Page 12 - profile-ok
P. 12
研究群 | Research Laboratories
Bioinformatics Laboratory
Research Faculty Group Profile
Ting-Yi Sung Our current research is classified into two main areas: systems biology and proteomics. In addition novel targets to study their involvement in viral pathogenesis. proach to predict various properties of proteins.
Research Fellow to research, we inaugurated the Bioinformatics Ph.D. Program in Taiwan International Graduate
Operations Research , New York University Program, Academia Sinica. As of Fall 2009, two students received their Ph.D. degrees and 34 stu- ● Metagenomics (1) Membrane protein structure prediction: We plan to develop novel
methods to address membrane protein structure prediction by
Jan-Ming Ho dents enrolled, including local students and foreign students from Canada, Germany, India, Malay- Based on whole genome shotgun sequencing data, we develop an in- dissecting the problem into several components in tertiary struc-
sia, the Philippines, Slovakia, the United States, and Vietnam. Our research and major collaborative
Research Fellow tegrated platform including various databases, gene expression analy- ture model construction. Building upon our previous works in
Electrical Engineering and Computer Science , projects are described below. sis, proteomic results and phylogenetic reconstruction to achieve a topologing and helix-helix interaction/contact prediction, we are
Northwestern University 1. Systems Biology comprehensive view of microbials. developing a highly accurate predictor for the solvent accessibility
Chun-Nan Hsu ● Post-transcriptional Gene Regulation ● Transcriptome Analysis of TM proteins, which in turn provides valuable information in the
Research Fellow rotational and exposure preferences of individual helices. In parallel,
Computer Science , University of Southern MicroRNAs play an important role in the posttranscriptional regulation of genes and diseases. The integrative bioinformatic platforms for high-throughput sequenc- we closely examine the physical constraints governing helix-pack-
California We have developed a reverse approach and predicted dozens of new human miRNA genes. ing of non-model species aims at supporting transcriptome studies of ing (i.e., crossing angles, closest point of contact, etc), and construct
Wen-Lian Hsu Some of these and their target genes have been validated through experiments, and the results specific domestic species, including A. hallerissp.gemmifera, Formosan a knowledge base for all known helix-helix interactions in currently
black bear, Formosan cypress, silvergrass, pteridophyte, and coral. Bi-
Distinguished Research Fellow provided strong evidence that our predicted microRNAs and their targets are indeed the direct ologists will analyze the transcriptome of these non-model species, available structures so as to facilitate 3D structure modeling.
Operations Research , Cornell University targets. In addition, we have been working on developing a systemic approach for integrating (2) Protein function prediction, subcellular localization prediction,
high-throughput next generation sequencing and other experimental data, in order to uncover though their genomes have not yet been fully assembled. Overall,
Chung-Yen Lin microRNA regulatory pathways in breast cancer metastasis, B-cell differentiation, and T-type designing effective and efficient algorithms and software to deal with and remote homology detection: Understanding protein function
Assistant Research Fellow Ca2+ channels involved in cardiac hypertrophy. In the study of microRNA evolution, we are go- these and related problems is a challenge. is one of the most basic problems in proteomics research. We are
Institute of Zoology , National Taiwan University particularly interested in predicting the functions of membrane
ing to develop a computational method to search and identify homologous miRNAs in distant 2. Proteomics proteins. Since the function of a protein is similar to its homologous
Arthur Chun-Chieh Shih genomes. We expect that newly generated data can provide us with more details on how the proteins, we will also work on remote homology detection to assist
Associate Research Fellow role of miRNAs in gene regulation has been expanded in evolution, especially in the lineage ● Mass Spectrometry-based Proteomics function prediction. Furthermore, since the function of a protein is
CSIE, National Central University leading to human. Mass spectrometry (MS) -based proteomic analysis involves protein related to its subcellular localization, we will also work on subcellu-
Huai-Kuang Tsai ● Regulatory Mechanism identification and protein quantitation so that differentially expressed lar localization prediction for various species and organelles.
Assistant Research Fellow proteins between different cell states, e.g., tumor cells and normal cells, ● Disease-centric Membrane Proteome Portal
Computer Science and Information Engineering , The transcription of genes is controlled by interactions between transcription factors (TFs) and can be identified to facilitate biomarker discovery. In this area, we will
National Taiwan University their binding sites (TFBSs) (or cis-regulatory elements). Inferring the function of a TF and identify- study the following two topics: Biological membranes are essential components of life and the struc-
ing its binding sites is helpful for understanding the mechanism of transcriptional regulation. In turing elements of living cells. They form a physical barrier between
past years, we developed two TFBS identification methods, TFBSfinder and MAGIIC, which utilize (1) Protein quantitation: We have previously developed automated the cells and their external environments, as well as with different in-
several data sources, including DNA sequences, phylogenetic information, microarray data and quantitation tools, including MaXIC-Q for iTRAQ-lableing (JPR 2006, tracellular organelles within eukaryotic cells. The basic structure and
ChIP-chip data. Empirical tests on known TFBSs show that our methods are highly accurate in
Postdoctoral identifying motifs, outperforming current methods and achieving high sensitivity and specificity NAR 2007), MaXIC-Q (NAR 2009) for ICAT- and SILAC-labeling, and function of biomembranes are provided by a lipid bilayer. However,
IDEAL-Q (MCP 2010) for label-free approach. We will further en-
for predicting experimentally verified TFBSs. In addition, we constructed a user-friendly interac- hance these tools. First, we will work on de-convoluting overlap- the proteins on the membrane are linked to many unique functions
and play a critical role in communication between separated compart-
Yao-Lin Chang tive platform (MYBS) for dynamic binding site mapping using ChIP-chip data and phylogenetic ping peptide peaks in spectra. Second, we will specifically tackle the ments, such as the signal transduction through surface receptors and
footprinting as the two filters. Based on MYBS, we further investigated the impact of DNA bind- issue of chromatographic shift, by providing an accurate retention the solutes exchanged by protein channels. Because of their location,
Yu-Jung Chang ing position variants on yeast gene expression. Our analysis shows that nucleotide variations, in time prediction method. Finally, we will integrate the above tools to abundance, and various functions, membrane proteins represent over
more than one-third of variable positions and in 20% of dependent position pairs, are highly cor- form an integrated platform for providing comprehensive quantita-
Lien-Chin Chen related to gene expression. We define such positions as functional. However, some positions are tive analysis for label-free and stable isotope labeling experiments. 50% of pharmaceutical drug targets. Therefore, we plan to develop
a disease-centric human membrane protein portal to facilitate bio-
Shu-Hwa Chen only functional as dependent pairs, but not individually. Our analysis supports the importance Friendly interfaces to visualize spectral data and to query proteomic medical research. This portal will integrate various existing databases
information, e.g., Gene Ontology and pathway annotations, will be
of nucleotide variants at variable positions of TFBSs in gene regulation. We now further study the
Allan Lo regulatory mechanisms in yeast and higher organisms (e.g., humans), including identifying tran- provided. and prediction tools, including our predictors on membrane protein
structure, in order to provide comprehensive proteomic information,
scription factor binding sites and discussing the functionality of degenerate positions in TFBSs (2) Protein identification and post-translational analysis: Protein iden- attained by prediction and databases, regarding structure, function,
Ke-Shiuan Lynn and the regulatory rule of adjacent genes. tification is a cornerstone of MS-based proteomics studies. Our sub-cellular localization, classification, interaction, and relationship
Tse-Yi Wang ● Network Biology quantitation tools use spectral raw data and protein identification with diseases.
results from existing identification tools as input. Though several
We will decipher the cellular interactomes of viruses (e.g., herpes viruses and the hepatitis C vi- prestigious commercial tools have been available, e.g., Mascot and 3. Major Collaborative Projects
rus) and the infected hosts, with the aim of identifying the protein complexes hijacked by patho- SEQUEST, they suffer several limitations. First, these search tools are
gen proteins and finding ways to block the mechanism of infection. Furthermore, key proteins usually inconsistent, and remain a problem for proteomic profiling. We are currently participating in various large-scale thematic projects,
and motifs in the virus-host network will be identified and hopefully provide virologists with In the case of post-translational modifications (PTMs), the identifi- including the Bioethanol from Cellulosics project (NSC project hosted by
Dr. Chi-Huey Wong, President of Academia Sinica), the C4 Rice project
cation problem becomes substantially more difficult, since not only (Academia Sinica project hosted by Dr. Wen-Hsiung Li, Director of Bio-
the sequence but also the modification sites of a modified protein diversity Research Center, Academia Sinica), the Bioinformatics Core for
need to be determined. Existing search tools can only identify a Genomic Medicine and Biotechnology Development project (National
few types of PTMs and may not correctly identify their modification Research Program for Genomic Medicine (NRPGM) hosted by Dr. I-Shou
sites. Moreover, many MS/MS spectra can not be identified. There- Chang), the Construction and Integration of Biological Resources project
fore, we started to work on protein identification. Furthermore, we (NSC project led by Director General Min-Liang Kuo, Department of Life
will study the identification of proteins with PTMs, including phos- Sciences, NSC, and Vice President Ching-Fong Chang, Nation Taiwan
phorylation, nitrosylation and glycosylation. Ocean University), and the Networks of the Two Component Systems
● Protein Prediction Problems project (NSC project hosted by Professor Wen-Ching Wang, Director of
Molecular and Cellular Biology Institute, National Tsing Hua University).
We will use a machine learning approach and a knowledge-based ap-
12 13