File-level malware detection using byte streams.

As more documents appear on the Internet, it becomes important to detect malware within the documents. Malware of non-executables might be more dangerous because people usually open them without worrying about inherent danger. Recently, deep learning models are used to analyze byte streams of the non-executables for malware detection. Although they have shown successful results, they are commonly designed for stream-level detection, but not for file-level detection. In this paper, we propose a new method that aggregates the stream-level results to get file-level results for malware detection. We demonstrate its effectiveness by experimental results with our annotated dataset, and show that it gives performance gain of 3.37-5.89% of F1 scores.

Short-term carcinogenicity study of N-methyl-N-nitrosourea in FVB-Trp53 heterozygous mice.

Carcinogenicity tests predict the tumorigenic potential of various substances in the human body by studying tumor induction in experimental animals. There is a need for studies that explore the use of FVB/N-Trp53em2Hwl/Korl (FVB-Trp53+/-) mice, created by TALEN-mediated gene targeting in Korea, in carcinogenicity tests. This study was performed to determine whether FVB-Trp53+/- mice are a suitable model for short-term carcinogenicity studies. To compare the carcinogenicity at different concentrations, 25, 50, and 75 mg/kg of N-methyl-N-nitrosourea (MNU), a known carcinogen, were administered intraperitoneally to FVB-Trp53+/- and wild-type male mice. After 26 weeks, the survival rate was significantly reduced in FVB-Trp53+/- mice compared to the wild-type mice in the 50 and 75 mg/kg groups. The incidence of thymic malignant lymphoma (TML) in the 50 and 75 mg/kg groups was 54.2 and 59.1% in FVB-Trp53+/- male mice, respectively. TML metastasized to the lungs, spleen, lymph nodes, liver, kidney, and heart in FVB-Trp53+/- male mice. Furthermore, the incidence of primary lung tumors, such as adenomas and adenocarcinomas, was 65.4, 62.5, and 45.4% in the FVB-Trp53+/- mice of the 25, 50, and 75 mg/kg groups, respectively. The main tumor types in FVB-Trp53+/- mice were TML and primary lung tumors, regardless of the dose of MNU administered. These results suggest that systemic tumors may result from malfunctions in the p53 gene and pathway, which is an important factor in the pathogenesis of human cancers. Therefore, FVB-Trp53 heterozygous mice are suitable for short-term carcinogenicity tests using positive carcinogens, and that the best result using MNU, a positive carcinogen, might have a single dose of 50 mg/kg.

Safety of Human Embryonic Stem Cell-derived Mesenchymal Stem Cells for Treating Interstitial Cystitis: A Phase I Study.

There are still no definite treatment modalities for interstitial cystitis (IC). Meanwhile, stem cell therapy is rising as potential alternative for various chronic diseases. This study aimed to investigate the safety of the clinical-grade mesenchymal stem cells (MSCs) derived from human embryonic stem cells (hESCs), code name MR-MC-01 (SNU42-MMSCs), in IC patients. Three female IC patients with (1) symptom duration >6 months, (2) visual pain analog scale (VAS) ≥4, and (3) one or two Hunner lesions <2 cm in-office cystoscopy within 1 month were included. Under general anesthesia, participants received cystoscopic submucosal injection of SNU42-MMSCs (2.0 × 107/5 mL) at the center or margin of Hunner lesions and other parts of the bladder wall except trigone with each injection volume of 1 mL. Follow-up was 1, 3, 6, 9, and 12 months postoperatively. Patients underwent scheduled follow-ups, and symptoms were evaluated with validated questionnaires at each visit. No SNU42-MMSCs-related adverse events including immune reaction and abnormalities on laboratory tests and image examinations were reported up to 12-month follow-up. VAS pain was temporarily improved in all subjects. No de novo Hunner lesions were observed and one lesion of the first subject was not identifiable on 12-month cystoscopy. This study reports the first clinical application of transurethral hESC-derived MSC injection in three patients with IC. hESC-based therapeutics was safe and proved to have potential therapeutic efficacy in IC patients. Stem cell therapy could be a potential therapeutic option for treating IC.

Development of transgenic models susceptible and resistant to SARS-CoV-2 infection in FVB background mice.

Coronavirus disease (COVID-19), caused by Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), is currently spreading globally. To overcome the COVID-19 pandemic, preclinical evaluations of vaccines and therapeutics using K18-hACE2 and CAG-hACE2 transgenic mice are ongoing. However, a comparative study on SARS-CoV-2 infection between K18-hACE2 and CAG-hACE2 mice has not been published. In this study, we compared the susceptibility and resistance to SARS-CoV-2 infection between two strains of transgenic mice, which were generated in FVB background mice. K18-hACE2 mice exhibited severe weight loss with definitive lethality, but CAG-hACE2 mice survived; and differences were observed in the lung, spleen, cerebrum, cerebellum, and small intestine. A higher viral titer was detected in the lungs, cerebrums, and cerebellums of K18-hACE2 mice than in the lungs of CAG-hACE2 mice. Severe pneumonia was observed in histopathological findings in K18-hACE2, and mild pneumonia was observed in CAG-hACE2. Atrophy of the splenic white pulp and reduction of spleen weight was observed, and hyperplasia of goblet cells with villi atrophy of the small intestine was observed in K18-hACE2 mice compared to CAG-hACE2 mice. These results indicate that K18-hACE2 mice are relatively susceptible to SARS-CoV-2 and that CAG-hACE2 mice are resistant to SARS-CoV-2. Based on these lineage-specific sensitivities, we suggest that K18-hACE2 mouse is suitable for highly susceptible model of SARS-CoV-2, and CAG-hACE2 mouse is suitable for mild susceptible model of SARS-CoV-2 infection.

Prediction of Postoperative Complications for Patients of End Stage Renal Disease.

End stage renal disease (ESRD) is the last stage of chronic kidney disease that requires dialysis or a kidney transplant to survive. Many studies reported a higher risk of mortality in ESRD patients compared with patients without ESRD. In this paper, we develop a model to predict postoperative complications, major cardiac event, for patients who underwent any type of surgery. We compare several widely-used machine learning models through experiments with our collected data yellow of size 3220, and achieved F1 score of 0.797 with the random forest model. Based on experimental results, we found that features related to operation (e.g., anesthesia time, operation time, crystal, and colloid) have the biggest impact on model performance, and also found the best combination of features. We believe that this study will allow physicians to provide more appropriate therapy to the ESRD patients by providing information on potential postoperative complications.

Developing an Individual Glucose Prediction Model Using Recurrent Neural Network.

In this study, we propose a personalized glucose prediction model using deep learning for hospitalized patients who experience Type-2 diabetes. We aim for our model to assist the medical personnel who check the blood glucose and control the amount of insulin doses. Herein, we employed a deep learning algorithm, especially a recurrent neural network (RNN), that consists of a sequence processing layer and a classification layer for the glucose prediction. We tested a simple RNN, gated recurrent unit (GRU), and long-short term memory (LSTM) and varied the architectures to determine the one with the best performance. For that, we collected data for a week using a continuous glucose monitoring device. Type-2 inpatients are usually experiencing bad health conditions and have a high variability of glucose level. However, there are few studies on the Type-2 glucose prediction model while many studies performed on Type-1 glucose prediction. This work has a contribution in that the proposed model exhibits a comparative performance to previous works on Type-1 patients. For 20 in-hospital patients, we achieved an average root mean squared error (RMSE) of 21.5 and an Mean absolute percentage error (MAPE) of 11.1%. The GRU with a single RNN layer and two dense layers was found to be sufficient to predict the glucose level. Moreover, to build a personalized model, at most, 50% of data are required for training.

Malware Detection of Hangul Word Processor Files Using Spatial Pyramid Average Pooling.

Malware detection of non-executables has recently been drawing much attention because ordinary users are vulnerable to such malware. Hangul Word Processor (HWP) is software for editing non-executable text files and is widely used in South Korea. New malware for HWP files continues to appear because of the circumstances between South Korea and North Korea. There have been various studies to solve this problem, but most of them are limited because they require a large amount of effort to define features based on expert knowledge. In this study, we designed a convolutional neural network to detect malware within HWP files. Our proposed model takes a raw byte stream as input and predicts whether it contains malicious actions or not. To incorporate highly variable lengths of HWP byte streams, we propose a new padding method and a spatial pyramid average pooling layer. We experimentally demonstrate that our model is not only effective, but also efficient.

Comparative Analysis on Machine Learning and Deep Learning to Predict Post-Induction Hypotension.

Hypotensive events in the initial stage of anesthesia can cause serious complications in the patients after surgery, which could be fatal. In this study, we intended to predict hypotension after tracheal intubation using machine learning and deep learning techniques after intubation one minute in advance. Meta learning models, such as random forest, extreme gradient boosting (Xgboost), and deep learning models, especially the convolutional neural network (CNN) model and the deep neural network (DNN), were trained to predict hypotension occurring between tracheal intubation and incision, using data from four minutes to one minute before tracheal intubation. Vital records and electronic health records (EHR) for 282 of 319 patients who underwent laparoscopic cholecystectomy from October 2018 to July 2019 were collected. Among the 282 patients, 151 developed post-induction hypotension. Our experiments had two scenarios: using raw vital records and feature engineering on vital records. The experiments on raw data showed that CNN had the best accuracy of 72.63%, followed by random forest (70.32%) and Xgboost (64.6%). The experiments on feature engineering showed that random forest combined with feature selection had the best accuracy of 74.89%, while CNN had a lower accuracy of 68.95% than that of the experiment on raw data. Our study is an extension of previous studies to detect hypotension before intubation with a one-minute advance. To improve accuracy, we built a model using state-of-art algorithms. We found that CNN had a good performance, but that random forest had a better performance when combined with feature selection. In addition, we found that the examination period (data period) is also important.

Development of a prediction model for hypotension after induction of anesthesia using machine learning.

Arterial hypotension during the early phase of anesthesia can lead to adverse outcomes such as a prolonged postoperative stay or even death. Predicting hypotension during anesthesia induction is complicated by its diverse causes. We investigated the feasibility of developing a machine-learning model to predict postinduction hypotension. Naïve Bayes, logistic regression, random forest, and artificial neural network models were trained to predict postinduction hypotension, occurring between tracheal intubation and incision, using data for the period from between the start of anesthesia induction and immediately before tracheal intubation obtained from an anesthesia monitor, a drug administration infusion pump, an anesthesia machine, and from patients' demographics, together with preexisting disease information from electronic health records. Among 222 patients, 126 developed postinduction hypotension. The random-forest model showed the best performance, with an area under the receiver operating characteristic curve of 0.842 (95% confidence interval [CI]: 0.736-0.948). This was higher than that for the Naïve Bayes (0.778; 95% CI: 0.65-0.898), logistic regression (0.756; 95% CI: 0.630-0.881), and artificial-neural-network (0.760; 95% CI: 0.640-0.880) models. The most important features affecting the accuracy of machine-learning prediction were a patient's lowest systolic blood pressure, lowest mean blood pressure, and mean systolic blood pressure before tracheal intubation. We found that machine-learning models using data obtained from various anesthesia machines between the start of anesthesia induction and immediately before tracheal intubation can predict hypotension occurring during the period between tracheal intubation and incision.

Systematic Proteogenomic Approach To Exploring a Novel Function for NHERF1 in Human Reproductive Disorder: Lessons for Exploring Missing Proteins.

One of the major goals of the Chromosome-Centric Human Proteome Project (C-HPP) is to fill the knowledge gaps between human genomic information and the corresponding proteomic information. These gaps are due to "missing" proteins (MPs)-predicted proteins with insufficient evidence from mass spectrometry (MS), biochemical, structural, or antibody analyses-that currently account for 2579 of the 19587 predicted human proteins (neXtProt, 2017-01). We address some of the lessons learned from the inconsistent annotations of missing proteins in databases (DB) and demonstrate a systematic proteogenomic approach designed to explore a potential new function of a known protein. To illustrate a cautious and strategic approach for characterization of novel function in vitro and in vivo, we present the case of Na(+)/H(+) exchange regulatory cofactor 1 (NHERF1/SLC9A3R1, located at chromosome 17q25.1; hereafter NHERF1), which was mistakenly labeled as an MP in one DB (Global Proteome Machine Database; GPMDB, 2011-09 release) but was well known in another public DB and in the literature. As a first step, NHERF1 was determined by MS and immunoblotting for its molecular identity. We next investigated the potential new function of NHERF1 by carrying out the quantitative MS profiling of placental trophoblasts (PXD004723) and functional study of cytotrophoblast JEG-3 cells. We found that NHERF1 was associated with trophoblast differentiation and motility. To validate this newly found cellular function of NHERF1, we used the Caenorhabditis elegans mutant of nrfl-1 (a nematode ortholog of NHERF1), which exhibits a protruding vulva (Pvl) and egg-laying-defective phenotype, and performed genetic complementation work. The nrfl-1 mutant was almost fully rescued by the transfection of the recombinant transgenic construct that contained human NHERF1. These results suggest that NHERF1 could have a previously unknown function in pregnancy and in the development of human embryos. Our study outlines a stepwise experimental platform to explore new functions of ambiguously denoted candidate proteins and scrutinizes the mandated DB search for the selection of MPs to study in the future.

Multimodal game bot detection using user behavioral characteristics.

As the online service industry has continued to grow, illegal activities in the online world have drastically increased and become more diverse. Most illegal activities occur continuously because cyber assets, such as game items and cyber money in online games, can be monetized into real currency. The aim of this study is to detect game bots in a massively multiplayer online role playing game (MMORPG). We observed the behavioral characteristics of game bots and found that they execute repetitive tasks associated with gold farming and real money trading. We propose a game bot detection method based on user behavioral characteristics. The method of this paper was applied to real data provided by a major MMORPG company. Detection accuracy rate increased to 96.06 % on the banned account list.

Analysis of context dependence in social interaction networks of a massively multiplayer online role-playing game.

Rapid advances in modern computing and information technology have enabled millions of people to interact online via various social network and gaming services. The widespread adoption of such online services have made possible analysis of large-scale archival data containing detailed human interactions, presenting a very promising opportunity to understand the rich and complex human behavior. In collaboration with a leading global provider of Massively Multiplayer Online Role-Playing Games (MMORPGs), here we present a network science-based analysis of the interplay between distinct types of user interaction networks in the virtual world. We find that their properties depend critically on the nature of the context-interdependence of the interactions, highlighting the complex and multilayered nature of human interactions, a robust understanding of which we believe may prove instrumental in the designing of more realistic future virtual arenas as well as provide novel insights to the science of collective human behavior.