Abstract

Poster Presentations

Day 4, June 25（Wed.）　

Room P (Maesato East, Foyer, Ocean Wing)

4P-PM-44
PDF

Phosphotyrosine Spectral Library-enhanced DIA-PRM Mass Spectrometry Enables Machine Learning-powered Companion Biomarker Discovery in Lung Cancer

(¹Dept. of Chemistry, NTU, Taiwan, ²IoC, AS, Taiwan, ³Dept. of Chemistry, NTNU, Taiwan, ⁴Dept. of BST, NTU, Taiwan)
^oShen-Shian Chan^1,2, Chiao-Chun Chang^1,2, Irene-Ya Tai^1,2, Yu-Hsuan Lin^2,3, Yi-Ju Chen², Sung-Liang Yu⁴, Yu-Ju Chen^1,2

Tyrosine phosphorylation plays a crucial role in cancer progression and a promising therapeutic target. However, its low abundance (< 2%) poses challenges for detection, limiting its clinical applications. To address this, we developed a phosphotyrosine spectral library-enhanced quantitative assay, integrating data-independent acquisition (DIA) and parallel reaction monitoring (PRM) mass spectrometry (MS) for simultaneous discovery and absolute quantification of druggable phosphosites. In non-small cell lung cancer (NSCLC) cells, our phosphotyrosine-enhanced spectral library was generated by pervanadate treatment using DIA-MS, resulting in a 5-fold increase (25%) in tyrosine phosphopeptides. Next, Incorporating phosphoproteomic datasets from 23 NSCLC cell lines resistant to EGFR therapies expanded the library to 202,285 phosphopeptides. Library-assisted DIA enables the construction of drug-response phosphoproteomic panels. Following the biomarker filtering to explore statistically significant targets, machine-learning decision tree prediction models hold promise in identifying the most suitable next-line drugs for patients. Additionally, calibration curves were established by synthetic heavy phosphopeptides in our strategy, demonstrating reliable targeted quantification of pY1197 on EGFR dynamics. This study advances the sensitive quantitation of phosphoproteomics and highlights its potential as a precision medicine pipeline toward nominating personalized treatment strategies.