 [95% CI: 1.1.01-1.21]; highest quartile: 1.17 [95% CI: 1.06-1.28]). These associations were stronger in female participants.Our study showed that long-term exposure to anthropogenic heat was associated with elevated BP and higher odds of hypertension. These findings suggest that mitigation strategies to reduce anthropogenic heat should be considered.

It is well observed that in deep learning and computer vision literature, visual data are always represented in a manually designed coding scheme (eg., RGB images are represented as integers ranging from 0 to 255 for each channel) when they are input to an end-to-end deep neural network (DNN) for any learning task. We boldly question whether the manually designed inputs are good for DNN training for different tasks and study whether the input to a DNN can be optimally learned end-to-end together with learning the weights of the DNN. In this paper, we propose the paradigm of {\em deep collective learning} which aims to learn the weights of DNNs and the inputs to DNNs simultaneously for given tasks. We note that collective learning has been implicitly but widely used in natural language processing while it has almost never been studied in computer vision. Consequently, we propose the lookup vision networks (Lookup-VNets) as a solution to deep collective learning in computer vision. This is achieved by associating each color in each channel with a vector in lookup tables. As learning inputs in computer vision has almost never been studied in the existing literature, we explore several aspects of this question through varieties of experiments on image classification tasks. Experimental results on four benchmark datasets, i.e., CIFAR-10, CIFAR-100, Tiny ImageNet, and ImageNet (ILSVRC2012) have shown several surprising characteristics of Lookup-VNets and have demonstrated the advantages and promise of Lookup-VNets and deep collective learning.

Accumulated evidences suggested that circular RNAs (circRNA) played critical roles in tumorigenesis and progression. To our knowledge, no study reported the function of circular RNA DGKB (circDGKB, circRNA ID: hsa_circ_0133622) on progression of neuroblastoma (NB). Here, we showed that circDGKB was upregulated in NB tissues compared to the normal dorsal root ganglia. Moreover, the expression level of circDGKB was negatively correlated with the survival rate of NB patients. Mechanically, overexpression of circDGKB promoted the proliferation, migration, invasion, and tumorigenesis of NB cells and reduced cell apoptosis, and vice versa. In addition, qRT-PCR and/or Western blot results showed that circDGKB overexpression inhibited the expression level of miR-873 and enhanced GLI1 expression. Moreover, miR-873 functioned an opposite role to circDGKB and significantly weakened circDGKB role in promoting NB progression. Furthermore, GLI1 upregulation also rescued the miR-873 role in inhibiting NB progression. In conclusion, our work proved that circDGKB promoted NB progression via targeting miR-873/GLI1 axis in vitro and in vivo. Our study provided a new target for NB treatment and indicated that circDGKB could act as a novel diagnostic marker for NB.

OBJECTIVE: Comparative analysis of the reduction effect of the evoked nystagmus in the non-affect side during Dix-Hallpike(D-H) or Roll-test in unilateral posterior semicircular canal benign paroxysmal positional vertigo(PC-BPPV) and PC-BPPV without above evoked nystagmus. METHOD: Retrospective analysis of 210 patients diagnosed with unilateral PC-BPPV by G-Force BPPV CRP system was made. Among them, 18 patients exhibited positive nystagmus only when the non-affected side was stimulated by D-H test(Group A), 30 was evoked only when stimulated by Roll-test(Group B), 26 was evoked when stimulated by both Roll-test and the non-affected side D-H test(Group C), 136 without nystagmus in the above positions(Group D). All the patients were diagnosed by G-Force BPPV and were treated through simulative Epley or Semont CRP. Compare the reduction effect among the groups. RESULT: At the first time, the reduction effect of nystagmus in Group D was superior to those in Group A and Group C(P < 0.05). There was no difference between Group D and Group B(P>0.05). The difference between Group A and Group C was also non-significant(P>0.05). The average CRP times of the four groups(CRP until nystagmus disappeared or no longer alleviated) were 1.44±0.63(Group A), 1.46±0.65(Group B), 1.52±0.87(Group C) and 1.48±0.73(Group D)respectively. There were no statistic difference between four groups(P>0.05). The differences of final reduction effect between groups were the same as those at the first time. CONCLUSION: The first time reduction effect was less effective when nystagmus evoked the non-affect side during D-H test in unilateral PC-BPPV, while it might be irrelevant to the nystagmus evoked only in Roll-test. Although the times of CRP were similar among the groups, the final reduction effect of groups with nystagmus evoked the non-affect side during D-H was poorer.

In recent years, the mechanism of cancer research has become hotspots of life science and medicine, especially due to the rapid development of molecular medicine and bioinformatics research. Similarly, the molecular mechanism also has received increasing attention in osteosarcoma (OS) research. Also, a considerable amount of research confirmed that circular RNAs (circRNAs) could regulate cancer cell growth and metastasis. This study aimed to explore the effect of a circRNA, circCCDC66, on OS and reveal its potential molecular mechanism. High circCCDC66 expression level was found in OS patient-derived tissue samples and OS cell lines by qRT-PCR. The abilities cell proliferation and metastatic of U2OS and SW1353 cells were then assessed by Cell Counting Kit-8 and transwell assay, respectively. The interaction between circCCDC66 and its target miRNAs were verified by the dual-luciferase reporter assay. Through functional experiments, we found that circCCDC66 knockdown promoted the inhibition of cell proliferation and metastatic of OS cell lines. From mechanistic perspective, circCCDC66 upregulated PTP1B by sponging miR-338-3p. Collectively, our findings demonstrated that circCCDC66 contributed to malignant behaviors of OS cells by miR-338-3p/PTP1B pathway, which suggested circCCDC66/miR-338-3p/PTP1B axis might be a potential therapeutic target.

Supplemental material, DS_10.1177_0363546519898691 for Effect of Lubricin Mimetics on the Inhibition of Osteoarthritis in a Rat Anterior Cruciate Ligament Transection Model by Daniel Nemirov, Yusuke Nakagawa, Zhexun Sun, Amir Lebaschi, Susumu Wada, Camila Carballo, Xiang-Hua Deng, David Putnam, Lawrence J. Bonassar and Scott A. Rodeo in The American Journal of Sports Medicine

Knowledge distillation (KD) is an effective technique to transfer knowledge from one neural network (teacher) to another (student), thus improving the performance of the student. To make the student better mimic the behavior of the teacher, the existing work focuses on designing different criteria to align their logits or representations. Different from these efforts, we address knowledge distillation from a novel data perspective. We argue that transferring knowledge at sparse training data points cannot enable the student to well capture the local shape of the teacher function. To address this issue, we propose locally linear region knowledge distillation ($\rm L^2$RKD) which transfers the knowledge in local, linear regions from a teacher to a student. This is achieved by enforcing the student to mimic the outputs of the teacher function in local, linear regions. To the end, the student is able to better capture the local shape of the teacher function and thus achieves a better performance. Despite its simplicity, extensive experiments demonstrate that $\rm L^2$RKD is superior to the original KD in many aspects as it outperforms KD and the other state-of-the-art approaches by a large margin, shows robustness and superiority under few-shot settings, and is more compatible with the existing distillation approaches to further improve their performances significantly.

AIM To investigate the predictive factors of lymph node metastasis (LNM) in poorly differentiated early gastric cancer (EGC); to guide the individual application of a combination of endoscopic submucosal dissection (ESD) and laparoscopic lymph node dissection (LLND) in a suitable subgroup of patients with poorly differentiated EGC. METHODS We retrospectively analyzed 138 patients with poorly differentiated EGC who underwent gastrectomy with lymphadenectomy between January 1990 and December 2015. The association between the clinicopathological factors and the presence of LNM was retrospectively analyzed by univariate and multivariate logistic regression analyses. Odds ratios (OR) with 95% confidence interval (95%CI) were calculated. We further examined the relationship between the positive number of the significant predictive factors and the LNM rate. RESULTS The tumor diameter (OR = 13.438, 95%CI: 1.773-25.673, P = 0.029), lymphatic vessel involvement (LVI) (OR = 38.521, 95%CI: 1.975-68.212, P = 0.015) and depth of invasion (OR = 14.981, 95%CI: 1.617-52.844, P = 0.024) were found to be independent risk factors for LNM by multivariate analysis. For the 138 patients diagnosed with poorly differentiated EGC, 21 (15.2%) had LNM. For patients with one, two and three of the risk factors, the LNM rates were 7.7%, 47.6% and 64.3%, respectively. LNM was not found in 77 patients that did not have one or more of the three risk factors. CONCLUSION ESD might be sufficient treatment for intramucosal poorly differentiated EGC if the tumor is less than or equal to 2 cm in size and when LVI is absent upon postoperative histological examination. ESD with LLND may lead to the elimination of unnecessary gastrectomy in poorly differentiated EGC.

Purpose To explore the surgical treatment and predictors of intestinal necrosis in children with intestinal obstruction through analyzing blood biochemical indicators, and to establish a predictive model and evaluate its predictive accuracy, sensitivity and specificity. Methods The data of children with intestinal obstruction hospitalized in Jiangxi Provincial Children's Hospital from January 2014 to June 2019 were retrospectively analyzed. Results Thirty-six substances in the blood of children with successful conservative management and those requiring surgical treatment were significantly different. The model composed of 7 variables, including age, white blood cell count, creatine kinase, troponin I, myoglobin, C-reactive protein and fibrinogen, can be used to predict the unsuccessful conservative management in children with intestinal obstruction, whom need further operation. The average prediction accuracy was 83.50%, the false positive rate was 16.67% (32/192), AUROC is 0.9160 (95% CI, 0.8930–0.9390), and the sensitivity and specificity were 83.20% and 92.70% respectively. A prediction model based on the white blood cell count, creatine kinase, troponin I and myoglobin could predict the occurrence of intestinal necrosis. The average prediction accuracy was 73.70%, false positive rate was 4.49% (15/334), AUROC was 0.7390 (95% CI, 0.6820–0.7960), and the sensitivity and specificity were 71.70% and 64.70%, respectively. Conclusions Combination of age, white blood cell count, creatine kinase, troponin I, myoglobin, C-reactive protein and fibrinogen can be used to predict whether the children with intestinal obstruction need surgical treatment or not. Leukocyte count, creatine kinase, troponin I and myoglobin are closely related to the condition of children with intestinal obstruction and can be used to predict whether intestinal necrosis occurs. Type of Study Retrospective Study Levels of Evidence Level I

// Hua Li 1, * , Shubo Chen 2, * , Hui Li 3, * , Jianxin Cui 4, * , Yunhe Gao 4 , Dianchao Wu 1 , Shangfeng Luan 1 , Yan Qin 1 , Tongshan Zhai 1 , Dengxiang Liu 5 and Zhibin Huo 1 1 Department of Surgical Oncology, Affiliated Xing Tai People Hospital of Hebei Medial University, Xingtai 054001, China 2 Department of Surgical Urology, Affiliated Xing Tai People Hospital of Hebei Medial University, Xingtai 054001, China 3 Department of General Surgery, Affiliated Xing Tai People Hospital of Hebei Medial University, Xingtai 054001, China 4 Department of General Surgery, Chinese People’s Liberation Army General Hospital, Beijing 100853, China 5 Institute of Cancer Control, Affiliated Xing Tai People Hospital of Hebei Medial University, Xingtai 054001, China * These authors have contributed equally to this work Correspondence to: Zhibin Huo, email: huozhibinxt@126.com Dengxiang Liu, email: rmyy666@163.com Keywords: PIK3CA; gastric cancer; prognosis; alteration; meta-analysis Received: August 09, 2017 Accepted: December 26, 2017 Published: January 02, 2018 ABSTRACT Background: Increasing evidence suggests that dysregulation of phosphatidylinositol-4, 5-bisphosphate 3-kinase, catalytic subunit alpha (PIK3CA) plays an important role in carcinogenesis. However, the relationship between PIK3CA expression and gastric cancer (GC) prognosis remains controversial. Methods: We searchedPubMed, Embase, Web of Science, and the Cochrane Library databases for relevant studies up to June 30, 2017. Primary outcomes were hazard ratio (HR), odds ratio (OR), and 95% confidence intervals (CI) for association with overall survival and clinicopathological features. Results: Eleven studies comprising 2481 GC patients were analyzed. Pooled analysis showed that PIK3CA upregulation was significantly associated with worse overall survival (HR = 1.79, 95% CI 1.42–2.27, p < 0.001) at the protein (HR = 1.94, 95% CI 1.52–2.47, p < 0.001) but not the gene (HR = 1.57, 95% CI 0.92–2.69, p = 0.097) level. PIK3CA gene mutation did not correlate with overall survival (HR = 1.05, 95% CI 0.83–1.34, p = 0.666) but was significantly associated with poor tumor differentiation (OR = 0.37, 95% CI 0.17–0.76, p = 0.011). Conclusion: High PIK3CA protein expression predicted poor prognosis in GC, whereas PIK3CA gene amplification or mutation did not. Moreover, PIK3CA mutation was an indicator of poorly differentiated tumors.

We aimed to construct a highly efficient, lentivirus-mediated vector for RNA interference (RNAi) silencing of the Id1 gene and to examine whether silencing of this gene promotes the neural differentiation of mesenchymal stem cells (MSCs). Three short hairpin RNA sequences targeting the coding region of the Id1 gene and one negative control (NC) sequence were designed. After monophosphorylation, these sequences were introduced into lentiviral vectors containing enhanced green fluorescent protein (EGFP) using the restriction enzyme BsmBI to construct the recombinant vectors pWSLV-EGF-shId1 (1-3) and pWSLV-EGF-NC. These recombinant lentivirus vectors were transfected into H9C2 cells to assess the silencing effect of the shId1 (1-3) sequences on the Id1 gene. Subsequently, pWSLV-EGF-NC and pWSLV-EGF-shId1 which exhibited the strongest silencing effect were transfected into the packaging cell line 293 FT. After determining the viral titers, the cultured MSCs were infected with the lentivirus (multiplicity of infection, MOI = 10). Real-time polymerase chain reaction (PCR) and Western blotting were used to examine the amounts of Id1 mRNA and protein, respectively, in the transfected MSCs, and the results revealed that Id1 expression was markedly inhibited. Finally, MSCs were treated with the growth factors EGF and bFGF in conjunction with ligustrazine to induce their differentiation into neuron-like cells. The expression of NSE/MAP-2/Nestin was detected using immunocytochemistry and real-time PCR, which revealed that the rate of neural differentiation in the pWSLV-EGF-shId1 group was higher than that of the other groups, as was NSE/MAP-2 expression. Thus, Id1 gene silencing promotes the neural differentiation of MSCs.