• ETRI Journal
• /
• v.33 no.5
• /
• pp.712-719
• /
• 2011

In this paper, we study the problem of domain adaptation for structural support vector machines (SVMs). We consider a number of domain adaptation approaches for structural SVMs and evaluate them on named entity recognition, part-of-speech tagging, and sentiment classification problems. Finally, we show that a prior model for structural SVMs outperforms other domain adaptation approaches in most cases. Moreover, the training time for this prior model is reduced compared to other domain adaptation methods with improvements in performance.

• Molecules and Cells
• /
• v.22 no.1
• /
• pp.13-20
• /
• 2006

Hepatitis C virus (HCV) non-structural protein 5A protein (NS5A), which consists of three functional domains, is involved in regulating viral replication, interferon resistance, and apoptosis. Recently, the three-dimensional structure of the domain 1 was determined. However, currently the molecular basis for the domains 2 and 3 of HCV NS5A is yet to be defined. Toward this end, we expressed, purified the domain 2 of the NS5A (NS5A-D2), and then performed biochemical and structural studies. The purified domain 2 was active and was able to bind NS5B and PKR, biological partners of NS5A. The results from gel filtration, CD analysis, 1D $^1H$ NMR and 2D $^1H-^{15}N$ heteronuclear single quantum correlation (HSQC) spectroscopy indicate that the domain 2 of NS5A appears to be flexible and disordered.

• Bioinformatics and Biosystems
• /
• v.1 no.2
• /
• pp.123-127
• /
• 2006

TIM barrel domain is widely studied since it is one of most common structure and mediates diverse function maintaining overall structure. TIM barrel domain's function is determined by local structural environment at the C-terminal end of barrel structure. We classified TIM barrel domains by local structural alignment tool, LSHEBA, to understand characteristics of TIM barrel domain's functionalvariation. TIM barrel domains classified as the same cluster share common structure, function and ligands. Over 80% of TIM barrels in clusters share exactly the same catalytic function. Comparing clustering result with that of SCOP, we found that it's important to know local structural environment of TIM barrel domains rather than overallstructure to understand specific structural detail of TIM barrel function. Non TIM barrel domains were associated to make different domain combination to form a different function. The relationship between domain combination, we suggested expected evolutional history. We finally analyzed the characteristics of amino acids around ligand interface.

• BMB Reports
• /
• v.40 no.1
• /
• pp.58-64
• /
• 2007

Similar to the other known structures of Bacillus thuringiensis Cry $\delta$-endotoxins, the crystal structure of the 65-kDa activated Cry4Ba toxin comprises three domains which are, from the N- to C-terminus, a bundle of $\alpha$-helices, a three-$\beta$-sheet domain, and a $\beta$-sandwich. To investigate the properties of the C-terminal domain III in isolation from the rest of the toxin, the cloned Cry4Ba-domain III was over-expressed as a 21-kDa soluble protein in Escherichia coli, which cross-reacted with anti-Cry4Ba domain III monoclonal antibody. A highly-purified domain III was obtained in a monomeric form by ion-exchange and size-exclusion FPLC. Circular dichroism spectroscopy indicated that the isolated domain III fragment distinctly exists as a $\beta$-sheet structure, corresponding to the domain III structure embodied in the Cry4Ba crystal structure. In vitro binding analysis via immuno-histochemical assay revealed that the Cry4Ba-domain III protein was able to bind to the apical microvilli of the susceptible Stegomyia aegypti larval midguts, albeit at lower-binding activity when compared with the full-length active toxin. These results demonstrate for the first time that the C-terminal domain III of the Cry4Ba mosquito-larvicidal protein, which can be isolated as a native folded monomer, conceivably participates in toxin-receptor recognition.

• Journal of Mechanical Science and Technology
• /
• v.15 no.5
• /
• pp.555-561
• /
• 2001

A new approach to analyze the multi-domain acoustic system divided and enclosed by flexible structures is presented in this paper. The boundary element formulation of the Helmholtz integral equation is used for the internal fields and the finite element formulation for the structures surrounding the fields. We developed a numerical analysis program for the structural-acoustic coupling problems of the multi-domain system, in which boundary conditions such as the continuity of normal particle velocity and sound pressure in the structural interfaces between Field 1 and Field 2 are not needed. The validity of the numerical analysis program is verified by comparing the numerical results with the experimental ones. Example problems are included to investigate the characteristics of the coupled multi-domain system.

• Journal of the Korean Magnetic Resonance Society
• /
• v.14 no.2
• /
• pp.105-116
• /
• 2010

SH3YL1, a novel protein containing one Src homology 3 domain at the carboxyl terminus was first detected in mouse anagen skin cDNA. This protein had a significant homology with YHRO 16c/Ysc 84, the yeast Src homology 3 domain-containing protein. The sequence identity was remarkable at the carboxyl and amino-terminal Src homology 3 domain, suggesting that the novel protein is a mouse homolog of the yeast protein and thus was termed as SH3YL1. SH3YL1 is composed of two domains, a DUF500 at N-termini and a SH3 domain at C-termini. In our study we cloned the SH3 domain in bacterial expression system in Escherichia coli using pET32a vector with TEV protease cleavage site and purified as a monomer using affinity chromatography. The N-terminal poly-Histidine tag was cleaved with TEV protease and target protein was used for backbone studies. Our study showed that SH3 domain primarily consists of $\beta$-sheet which is in consistence with previous result performed on the truncated SH3 domain of SH3YL1.

• Journal of the Korean Magnetic Resonance Society
• /
• v.13 no.1
• /
• pp.56-63
• /
• 2009

SWIRM domain, a core domain of SRG3 is well conserved in SW13, RSC8, and MOIRA family proteins. To understand structural basis for cellular functions of the SWIRM domain, we have initiated biochemical and structural studies on SWIRM domain and mutants using gelfiltration chromatography, circular dichroism and NMR spectroscopy. The structural properties of the mutant SWIRM domains (K34A and M75A) have been characterized, showing that the structures of both wild-type and mutant proteins are a-helical conformation. The data conclude that mutations at interaction sites of its binding partner protein do not affect its secondary and tertiary structure.

• Proceedings of the Korean Society for the Gifted Conference
• /
• 2005.05a
• /
• pp.173-180
• /
• 2005

The purpose of this study was to identify common and domain-specific cognitive characteristics of gifted students based on a hierarchical structural model of human abilities. This study is based on the premise that abilities identified by tests can appear as observable characteristics in test or school situations. Abilities proposed by major models of intelligence were reviewed in terms of their power to explain cognitive characteristics of gifted students. However, due to the lack of their explanatory power and disagreement on common and domain-specific cognitive abilities, a new hierarchical structural model was conceptualized in a unique way based on interrelationships between abilities proposed by the models. The newly established model hypothesizes a cognitive mechanism that accounts for how domain-specific knowledge is formed, as well as which abilities are common and domain-specific, how they are related functionally, and how they account for common and domain-specific cognitive characteristics of gifted students. The cognitive mechanism has important implications for our understanding of the chronically controversial concepts, 'intelligence' and 'knowledge.' Clearer definitions of what intelligence is (g or multiple), what knowledge is, and how knowledge develops ('genetic or environmental,' 'rationalistic or empiricist') may result from this model.

• BMB Reports
• /
• v.35 no.3
• /
• pp.343-347
• /
• 2002

The human protein tyrosine kinase-6 (PTK6) polypeptide that is deduced from the cDNA sequence contains a Src homology (SH) 3 domain, SH2 domain, and catalytic domain of tyrosine kinase. We initiated biochemical and NMR characterization of PTK6 SH3 domain in order to correlate the structural role of the PTK6 using circular dichroism and heteronuclear NMR techniques. The circular dichroism data suggested that the secondary structural elements of the SH3 domain are mainly composed of $\beta$-sheet conformations. It is most stable when the pH is neutral based on the pH titration data. In addition, a number of cross peaks at the low-field area of the proton chemical shift of the NMR spectra indicated that the PTK6 SH3 domain retains a unique and folded conformation at the neutral pH condition. For other pH conditions, the SH3 domain became unstable and aggregated during NMR measurements, indicating that the structural stability is very sensitive to pH environments. Both the NMR and circular dichroism data indicate that the PTK6 SH3 domain experiences a conformational instability, even in an aqueous solution.

• Smart Structures and Systems
• /
• v.8 no.4
• /
• pp.343-365
• /
• 2011

Structural system identification (SSI) is an inverse problem of difficult solution. Currently, difficulties lie in the development of algorithms which can cater to large size problems. In this paper, a parameter estimation technique based on evolutionary strategy is presented to overcome some of the difficulties encountered in using the traditional system identification methods in terms of convergence. In this paper, a non-traditional form of system identification technique employing evolutionary algorithms is proposed. In order to improve the convergence characteristics, it is proposed to employ immune algorithms which are proved to be built with superior diversification mechanism than the conventional evolutionary algorithms and are being used for several practical complex optimisation problems. In order to reduce the number of design variables, domain decomposition methods are used, where the identification process of the entire structure is carried out in multiple stages rather than in single step. The domain decomposition based methods also help in limiting the number of sensors to be employed during dynamic testing of the structure to be identified, as the process of system identification is carried out in multiple stages. A fifteen storey framed structure, truss bridge and 40 m tall microwave tower are considered as a numerical examples to demonstrate the effectiveness of the domain decomposition based structural system identification technique using immune algorithm.