Research Fellow  |  Sung, Ting-Yi  
Software Name:MAGIC-web
Inventors:Ting-Yi Sung, Wen-Lian Hsu, T. Mamie Lih, Wai-Kok Choong, Chen-Chun Chen, Cheng-Wei Cheng, Hsin-Nan Lin, Ching-Tai Chen, Hui-Yin Chang

MAGIC-web is the first web server, to the best ofour knowledge, that performs both untargeted andtargeted analyses of mass spectrometry-based glycoproteomicsdata for site-specific N-linked glycoproteinidentification. 

Software Name:iMET-Q
Inventors:Ting-Yi Sung, Wen-Lian Hsu, Hui-Yin Chang, Ching-Tai Chen, T. Mamie Lih, Ke-Shiuan Lynn,

iMet-Q (intelligent Metabolomic Quantitation) is an automated tool for metabolomics quantitation from high-throughput LC-MS data.  By performing peak detection and peak alignment, iMet-Q provides a summary of quantitation results and reports ion abundance at both LC-MS run level and sample level. Furthermore, it gives the charge states and isotope ratios of detected metabolite peaks to facilitate metabolite identification.

Software Name:MAGIC: Automated system for large-scale intact glycoprotein identification by a novel pattern matching algorithm for Y1 ion detection
Inventors:Ting-Yi Sung, Wen-Lian Hsu, Ke-Shiuan Lynn, Mamie Lih, Cheng-Wei Cheng, Chung-Hao Chang, Chia-Ying Cheng, and in collaboration with Dr. Yu-Ju Chen

Glycosylation is a prevalent and highly complex modification of proteins, influencing their functions and activities. Thus interpretation of intact glycopeptide spectra is crucial and, however, remains challenging. In this paper, we present a MS-based Automatic Glycopeptide IdentifiCation platform, called MAGIC, to identify peptide sequences and glycan compositions directly from intact glycopeptide spectra. For unknown glycoproteomic analysis without given protein sequence or glycan database information, correct determination of the Y1 (peptideY0+GlcNAc) ion is most critical. To ensure accurate and efficient Y1 ion detection, we propose a novel algorithm, called Trident, that detects a minimal of triplet peaks with two consecutive mass differences corresponding to the pattern [Y0, Y1, Y1+GlcNAc] or [Y0-NH3, Y0, Y1] from fragmentation of common trimannosyl core of N-linked glycopeptides. To facilitate subsequent peptide sequence identification by common database search engines, MAGIC generates in silico spectra by reassigning the correct precursor m/z and removing all glycan-related ions. Finally, MAGIC solves the glycan compositions by the glycan mass and ranks compositions by the detected glycan-related ions. On the model glycoprotein HRP and a 5-glycoprotein mixture, 2-31 fold increase on relative intensities of peptide fragment ions were achieved, leading to identification of 7 tryptic glycopeptides from the HRP and 16 glycopeptides (five glycoproteins) from the 5-glycoprotein mixture via Mascot database search. On the HeLa cell proteome dataset, MAGIC is capable of simultaneously processing over a thousand of tandem mass spectra in three minutes on a PC. In this dataset, MAGIC successfully identified 112 spectra by Mascot, yielding 36 glycopeptides with 23 distinct top-1 ranked glycan compositions from 26 glycoproteins. Finally, a remarkable false-positive identification rate of 0 was achieved on a dataset of N-linked glycosylation-free Escherichia coli. MAGIC is freely available for download at

Software Name:ProDIA: an automated tool to effectively generate in silico MS/MS spectra from SWATH-MS datasets for untargeted protein identification
Inventors:Ting-Yi Sung, Wen-Lian Hsu, Laura Chang, Nai-Yuan Chiang, Ke-Shiuan Lynn, and in collaboration with Dr. Yu-Ju Chen

We present a fully automated tool, called ProDIA, to generate in silico MS/MS spectra from SWATH-MS datasets for untargeted protein identification. It accepts raw data in either standard mzXML or mzML format as input data. To effectively generate in silico MS/MS spectra, ProDIA decomposes scans in a SWATH-MS dataset to several subsets according to an experiment number of each scan. After constructing precursor/fragment XICs and determining their charge states, ProDIA apply the FWHM of precursors to link precursor and fragment, decreasing the possibility of false-positive linkage. We apply three protein experiments for evaluating the performance of ProDIA. Firstly, a public large-scale yeast lysate experiment showed high accuracies of ProDIA in terms of constructing precursor/fragment XICs (i.e. above 80% accuracy) and determining the charge states of those constructed precursor/fragment XICs (i.e. above 90% accuracy). Secondly, by using the generated in silico MS/MS spectra, two peptides of Enolase protein and 102 peptides of E. coli proteins are additionally confidently identified, respectively. Furthermore, 18 E. coli proteins are additionally discovered, providing more protein candidates. To provide a comfort system environment, several well-designed user interfaces are provided to facilitate convenient inspection and operation. In summary, ProDIA is an effective, user-friendly, and robust tool for generating in silico MS/MS spectra.

Software Name:IDEAL-Q+: a computational tool for labeling and label-free mass spectrometry-based proteomics
Inventors:Ting-Yi Sung, Wen-Lian Hsu, Chih-Chiang Tsou, Han-Ying Yang, and in collaboration with Dr. Yu-Ju Chen

IDEAL-Q+ is a comprehensive user-friendly software tool for mass spectrometry-based quantitative proteomics analysis. IDEAL-Q+ couples an enhanced elution time prediction algorithm and peptide alignment with a novel method called “Spectral Recovery” to improve the performance of identification and quantitation in labeling and label-free experiments with multiple replicate runs. It is suitable for most quantitative proteomics experiments using different instruments and search engines. IDEAL-Q+ is available at

Software Name:iScore: an automated tool for phosphorylation site assignment
Inventors:Ting-Yi Sung, Wen-Lian Hsu, Sheng-Jay Lu, Lien-Chin Chen, Chih-Chiang Tsou, and in collaboration with Dr. Yu-Ju Chen
Software Name:IDEAL-Q: An Automated Tool for High-Performance label-free quantitation with an efficient elution time prediction algorithm
Inventors:Ting-Yi Sung, Wen-Lian Hsu, Chih-Chiang Tsou, Ethan Yin-Hao Tsui, Ke-Shiuan Lynn, and in collaboration with Dr. Yu-Ju Chen

IDEAL-Q is an automated analysis tool for label-free quantitative proteomics. It accepts generic input format including mzXML raw data format and Mascot, SEQUEST, PeptideProphet/ProteinProphet for search result. IDEAL-Q uses an algorithm, called IDEAL (ID-based Elution time prediction by frAgmentaL regression), to predict the elution time based on confident peptide identification (ID) result, and thus the predicted elution time together with precursor m/z to can be used to locate the peptide signal in other LC-MS runs. In comparison with conventional identity-based label-free quantitation analysis, the quantitation coverage in terms of percentage of identified peptides and proteins can be much increased. Furthermore, the tool adopts an stringent validation stelp on Signal-to-noise ratio, Charge state, Isotopic distribution (SCI validation) and modification type using computational and statistical methods so that quantitation accuracy can be ensured even with increased quantitation coverage. IDEAL-Q provides variously optional normalization tools for flexible workflow design such as addition of fractionation strategies and multiple spiked internal standards. Furthermore, many user-friendly interfaces and statistical charts are provided in IDEAL-Q for user to conveniently inspect and validate the quantitation result and also enable to manually re-quantify data. It also provides comprehensible output reports in various formats with useful visualization diagrams and statistical analysis. IDEAL-Q can be downloaded from

Software Name:Multi-Q: a fully automated tool for multiplexed protein quantitation
Inventors:Ting-Yi Sung, Wen-Lian Hsu, Wei-Neng Hung, Wen-Ting Lin, Yi-Hwa Yian, Kun-Pin Wu, Paul CY Yu, Ethan Tsui and in collaboration with Dr. Yu-Ju Chen
Abstract:Multi-Q is developed for multiplexed iTRAQ-based quantitation in protein profiling. It is designed as a generic platform that can accommodate various input data formats from search engines and mass spectrometer manufacturers. To calculate peptide ratios, the software automatically processes iTRAQ’s signature peaks, including peak detection, background subtraction, isotope correction, and normalization to remove systematic errors. Furthermore, Multi-Q allows users to define their own data-filtering thresholds based on semi-empirical values or statistical models so that the computed results of fold changes in peptide ratios are statistically significant. This feature facilitates the use of Multi-Q with various instrument types with different dynamic ranges, which is an important aspect of iTRAQ analysis. Multi-Q is executable on the Windows Platform and available for download at
Software Name:MaXIC-Q: A Fully Automated Generic Tool Using Statistical and Computational Methods for Protein Quantitation Based on Stable Isotope Labeling and LC-MS
Inventors:Ting-Yi Sung, Wen-Lian Hsu, Ethan Yin-Hao Tsui, Yi-Hwa Yian, Chih-Chiang Tsou, Paul Chuan-Yih Yu and in collaboration with Dr. Yu-Ju Chen
Abstract:MaXIC-Q is designed as a generic tool for quantitation using differential isotope-labeling techniques, e.g., SILAC, ICAT, ICPL labelings, and user-developed labeling methods. The tool can also accommodate search results from SEQUEST and Mascot, as well as mzXML files converted from raw files produced by various mass spectrometers. MaXIC-Q contains a filtering module that allows users to filter out low-confidence search results. Statistical and computational methods are applied to construct two kinds of elution profiles for each ion, namely, PIMS (projected ion mass spectrum) and XIC (extracted ion chromatogram); the latter is used to quantify the ion ratio. To ensure that the data used for quantitation analysis is of high quality, MaXIC-Q defines validation criteria for PIMSs so that ions interfered with by co-eluting peptides or noise can be detected. MaXIC-Q can be downloaded from
Software Name:Bioinformatics online web services
Inventors:Wen-Lian Hsu, Ting-Yi Sung, Hua-Sheng Chiu, Allan Lo, Jia-Ming Chang, Chia-Yu Su, Yi-Wen Yang, Chin-Tai Chen, Hsin-Nan Lin, Yi-Yuan Chiu
Abstract:Please visit the webpage of the IASL Lab at