PSExplorer: whole parameter space exploration for molecular signaling pathway dynamics

Tung, Thai Quang,Lee, Doheon Oxford University Press 2010 Bioinformatics Vol.26 No.19

<P><B>Motivation:</B> Mathematical models of biological systems often have a large number of parameters whose combinational variations can yield distinct qualitative behaviors. Since it is intractable to examine all possible combinations of parameters for non-trivial biological pathways, it is required to have a systematic strategy to explore the parameter space in a computational way so that dynamic behaviors of a given pathway are estimated.</P><P><B>Results:</B> We present PSExplorer, a computational tool for exploring qualitative behaviors and key parameters of molecular signaling pathways. Utilizing the Latin hypercube sampling and a clustering technique in a recursive paradigm, the software enables users to explore the whole parameter space of the models to search for robust qualitative behaviors. The parameter space is partitioned into sub-regions according to behavioral differences. Sub-regions showing robust behaviors can be identified for further analyses. The partitioning result presents a tree structure from which individual and combinational effects of parameters on model behaviors can be assessed and key factors of the models are readily identified.</P><P><B>Availability:</B> The software, tutorial manual and test models are available for download at the following address: http://gto.kaist.ac.kr/∼psexplorer</P><P><B>Contact:</B>tqtung@kaist.ac.kr; tqtung@gmail.com</P>

K-NN 분류에 투표 방법을 이용한 단백질 부세포의 위치 예견에 관한 연구

Thai Quang Tung,Lim, JongTae 公州大學校 基礎科學硏究所 2004 自然科學硏究 Vol.11 No.-

단백질의 위치가 생물학적 기능과 밀접한 관계가 있다. 부세포의 실험적 위치 결정은 많은 시간과 비용이 요구된다. 밝혀진 단백질 서열이 데이터뱅크에 빠르게 증가함에 따라 단백질 부세포의 위치를 자동적으로 파악하는 일이 아주 중요하다. 단백질 부세포의 위치 결정에 대한 많은 방법과 시스템이 개발되었지만 어느 방법도 단백질 부세포의 위치를 정확하게 예측하지 못하고 있다. 본 논문에서는 단백질 서열을 아미노산 쌍을 정규화한 결합 형태로 표현하는 새로운 방법을 제안한다. 이 방법은 아미노산 결합 데이터들이 중복되어 나타나고 있다는 성질에 근거한다. 랜덤하게 선택한 부분 집합에 가장 인접한 아미노산 쌍을 근거로 분류하여 부정확한 분류를 줄일 수 있다. 이 분류에 투표 강식을 도입하여 예견성의 정확성을 높이고자 한다. 본 논문에서 제안한 방법의 성능을 평가하기 위해서 이전 방법에서 사용한 동일한 실험 데이터 집합을 사용하여 시뮬레이션 한 결과 제안한 방법의 우수성을 파악할 수 있었다. The localization of a protein in a cell is closely correlated with its biological function. Experimental determination of sub cellular location is time-consuming and costly. With the number of sequences entering databanks rapidly increasing, the importance of developing a powerful tool to identify protein sub-cellular location automatically has become self-evident. Several methods and systems have been developed, but it has become clear that there is no single method of prediction can achieve high predictive accuracy for all localization. In this thesis, I introduce a new prediction method based on multiple subsets of normalized compositions of amino acid pair. The idea comes from the observation that amino acid pair composition data is quite redundant. Building nearest neighbor classifiers based on randomly selected subsets, I force the classifiers to make different and hopefully un-correlated errors. Therefore by applying a voting scheme the prediction accuracy can be improved. The experiment was conduct on several data sets which had been used by other methods. In all cases, our approach can give a significant improvement.

Wifi Fingerprint Calibration Using Semi-Supervised Self Organizing Map

Quang Tung Thai(타이광퉁),Ki-Sook Chung(정기숙),Changsup Keum(금창섭) 한국통신학회 2017 韓國通信學會論文誌 Vol.42 No.2

A method to improve protein subcellular localization prediction by integrating various biological data sources

Tung, Thai Quang,Lee, Doheon BioMed Central 2009 BMC bioinformatics Vol.10 No.suppl1

<P><B>Background</B></P><P>Protein subcellular localization is crucial information to elucidate protein functions. Owing to the need for large-scale genome analysis, computational method for efficiently predicting protein subcellular localization is highly required. Although many previous works have been done for this task, the problem is still challenging due to several reasons: the number of subcellular locations in practice is large; distribution of protein in locations is imbalanced, that is the number of protein in each location remarkably different; and there are many proteins located in multiple locations. Thus it is necessary to explore new features and appropriate classification methods to improve the prediction performance.</P><P><B>Results</B></P><P>In this paper we propose a new predicting method which combines two key ideas: 1) Information of neighbour proteins in a probabilistic gene network is integrated to enrich the prediction features. 2) Fuzzy k-NN, a classification method based on fuzzy set theory is applied to predict protein locating in multiple sites. Experiment was conducted on a dataset consisting of 22 locations from Budding yeast proteins and significant improvement was observed.</P><P><B>Conclusion</B></P><P>Our results suggest that the neighbourhood information from functional gene networks is predictive to subcellular localization. The proposed method thus can be integrated and complementary to other available prediction methods.</P>

ezBioNet: A Modeling and Simulation System for Analyzing Biological Reaction Networks

유석종,Thai Quang Tung,박준호,임종태,유재수 한국물리학회 2012 THE JOURNAL OF THE KOREAN PHYSICAL SOCIETY Vol.61 No.8

To achieve robustness against living environments, a living organism is composed of complicated regulatory mechanisms ranging from gene regulations to signal transduction. If such life phenomena are to be understand, an integrated analysis tool that should have modeling and simulation functions for biological reactions, as well as new experimental methods for measuring biological phenomena, is fundamentally required. We have designed and implemented modeling and simulation software (ezBioNet) for analyzing biological reaction networks. The software can simultaneously perform an integrated modeling of various responses occurring in cells, ranging from gene expressions to signaling processes. To support massive analysis of biological networks, we have constructed a server-side simulation system (VCellSim) that can perform ordinary differential equations (ODE) analysis, sensitivity analysis, and parameter estimates. ezBioNet integrates the BioModel database by connecting the european bioinformatics institute (EBI) servers through Web services APIs and supports the handling of systems biology markup language (SBML) files. In addition, we employed eclipse RCP (rich client platform) which is a powerful modularity framework allowing various functional expansions. ezBioNet is intended to be an easy-to-use modeling tool, as well as a simulation system, to understand the control mechanism by monitoring the change of each component in a biological network. A researcher may perform the kinetic modeling and execute the simulation. The simulation result can be managed and visualized on ezBioNet, which is freely available at http://ezbionet.cbnu.ac.kr.

A Unified Biological Modeling and Simulation System for Analyzing Biological Reaction Networks

유석종,Thai Quang Tung,박준호,임종태,유재수 한국물리학회 2013 THE JOURNAL OF THE KOREAN PHYSICAL SOCIETY Vol.63 No.11

In order to understand the biological response in a cell, a researcher has to create a biological networkand design an experiment to prove it. Although biological knowledge has been accumulated,we still don’t have enough biological models to explain complex biological phenomena. If a new biologicalnetwork is to be created, integrated modeling software supporting various biological modelsis required. In this research, we design and implement a unified biological modeling and simulationsystem, called ezBioNet, for analyzing biological reaction networks. ezBioNet designs kinetic andBoolean network models and simulates the biological networks using a server-side simulation systemwith Object Oriented Parallel Accelerator Library framework. The main advantage of ezBioNetis that a user can create a biological network by using unified modeling canvas of kinetic andBoolean models and perform massive simulations, including Ordinary Differential Equation analyses,sensitivity analyses, parameter estimates and Boolean network analysis. ezBioNet integratesuseful biological databases, including the BioModels database, by connecting European BioinformaticsInstitute servers through Web services Application Programming Interfaces. In addition, weemploy Eclipse Rich Client Platform, which is a powerful modularity framework to allow variousfunctional expansions. ezBioNet is intended to be an easy-to-use modeling tool and a simulationsystem for understanding the control mechanism by monitoring the change of each component ina biological network. The simulation result can be managed and visualized on ezBioNet, which isavailable free of charge at http://ezbionet.sourceforge.net or http://ezbionet.cbnu.ac.kr.

Design of an Advanced Wearable Sensor Platform for Multi Applications

Anh-Tuan Nguyen,Tung Hoang,Quang-Vinh Thai,T.T. Quyen Bui 제어로봇시스템학회 2015 제어로봇시스템학회 국제학술대회 논문집 Vol.2015 No.10

In the emerging Internet-of-Things (IoT), wireless sensors are the vital mediators between the physical world and the cyber space. In many telemonitoring and interactive applications, high-rate data streams up to MB/s may need to be transported from the wireless sensors to the cloud computing servers over the Internet and processed by the remote servers in quasi-real time fashion. These applications generally demand substantial reduction of communication bandwidth, response latency and power consumption of these Internet-based cyber-physical systems. To tackle these demands for efficient use of communication, computing and power resources, we developed an advanced wearable sensor platform (AWSP) which integrated a powerful system-on-chip (SoC) processor, a smart power management unit, a highly accurate real-time clock and multi-function peripherals into a miniature module. This AWSP designed to function is capable of performing sophisticated data pre-processing including real-time artifact and noise removal, data compression, and even feature extraction before uploading the data to other devices like mobile phone, computer, etc. Furthermore, with the installation of embedded real-time Linux operating system, this sensor provides a familiar and powerful software development environment for system developers to build their computation-intensive real-time applications. This paper presents the design and development process.

생물학적 반응 네트워크 분석을 위한 불리언 네트워크 모데링 및 시뮬레이션 시스템

박준호,임종태,김동주,이윤정,안민제,류은경,차재홍,유석종,Thai Quang Tung,유재수 한국정보과학회 2012 데이타베이스 연구 Vol.28 No.3

A living body is composed of complicated regulation mechanisms ranging from gene regulations to signal transduction to sustain its life. For understanding such life phenomena, an integrated analysis tool that performs techniques and computer simulations on the regulation mechanisms of biological reactions as well as new experimental methods for measuring biological phenomena and high-precision analysis methods is absolutely required the increase. In this paper, we design and implement a boolean network modeling and simulation system using server-side simulation system with OPAL framework. The main advantage of the proposed system is that a researcher can create a biological network using unified modeling canvas of boolean model and perform the massive boolean network analysis. In addition, its plug-in-based modular design framework allows various functional expansions. A researcher may perform the integrated modeling and information management of the regulation mechanisms from genome information to signaling networks through the proposed system. Moreover, it can be utilized as a tool capable of analyzing the interactions between regulation mechanisms as well as the understanding of the regulation mechanisms using computer simulations in the future. 생명 현상을 지속하기 위해 생체는 유전자 발현 조절에서부터 신호전달 작용까지 복잡한 조절기작으로 구성된다. 이와 같은 생명현상을 이해하기 위해서는 생명현상을 측정하기 위한 새로운 실험 방법과 정밀한 분석법이 필요할 뿐만 아니라 조절기작에 대한 반응 모델링 기법과 전산모사에 대한 통합적인 분석도구의 요구도증가하고 있다. 본 논문에서는 유전체에서부터 유전자 발현 및 신호전달과정에 이르는 다양한 세포내의 반응을 불리언 네트워크 모델링하고, 서버 기반의 시뮬레이션을 동시에 수행할 수 있는 시스템을 설계하고 구현한다. 제안하는 시스템 불리언 네트워크의 모델링을 수행할 수 있도록 구현하고, 설계된 네트워크를 서버 기반시뮬레이션을 수행 할 수 있도록 개발한다. 또한 향후 다양한 기능 확장을 보장하기 위해서 플러그인 기반의설계를 적용하여 새로운 모듈을 추가할 수 있도록 설계한다. 이를 통해 연구자는 유전체 정보로부터 신호전달네트워크까지 세포 내 조절기작을 통합적으로 모델링하고 정보를 관리할 수 있다. 또한 전산모사를 통해 조절기작의 이해는 물론 향후 상호작용을 분석할 수 있는 도구로 활용될 수 있다.

Design and Implementation of a Kinetic Network Modeling and Simulation System in HPC

Junho Park,Jongtae Lim,Dongjoo Kim,Yunjeong Lee,EunkyungRyu,MinjeAhn,Jaehong Cha,Seokjong Yu,Thai Quang Tung,HakyongKim,JaesooYoo 한국콘텐츠학회 2012 ICCC International Digital Design Invitation Exhib Vol.2012 No.12

Design and Implementation of a Kinetic Network Modeling and Simulation System in HPC

Junho Park,Jongtae Lim,Dongjoo Kim,Yunjeong Lee,EunkyungRyu,MinjeAhn,Jaehong Cha,Seokjong Yu,Thai Quang Tung,HakyongKim,JaesooYoo 한국콘텐츠학회 2012 한국콘텐츠학회 종합학술대회 논문집 Vol.2012 No.12