Overcoming the Common Challenges in Differential Gene Expression Analysis Studies

Huang, Yan ProQuest Dissertations & Theses University of Cali 2019 해외박사(DDOD)

소속기관이 구독 중이 아닌 경우 오후 4시부터 익일 오전 9시까지 원문보기가 가능합니다.

The ability to analyze gene expression data has had a fundamental impact in the biological sciences and on our understanding of the causes and mechanisms of disease. However, a significant statistical challenge is posed by the combination of the small number of replicates together with the large number of genes leading to an undesirable level of misclassified genes when identifying genes with differential expression levels. When multiple gene expression data sets are generated under the same set of experimental conditions, the ques- tion arises as to how to efficiently combine this information. Several methods in the literature have been suggested to aggregate ranked data from multiple sources. We introduce a new classifier, underpinned by Bayesian principles, called Peer Reinforced Ranker (PR-Ranker) which uses density estimation to approximate the probability that a gene is differentially expressed given a collection of ranked lists.Our classifier is amenable to theoretical analysis when the number of genes and lists is large using the theory of large deviations. Under modest technical assumptions we show that asymptotically PR-Ranker has the smallest loss of any rank aggregation procedure. Moreover, we prove that other more ad hoc methods, such as Borda, have a strictly higher asymptotic rate of loss.While the theoretical results are asymptotic, we perform a series of simulation studies that demonstrate that our classifier outperforms existing methods on datasets of realistic size for biological data. Furthermore, we show that the outperformance is even greater when the lists exhibit varying levels of noise or when some sources are corrupted. PR-Ranker automatically adapts to varying data quality and efficiently combines the data from different sources. Finally we apply PR-Ranker to a gene expression data set in a preeclampsia study. The top ranked genes identified were known to be biologically relevant to preeclampsia and our method achieved a substantially higher Consistency Index than other rank aggregation procedures.

Combining Theory and Experiment in Pt-Based Catalysts Design for Energy Conversion

Huang, Jin ProQuest Dissertations & Theses University of Cali 2021 해외박사(DDOD)

소속기관이 구독 중이 아닌 경우 오후 4시부터 익일 오전 9시까지 원문보기가 가능합니다.

Investigation of Structure-Mechanics-Property Relations in Heart Valve Tissues

Huang, Siyao North Carolina State University 2014 해외박사(DDOD)

소속기관이 구독 중이 아닌 경우 오후 4시부터 익일 오전 9시까지 원문보기가 가능합니다.

The microstructure and the mechanical property of the heart valve are physically critical factors in modeling the valve tissue equivalent. The heart valve tissue has a complex microstructure primarily containing collagen fibers and valve interstitial cells (VICs). The extracellular matrix (ECM) of the heart valve tissue, where collagen is the main component, provides strength to support the structure of the tissue. The inhomogeneity of the collagen fiber architecture gives rise to the nonlinear anisotropic material property of the tissue. Besides, to respond mechanical loads, VICs mediate a series of bioregulations for ECM remodeling such as homeostasis, collagen synthesis, and collagen degradation through collagen fibers, which further result in changes of the material behavior corresponding to mechanical stimulation. Due to the inhomogeneous architecture of collagen fibers and randomly distributed cell population, the mechanical behavior of the heart valve tissue becomes more complicated during cardiac cycles. Heart valves open and close correctly and persistently during cardiac systole and diastole for blood circulation. It is a widely acknowledged health concern that heart valve diseases lead to defective structures and improper functions of heart valves to directly influence the blood circulation and the heart workload. To date, at least 250,000 people are suffering heart valve diseases in the United States, and the population keeps rising. Studies have indicated that heart valve diseases are caused by the disrupted tissue homeostasis under a variety of pathological conditions, resulting in alterations in heart valve microstructures, mechanical properties, and other biomechanical regulations. Severe collagen depletion is one of disordered tissue remodeling caused by matrix metalloproteinases (MMPs) that pathologically induces matrix destruction, changes the viscoelastic property of the heart valve tissue, further affects cellular regulations mediated by VICs, and even leads to heart valve diseases. With application of collagenase simulating collagen degradation by MMPs, this study focuses on characterizing stress-relaxation behaviors of fresh porcine heart valve tissues and collagenase-treated ones under different stretching conditions. Moreover, the collagen concentration is measured to provide biochemical information related to the mechanical stress relaxation. The results reveal the sensitivity of collagen fibers to proteolytic degradation. The decrease in the stress state of the heart valve tissue is associated with the stretching level and the collagenase concentration. Stress is further decreased after applying collagenase, and a larger stress drop results from a higher strain level and/or a higher collagenase concentration. Therefore, the current study provides important links between several factors: collagen degradation, activities of matrix metalloproteinases, collagen fiber directions, and mechanical stimulations. It is known that heart valves constantly experience different stress states during cardiac cycles; however, how these mechanical stimuli translate into extracellular matrix remodeling, cellular mechanotransduction, cell migration, and collagen synthesis are still unclear. Although the computational simulations of heart valve tissue have been widely studied via the homogenization of collagen fiber distribution or the simplified representation of the highly heterogeneous collagen fiber network excluding the cell population, the matrix-to-cell stress transfer is underestimated. Meanwhile, VIC regulations have been investigated from cells generally isolated from the matrix prior to adhesion molecule characterization; thus, tissue-cell mechanical interactions have not been fully characterized in the native in vivo environment in which they normally operate. To demonstrate heterogeneously distributed collagen fibers responsible for transmitting forces into cells, this study introduces a virtual experiment via an image-based finite element analysis incorporating a histological photomicrograph of a porcine heart valve tissue. Furthermore, the evolution of stress fields at both the tissue and cellular levels is reported to contribute toward refining our collective understanding of valvular tissue micromechanics while a computational tool is provided to further study of valvular cell-matrix interactions.

Modeling Diffusion Induced Stresses for Lithium-Ion Battery Materials

Chiu Huang, Cheng-Kai North Carolina State University 2014 해외박사(DDOD)

소속기관이 구독 중이 아닌 경우 오후 4시부터 익일 오전 9시까지 원문보기가 가능합니다.

Advancing lithium-ion battery technology is of paramount importance for satisfying the energy storage needs in the U.S., especially for the application in the electric vehicle industry. To provide a better acceleration for electric vehicles, a fast and repeatable discharging rate is required. However, particle fractures and capacity loss have been reported under high current rate (C-rate) during charging/discharging and after a period of cycling. During charging and discharging, lithium ions extract from and intercalate into electrode materials accompanied with the volume change and phase transition between Li-rich phase and Li-poor phase. It is suggested that the diffusion-induced-stress is one of the main reasons causing capacity loss due to the mechanical degradation of electrode particles. Therefore, there is a fundamental need to provide a mechanistic understanding by considering the structure-mechanics-property interactions in lithium-ion battery materials. Among many cathode materials, the olivine-based lithium-iron-phosphate (LiFePO4) with an orthorhombic crystal structure is one of the promising cathode materials for the application in electric vehicles. In this research we first use a multiphysic approach to investigate the stress evolution, especially on the phase boundary during lithiation in single LiFePO4 particles. A diffusion-controlled finite element model accompanied with the experimentally observed phase boundary propagation is developed via a finite element package, ANSYS, in which lithium ion concentration-dependent anisotropic material properties and volume misfits are incorporated. The stress components on the phase boundary are used to explain the Mode I, Mode II, and Mode III fracture propensities in LiFePO4 particles. The elastic strain energy evolution is also discussed to explain why a layer-by-layer lithium insertion mechanism (i.e. first-order phase transformation) is energetically preferred. Another importation issue is how current rate (C-rate) during charging/discharging affects diffusion induced stresses inside electrode materials. For the experimental part we first conduct charging/discharging under different C-rates to observe the voltage responses for commercial LiFePO4 batteries. Then Time-of-Flight Secondary Ion Mass Spectrometry technique is applied to measure the lithium ion intensities in different C-rate charged/discharged samples. These experimental results could be used to support that a more significant voltage fluctuation under high C-rates is due to different lithium insertion mechanisms, rather than the amount of lithium ions intercalated into electrode materials. Thus the investigation of C-rate-dependent stress evolution is required for the development of a more durable lithium ion battery. In this dissertation, we extend the single particle finite element model to investigate the C-rate-dependent diffusion induced stresses in a multi-particle system. Concentration dependent anisotropic material properties, C-rate-dependent volume misfits and concentration dependent Li-ion diffusivity are incorporated in the model. The concentration gradients, diffusion induced stresses, and strain energies under different C-rates are discussed in this study. Particle fractures have been observed in many experimental results, in this study we further discuss the effect of the crack surface orientation on the lithium concentration profile and stress level in cathode materials. The results of this dissertation provide a better understanding of diffusion induced stresses in electrode materials and contribute to our fundamental knowledge of interplay between lithium intercalations, stress evolutions, particle fractures and the capacity fade in lithium-ion batteries.

Gold nanorod-based assemblies and composites cancer therapeutics, sensors and tissue engineering materials

Huang, Huang-Chiao ProQuest LLC. 2012 해외석사

High-dimensional gene expression classification and genome-wide association studies of complex traits

Huang, Song Yale University 2010 해외박사(DDOD)

소속기관이 구독 중이 아닌 경우 오후 4시부터 익일 오전 9시까지 원문보기가 가능합니다.

The availability of high-throughput technologies has greatly advanced biomedical research. The resulting high-dimensional data require the development of novel computational and statistical methodologies. In this dissertation, I present such methods in Statistical Genetics, Bioinformatics and the combination of the two fields, i.e., "Genetical Genomics", to aid high-dimensional genetic and gene expression data analyses. The dissertation is organized as follows. In Chapter 1, I introduce related background in genetic and gene expression data analyses. Chapter 2 focuses on gene mapping in animals. I introduce a more powerful statistical approach that efficiently utilizes available pedigree information in mapping complex traits in heterogenous stock of mice (Huang and Zhao, 2010). The main focus of Chapter 3 is gene expression data analysis. I propose bias-corrected diagonal discriminant rules (BDDRs) for high-dimensional gene expression classification and analytically show that BDDRs outperform the existing methods (Huang et al., 2010). In Chapter 4, I present an expression quantitative trait loci mapping study (Huang et al., 2007). Finally, I discuss the current challenges in the related fields and outline future research directions in Chapter 5.

Artificial Intelligence Open-Domain News Event Extraction Method Based on BERT Huang Hairui

HUANG HAIRUI 아주대학교 일반대학원 2024 국내석사

News is an important way for people to obtain information, and in the open news environment, the type of news is increasingly diversified and the scale of news is huge, which causes problems such as information overload and redundancy. The open-domain event extraction task aims to identify and extract various types of event information from predefined text, The task is usually based on methods such as pre-training or neural topic modeling. However, there are a number of problems with existing methods. First, Existing pre-trained models suffer from insufficient feature vector extraction and excessively high embedding dimensions. Second, Existing methods are not rich enough in semantics and lack syntactic structural information, resulting in poor readability of results and insufficient extraction accuracy. Therefore, to address these issues, this paper first improves the open-domain event extraction method based on the neural topic model of BERT, and then dynamically in- tegrates semantic and syntactic dependency information to obtain rich semantic and syn- tactic features, in order to further improve the model performance. The main research is as follows: Proposed an improvement method of neural topic modeling based on BERT. First, BERT is used in the coding layer for pre-training to obtain the contextual representation of the feature sequences. Second, the Umap dimensionality reduction method is used to obtain more extensive local and global information, and the joint distribution of variables is combined with the deep hidden variable probabilistic graph model to further optimize the parameter inference learning process. Finally, the self-attention mechanism is introduced to assign weights to different nodes to reduce the influence of noisy data, so that the model can pay attention to the more critical features, and further improve the performance of the open-domain event extraction model. Keywords: Event extraction, Open-domain event extraction, Neural topic model.

Cross-regional study of Employees' Work Goals in China

Huang, Tai Kangwon National University 2011 국내석사

This article investigates the work goals of Chinese workers. The questionnaire used to analyze work goal difference contains 11 items, which are divided into two categories, low level goals and high level goals. Five clusters are formed and geographical division map of China according to work goal discrepancy are drawn. Differences among age and gender categories are explored. The results show first, “variety” is the least important work goal identified among Chinese workers. Second, a distinct decreasing trend is observed for low level goals among age categories. Third, generally no gender differences are observed except for goals such as “learn”, “relations” and “job security”. Managerial implications are given for those who are going to develop a business relationship with Chinese while not familiar with the work goal diversity of them.

Polybrominated Diphenyl Ethers, Thyroid Hormones, and Risk of Papillary Thyroid Cancer

Huang, Huang ProQuest Dissertations & Theses Yale University 2021 해외박사(DDOD)

소속기관이 구독 중이 아닌 경우 오후 4시부터 익일 오전 9시까지 원문보기가 가능합니다.

The maintenance ability and Ca2+ availability of skeletal muscle are enhanced by sildenafil

Huang, Mei 가톨릭대학교 대학원 2017 국내박사