High-Accuracy De Novo Prediction for N- and O-linked Glycopeptides Across Multiple Fragmentation Techniques

N- and O-glycosylation, the most intricate post-translational modifications of proteins, is essential for regulating biological functions. Glycoproteomics faces substantial challenges, particularly in the analysis of O-glycans due to its diversity compared to N-glycans. Additionally, tandem mass spectrometry data patterns exhibit variability among different fragmentation methods, such as stepped collision energy higher-energy collisional dissociation (sceHCD) and electron transfer higher-energy collisional dissociation (EThcD). Existing algorithms often lack sensitivity and are limited to sceHCD fragmentation, restricting their practical application. To address these limitations, we introduce GlycopepECHO, the first deep-learning-based de novo glycopeptide algorithm that captures correlations between spectrum and glycopeptide fragmentation ions among N- and O-glycans. GlycopepECHO achieves over 92% glycan recall and around 95% glycan precision on N-glycan fragmented by both sceHCD and EThcD. It additionally allows O-glycan de novo sequencing benefits from zero-shot learning. GlycopepECHO expands the analytical capabilities of glycoproteomics, shedding light on the diverse roles of glycosylation.