Chih-Chiang Tsou

Chih-Chiang Tsou, Ph.D.

Ph.D. Program
Senior Bioinformatics Scientist
Immatics US


Dissertation Title

Computational Framework for Data-Independent Acquisition Proteomics.

Research Interest

Mass spectrometry (MS) is one of the main techniques for high throughput discovery- and targeted-based proteomics experiments. The most popular method for MS data acquisition has been data dependent acquisition (DDA) strategy which primarily selects high abundance peptides for MS/MS sequencing. DDA incorporates stochastic data acquisitions to avoid repetitive sequencing of same peptide, resulting in relatively irreproducible results for low abundance peptides between experiments. Data independent acquisition (DIA), in which peptide fragment signals are systematically acquired, is emerging as a promising alternative to address the DDA's stochasticity. DIA results in more complex signals, posing computational challenges for complex sample and high-throughput analysis. As a result, targeted extraction which requires pre-existing spectral libraries has been the most commonly used approach for automated DIA data analysis. However, building spectral libraries requires additional amount of analysis time and sample materials which are the major barriers for most research groups. In my dissertation, I develop a computational tool called DIA-Umpire, which includes computational and signal processing algorithms to enable untargeted DIA identification and quantification analysis without any prior spectral library. In the first study, a signal feature detection algorithm is developed to extract and assemble peptide precursor and fragment signals into pseudo MS/MS spectra which can be analyzed by the existing DDA untargeted analysis tools. This novel step enables direct and untargeted (spectral library-free) DIA identification analysis and we show the performance using complex samples including human cell lysate and glycoproteomics datasets. In the second study, a hybrid approach is developed to further improve the DIA quantification sensitivity and reproducibility. The performance of DIA-Umpire quantification approach is demonstrated using an affinity-purification mass spectrometry experiment for protein-protein interaction analysis. Lastly, in the third study, I improve the DIA-Umpire pipeline for data obtained from the Orbitrap family of mass spectrometers. Using public datasets, I show that the improved version of DIA-Umpire is capable of highly sensitive, untargeted analysis of DIA data for the data generated using Orbitrap family of mass spectrometers. The dissertation work addresses the barriers of DIA analysis and should facilitate the adoption of DIA strategy for a broad range of discovery proteomics applications. 

Current Placement

Immatics US