GIST Scholar
Not a member yet
30271 research outputs found
Sort by
Advancement of Urea Elimination Using Pre-Halogenation Processes in UV254/Bromine and UV254/Chlorine Systems
No full textDevelopment of Simplified Assay for the Detection of Agmatine to Assess Arginine Decarboxylase Activity
No full textArginine deprivation therapy represents a promising strategy for targeting arginine-auxotrophic cancers, exploiting the metabolic vulnerability of tumor cells deficient in argininosuccinate synthetase (ASS1) and argininosuccinate lyase (ASL). Among arginine-depleting enzymes, arginine decarboxylase (RDC) offers a unique advantage due to its pH-dependent activation in acidic tumor microenvironments. However, the relatively low enzymatic activity of RDC compared to other arginine-depleting enzymes limits its therapeutic potential. Directed evolution has shown promise in enhancing enzyme activity, but its application to RDC has been constrained by the lack of a robust, high-throughput screening (HTS) assay tailored to measure RDC activity. This study addresses this limitation by developing an innovative enzymatic cascade assay that uses diamine oxidase (DAO) and horseradish peroxidase (HRP) in a cascade reaction with TMB as the substrate. This system allows for rapid, sensitive, and high-throughput quantification of RDC activity. The assay was successfully validated with RDC wild-type (RDC-WT) and mutant (RDC-T39W) forms, demonstrating its ability to detect activity differences in both purified enzymes and cell lysates. By eliminating separation steps and streamlining the workflow, this mix-and-measure assay significantly simplifies RDC activity screening and paves the way for directed evolution to optimize RDC functionality. The establishment of this HTS-compatible assay marks a critical step toward enhancing RDC's therapeutic potential as a tumor-targeted arginine depletion enzyme. These findings contribute to the broader development of innovative enzymatic therapies for cancer treatment.MasterAbstract
Contents
List of Figures
I. Introduction
II. Materials and Methods
2.1. Materials
2.2. Preparation of RDC variants
2.3. Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis (SDS-PAGE) Analysis
2.4. Matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass analysis
2.5. RDC variants enzymatic activity measurement through butanol separation assay
2.6. H2O2 calibration of DAO-HRP enzymatic cascade assay
2.7. Agmatine calibration of DAO-HRP enzymatic cascade assay
2.8. Comparative evaluation of RDC activity measurement methods
2.9. RDC variants enzymatic activity measurement through DAO-HRP enzymatic cascade assay
2.10. RDC variants enzymatic activity measurement through DAO-HRP enzymatic cascade assay in crude lysate
III. Results and Discussion
3.1. Preparation of RDC variants
3.2. H2O2 quantification by HRP
3.3. Agmatine quantification by DAO-HRP
3.4. RDC reaction product quantification using DAO-HRP enzymatic assay
3.5. Enzymatic activity comparison of RDC mutants using DAO-HRP assay
3.6. Activity comparison of RDC mutants in crude lysates using DAO-HRP assay
IV. Conclusions
Summary
Reference
Acknowledgemen
Vitrimer with Dynamic imine bond based Self-healing Solid Polymer Electrolyte for dendrite-free Lithium metal batteries
No full textAll-solid-state lithium (Li) metal batteries have garnered candidates as advanced electric vehicles and energy storage systems due to their high energy density and theoretical capacity (3680 mAh g⁻¹). However, Li metal anodes suffer from uncontrolled dendrite growth during the charge-discharge process, and electrolyte decomposition because large volume changes generate unstable electrode|electrolyte interfaces. As a result, the consumption of electrolytes would increase the fracture between electrolytes and active materials, which causes solid-state electrolyte cracking. Vitrimers possess dynamic covalent adaptable networks (CANs), which can reform their bond under external stimuli. Consequently, the crosslink density of materials is constant when the temperature sharply increases, and their self-healing properties enhance stable interface contact. Herein, we present self-healing vitrimer solid polymer electrolytes (SPEs) with dynamic imine bonds via catechol imine ligand
In-situ electrochemical surface engineering of electrodes for the advanced Li metal batteries
No full textThe growing concern of climate change have led to the worldwide investigation of renewable energy sources and energy storage system which play important role for carbon neutrality. Driven by such paradigm shift from fossil fuel to renewable energy, the emerging market growth of the electric vehicles (EVs) and energy storage systems (ESS) occurred. This rapid market growth led to the accelerated exploration of energy storage system with high power and energy density. Among various energy storage systems, Li-ion batteries (LIBs) are widely used battery system which has been already commercialized since 1991. However, state-of-art LiBs utilizing graphite as anode have almost reached its theoretical energy density limit due to material limitation, still insufficient to meet the growing energy demands. To overcome this challenge, replacing the graphite anode with a lithium metal anode (LMA) offers a promising solution, as LMA provides about ten times higher theoretical capacity (3,860 mAh g-1) and operates at a lower potential (-3.04 V vs. S.H.E.) compared to graphite (372 mAh g-1, -2.84 V vs. S.H.E.). Consequently, lithium metal batteries (LMBs) incorporating LMAs are regarded as the "holy grail" of battery systems, with the potential to deliver twice the energy density of conventional LiBs. However, practical application of LMBs are hindered by uncontrollable Li dendritic growth and formation of fragile native solid electrolyte interphase (SEI) layer. Li dendrite is provoked by inhomogeneous initial nucleation of Li and further deposition due to the lithiophobic nature of current collector. Moreover, fragile native SEI layer could be broken during dendrite formation which can proliferate dendrite growth onto the crack, by acting as ‘hot spot’. As grown Li dendrite could penetrate the separator to lead ‘short circuit’ leading to cell explosion. Moreover, Li dendrite can break the fragile nature solid electrolyte interphase (SEI) layer and results in rapid consumption of electrolyte. Even more, such dendritic Li can be electrically isolated to become ‘dead Li’ during the Li deposition/stripping process. Those side reactions lead to safety issues and rapid failure of LMBs, hence suppression of Li dendrite and reinforcement of native SEI layer is urgent issue in this society. To suppress Li dendrite growth and reinforcement of native SEI layer, in this thesis, various surface engineering strategies of electrode (current collector) will be suggested. Specifically, surface modification approaches based on the in-situ electrochemical methods to form favorable interface for Li deposition/stripping are intensively studied. In-situ electrochemical surface engineering has several advantages as follows; 1) it is cost-effective over other methods as it only requires electrochemical signals (current, potential) for surface treatment, 2) we can design the electrode surface by controlling the redox reaction of electrolyte component, 3) it is universal method for various electrolyte systems and electrodes. The following contents will be presented throughout this thesis. 1. In the introduction section (section 1), the brief review of Li metal battery systems, problems of those systems and literature survey of current solutions will be presented. 2. In the section 2, a topic related to the formation of artificial SEI layer (ASEI) layer by electrochemical method will be presented. In particular, this topic will present the in-situ method to form inorganic-rich ASEI layer on current collector (CC) via catalyzing reduction of certain electrolyte additive, and mechanism of catalyzed reduction will be briefly presented. 3. In the section 3, surface modification method introducing lithiophilic materials on CC by in- situ method will be presented. Specifically, the one-pot fabrication method of the lithiophilic LMA which unifying surface treatment and lithiation via in-situ oxidation reaction of functional additive will be presented. 4. In the last section, in-situ surface modification method in the anode-free Li metal batteries (AFLMBs) will be presented. This topic will present the electrochemical pretreatment method to form lithiophilic materials and robust SEI on the CC in the working AFLMBs.DoctorAbstract i
Contents iii
1. Introduction 1
2. A study on the in-situ electrochemical formation of inorganic-rich artificial solid electrolyte
interphase (ASEI) layer for stable Li metal anode 11
2.1. Background 11
2.2. Experimental section 13
2.3. Results & discussion 15
2.4. Conclusion 35
3. A study on the one-pot fabrication method of lithiophilic Li metal anode by in-situ electrochemical oxidation and Li electroplating 40
3.1. Introduction 40
3.2. Experimental section 42
3.3. Results & discussion 45
3.4. Conclusions 74
4. A study on the in-situ lithiophilic surface engineering for the anode-free Li metal battery 79
4.1. Introduction 79
4.2. Experimental section 81
4.3. Results & discussion 84
4.4. Conclusion 97
5. Conclusion 10
Quantum Approximate Metaheuristic Optimization: Formulation, Development, and Examination
No full text신뢰성 있고 실용적인 양자 알고리즘을 개발하는 일이 어렵고 불확실한 과정임은 일전의 여러 연구에 의해 언급되어 왔다. 이에 대해 본 연구에서는 메타휴리스틱 최적화 전략에 기반한 일련의 양자 알고리즘 연구들을 소개하며, 특히 연산 효율, 양자 회로 설계, 큐빗 사용량 등의 측면들을 기준으로 양자 메타휴리스틱 최적화의 실현성을 검 토하는데 중점을 두었다. 관련 실험들을 통해 제안된 방법들을 검증하였고, 그 결과는 메타휴리스틱전략이양자최적화분야에서유효하고실용성있는접근법이될수있음을 암시한다.|Apart from the heated expectation upon the advent of quantum computers, re- lated studies have noted that developing quantum algorithms with plausible credibility and utility could likely be a difficult and uncertain venture. This study introduces a series of investigations upon the topic of quantum optimization, especially focusing on implementing quantum algorithms based on metaheuristic optmization strategies. Specifically, the feasibility of realizing quantum metaheuristic optimization is examined in various aspects including computational efficiency, circuit configuration, and qubit usage. The proposed approaches are verified with corresponding experiments, the re- sults of which imply the validity and practical utility of the metaheuristic strategy in the field of quantum optimization.DoctorAbstract (English) i
Abstract (Korean) ii
List of Contents iii
List of Tables vi
List of Figures viii
List of Algorithms xv
1 Introduction 1
2 Backgrounds of Quantum Computing 4
2.1 Fundamental Composition 4
2.1.1 Qubit and Quantum State 4
2.1.2 Quantum Gates 9
2.1.3 Quantum Circuit 12
2.2 Important Algorithms 13
2.2.1 Quantum Speedup Capabilities 13
2.2.2 Unordered Data Search 15
2.2.3 Quantum Fourier Transform 19
3 Quantum Computing in Optimization 21
3.1 General Statements 21
3.1.1 Basic Definition 21
3.1.2 Important Notions 23
3.1.3 Exact and Heuristic 25
3.2 Quantum Approaches 27
3.2.1 Purely Quantum Strategies 28
3.2.2 Quantum-Classical Hybrids 30
4 Toward Quantum Meta-Heuristic Optimization 34
4.1 Genetic/Evolutionary Algorithm 34
4.2 Quantum Genetic/Evolutionary Algorithm 38
4.3 Expendable Quantum Computation 43
4.3.1 Deutsch Algorithm: Review 44
4.3.2 Application to Machine Learning 45
4.3.3 Analysis 48
4.4 Partial Application of Quantum Search 50
4.4.1 Premature Convergence and Crowding 51
4.4.2 Adaptation of Quantum Algorithm 54
4.4.3 Implementation 60
4.4.4 Simulation Result 65
4.4.5 Analysis 68
4.5 Assertive Quantum Accelerator 70
4.5.1 Quantum Counting Algorithm 71
4.5.2 Quantum Population Initialization 73
4.5.3 Quantum Adaptation Strategy 76
4.5.4 Complexity and Accuracy Analysis 80
4.5.5 Experiment 81
4.5.6 Analysis 84
4.6 Effectively Mapping Solution Candidates 89
4.6.1 Review: Reduced Quantum Genetic Algorithm 90
4.6.2 Strictly Structured Quantum Genetic Algorithm 92
4.6.3 Proposed Improvement 96
4.6.4 Excess in Population 97
4.6.5 Removing Trivial Individuals 99
4.6.6 Experiment 104
4.6.7 Analysis 109
4.7 Practical Quantum Design and Application 117
4.7.1 Quantum Approximate Optimization Algorithm Revisited 119
4.7.2 Gravitational Search Algorithm 121
4.7.3 Quantum Quasi-Divide-and-Conquer Strategy 122
4.7.4 Algorithmic Procedure 127
4.7.5 Verification 130
4.7.6 Analysis 135
5 Quantum Metaheuristic Algorithm with Approximate Optimization
Strategy 142
5.1 Summary on Preliminaries 142
5.1.1 Partial Quantum Application 143
5.1.2 Quantum Add-on 143
5.1.3 Efficient Mapping of Quantum States 144
5.1.4 Simplified Search Process 145
5.2 Implications 146
5.2.1 Data Addressing Issues 146
5.2.2 Approximation of Search Domain 147
5.3 Proposition 148
5.3.1 Configuration 148
5.3.2 Approximate Mapping 150
5.3.3 Gradual Search Reduction 154
5.3.4 Dynamic Iteration 158
5.4 Verification 158
5.4.1 Setup 159
5.4.2 Results 164
5.5 Analysis 165
5.5.1 Optimization Performance 167
5.5.2 Exploiting Quantum Parallelism 170
5.5.3 Scalability 172
6 Conclusion and Prospects 173
Summary 17
Microbial collagenase activity is linked to oral-gut translocation in advanced chronic liver disease
No full textMicrobiome perturbations are associated with advanced chronic liver disease (ACLD), but how microorganisms contribute to disease mechanisms is unclear. Here we analysed metagenomes of paired saliva and faecal samples from an ACLD cohort of 86 individuals, plus 2 control groups of 52 healthy individuals and 14 patients with sepsis. We identified highly similar oral and gut bacterial strains, including Veillonella and Streptococcus spp., which increased in absolute abundance in the gut of patients with ACLD compared with controls. These microbial translocators uniquely share a prtC gene encoding a collagenase-like proteinase, and its faecal abundance was a robust ACLD biomarker (area under precision-recall curve = 0.91). A mouse model of hepatic fibrosis inoculated with Veillonella and Streptococcus prtC-encoding patient isolates showed exacerbation of gut barrier impairment and hepatic fibrosis. Furthermore, faecal collagenase activity was increased in patients with ACLD and experimentally confirmed for the prtC gene of translocating Veillonella parvula. These findings establish mechanistic links between oral-gut translocation and ACLD pathobiology.TRUEsciescopu
Data-Driven Prediction of Controllability of Fighter Aircraft and Real-time Aerodynamic Analysis using Physics-Informed Neural Network
No full textPart 1에서는 물리지식기반 신경망을 활용하여 형상(받음각)에 따른 실시간 유동 장예측 대리모델을 구축하였습니다. 높은 레이놀즈 수에서 Physics Informed Neural Network (PINN)가 물리적으로 의미 있는 해를 도출하는 데 한계를 보이는 문제를 해 결하기 위해, 제한된 수의 유동 데이터를 PINN에 학습시키는 데이터 기반 접근 방식을 결합하여 모델 성능을 개선하였습니다. 또한 데이터 기반 Conditional U-Net 모델과 물리정보기반신경망의내,외삽영역유동장예측성능비교를통해,본연구에서구축 한 Data-Assisted PINN이 학습 데이터 수 측면에서 효율적이며 일반화 성능이 뛰어난 모델임을 밝혔습니다. Part 2에서는 F-16 전투기와 Bio-Inspired Rotating Empennage(BIRE) 항공기의 조 종성을 평가하기 위해 다양한 비행 조건과 조종면 편향에 따른 달성 가능한 모멘트 세트 (Attainable Moment Set, AMS)를 도출하였습니다. 받음각, 사이드슬립 각, 3축 회전율, 조종면 편향 각도를 포함한 8개의 입력 변수를 고려하였으며, 입력 변수 범위에 대한 라틴 하이퍼큐브 샘플링(LHS)을 통해 샘플링 데이터를 생성하였습니다. 생성된 샘플링 지점에 대해 SU2 솔버를 활용하여 RANS 시뮬레이션을 수행하여 공력 데이터베이스 를 구축하였고, 이를 기반으로 가우시안 프로세스 회귀(GPR) 모델을 학습하였습니다. – iii – 학습된 GPR 모델은 학습되지 않은 조건에서도 공기역학적 모멘트를 정확히 예측할 수 있어 두 항공기의 AMS를 효과적으로 분석할 수 있었습니다. 본 연구는 전통적인 공력 해석 방법론에 머신러닝과 딥러닝을 접목함으로써, 공력 해석 및 최적 설계 과정에서 시간과 비용을 절감할 수 있는 효율적이고 혁신적인 접근 방식을 제시합니다. ©2025 김 성 연 ALL RIGHTS RESERVED|Part 1 focused on developing a real-time surrogate model for flow field prediction using Physics-Informed Neural Networks (PINNs) with respect to design variables. To address the limitations of PINNs in producing physically meaningful solutions at high Reynolds numbers, a data-driven approach was integrated into PINNs, by utilizing a limited amount of flow data to significantly improve the model’s performance. Addition- ally, a comparison of flow field prediction accuracy in interpolation and extrapolation regions between the data-driven Conditional U-Net model and the PINN-based model revealed that the Data-Assisted PINN is more efficient in terms of data requirements and demonstrates superior generalization capabilities. Part 2 evaluated the controllability of the F-16 and Bio-Inspired Rotating Em- pennage (BIRE) aircraft by deriving Attainable Moment Sets (AMS) under various flight conditions and control surface deflections. Eight input variables, including angle of attack, side slip angle, three-axis rotation rates, and control surface deflection an- gles, were considered. Samples were generated using Latin Hypercube Sampling (LHS) within the input variable ranges. Using these sampled points, RANS simulations were conducted with the SU2 solver to build an aerodynamic database, which was then used to train a Gaussian Process Regression (GPR) model. The trained GPR model accu- rately predicted aerodynamic moments even for untested conditions, enabling compre- hensive AMS analysis for both aircraft. This study proposes an efficient and innovative approach by integrating machine learning and deep learning into traditional aerody- namic analysis methodologies, enabling time and cost savings in aerodynamic analysis and optimal design processes. ©2025 Sungyeon, Kim ALL RIGHTS RESERVEDMasterAbstract (English) i
Abstract (Korean) iii
List of Contents v
List of Tables vii
List of Figures viii
1 Introduction 1
1.1 Research Background and Objectives 1
1.2 Literature review 4
1.3 Contribution 6
2 Methods of Part1 8
2.1 Physics Informed Neural Network 8
2.1.1 Theoretical Background of Artificial Neural Network 8
2.2 Physics-Informed Neural Network 12
2.2.1 Sampling points for surrogate PINN 14
2.2.2 Data-assisted PINN 16
2.3 CNN based data-driven surrogate model 19
2.3.1 Preprocessing of Flow Field Data for CNN based model 19
2.3.2 Theoretical Background of Convolutional Neural Network 20
2.3.3 Conditional U-Net 20
2.3.4 Customized Loss Function 21
3 Methods of Part2 23
3.1 Trim Condition 23
3.2 Aerodynamic database and Operating conditions 26
3.3 Aerodynamic Analysis and Validation 28
3.4 Attainable Moment Set 29
3.5 Gaussian Process Regression 32
– v –
4 Validation 35
4.1 Numerical analysis for training surrogate models 35
4.2 Prediction at Laminar flow 37
4.3 Prediction at Turbulent flow 41
4.4 Sensitivity Analysis for Amount of Sampling points 43
4.5 Comparison of DA-PINN Models Using RANS and Navier-Stokes Equa-
tions 44
4.6 Comparison of Convergence between DA-PINN and PINN according to
Iterations 46
4.7 Accuracy Comparison of Unet and DA-PINN 47
5 Results of Part2 50
6 Conclusion 59
6.1 Conclusion of part1 59
6.2 Conclusion of part2 60
References 61
– vi
Hydrogen Peroxide-Releasing Hydrogel-Mediated Cellular Senescence Model for Aging Research
No full textCellular senescence, a process that induces irreversible cell cycle arrest in response to diverse stressors, is a primary contributor to aging and age-related diseases. Currently, exposure to hydrogen peroxide is a widely used technique for establishing invitro cellular senescence models; however, this traditional method is inconsistent, laborious, and ineffective in vivo. To overcome these limitations, we have developed a hydrogen peroxide-releasing hydrogel that can readily and controllably induce senescence in conventional 2-dimensional cell cultures as well as advanced 3-dimensional microphysiological systems. Notably, we have established 2 platforms using our hydrogen peroxide-releasing hydrogel for investigating senolytics, which is a promising innovation in anti-geronic therapy. Conclusively, our advanced model presents a highly promising tool that offers a simple, versatile, convenient, effective, and highly adaptable technique for inducing cellular senescence. This innovation not only lays a crucial foundation for future research on aging but also markedly accelerates the development of novel therapeutic strategies targeting age-related diseases. Copyright © 2025 Shibo Wei et al.TRUEsciescopuskc
Effects of atorvastatin-loaded PEGylated liposomes delivered by magnetic stimulation for stroke treatment
No full textBackground: Focused magnetic stimulation (MagStim) can temporarily and safely open the blood-brain barrier (BBB) for target delivery. We investigated whether opening the BBB with MagStim and delivering atorvastatin-loaded PEGylated liposomes (LipoStatin) would work synergistically for subacute post-stroke treatment. Methods: Two weeks after middle cerebral artery occlusion (MCAO), an injection of 15 mg/ml magnetic nanoparticles (MNPs) was performed, followed by 30 min of MagStim, in subacute stroke models. The procedure was conducted over a week, during which MagStim and MNPs were administered three times at two-day intervals, and LipoStatin (10 mg/kg) was injected immediately after each MagStim treatment. We investigated the motor function, BBB integrity, neuroinflammation, and neurogenesis three weeks after stroke (Sham vs. Control vs. LipoStatin vs. MagStim + LipoStatin). Results: The MagStim + LipoStatin group showed improved motor function compared to the Control (p = 0.007) group. The MagStim + LipoStatin group significantly reduced infarct volume and improved BBB integrity compared to the control and LipoStatin groups. In the MagStim + LipoStatin group, the expression of TNF-α was reduced (p = 0.020) compared to the LipoStatin group, and eNOS was enhanced (p = 0.037) compared to the Control group. Markers for neurogenesis were also considerably increased in the MagStim + LipoStatin group compared to the Control and LipoStatin groups (p < 0.0001). Conclusions: Our study demonstrates the beneficial synergistic effects of MagStim and the target delivery of LipoStatin in subacute ischemic stroke. These findings underscore the need for future advancements in promising novel non-invasive MagStim methods and nanotherapeutic hybrid approaches for target drug delivery and treatment in post-stroke recovery. © 2025 Elsevier B.V., All rights reserved.TRUEsciescopu
Scalable Purification of Gold Nanocubes via Centrifugal Depletion-Induced Flocculation: A Pathway to High-Precision Nanomaterials
No full textMetal nanoparticles, especially when synthesized on a large scale, often exhibit significant heterogeneity in size and shape, which can limit their reliability and reproducibility in various applications, including plasmonics. This study introduces a centrifugal depletion-induced flocculation (CDF) method for the scalable and efficient purification of Au nanocubes (AuNCs). Unlike traditional flocculation methods, CDF leverages controlled centrifugation and depletion interactions to achieve high particle homogeneity and recovery efficiency while significantly reducing processing time. Systematic optimization of critical parameters, such as nanoparticle concentration, centrifugal force, and flocculation time, enabled recovery efficiency exceeding 98% and yields of approximately 98-99% for 60 nm AuNCs. The adaptability of this method was also demonstrated for NCs of varying sizes and larger volumes, underscoring its versatility and scalability for diverse nanoparticle systems. The superior monodispersity and shape uniformity of purified AuNCs achieved through this approach hold significant potential for applications requiring precise control over plasmonic properties, such as high-performance optical sensors and integrated nanophotonic systems. © 2025 American Chemical Society.FALSEsciescopu