Zhongzhi Luan

BiATNovo improves de novo peptide sequencing from DIA mass spectrometry data by using attention-based bidirectional models and post-processing, achieving better accuracy for long peptides and amino acid-level predictions.

The paper introduces DPHL, a pan-human protein mass spectrometry library built using DIA and PRM, and demonstrates its utility in identifying differentially expressed proteins in prostate cancer and DLBCL.

This paper develops a deep learning method using DIA tensor for phenotype prediction from mass spectrometry data, achieving high accuracy in cancer classification.