Full-DIA enables complete single-cell proteomics from diaPASEF using deep learning

diaPASEF improves ion utilization and sensitivity by synchronizing quadrupole isolation with trapped ion mobility separation, making it suitable for single-cell proteomics. We present Full-DIA, a deep learning-driven software that enhances proteome coverage, quantitative accuracy, and analysis speed over DIA-NN for single-cell diaPASEF data. Notably, Full-DIA generates a missing-value-free protein matrix under stringent global FDR control, enabling downstream analyses without data gaps. Applied to LPS-treated and cell-cycle datasets, this matrix yields pathway enrichment results with fewer off-target and more biologically relevant pathways. Full-DIA highlights the potential of deep learning for four-dimensional diaPASEF analysis and offers a solution to missing values.