Search

Supplemental Material, Table_S1_and_S2 for A Logistic Regression Model for Noninvasive Prediction of AFP-Negative Hepatocellular Carcinoma by Chang-Liang Luo, Yuan Rong, Hao Chen, Wu-Wen Zhang, Long Wu, Diao Wei, Xiu-Qi Wei, Lie-Jun Mei and Fu-Bing Wang in Technology in Cancer Research & Treatment

Deep convolutional neural networks (CNNs) have been widely used to obtain high-level representation in various computer vision tasks. However, for the task of remote scene classification, there are no sufficient images to train a very deep CNN from scratch. Instead, transferring successful pre-trained deep CNNs to remote sensing tasks provides an effective solution. Firstly, from the viewpoint of generalization power, we try to find whether deep CNNs need to be deep when applied for remote scene classification. Then, the pre-trained deep CNNs with fixed parameters are transferred for remote scene classification, which solve the problem of time-consuming and parameters over-fitting at the same time. With five well-known pre-trained deep CNNs, experimental results on three independent remote sensing datasets demonstrate that transferred deep CNNs can achieve state-of-the-art results in unsupervised setting. This chapter also provides baseline for applying deep CNNs to other remote sensing tasks.

Chang-Liang Luo,1,* Zhi-Gao Xu,2,* Hao Chen,2 Jia Ji,1 Yu-Hui Wang,1 Wei Hu,1 Kun Wang,3 Wu-Wen Zhang,1 Chun-Hui Yuan,4 Fu-Bing Wang1 1Department of Laboratory Medicine, Zhongnan Hospital of Wuhan University, Wuhan, China; 2Department of Pathology, Zhongnan Hospital of Wuhan University, Wuhan, China; 3Department of Laboratory Medicine, Hubei Cancer Hospital, Wuhan, China; 4Department of Laboratory Medicine, Wuhan Children’s Hospital, Huazhong University of Science and Technology, Wuhan, China *These authors contributed equally to this work Background: Tissue biopsy-based cancer diagnosis has limitations because of the fact that tumor tissues are in constant evolution and extremely heterogeneous. The current study was aimed to examine whether tumor-educated blood platelets (TEPs) might be a potential all-in-one source for blood-based cancer diagnostics to overcome the limitations of conventional cancer biopsy.Methods: In the present study, we evaluated the expression pattern of MAGI2 antisense RNA 3 (MAGI2-AS3) and ZNFX1 antisense RNA 1 (ZFAS1) in both plasma and platelets of 101 non-small-cell lung cancer (NSCLC) patients. Receiver operating characteristic (ROC) curve was generated to evaluate their diagnostic potential. In addition, epidermal growth factor receptor (EGFR) mutations were detected in DNA and RNA samples of platelets for companion diagnostics.Results: Our results showed that the levels of MAGI2-AS3 and ZFAS1 in both plasma and platelets of NSCLC patients were significantly downregulated than those in healthy controls. A positive correlation of long noncoding RNA expression was observed between platelets and plasma (r=0.738 for MAGI2-AS3, r=0.751 for ZFAS1, respectively). By ROC analysis, we found that molecular interrogation of MAGI2-AS3 and ZFAS1 in TEPs and plasma can offer valuable diagnostic performance for NSCLC patients (area under the ROC curve [AUC]MAGI2-AS3=0.853/0.892, and AUCZFAS1=0.780/0.744 for diagnosing adenocarcinoma and squamous cell carcinoma cases from controls, respectively). Clinicopathologic characteristic analysis further revealed that MAGI2-AS3 level significantly correlated with tumor–node–metastasis (TNM) stage (p=0.001 in TEPs, p=0.003 in plasma), lymph-node metastasis (p=0.016 in TEPs, p=0.023 in plasma), and distant metastasis (p=0.045 in TEPs, p=0.045 in plasma), while ZFAS1 level was only correlated with TNM stage (p=0.005 in TEPs, p=0.044 in plasma). Furthermore, EGFRvIII RNA existed in both TEPs and plasma, but EGFR intracellular mutations cannot be detected in DNA of TEPs isolated from NSCLC.Conclusion: Our data suggested that TEP is a promising source for NSCLC diagnosis and companion diagnostics. Keywords: tumor-educated platelets, lncRNA-MAGI2-AS3, lncRNA-ZFAS1, EGFR mutations, NSCLC

Considering the asymmetric dependency structure and regime switching process, we construct the dynamic Markov Regime Switching Copula (MRS-Copula) models to measure the financial risk contagion. The dynamic MRS-Copula models consist of the marginal model and dynamic MRS Rotated-Gumbel function, and they are examined by the goodness-of-fit test method. Using the dynamic MRS-Copula models, we calculate the daily lower tail dependency by adopting the international stock market index data from January 1997 to June 2015, and provide evidence of financial risk contagion effects between Chinese stock market and other international stock markets after the reform of the RMB exchange rate system in China, and this is particular the case after in the U.S. sub-prime mortgage crisis and the European debt crisis. When conducting robust tests with weekly and monthly data, the empirical result basically holds. As for the financial risk contagion channels in Chinese stock market, the fundamental economic linkages play a more important role than liquidity, information and other factors.

SONG (Stellar Observations Network Group) is an international collaboration project involving many countries. Chinese stellar physics community is participating the project as a national member, contributing a full set of instruments and a site. The key instrument is a high resolution spectrograph which, when adopting iodine cell for wave length calibration, can deliver radial velocity measurements at an accuracy of 1 m•s-1. Such a high accuracy together with the long term stability is the keys to the science goals of SONG, one of which is to study the internal structure of stars directly using the technique of Asteroseismology. In this paper, we are going to present the design, parameters and performance of the spectrograph by using spectra taken during the commissioning of the system. As long term thermo-mechanic stability of the system is extremely important for the science goals, we will also describe the laboratory environments to realize such a condition.

Many studies have confirmed that the role of online learning collaboration group is very important. For large-scale online learning, how to effectively find the learning collaboration group is a difficult problem. The online learning forum is the main place for learners to learn and communicate, so it is the main venue for learning collaboration groups to implement collaborative learning. In the implementation process of learning collaboration, there are two characteristics between learning team members. First, there is interaction between them. Second, the contents of their discussion have high relevance. In this paper, the study uses these two important characteristics and carries out two-stage clustering algorithm based on the interaction structure and interactive contents of learners to find the potential learning collaboration groups in the large-scale online learning forum. The experimental results show that the method proposed in this paper is effective. This has practical significance for large-scale online learning support services.